Eindhoven University of Technology

MASTER

Energy analysis and plant design for ethylene production from naphtha and natural gas

van Gijzel, R.A.

Award date: 2017

Link to publication

Disclaimer This document contains a student thesis (bachelor's or master's), as authored by a student at Eindhoven University of Technology. Student theses are made available in the TU/e repository upon obtaining the required degree. The grade received is not published on the document as presented in the repository. The required complexity or quality of research of student theses may vary by program, and the required minimum study period may vary in duration.

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

Process Engineering Multiphase Reactors group (SMR) Department of Chemical Engineering and Chemistry

Den Dolech 2, 5612 AZ Eindhoven P.O. Box 513, 5600 MB Eindhoven The Netherlands www.tue.nl

Graduation committee Prof. Dr. Ir. M. van Sint Annaland Dr. F. Gallucci Dr. V. Spallina (supervisor) Energy analysis and plant design for ethylene production Dr. T. Noël (external member) I. Campos Velarde from naphtha and natural gas

Author R.A. van Gijzel

MSc. thesis

Date June 2016

Department of Chemical Engineering and Chemistry – Multiphase Reactors Group

Technische Universiteit Eindhoven University of Technology

ii

Technische Universiteit Eindhoven University of Technology

Abstract

Global ethylene demand, especially in the Asia-Pacific region, is forecasted to grow with a CAGR of 3.5 % for the period from 2013 to 2019. In order to supply this increasing demand, ethylene production capacity has to be increased. Therefore existing plants have to be scaled up and or additional ethylene production plants have to be built. On the other hand further developments, such as advancements in research, strict environmental regulations, and the need for sustainable and energy efficient processes, result in the search for alternative ethylene process technologies to fill this capacity gap.

At first an extensive literature study was performed on commercially available and potential ethylene production process technologies. This literature study showed that a lot of research is performed on this subject and that there are numerous options. These process technologies can be classified into fossil fuel based routes and ‘green’ renewable feedstock based routes. These green routes are thoroughly researched due to fossil fuel, mainly crude oil, depletion and the fact that the conventional ethylene production technology, naphtha steam cracking, has a high CO2 emission. From these green routes only bio-ethanol from biomass, e.g. sugarcane, to ethylene (BETE) is momentarily commercially available, profitable and capable of producing high quantities of ethylene. The highest capacity BETE plant in the world produces 200,000 t/y, which is still only 20 % of the capacity of an average naphtha steam cracker plant. Another disadvantage is the huge agricultural land requirement to grow the crops. Fossil fuel based routes deliver multiple commercially available processes, such as coal to olefins (CTO) via syngas to methanol production. From an economical point of view, together with coal abundance this is a good option. However, from an environmental point of view, this is a step back due to the higher

CO2 emissions. That leaves natural gas and crude oil, of which the latter is refined into naphtha. At the moment 74 % of the ethylene produced in Europe comes from naphtha steam cracking, 71 % of the ethylene produced in the U.S. comes from natural gas via ethane steam cracking. As mentioned before, both these technology processes are very energy intensive and have high CO2 emissions. Shale gas revolution and consequently low prices of natural gas, and the large natural gas reserves result in several studies regarding the exploit of methane. One of promising process technologies, is the oxidative coupling of methane (OCM) to ethylene as an alternative. Two main obstacles have prevented OCM to be commercialized in industrial application, the first is the relatively low ethylene concentration in output gases and determining how to convert methane to C2+ with high selectivity. Despite extensive research, still many fundamental aspects about this process remain unknown.

This work is part of the European project MEthane activation via integrated MEmbrane REactors (MEMERE), which is a four year research and innovation project in which eleven industrial and higher education partners work together at methane activation towards C2+. The main focus of MEMERE is the air separation through novel MIEC membranes within a reactor operated at high temperature, and hopefully achieving higher yield compared with a conventional reactor, combined with improved energy iii

Technische Universiteit Eindhoven University of Technology

efficiency in this multifunctional unit. Therefore this thesis assesses the energy analysis, environmental and economic performance (in terms of ethylene yield and energy efficiency) of the main conventional ethylene production process and the same for a non-membrane OCM reactor with a separate cryogenic air separation unit on site. This assessment consists of the basic design and optimization of complete, full-scale plants, built around the reactor design at different operating conditions (e.g. temperature, reactant flow and energy balance).

In this thesis the following main plant configurations are modelled in Aspen Plus, each with adjustments to investigate different parameters and perform sensitivity analyses:

- Model 1: The benchmark technology of the naphtha steam cracker - Model 2: The oxidative coupling of methane with air separation unit (ASU), without fuel recycle - Model 3: The oxidative coupling of methane with air separation unit, with fuel recycle

iv

Technische Universiteit Eindhoven University of Technology

Contents

Abstract iii

1 General Introduction 1

1.1 Project Motivation 1

1.4 Objectives and aim of this work 2

1.5 MEMERE 2

2 Literature review 3

2.1 Background 3

2.2 Research questions 3

2.3 State of the art ethylene production processes 5

2.4 Discussion 17

3. Energy analysis and plant design naphtha cracker 21

3.1 Theory 21

3.2 Process Flow Diagram 27

3.3 PFD section I 29

3.4 PFD section II 35

3.5 PFD section III 37

3.6 PFD section IV 39

3.7 Recycle, combustion and power generation 46

3.8 Emissions and waste 48

3.9 Coke deposition and storage 48

3.10 Energy analysis 50

3.11 Aspen Plus model 53

3.12 Steam cracking of ethane 54

4 Energy analysis and plant design for OCM 55

4.1 Introduction 55

4.2 Theory 56

v

Technische Universiteit Eindhoven University of Technology

4.3 Aspen Plus model 61

4.4 ASU model 64

4.5 OCM part I 70

4.6 Energy analysis and plant design OCM part II 80

4.7 Emissions and waste 92

4.8 Sensitivity analysis recycling methane and ethane 93

4.9 Pressure sensitivity analysis 98

4.9 Temperature sensitivity analysis 104

4.10 Pinch analysis 105

4.11 Aspen Plus model 106

5 Techno-economic analysis 108

5.1 Base case economic analysis 109

5.2 Sensitivity analysis on economics 110

6 Summary and Conclusions 112

7 Recommendations 117

8 Acknowledgements 119

9 Bibliography 120

10 Appendix 124

vi

Technische Universiteit Eindhoven University of Technology

1 General Introduction

One of world’s largest produced raw materials in the petrochemical industry is ethylene, which is the simplest olefin (C2H4). Ethylene has various direct and indirect applications, for instance HDPE, LLDPE and LDPE or as petrochemical intermediate in order to produce PET, solvents, cosmetics, paints, packaging, etc. Also various chemical components such as ethylene dichloride, ethylene oxide and ethylbenzene are extracted from ethylene.

Global ethylene capacity is estimated to rise from 167 million metric tons per year (Mtpy) in 2014 to 208.5 Mtpy in 2017, with the US and China producing one-third of the additional capacity over the forecasted coming years [1]. As the demand for ethylene is still growing, this secures ethylene production and its optimization to be an ever-interesting topic.

1.1 Project Motivation

Currently, the main operation (>95 % in Western Europe) involving the production of ethylene is performed by steam cracking of larger hydrocarbons as feedstock, mainly ethane and naphtha [2]. Steam cracking introduces smaller hydrocarbons and unsaturation by the use of very high temperatures between 750-1000 °C. Reaction 1.1 below shows a simplification of ethane cracking to ethylene: CHCHH 2 6o 2 4  2 (1.1)

Ethane has, besides methane, the highest endothermic heat of cracking, +4893 kJ/kg. For propane the corresponding amount is +4295 kJ/kg and for a high-molecular-mass n-alkane around 1364 kJ/kg [3]. The breaking of the bonds of the higher hydrocarbons, and thus the endothermic reactions makes the steam cracking a highly energy intensive process. Specifically, the pyrolysis section of a naphtha steam cracker consumes approximately 65% of the total process energy required for the plant [4]. According to Ren et al. (2006) the overall steam cracking process is famous for being the most energy-consuming process in chemistry because it globally uses 8% of the total primary energy use in the chemical sector. Besides cracking, other energy-intensive and costly processes such as repeated compression, cryogenic distillation and multiple water quenching steps are required to eventually produce PE grade purity. Another disadvantage of steam cracking is that the process contributes to environmental pollution by producing approximately 180-200 Mtpy of CO2 worldwide [4]. As further improvements are rather difficult due to the fact that the steam cracking process has been studied extensively and has been improved and optimized for more than 50 years, it is more realistic to develop or improve another (new) process which could have the same or higher potential.

1 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

1.4 Objectives and aim of this work

First an extensive literature survey of the state of the art technology was performed to assess the current performance of the system and potential technologies, specifically the oxidative coupling of methane (OCM) with natural gas as feedstock. Hereafter, naphtha steam cracking was selected as benchmark technology. The main aim of this work is thus to assess the energy analysis, environmental and economic performance (in terms of ethylene yield and cost of production) of the conventional ethylene production technology compared to OCM. This assessment consists of the design and optimization of complete, full-scale plants, with different operating conditions (such as temperature, reactant flows, energy balance). In particular, the efficiency prediction and the mass and energy balances will be predicted by means of a commercial computer model (Aspen Plus v8.6 ®). Eventually, the following three main plant configurations were modelled and studied in detail:

- Model 1: The benchmark technology of the naphtha steam cracker - Model 2: The oxidative coupling of methane with air separation unit (ASU), without fuel recycle - Model 3: The oxidative coupling of methane with air separation unit, with fuel recycle

The third part of this project consists of sensitivity and economic analysis on the different plant configurations.

1.5 MEMERE

This thesis is part of the European project MEthane activation via integrated MEmbrane REactors (MEMERE), which is a four year research and innovation project in which eleven industrial and higher education partners work together at methane activation towards C2+ [5]. MEMERE received a European grant from the Vision2020: The Horizon Network to carry out this project. The focus of MEMERE is on the air separation through novel MIEC membranes integrated within a reactor operated at high temperature for OCM allowing integration of different process steps in a multifunctional unit and hopefully achieving higher yields compared with a conventional reactor, combined with improved energy efficiency. Therefore this thesis assesses the energy analysis, environmental and economic performance of the main conventional ethylene production process and the same for a non-membrane OCM reactor with a separate cryogenic air separation unit on site. This work will be used as a comparison to the novel oxygen membrane technology.

2 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

2 Literature review

2.1 Background

Global ethylene production capacity growth was higher in 2013 compared with the years before, and production capacity for 2014 was again increased with another 2.6 Mtpy to more than 146 Mtpy [6]. According to the Oil and Gas Journal ethylene production capacity for North America and Europe is anticipated to be quite steady, while it is forecasted that the Asia-Pacific region will be the fastest increasing ethylene market in the next five years resulting in a compound average growth rate (CAGR) of 3% for global ethylene demand from 2014 to 2020 [7]. According to the Ministry of Economy, Trade and Industry of Japan the forecasted CAGR for the period 2006 to 2016 was 2.6%, and for the period from 2013 to 2019 is expected to be 3.5% [8]. However there are also sources which claim an even higher forecasted CAGR of 6% [9],[10]. Nonetheless it is certain a further growth of ethylene production capacity will have to take place in the future, in which China will remain the fastest growing market for ethylene and ethylene derivatives. Despite that China will remain dependent on import, it has also increased its own production capacity both through naphtha cracking and coal-to-olefins [11].

Oil and Gas Journal survey data shows that Asia-Pacific has the biggest ethylene capacity with 45.7 Mtpy and that the two biggest ethylene complexes are situated in Singapore and China. The top ten of ethylene production complexes, as of January 2014, have a capacity between 1.705 and 3.5 Mtpy [6],[7], which comes down to nearly 200,000-400,000 kg of ethylene produced per hour at constant production. However from the list of top ethylene producers which includes information about their total capacity and the number of plants available, it can be concluded that most ethylene producing plants produce around 700,000 tpy in a single-train plant. Interesting is the fact that the top 10 ethylene producers are almost all companies well-known for their petrochemical products. This insinuates that ethylene plants are probably integrated at the same place as crude oil distillation and that the feedstock mainly consists of naphtha and natural gas.

2.2 Research questions

This leads to one of the main research questions, from which feedstocks can ethylene be produced? Other interesting questions are what the advantages and disadvantages of these processes are, e.g. most optimized process, highest yield, milder process conditions, etc.? Eventually it always comes down to money, saftey and environmental issues meaning it is important to research which process is profitable but also if this process is also energy efficient, safe and environmental friendly, and eventually if this process could be scaled up. In the general introduction it is already implied that the main focus will be on the prospects of oxidative coupling of methane coming from natural gas and the

3 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

conventional technology of naphtha steam cracking. See figure 1 for an overview of the ethylene process technologies.

.

Abbreviations family portrait:

BATH: Bio-acid acetone to hydrocarbons (e.g. olefins) CC: Catalytic Cracking or Catalytic Pyrolysis DCC: Deep Catalytic Cracking DH: De-Hydration process (e.g. methanol to olefins (MTO), ethanol dehydration) FEM: Fermentation FP: Flash Pyrolysis, sometimes in the presence of methane FT: Fischer-Tropsch synthesis GAS: Gasification and liquefaction GS: Gas Stream reactor technologies, e.g. shockwave reactors HG: Hydrogenation+ other options are via electrolysis and photosynthesis HP: Hydro-Pyrolysis HTUL: Hydro-Thermal Upgrading Liquefaction which produces naphtha from biomass feedstock LIQ: Liquefaction OC: Oxidative Coupling of methane OD: Oxidative Dehydrogenation of ethane OM: Olefin Metathesis, e.g. ABB-Lummus Olefin Conversion Technology, IFP-CPC meta-4 OU: Olefins Upgrading (conversion of C4- C10) to light olefins PD: Propane Dehydrogenation RCY: Re-cycling pyrolysis using waste, such as discarded plastics, used rubber, etc. REC: Recovery of refinery off gases, which contains ethylene, propylene, propane, etc. REF: Refinery processes. Distillation of crude oil produces naphtha and heavy oil. Catalytic cracking produces off gases. Cryogenic separation and absorption produces ethane and LPG; SC: Steam Cracking (conventional) SEP: Gas separation process which produces methane, ethane and propane SR: Steam Reforming of natural gas to produce methanol

Figure 1: Family portrait of olefins production and their definitions based on Ren et al. (2006)

4 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

2.3 State of the art ethylene production processes

This chapter will show an overview of the different routes for ethylene production, as well as a more in depth view on a selection of these technologies. The highlighted processes are modelled in this thesis.

Ren et al. (2006) provides a ‘family portrait’ which gives a schematic overview of various process technologies to produce olefins such as ethylene, using different feedstocks. This family portrait is depicted in figure 1 on the previous page. According to Thinnes (2012) bio-based chemicals such as bio- ethylene could be the next new thing [12]. This statement was made based on a just published article about bio-plastics from ChemSystems Nexant. ChemSystems proposes a more detailed scheme for renewable feedstocks and ‘green routes’ as an alternative to fossil fuels for ethylene and polymer production, this scheme can be found in figure 2 [13].

Figure 2: Potential green processes for ethylene and other HVCs from ChemSystems (2011).

Ethylene, and the polymer poly ethylene (PE) produced from this chemical, are highly dependent on large volume production, so even if ethylene can be produced by bio-based routes scaling-up is of great importance. Strong agricultural-based economies such as Brazil, the U.S. or India could be promising sources for bio-based ethylene instead of the Middle East or U.S. for ethylene from conventional petrochemical routes, using crude oil and its derivative naphtha.

5 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

2.3.1 CO2 as feedstock

Photosynthetic conversion of CO2 to C2H4

There are a few sources which review the conversion of CO2 into ethylene, for example via photosynthesis by using the cyanobacterium Synechocystis sp. PCC 6803. See equation 2.1 for the reaction and figure 3 for an overview of the process. 2COHOCHO 2 3 2 2 o 2 4  2 (2.1) In short, the ethylene-forming-enzyme (EFE) catalyses the conversion of the TCA cycle intermediate, a- ketoglutarate, into ethylene. A side reaction of the EFE enzyme generates succinate from arginine, for completion of the TCA cycle in cyanobacteria.

Figure 3: Ethylene production process in Synechocystic from Ungerer et al. (2012) [14].

It is easier to produce bio-ethylene than other biofuels due to the gaseous phase of ethylene, because it does not suffer from technical difficulties that are associated with liquid biofuels such as cell harvesting and oil extraction. Ethylene is not toxic to the host or a food source for contaminating microbes compared to many other liquid biofuels. A peak production rate of 7125 µg L-1 h-1 (171 mg L-1 day -1) was achieved and a continuously production rate of 3100 µg L-1 h-1 (92 mg L-1 day -1). This production rate is higher than that reported for other algal or cyanobacterial petrochemicals. However the increased productivity requires continuously, every three weeks, refreshing the media for sustained production. The same ethylene production rate was achieved across four sub-cultures without apparent loss of ethylene-forming ability, one culture is however only 35 mL. The requirement of continuous production can be met by using a continuous or semi-continuous photo bioreactor, with the additional benefit of being able to harvest biomass for processing into additional biofuel. Eventually up to 5.5 % of the fixed carbon was directed to ethylene synthesis, which is higher than previous carbon-partitions rates published into the TCA cycle [14].

6 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Photo-biological ethylene production gas has potential as an alternative technology for ethylene production, for reduction of greenhouse gas emission and for treating wastewater. However, scaling up will face big future challenges such as large capital costs, productivity of the organism, harvesting of the biofuel and joint ventures between engineering and biology.

Artificial photosynthesis

Another technology, which also uses CO2 as feedstock, performs electrochemical reduction of CO2 to ethylene which can be described as artificial photosynthesis. CO2 electro-reduction process is run by renewable energy in comparison with natural photosynthesis which converts CO2 and water into glucose and molecular oxygen under sunlight. The electrochemical conversion of CO2 to ethylene is thermodynamically uphill, which means it requires intense input energy and will only be a good alternative when renewable energy sources such as sunlight, wind or terrestrial heat are applicable to the purpose. Well known is that the generation of electricity by renewable source is not very dependable due to unstable functioning and electricity storing difficulties. However the latter can be solved due to electrochemical conversion of CO2 to energy-rich substances that are driven by the electricity originated in renewable energy.

Reaction 2.2 shows the overall reaction which selectively converts CO2 to ethylene, while solar energy is stored in the chemical bonds of ethylene (which has a much higher energy density 50 MJ/kg compared to 0.63 MJ/kg as the energy density of a Li-ion battery) [15].

2CO 12 H 12 e  C H4 H O Cathode: 2   o2 4  2 2H O O4 H 4 e  Anode: 2o 2   (2.2)

2COHOCHO 2 3 Total: 2 2 o 2 4  2 This process resulted in a maximum conversion of 8.0% with a maximum selectivity of 57.7%, under these conditions increased concentrations of impurities, e.g. heavy metals were produced [15].

Other alternative methods for CO2 to C2H4 production

Other methods to convert CO2 to ethylene, consist of hydrogenation in combination with a heterogeneous catalyst at high temperature to methanol or converting CO2 under supercritical conditions. The driving force of each of these green routes, including the ones which use biomass, is based on the environmental impact.

7 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

2.3.2 Biomass as feedstock

The recent increase in the fuel ethanol market, particularly in Brazil and the United States, has resulted in renewed interest in the production of high value chemicals from ethanol. Environmental concerns over fossil fuel depletion and the concept of sustainability, are also reasons the attention has switched to renewable sources for both chemical feedstocks and for energy. The use of biomass can contribute to the reduction of CO2 emissions and partially fulfil the growing demand for PE in the near future.

Anaerobic fermentation of sugars to ethanol

Ethanol can also be obtained by the anaerobic fermentation of sugars (mostly glucose) which in turn are the products of acid or enzymatic hydrolysis of cellulosic compounds. These compounds are derived from biomass, particularly wood materials. Diethylether is considered an intermediate and not a by- product, it is produced between 150-300°C while ethylene is predominantly produced between 320-

500°C. However under these operational conditions acetaldehyde (CH3CHO), hydrogen and other minor by-products are formed. On the other hand with the right catalyst, mostly activated alumina, the ethanol conversion is usually higher than 95% and could even reach 99.5% with a reaction molar selectivity ranging between 95-99% [16].

Ethanol dehydration

Only a limited number of petrochemical commodities can be produced from biomass. Ethylene, propylene and BTX (benzene, toluene and xylene) can be produced from different types of biomass [17]. Biomass and organic waste (wood, food, etc.) such as corn stover, sugarcane, cassava, and wheat can be used to produce bio-ethanol and eventually ethylene after dehydration of ethanol. The production costs of bio-ethanol have decreased significantly over the years as a result of productivity improvements and economies of scale. Ethanol dehydration translates into a utilization of 63.3 wt-% of ethanol for ethylene production due to that ethanol contains oxygen which is eliminated as water in the dehydration reaction. The dehydration reaction of ethanol involves acidic catalysts at temperatures above 200 °C, which can result in a selectivity over 99 % due to tailored catalysts. See reaction 2.3 for the dehydration reaction of ethanol, which is an endothermic reaction that requires 390 cal/g of ethylene formed: C H OH H O C H 2 5 l2  2 4 (2.3)

8 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

A PFD of the generic bio-ethanol-to-ethylene (BETE) process diagram is depicted in figure 4.

Figure 4: PFD of generic BETE process from Morschbacker (2009).

Profitability in bio-ethanol for BETE is rather difficult due to today’s low cost of petroleum ethanol, (produced via hydration of ethylene) and the relatively high production cost of fermentation ethanol, which is 10-40% higher than petroleum ethylene [18]. According to Haro et al. (2013) the production of ethylene from ethanol should be below 0.45€/L in order to achieve profitability from the BETE process regardless of the origin of the ethanol. Bio-ethanol production costs, are highly dependent on the feedstock price (minimum 85% of the production cost). BETE with Brazilian ethanol can be profitable or via ethanol produced via the indirect synthesis of syngas [17].

Besides profitability, the availability of biomass is a bottle neck for all biomass-derived products, along with high production costs, scaling up and future regulations of biofuels. A BETE plant with the same ethylene production capacity requires huge land requirement for producing the biomass necessary for the bioethanol. As 95% of the feedstock for bio-ethanol comes from agricultural crops. Standard bio- ethanol plants produce around 40,000 t/y bio- ethanol, which means multiple of those plants are required to produce sufficient feedstock for one BETE plant of standard size. An ethanol productivity of 5,500 L/ha, is equivalent to 2.7 t/ha ethylene. A BETE plant would require 165,000 ha of agricultural land (approximately 40 x 40 km) to produce 450,000 t/y of ethylene [3]. Moreover, the alternating price of ethylene prices brings uncertainty regarding the ethanol-to-ethylene plant. Braskem is the world’s leading producer of biopolymers with its 200,000 t/y BETE plant from sugarcane-based ethanol. Bio- ethylene from ethanol reduces CO2 emissions by 70-80 %. Greenhouse gas emissions per kilogram bio- based ethanol range from 0.7 to 1.5 kg CO2 eq. per kg ethanol and from 1.3 to 2 kg per kg if emissions at end-of-life are included [19].

Other ethylene production technologies via biomass

For ethylene production from biomass different technologies are possible, for instance via dehydration of bioethanol, the conversion of bio-dimethyl ether (bio-DME), (flash) pyrolysis of biomass, enzymatic hydrolysis, gasification and fermentation of syngas [20].

9 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

2.3.3 Coal as feedstock

Coal is a more abundant than oil and gas, and it is composed mainly of carbon, hydrogen and oxygen. According to the International Energy Agency (IEA) reserves of coal reaches 1040 billion tonnes, equal to 142 years of production at current rates. In comparison, to 42 years of production at current rates with crude oil reserves [21].

From coal to syngas to methanol production

China is one of largest producers of coal to olefins (CTO), e.g. ethylene and propylene [22]. This technology consists of multiple steps and various process technologies, a process flow sheet of the CTO process can be found in figure 5. In this process, first coal is converted to syngas by gasification and liquefaction at high pressure of 5 MPa and high temperature of 1400-1600 °C with controlled amount of compressed oxygen, nitrogen and steam which serve as reactants. Syngas is a mixture which primarily consists of CO (c. 63%) and H2 (c. 27%) and can be produced from various feedstocks (as can be seen in figure 1). Hereafter syngas is purified and converted to methanol via Fischer-Tropsch. Thereafter the methanol is cooled, further refined, purified and eventually stored until it will be catalytically dehydrated and partially converted to ethylene over alumina and zeolite catalysts at 350°C and 0.1-0.3 MPa. This production process requires an abundance of water and simultaneously produces an abundance of CO2 emissions.

Figure 1: Process flow sheet diagram of coal-to-olefins process from Xiang et al. (2014).

10 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Approximately 1.4 tons of coal is required to produce 1 ton of methanol, and 2.6 tons of methanol consumed per ton of light olefin produced. In CTO the propylene/ethylene product ratio is between 1.2- 1.8 [23]. 70% of global methanol production comes from synthesis of natural gas into syngas while only 11% comes from the synthesis of coal to syngas.

Currently 100% of the world’s coal-to-methanol production is based in China. A 0.6 Mtpy CTO facility in

Beijing would increase provincial CO2 emissions by 14%. China’s CTO capacity represents only 0.92% of the world’s total ethylene and propylene capacity (256.2 Mtpy). Yet, China’s government has approved an additional 6.9 Mtpy and another 13.4 Mtpy of CTO capacity. So, the main driving force of this method is the coal abundance and its feedstock price.

2.3.4 Natural gas as feedstock

Natural gas consists primarily of methane (90%) and is a very popular feedstock for ethylene production in the US due to the much lower olefin production price as a result of the low natural gas feedstock price. As mentioned before natural gas prices in North America and the Middle East are very low and natural gas-to-olefins production price is the lowest compared to other feedstock such as coal and naphtha from crude oil. Momentarily a great switch from naphtha cracking to ethane (from natural gas) cracking is taking place in the U.S. Gas feedstocks like natural gas are less used in Europe because they are rarely economically available (due to difficulty and high costs of transportation) [24].

Natural gas to syngas to olefins production

Natural gas to ethylene processes could involve steam reforming to syngas, sometimes referred to as steam methane reforming, hereafter the Fischer-Tropsch process converts syngas into liquid hydrocarbons via the following reaction.

2n 1 H nCO C H nH O (2.4)  2  on 2n 2  2

Where n is typically 10-20, the product is thus a heavy naphtha which could in turn be steam cracked into ethylene. Another technology to convert natural gas into ethylene involves gas separation of the natural gas coming directly from a (oil, gas or condensate) well. This raw natural gas contains many natural gas liquids (NGL) such as ethane, propane, butane which can be very valuable. Some olefins are also produced directly as by-product of gasoline and diesel fuel production from syngas Fischer-Tropsch processes. There are two principles for removing NGLs from the natural gas stream, via absorption oil or via cryogenic expansion. The resulting products LPG and ethane are hereafter converted into ethylene. This process involves hydro-pyrolysis of a hydrocarbon oil in the presence of hydrogen followed by cracking of the hydro-pyrolysis effluent. Hereafter separation and recycling of hydrogen and ethane, and recovering of methane and ethylene takes place [25].

11 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Natural gas to methanol to olefins production

Economics of the new process developed by UOP/Norsk Hydro, which converts methane to methanol in a first stage, and converts methanol to olefins in the second stage, turns out to be competitive with conventional processes. The process is based on 80 % conversion of the carbon content of methanol into ethylene and propylene, which takes place in a fluidized bed reactor [26].

Oxidative coupling of methane

A third option to produce ethylene from natural gas involves the direct conversion of natural gas to ethylene via oxidative coupling of methane (OCM). See reaction 2.5 for the catalytic exothermic reaction to ethylene. 2 2 CHOCHHO4 2 o 2 4  2 (2.5)

The complete reaction mechanism is more complicated and will be discussed in chapter 4. Methane is actually first oxidatively converted into ethane, and then into ethylene. The fundamental aspects of this process is based on a heterogeneous process due to methane activation on a metal oxide surface, as well as a homogeneous gas-phase process because of the free radical reaction mechanism. Ethane is produced mainly by the coupling of methyl radicals in the gas phase, generated on the surface. Hereafter ethane can be dehydrogenated to ethylene. The yield of ethane and ethylene are limited by secondary reactions of methane with the surface and by the further oxidation of ethylene. Other sequential reactions could produce small amounts of higher hydrocarbons, these products are often reported as C2+. In addition to these hydrocarbons CO and CO2 are formed non-selectively [27]. The conversion of methane to CO2 is an undesired reaction due to reduced yield of ethylene and the fact that the formation of CO2 is highly exothermic. The dissipation of heat during the OCM process is a big engineering problem for commercial operation.

OCM to ethylene occurs at high temperatures between 750-950 °C, at the limits of catalytic combustion where the availability of oxygen largely dictates the C2+ selectivity [27]. At this temperature methane is activated by the catalyst, forming methyl free radicals. In many cases, the catalysts used in OCM are uncommon in oxidation catalysis, as they do not contain transition metal ions. Since the first research on OCM catalysts in 1984 [28], hundreds of catalysts have been found to be more or less active and selective for the coupling of methane. Maximizing C2 yield by varying catalyst composition was popular in earlier OCM research, however from an economic point is was shown that C2 selectivity is more important than the yield, provided one can obtain a decent CH4 conversion. The maximum selectivity and the conversion are related, and the sum of those two are useful to consider. The conversion and selectivity generally increases, as the partial pressures of the reagents decreases. However the separation of heavily diluted ethylene is undesired. It is also mentioned that specific activity is a less important factor, because of the heat removal problem. One of the best results to date, is 30 % conversion of methane and about 80 % selectivity to C2+ hydrocarbon products, ethane and ethylene [29]. The yields obtained so far (20-25 %) are not competitive with conventional routes since the

12 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

ethylene concentration in the product gas is only ca. 10 vol-% [30], [31]. Application of a special separative reactor has led to yields up to 50-60 %.

OCM is a very much researched topic, however until now none of the big companies came close to a process which could compete with the steam cracking process. One of the companies researching this technology is Siluria Technologies, a scheme of their OCM process can be found in figure 6 [32].

Figure 6: Siluria technologies’ OCM process, dashed line represents an optional stream.

In the USA, especially on the Gulf Coast, a great switch from naphtha to natural gas is observed. This change is mainly due to the drop of natural gas feedstock prices due to the new shale gas revolution. Another reason for the recent switch are the price fluctuations to which crude oil have been exposed. Besides the recent drop due to shale gas production, natural gas prices have been quite stable. However this profitability may be eroded by low oil prices and increased value of co-produced high value components produced during naphtha steam cracking, e.g butadienes and propylene. Proper access to pipeline supplies from North Sea production fields or being located in area with sufficient concentration of gas, makes natural gas as feedstock a lot more profitable. The US also has lower utility and labour costs compared to Europe [24]. The driving forces for these different production technologies, are the feedstock price and the upgrading of natural gas to a more valuable product.

2.3.5 Crude oil as feedstock

Crude oil is composed of a great variety of hydrocarbons in various molecular structures, namely saturated hydrocarbons (paraffins or alkanes), cycloalkanes (naphtenes), unsaturated (olefins or alkenes), ring systems with paraffinic or olefinic side chains (aromatics). Crude oil also contains a relative small number of heterorganic compounds such as sulphur, nitrogen and oxygen-containing organic compounds and metal compounds. Classification of the quality of crude oil can be done according to hydrocarbon type, product fractions or sulphur content, for example paraffin-based crude oils containing the low and medium molecular mass paraffins are more suitable for all kinds of catalytic conversion into gasolines and middle distillates such as naphtha [33].

13 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Deep catalytic cracking and off-gas recovery

Crude oil distillation results in multiple products which could be used for ethylene production, such as off gases, heavy oil, gas oil, liquid naphtha and propane as can be seen in figure 1. The composition of the crude oil is highly dependent on the country of origin. After crude oil processing separation of the different products takes place, hereafter off-gases (i.e. ethylene, propylene etc.) recovery could take place or deep catalytic cracking of heavy oil to ethylene. Recovery from FCC off-gas has gained importance, but due to NOx content and risk of formation of explosive resins, additional (safety) measurements have to be implemented.

Hydropyrolysis and oxidative dehydrogenation

In case of C4+ components further cracking of olefins could be performed to produce ethylene. Ethane or LPG resulting from crude oil can be converted into ethylene via hydro-pyrolysis, steam cracking or oxidative dehydrogenation of ethane. Gas oil and naphtha are converted into ethylene via steam cracking, catalytic cracking, thermal cracking or hydrocracking. On the next page shows a simplified overview of a naphtha cracking plant, which will be discussed in detail in the next chapter.

Naphtha

Naphtha steam cracking will be explained in detail in the next chapter, however before doing so, naphtha as a mixture should be first discussed. Naphtha can be produced from crude oil as petroleum distillates as mentioned above, however there are also other options to produce naphtha which are not displayed by Ren et al. For instance naphtha can be produced from natural gas condensates, or by distillation of coal tar, peat, as well as the destructive distillation of wood. To remain within the context of this research, there will only be dealt with naphtha produced by the processing of crude oil in petroleum refineries. This petroleum naphtha or petroleum solvent is processed in different grades, of which the grades are normally based on boiling point range due to distillation cuts. These boiling point ranges can vary from refinery to refinery due to different crude oil compositions or blends, production technology (e.g., crude oil distillation or fluid catalytic cracking), different operating conditions or product specifications of the producing company. Therefore is difficult to specify a general naphtha compositions.

On the other hand naphtha is sometimes defined by the carbon chain length range, which can be considered as an indirect indication of a boiling point range [3]. For instance from distillation crude oil, the light fraction is referred to as light virgin naphtha and has a boiling point range between 0-100 °C, a heavy virgin naphtha has a boiling point range between 100-250°C. In the fluid catalytic cracking (FCC) process the product stream is split into three fractions with corresponding boiling point ranges; light (<105°C), intermediate/medium (105-160°C) and heavy (160-200°C) [34].

14 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

It is thus evident that all kinds of naphtha could be used for cracking to olefins, however light naphtha is more suitable because of the presence of paraffinic hydrocarbons. Heavy naphtha is the bottom product and is more suitable for the production of high octane gasoline and valuable petrochemical like benzene, toluene, and xylenes (BTX) [34], [35]. The fact that there is no such thing as thé composition of naphtha, could however work in your advantage by producing or buying the naphtha ideal for ethylene production in terms of operating conditions, highest yield and economical point of view. Design specifications on naphtha such as aromatic free or sulphur free can also have a great influence on the naphtha product. Which also influences the danger of the fluid itself, for instance aromatic naphtha can be smoky, toxic and carcinogenic.

It is found that straight hydrocarbons, i.e. paraffins/alkanes, yield olefins when heated at a high temperature (cracked). In contrary, branched and aromatic hydrocarbons become heavier upon heating or decomposed to carbon, coke, which fouls the reactor and equipment. Since aromatic feedstocks contribute little to ethylene yields and are a precursor to coking, quality characterization factors have been developed by Watson et al. [36] For example the K factor which indicate the aromatics content by empirical correlation. The K factor is defined as formula 2.6:

1.8T 1 3 k K (2.6) d

Where Tk is the molal average boiling point in K and d is the relative density. Naphtha with a K factor of 12 or higher are considered saturated, those below 12 are considered aromatic or naphthenic. In the next chapter the steam cracking process of naphtha will be modelled and discusses in detail. Figure 7 shows an overview of this naphtha steam cracking process. The reason for this method being the conventional process used, are the feedstock prices, high yield compared to most technologies, and the fact that the technology is available and economically viable.

Figure 2: Schematic overview of a simplified naphtha cracking unit from Speight et al. (2015).

15 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

2.3.6 Other ethylene production methods

Apart from the ethylene production technologies discussed in the previous paragraphs, two other processes are discussed in this paragraph due to recent success in research or commercialization.

Metathesis of propene

Metathesis of propene leads to 1-butene and ethylene, see the following reaction: 2 CHCHCH3 6l 2 4  4 8 (2.7)

Since this reaction is reversible, the ethylene and propylene production of a plant can be adjusted to market conditions. This technique is commercialized in a plant in Canada and the U.S. Gulf Coast. Due to the higher demand for propylene and the lack of propylene production in ethane crackers, metathesis has gained attention for the production of propylene via the reverse reaction. However, contents of butene and butadiene in the feed create problems, as coking rates of the applied catalysts are increasing.

Oxidative dehydrogenation of ethane

Dehydrogenation of ethane (DH) is limited by equilibrium and has a very poor yield which is not competitive with conventional routes. On the other hand, some researchers mention oxidative dehydrogenation (ODH) of ethane to ethylene in the presence of a suitable catalyst as the most promising alternative to naphtha steam cracking [37]. In ODH the reactions are irreversible and exothermic, no external heating is therefore required and no equilibrium limitations are observed. The reactions proceed via a triangular series/parallel scheme with undesired complete combustions of both ethane and ethylene [38]. See the following three reactions in formula 2.8-2.10: 0.5 CHOCHHO2 6 2 o 2 4  2 ̴H=-105 kJ/mol (2.8) 3.5 2 3 CHOCOHO2 6 2 o2  2 ̴H=-1428 kJ/mol (2.9) 3 2 2 CHOCOHO2 4 2 o2  2 ̴H=-1323 kJ/mol (2.10)

The main disadvantage of ethane ODH, is that the yield of ethylene is limited by the undesired total oxidation reactions of both ethane and produced ethylene to carbon dioxide. These oxidation reactions generate a large amount of heat that can cause a runaway in the reactor and even explosions. This requires careful operation.

In the case of ethane ODH, the ethylene yield values reported with some catalysts are comparable to, or even better than, the corresponding values obtained from steam cracking. ODH has been studied over a wide range of catalysts, of which the majority exhibits low selectivity to the desired reaction at high conversion levels. However, on the best catalyst, selectivities higher than 80 % at conversion levels higher than 80 % (75 % ethylene yield) were obtained at relatively low reaction temperature of 340-400

16 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

°C [38], [39]. Multi-tubular reactors are commonly used in the industry to carry out the exothermic ODH reactions, since the control of the reaction temperature plays a key role in the process. An attractive alternative to these fixed bed reactors is the use of a membrane reactor (see figure 8) to axially distribute the oxygen in the reaction medium.

Figure 8: Representation of a membrane reactor with distributed oxygen feeding from Rodriguez et al. (2010) [38].

In comparison to a conventional wall-cooled fixed-bed reactor, the multi-tubular membrane reactor has significant ethylene production per tube and milder temperature profiles. This is a consequence of the low partial pressures of oxygen along the membrane reactor, which results in improved selectivity to ethylene (lower heat generation rates), and has the advantage of high heat transfer area per unit of ‘‘effective’’ catalyst volume [38]. These improvements result in higher obtained ethylene yield. Under certain operating conditions, an undesirable oxygen accumulation phenomenon inside the catalyst bed may appear which can be minimized by carefully adjusting the operating conditions. The presence of O2 at the reactor inlet significantly improves the ethylene production rates. However, if higher amounts of

O2 are fed to the reactor inlet, the selectivity could drop, which may lead to hot spots near the inlet.

On the other hand, the ODH process is mainly used due to the increased demand for propylene, while ethylene is the preferred product in this study. Because the current production of olefins is not in line with the forecasted market demand for olefins according to Cavani et al. (2007) [39]. Therefore, the research of ethane (and propane) ODH is mainly based on finding a catalyst capable of combining industrially acceptable conversion and productivity with specificity to propylene. At present, ODH is not commercialized however very extensive research activities are being carried out to develop an economically feasible process [3], [37].

2.4 Discussion

In Western Europe, naphtha is by far the most important raw material as feedstock and accounts for 74 % of ethylene production. An overview of ethylene production per region and percentage of naphtha or ethane feedstock per region can be found in table 1. Other feedstocks are less significant in Western Europe, but ethylene is also produced from gas-oil (10 %), butane (6 %), ethane (5 %), propane (4 %) and other sources (2 %) [24]. Besides naphtha cracking, ethane (from natural gas) cracking is the biggest production method for ethylene production.

17 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 1: Worldwide ethylene production by region and feedtstock from Godini (2014) [2].

Region Percentage of the world Percentage of ethane in Percentage of naphtha production capacity total feedstock in total feedstock

Asia-Pacific 30 15 70

North America 24 71 15

West Europe 17 7 74

Middle East 17 59 24

Naphtha feed has an average yield between 29-34 wt-% ethylene, depending on the severity of the operating conditions. Ethane feedstock yields around 80-84 wt-% ethylene. Besides the naphtha and ethane to ethylene yield, the crude oil to naphtha and natural gas to ethane yield should be taken into account. Crude oil distillation has an output of 10% naphtha, while the first separation of natural gas results in 1-14 wt-% ethane [4].

Comparing ethane and naphtha cracking

However, it is not completely fair to compare naphtha to ethane cracking just by comparing yield, because naphtha cracking also yields considerable amounts of other valuable by-products, high value components (HVCs). For this reason, Ren et al. uses GJ/t per high value components. They argue that ethane cracking typically yields 82 wt-% of HVCs, while naphtha typically yields 55 wt-% of HVCs. Their energy analysis results show process energy (sum of fuel, steam and electricity for reactions and subsequent processes) for ethane cracking is typical 17-21 GJ/t ethylene, and for naphtha cracking 26- 31 GJ/t ethylene. However for naphtha cracking the specific energy consumption (SEC) is typical 14-17 GJ/t HVC compared to 16-19 GJ/t HVC for ethane cracking. Ethane cracking thus has a higher SEC per HVC. Ethane and naphtha cracking have a similar energy consumption distribution, but ethane cracking has less SEC in pyrolysis (50% instead of 65%) and more energy is required in compression and difficult separations such as ethane and ethylene due to very similar boiling points. According to Ren et al. ethane cracking is not self-sufficient in terms of energy and therefore requires energy import, in contrast to naphtha cracking. The energy import accounts for 15% of the total SEC.

Feedstock prices and ethylene production prices

Besides the coke formation and energy intensive process of naphtha cracking, another reason to switch feedstock to natural gas or ethane is the fluctuating price of crude oil. Since half the price of ethylene is due to the feedstock (crude oil), the price of ethylene is highly dependent on the price of crude oil as can also be seen in figure 9. The crude oil price however dropped considerably last two years to only $45/bbl in November 2015 and even dropped to $30/bbl in January 2016 [40], while ethylene prices

18 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

per metric tonne have risen 7% up to $819/t in November 2015, according to McKinsey ethylene prices are $1200/t at the moment [41], [42]. On the other hand, the natural gas spot price was only 3.93 $/MBtu in November 2015 [42], [43]. As one barrel of crude oil is equivalent to 5.55 MBtu, this means that 1$ of natural gas is approximately good for 500,000 units of energy compared to 1$ of crude oil which is good for approximately 120,000 units of energy (in November 2015). This is more than a 400% energy content price gap [44]. Coal prices have dropped to less than $50/t and have been around $48/t from March 2015 till this day [45]. Coal-to-olefins (CTO) production process costs $640/t ethylene produced, compared to $1,185/t in case of naphtha feedstock (with a barrel price of 110$, which is more likely to be $35 as of today). On the other hand, natural gas at a price of $5/MBtu with an associated ethylene production price of $338/t [46]. It is thus obvious that price fluctuations of feedstock could have huge consequences for the production price per ton ethylene and eventually the profit margins.

Figure 9: Price correlations ethylene bulk price and Brent crude oil price 2006-2012, from Haro et al. (2013).

Availability of feedstock

Next to the price fluctuations, a constant supply of feedstock is important, for instance it is much easier to get a constant supply of crude oil because of the global supply and its transport possibilities. Natural gas on the other hand is dependent on the national supply due to transport limitations, which restricts the location of the production plant. Quantity and availability of oil and natural gas resources should be taken into consideration. Also it is convenient to integrate an ethylene plant at a crude oil refinery.

Different ethylene production technologies

Concluding, renewable feedstocks are promising and have potential however at the moment, the processes which have organic waste or carbon dioxide as feedstock are nowhere near a commercially viable production process. Another problem of these bio feedstock options is that they are not able to scale up their process to meet the required ethylene demand to be an alternative to steam cracking.

19 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

On the other hand, biomass as feedstock is growing rapidly and a competing plant in terms of capacity could perhaps be built in the next 5-10 years, if this growth continues. At the moment there is no real competition from green routes, therefore these processes will not be taken into account for in detail comparison.

Compared to OCM, so far yields up to 20-25 % have been obtained. However, application of special separation reactors has led to improved yields up to 50-60 %. At the moment energy use is also a problem, as this process is said to require twice as much as the conventional route of naphtha steam cracking and has more CO2 emission. Ethanol dehydrogenation compared to naphtha cracking ethylene production, requires a much lower temperature of 300 °C and has a high selectivity which results in a much smaller amount of unwanted by-products. The energy consumption for the dehydration varies according to the required purification grade. However, in total, the ethanol based process requires less than half of the energy consumption of a modern steam cracker to produce ethylene. Also CO2 emissions are almost 60 % lower as a result of the selectivity. The amount of CO2 produced during ethanol combustion and production has already been absorbed from the atmosphere during the growth of crop due to the photosynthesis.

In summary, from an environmental point of view bio-ethanol based ethylene production is much better than steam cracking due to lower emissions of CO2, NOx and CO, fewer by-products due to selectivity and lower production temperatures. From an economical point of view, the preliminary conclusion could be that steam cracking of ethane and naphtha is more profitable due to the very low crude oil prices at the moment. CTO could be competitive with naphtha-to-olefin industry but it is not cost competitive to natural gas or shale gas-to-olefins. As Asia mainly uses naphtha as feedstock for ethylene production, and produces more than 30% of global ethylene capacity it is strange to use such a strategy in which it forces out its own naphtha-to-olefins industry. At the same time this CTO process could be forced out by an even lower-cost North American or Middle Eastern natural gas-to-olefins industry [47]. OCM could be also be cost competitive to naphtha steam cracking due to the low natural gas prices. The next chapters will elaborate more on the potential of OCM with natural gas compared to naphtha steam cracking.

20 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3. Energy analysis and plant design naphtha cracker

Chapter 2 provided an overview of the different sections in a naphtha cracking plant in figure 7, while this chapter will provide a more detailed scientific analysis and literature study on the naphtha cracking process. The theory and collected data will be used to replicate the plant in Aspen Plus v8.6 [48]–[50] and reproduce the results from a commercial naphtha steam cracking plant.

The steam cracking of naphtha with steam is a well-known process which has been exposed to optimization operations for decades. Various research papers are dedicated to coking kinetics, modelling and simulation of the cracking in the furnace and optimal operation of the reactor. For confidentiality reasons, there are no papers which present the detailed setup of the plant with the exact operating conditions and production results. Only a few public articles have been published which have presented some results or performed themselves a quantitative energy analysis of steam cracking. However various (incomplete) documentation can be found with schematic process overviews such as in the previous chapter. Ullmann’s Encyclopaedia of Industrial Chemistry has been a good reference work for data about the general process and its different components [3].

3.1 Theory

3.1.1 Different cracking processes

As mentioned before cracking involves breaking hydrocarbons into smaller hydrocarbon molecules and introduces unsaturation, olefins. There are different types of cracking, e.g. catalytic cracking, hydrocracking, thermal cracking and steam which are performed under the influence of heat, with or without catalysts and solvents. Independent of the type of cracking, the reaction is always endothermic due to the energy required for the breaking of bonds which makes cracking in general a highly energy intensive process. The cracking processes differ in operating conditions and thus resulting in different products and by-products and thus different yields of these products. Table 2 gives an overview of the collected data on the different types of cracking.

- Thermal cracking or pyrolysis involves breaking C-C bonds under the influence of heat, without any catalysts or solvents and leaves a heavy residue. Fluid catalytic cracking (FCC), as the name already suggests, involves a fluidized zeolite catalyst which makes it possible to carry out the process under milder conditions. - FCC is performed at a lower temperature and atmospheric pressure, and produces a high yield of petrol and LPG. FCC has replaced thermal cracking almost completely over the years, because it produces more gasoline with a higher octane rating. Traditional FCC operations produce less than 2 wt-% ethylene and 6 wt-% propylene, but could be maximized for ethylene and propylene production to respectively 8 wt-% and 20 wt-% [51].

21 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

- Hydrocracking (HC) is catalytic cracking in the presence of hydrogen, which is used to break C- C bonds and for addition to aromatics and olefins in order to produce naphthenes and alkanes. Hydrocracking thus cracks big hydrocarbon chains and then hydrogenates them. This process produces high yields of jet fuel, kerosene, and diesel, but also naphtha is produced. - Deep catalytic cracking (DCC) is a fully commercialized process for producing light olefins (C3 to C5) from heavy feedstocks such as gas oils and paraffinic residuals. DCC is also a fluidized catalytic process as is FCC. DCC optimized (ideal feedstock and operating conditions) for maximized ethylene production still results in only 6.1 wt-% ethylene [51]. - Steam cracking (SC) involves cracking of typically shorter hydrocarbon chains to begin with. It is performed at very short residence times depending on severity in the presence of steam at a high temperature and atmospheric pressure. The steam is introduced to dilute the feedstock, mainly naphtha, in order to increase the feed temperature but also to reduce coking due to aromatics.

To conclude, steam cracking is compared to the other cracking methods, a more effective method for producing smaller hydrocarbons, since it already uses smaller hydrocarbons and results in the highest ethylene yields.

Table 2: Overview operating conditions and ethylene yields different cracking processes.

Thermal FCC HC DCC SC

Reactor temperature, [°C] 750-900 510-550 260-450 525-550 760-870

Reaction time, [s] 1 2 (riser) 10 2 (riser) 0.1-0.5

Reactor pressure, [MPa] 0.7 0.1 3.5-20.0 0.1—0.14 0.1

Catalyst/oil, [wt/wt] - 5-8 * 7-11 -

Steam injection, [wt-% feed] - 2-7 - 10-15 30-80

Ethylene yield, [wt-%] 28 8 15 6.1 25-32

*Value not found

3.1.2 Cracking reactions

Steam cracking involves the production of radicals due to various reactions such as decomposition, dehydrogenation, isomerisation, polymerisation, etc. In the case of cracking of petroleum distillates such as naphtha, the decomposition reaction is the most important reaction and the residence time since the yield of the coke increases with increasing residence time. Therefore a short residence time and immediate quenching of the cracked gas is required to prevent ‘choking’ of the furnace tubes [35].

Principal reactions taking place during pyrolysis, include dehydrogenation of an alkane due to cracking of a C-H bond and the cracking reaction of one or more covalent C-C bonds, see reactions (3.1-3.2):

CHCHH (3.1) n2 n 2 o n2 n  2

CHCHCH (3.2) n2 n 2 o m2 m  q2 q 2

22 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Consequently, a large number of smaller molecules is formed. The radical mechanism of naphtha cracking will be explained based on the following example of ethane cracking [3]. Which is used because naphtha cracking has a similar reaction mechanism, however it is difficult to predict due to the diversity and numerous chemical species present and varying composition. All these different components can be cracked and the resulting radicals could react with each other and thus result in a much more complex free radical mechanism with up to 400 reactions.

The first step of the radical mechanism is called initiation, which involves the breaking of C-C bonds and occurs at elevated temperatures with the formation of radicals of various molecular weights. As can be seen in the chemical formula below, an ethane molecule is split into two methyl radicals.

2 CHCH2 6 o 3 (3.3)

Radicals are unstable and are relatively short-lived. Hereafter propagation takes place in which free radicals of higher molecular weight are produced. The C-C bonds are ŵĂŝŶůLJďƌŽŬĞŶďLJɴ-scission. The formation of the new radicals depends on the position of the hydrogen atom (primary, secondary, tertiary), see formulas (3.4-6). The methyl radical reacts with ethane and forms an ethyl radical in combination with methane. The ethyl radical decomposes into ethylene and a hydrogen atom, the latter than reacts with another ethane molecule to give a hydrogen molecule and a new ethyl radical.

CHCHCHCH3 2 6 o4  2 5 (3.4)

CHCHH2 5o 2 4  (3.5)

HCHHCH2 6 o 2  2 5 (3.6)

Reactions (3.4) and (3.5) terminate if either an ethyl radical or hydrogen atom reacts with another ethyl radical or hydrogen atom by reactions as can be seen in formulas (3.7-3.12).

2 HH o 2 (3.7)

CHHCH3  o 4 (3.8)

* * HCHCH2 5 o 2 6 (3.9)

CHCHCH3 2 5 o 3 8 (3.10)

2 * CHCH2 5o 4 10 (3.11)

Another possibility is termination by disproportionation:

CHCHCHCH2 5 3 o 2 4  4 (3.12)

23 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Radical decomposition is an important reaction type because it directly produces ethylene according to formula (3.1) and (3.2). These reactions are highly endothermic, reported values of 'H at 827 °C are +144.53 kJ/mol for equation (7; with n=2) and +232.244 kJ/mol for butane decomposition into two ethylene molecules and hydrogen [3]. New methyl or ethyl radicals or a new hydrogen atom must be generated upon termination of chain propagation in order to start a new chain. Thus, every time a new chain is initiated, a molecule of ethane is formed and a molecule of ethylene is produced. A simple representation of the feed composition and resulting primary and secondary products from naphtha steam cracking can be found in table 3.

Table 3: Possible products resulting from the primary and secondary reactions in naphtha steam cracking.

Feedstock Products Products Naphtha + Steam Primary Reactions Secondary Reactions

C3-C10 Paraffins Ethylene C4 (e.g. butane, propane) Propylene C5 C5-C10 Naphthenes Methane C6 (e.g. cyclopentane) Acetylene Aromatics (e.g. BTX) Cyclo olefins (e.g. BTX) Hydrogen C7+ Etc. Etc. Etc.

Pyrolysis Model

Nowadays there are various options to model the furnace and predicting the products, besides the molecular model presented above. The main options are empirical, regression, molecular, and mechanistic models. Mechanistic computer models, which are available from various companies, are used for design, optimization and operation of modern olefin plants. These are however not implementable in Aspen Plus. Regression models are based on a data set, which can consist of historical data (which is not available due to confidentiality) or calculated data. Molecular kinetic models that use only molecular reactions and thus describe the main products as a function of feedstock consumption have been applied with some success to the cracking of simple components such as ethane, propane, and butanes [3]. Different kinetic molecular models were implemented in Aspen Plus v8.6, using molecular reactions and simplification of possible components and products. Kinetic molecular models and data from Sundaram et al., Froment et al., Renjun et al., Kumar et al., Tian et al. and Haghighi et al were used [52]–[58] for implementation. However these kinetics did not result in the same results as described in the papers, therefore a different approach (a user model) was used to model the reaction mechanism in Aspen Plus, this model will be explained in paragraph 2.

24 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.1.3 Furnace design and cracking conditions

Naphtha cracking or pyrolysis takes place in a tubular reactor (radiant tube or radiant coil) inside a steam cracker, which is a furnace with various heating sections. A simplified design of the furnace with its different heating sections is shown in figure 10.

Figure 10: Design of the furnace for naphtha cracking from Chauvel et al. (1989) [59].

The furnace exists of a convection section for preheating, radiation and multiple zones with independent heating to achieve the desired temperature profile and stack. The convection section inside the cracking furnace can exist out of multiple separate parts with heat exchangers in series, for instance one part to preheat naphtha and the second part to preheat steam to get superheated steam before mixing.

Burners are placed along the walls and/or at the bottom of the cracker, which provides the heat by combustion of fuel gas. In the convection section, the naphtha and steam are preheated and eventually mixed in the heat exchangers and further heated to the incipient cracking temperature (500-680 °C). The mixture then enters the radiant coils where, under controlled residence time, temperature profile, and partial pressure, it is heated to 760-870 °C for 0.1-0.5 s. The cracked gas leaving the radiant coils at 800-850 °C is cooled within 0.02-0.1 s to 550-650°C in the first transfer line heat exchanger (TLE) to prevent secondary reactions [3].

Since the conversion of saturated hydrocarbons to olefins is highly endothermic, high energy input rates are required. To optimize the energy input, for example a combination of floor and wall firing could offer the best distribution of the released heat over the whole firebox and an excellent radiant

25 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

efficiency. Special temperature profiles are applied along the cracking coil to avoid long residence times at low temperature, because the latter favours secondary reactions. An isothermal temperature profile is thus preferred. The number of coils required for a given ethylene capacity is determined by the radiant coil surface, which is in the range of 10-15 m2/t naphtha. The average heat flux for each coil should be maximum 85 kW/m2. Short residence time coils are typical between 10-16 m per coil, compared to 60-100 m for long residence time. Shorter coils however result in more coils to create the same surface, 4-200 coils and typical heights of up to 15 m and widths of 2-3 m are standard. The firebox length is thus influenced by the degree of severity. In various ethylene plants, multiple furnaces operate in series [3].

Ethylene yield

Amongst others the ethylene yield in different processes will be analysed in this thesis, that’s why maximizing the ethylene yield is required to compare it properly to other processes such as oxidative coupling of methane. Thus for this case high severity conditions are preferred during pyrolysis. For liquid feedstocks such as naphtha, severity is defined as the ratio of propylene to ethylene P/E (on weight basis) or as the ratio of methane to propylene M/P, in this thesis the P/E ratio will be used. For instance high severity conditions with steam ratio of 0.5, P/E of 0.45 and residence time of 0.1170 s can result in 32.38 % ethylene yield. In contrast, at low severity conditions with a steam ratio of 0.4, P/E ratio of 0.65 and residence time of 0.1096 s, an ethylene yield of 26.43 % is achieved [3].High severity conditions can thus result in 5 wt-% more ethylene yield. This increased yield is reached by a higher temperature (higher than 850 °C at the coil outlet), a highly saturated feedstock, a shorter residence time (<0.5 sec) and rapid quenching. Infinitely increasing of operating conditions is however not possible, for example the temperature is limited by the plant materials i.e. metallurgy of tubes and rapid coking tendency in the coils under these conditions.

26 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.2 Process Flow Diagram

Figure 11 shows a schematic process flow diagram (PFD) for naphtha steam cracking to produce ethylene. The process can be divided into different sections to prevent losing the overview of the entire system, namely (1) the cracking and combustion, (2) quenching, (3) compression and chilling and (4) cryogenic separation. Within these sections each step and their results will be discussed in detail in the next few paragraphs. This schematic overview, combined with information about the operating conditions from the literature study, was used to build a naphtha steam cracking plant in Aspen Plus v8.6. The PFD of this naphtha cracker plant can be found on the next page in figure 12. The different components in this PFD and its results will be discussed in the next paragraphs.

Figure 11: Schematic overview of naphtha steam cracking for ethylene production.

27 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 12: Aspen Plus PFD of naphtha steam cracker.

28 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.3 PFD section I

Figure 13 shows the first section of the naphtha steam cracking which will be discussed.

Figure 13: Aspen Plus PFD naphtha steam cracker section I.

3.3.1 Naphtha composition

Naphtha feedstock will not be produced on site but acquired from a supplier such as SABIC or ExxonMobil. As follows a probable crude oil refinery integrated into the ethylene plant will not be necessary, which will lead to a simplification at the beginning stage. Unfortunately there is little detailed data available about naphtha composition due to the varying crude oil compositions and confidentiality reasons. However after extensive searching, data was found for different naphtha (light virgin, straight run, light aromatic, medium, heavy, etc.) compositions and/or their PIONA analyses which was quite often the only available data. Also data on possible outcome compositions was collected and compared. With the results from the comparison and data from one specific coupled inlet and outlet composition, a detailed inlet composition was composed and corresponding naphtha properties were calculated according to the PIONA analysis. Again, Ullmann’s Encyclopaedia is the only literature source which provides detailed information about the outlet composition at different operating conditions in combination with PIONA analysis data from the feedstock. Unfortunately it didn’t provide a detailed component inlet composition, that’s why a composition matching this specific PIONA analysis was made. Table 4 shows the PIONA analysis data from Ullmann and the results from the composed inlet.

As can be seen, the results from the detailed inlet was composed according to the PIONA data (they both have the exact same PIONA data), and the various components within the PIONA were then altered in such a way that the hydrogen and carbon content of the naphtha were similar, with only a slight error in difference.

29 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 4: Naphtha properties (e.g. PIONA) from Ullmann and calculated from detailed inlet composition.

PIONA Unit Ullmann Calculated #1 Error #1 Calculated #2 Error #2 Light Naphtha [%] FR Naphtha [%] n-Paraffins wt-% 36.13 0 0 72.75 72.75 i-Paraffins wt-% 36.62 0 0 Olefins wt-% 0.21 0.21 0 0.21 0 Naphtenes wt-% 21.06 21.06 0 21.06 0 Aromatics wt-% 5.98 5.98 0 5.98 0 H content wt-% 15.17 15.665 3.26 15.347 1.17 C content wt-% 84.8 84.335 -0.55 84.653 -0.17 Av. mole mass g/mol 92 79.052 -14.074 91.994 -0.00695 Outlet H content wt-% 15.17 15.665 5.834E-10 15.347 -7.095E-08 Outlet C content wt-% 84.8 84.301 -0.0408 84.618 -0.0405

As it was not mentioned what kind of naphtha Ullmann used a light (#1) and full range naphtha (#2) were composed. The material balance was obtained by matching the hydrogen, carbon and sulphur content of the feedstock with that of the cracked gas. This type of analysis is time-consuming, therefore no more variations than composition 1 and 2 were made. From the average mole mass it can be deduced, that the naphtha from Ullmann is a full range naphtha. From now on, the full range naphtha (#2) will be used. In Appendix A the detailed composition of the naphtha compositions and the used Aspen Plus composition, can be found. For the Aspen composition it was chosen to use less components, first to simplify the kinetics and later on to simplify the user model.

3.3.2 Inlet conditions, mixing and preheating

Figure 13 shows section 1 of the naphtha steam cracking plant up to the low pressure steam generation. At the start naphtha is processed as a liquid feed, which is fed from a storage tank at a temperature of 25 °C and a pressure around 1 bar with a flow rate of 350,000 kg/hr in order to create a 1 Mt/y ethylene production plant. Dilution steam is required to prevent serious coking in the reactor and other equipment downstream. This would otherwise result in a short run length and shutting down the whole plant for a decoking process is of course not preferred. Dilution steam is also added to reduce the partial pressure of naphtha (the hydrocarbons) which drives the reaction of the feed towards ethylene and inhibits secondary reactions that may occur (e.g. polymerization, producing undesired by-products). Coke formation is slowed down due to the reaction with steam and forms carbon monoxide, carbon dioxide, and hydrogen in a water gas shift reaction [60]. The desired steam:naphtha ratio lies between 0.4 and 0.6, ideally it is 0.5. The dilution steam, is therefore fed with a flow rate of 175,000 kg/hr to one

30 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

part of the convection section of the cracker. In order to calculate the energy required to preheat the feedstock, first water is fed to the system at 1 bar and 25 °C and preheated to 200 °C. In order to imitate the convection and radiation section in the PFD, the steam and naphtha are mixed after preheating to 400 °C. The preheating is performed by residual heat from combustion of the fuel. Hereafter the mixture is preheated to the cracking temperature of 850 °C, the heat duty required for reaching this cracking temperature is 202.4 MW and provided by the combustion of fuel.

3.3.3 Cracking furnace and tubular reactor

As mentioned before different kinetic molecular models were implemented in Aspen Plus, however the results did not match the yields in the reference. After trying various other solutions such as pseudo components, user defined components, petroleum assays and even using different software like Aspen HYSYS with hypothetical components, it was decided to use a user model. In this used model composition 2 from table 2 is used, with the specific outlet composition. During the runs in Aspen, the inlet and outcome of the cracker is not changed. The disadvantages of the user model are for example that it is not easy to change the feed or outlet, also you cannot change reactor dimensions or other reactor specifications such as a temperature profile.

As discussed during the cracking reactions, a rapidly rising temperature profile will favour primary decomposition to ethylene over other steps leading to higher hydrocarbons and coke. The temperature in the reactor in Aspen Plus is however kept constant at 850 °C. The firebox temperature is typically between 1000-1200 °C, therefore a combustion temperature of 1100 °C has been chosen.

Cracking coils

Pyrolysis mainly takes place in the radiant section of the furnace, where tubes are normally externally heated to 750-900°C by fuel oil or gas fired burners. The fired tubular reactor can exist of radiant tubes or coils, with different geometrics (each with their own advantages). In case only wall burners are used, there would be a non-uniform heat flux. The non-uniform heat flux results in more coke and hot spots, thus floor and roof burners should be added. As the product gas leaves the coils, the outlet temperature of the reactor has to be between 800-850 °C to make sure the primary reaction takes place under optimal conditions. Then the gas mixture should be cooled down immediately to prevent secondary reactions, while at the same time generating high pressure steam for driving compressors, etc.

Coil surface and number of furnaces

In this plant, a radiant coil surface between 3500-5250 m2 is required for cracking 350,000 kg/hr of naphtha. This coil surface has a heat flux domain between 297.5 and 446.3 MW allowed for all the coils combined. The heat required in the naphtha cracker is 371 MW, which lies in between the domain

31 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

suggesting the maximum heat flux allowed per coil is not reached with ideal heat distribution. The maximum number of coils of 200 suggests, that the diameter for one coil is 0.14 m. This is the case if an average fire box is assumed with a maximum length of 15 m and 2 m of width, space between coils and walls is taken into account and the fact that a typical coil is in the range of 25-180 mm. This also takes the short run length of short residence time coils (10-16 m) into account. The area calculated for one coil results in 6.62 m2, using a simple cylinder area calculation, see equation 3.13 below:

2 A 2SS r  2 hr (3.13)

This results in a maximum area of 1325 m2 of coil surface available per furnace. This means that three or four furnaces are necessary for the corresponding coil surface of 3500-5250 m2. Literature also states that the maximum ethylene capacity per furnace ranges between 50-250 x 103 t/a, which also suggests four furnaces are required. In the model just one furnace is implemented, since the geometry of the coils are not taken into account due to the user model for the furnace and thus will not make any difference for the results. This should however be taken into account for the economic analysis.

This result however also assumes maximum capacity and full utilization of the firebox. This means tubes need to be staggered which results in non-uniform heat flux due to shadowing effects of one coil on the neighbouring coil. Fortunately tube staggering can be adjusted so that the shadowing effect is only marginally greater than with a single row arrangement [3]. Multiple furnaces used at one plant for simultaneous cracking means one could be used only for cracking the recycle stream or as spare furnace to make it possible to put a furnace offline. However this approach is very expensive.

3.3.4 Combustion

The heat necessary for the heat exchangers in the convection and radiant section is provided by combustion of fuel gas. During the cracking, methane and hydrogen are produced which thus will be used as fuel to regulate the temperature in the furnace, after separation. With the help of a calculator in Aspen Plus the required heat is calculated. The calculator imports the mass flow, mass enthalpy of the stream going in and out, and with this information the heat required for the reaction is exported. See formula 3.14 below:

Q m() Hin  H out (3.14)

Together with the results from formula 3.14, another design specification is implemented to calculate how much of the hydrogen and methane produced in the system needs to be combusted as fuel to provide this heat. The combustion takes place in a Gibbs reactor, which calculates the combustion flow rates when the possible products are provided in the reactor. The design specification performs this calculation by varying the split fraction and thus flow rate of the fuel, while keeping the temperature of the exhaust gas at 1100 °C. Another calculator is used to calculate the amount of air (nitrogen and oxygen) required for this amount of fuel being burned properly to prevent incomplete combustion or

32 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

too much fuel being used. This is calculated with a Fortran statement, which is based on the combustion reactions of hydrogen, methane and ethylene which are displayed in the reactions 3.15 to 3.17.

C 4HO2 2 o 2  CO2 2 HO (3.15)

2HOHO2 2 o 2 2 (3.16)

CHOCOHO2 21.5 2 o 2 2  2 (3.17)

The effects of excess combustion air and combustion air preheating on radiative heat transfer have been determined [61], [62]. Excess air rates as low as 5% are operated in modern furnaces and combustion air pre-heating to 270 °C, results in reduced fuel consumption. That’s why the air is preheated to 300 °C, and the fuel is preheated to 400 °C against combustion heat in the model and the amount of oxygen in the exhaust gas is kept at 5 %. Higher preheating (up to 500-600 °C) would result in increased thermal efficiency.

Results combustion

These calculators and design specifications, result in 0.80426 split fraction which means 80.43% of the methane and hydrogen produced is required for providing the heat duty necessary for the cracking. This split fraction accounts for 1256 kg/hr H2 and 46035 kg/hr CH4 needed to be burned. Due to the separation also some ethylene is present in the stream (712 kg/hr), which is taken into account for the heat and air fed to the combustion. For this amount of fuel 1,094,200 kg/hr air needs to be fed for complete combustion. The heat duty left in the exhaust gas is used for preheating the air to 300 °C, preheating the steam to 200 °C and preheating the mixture and fuel to 400 °C. The temperature of the exhaust gas after preheating is 137.2 °C. Due to combustion 128,798 kg/hr CO2 and 115,531 kg/hr H2O are produced. The residual fuel will be burned in another combustion reactor together with the recycle stream further on in the process for additional power production, this will be explained in the last section.

3.3.5 Heat requirement, heat efficiency and heat of reaction

Heat requirement for cracking is determined by feedstock and cracking conditions and can be divided into three parts [3]:

1. The enthalpy required to heat the feedstock, including the latent heat of vaporization of liquids 2. The endothermic heat of cracking or heat of reaction 3. The enthalpy required to heat the cracked gas from the radiant coil inlet temperature to the radiant coil outlet temperature.

The first is accomplished in the convection section of the furnace, the other two in the radiant section. The enthalpies for the first and third part can be calculated from standard heat-capacity data. A typical energy balance for a modern naphtha furnace can be found in table 4 [3]:

33 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 4: Energy balance in cracker.

Part Function Energy Required [%] Radiant Reaction 30 Process 12 Losses 1.2 Convection Process 35.5 HPSS 16 Losses 0.3 Stack Losses 5 Fired duty 100%

The losses are for example losses due to the wall or radiation loss through the crossing of steel shells from the radiant and convection sections, or coking. Energy is saved in the convection section by preheating the feedstock, preheating the combustion air and dilution steam preheating. Stack heat loss is the portion of heat released to the atmosphere in the flue gas and is sensitive to excess combustion air in the furnace and thus fuel composition, and air leakage. Process heat duty is the energy required to heat the feedstock and dilution steam from the temperature at which they enter the convection section till the temperature at which they leave the radiant coil. Overall energy efficiencies of up to 94- 95 % are typical for Linde furnaces.

Results heat of reaction

The endothermic heat of cracking is calculated from heats of formation by the following equation:

'HHHr ' p  ' f (3.18)

The heat duty required for cracking and heating it to 850 °C together is 372.22 MW. The heat duty for preheating is 202.44 MW or 53.90 kJ/mol feed. The heat of reaction requires 169.98 MW, which is 160.54 kJ/mol of naphtha. The same value was manually calculated with individual heats of formation for each products and reactant. This amount is the second heat duty from the list. The enthalpy required to heat the cracked gas from the radiant coil inlet temperature to the radiant coil outlet, cannot be calculated in Aspen, as the temperature in the reactor is kept at 850 °C.

34 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.4 PFD section II

The following steps can be seen in figure 14 and consist of the removal of heat contained in the cracked gas while producing high pressure (HPS) and medium pressure steam (MPS), oil quenching, water quenching and flashing.

Figure 14: Aspen Plus PFD naphtha steam cracker section II.

3.4.1 Transfer Line Heat Exchangers and turbines

The first step after cracking, is cooling of the cracked gas. The temperature of the cracked gas leaving the radiant coils is 850 °C. The cooling down is done rapidly in multiple steps, the mixture is cooled to 305 °C in the first transfer line heat exchangers (TLEs), the second TLE cools down to 230 °C. The TLE’s are modelled as shell and tube, but could also be modelled as linear tubes, and are coupled to the radiant coil outlet. In the first TLE HPS is produced at 80 bar, by first feeding pumped water (80 bar) just below saturation temperature (290 °C) to the TLE, which becomes HPS in the TLE. In the second TLE, MPS is produced by first feeding pumped water (10 bar) just below saturation temperature (175 °C) to the TLE, which becomes MPS in the TLE. Both of these streams are coupled to a turbine in order to produce electricity to provide electricity for the compressors and pumps in the process. The high pressure stream is fed to the first turbine together with the high pressure steam produced further on in the process (from residual fuel and recycle stream combustion). The outlet stream from the first turbine is then mixed together with the low pressure steam, and expansion of those two streams takes place in the second turbine at 0.08 bar.

Results steam generation and power production

The results from the steam and power generation can be found in table 5. The expansion of the HPS produced during the combustion of the residual fuel and recycle combustion, results in an additional 124.05 MW. This results in a total of 255.30 MW of power production together with the electricity from the HPS and MPS from the TLEs.

35 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 5: Transfer line heat exchanger and turbine results.

Unit TLE 1 TLE 2

Aspen component - HEATX1 HEATX2

Temperature hot stream in °C 850 305

Temperature hot stream out °C 305 230

Temperature cold stream in °C 290 175

Temperature cold stream out °C 496.4 278.4

Pressure steam bar 80 10

Heat duty exchanged MW 228.63 25.25

Preheating heat duty MW 148.91 7.58

Pump duty MW 1.06 0.01

Unit Turbine 1 Turbine 2

Steam flow rate HPS and MPS kg/h 403,200 39,240

Total steam flow rate kg/h 807,480 846,720

Power production without MW 51.27 79.98 combustion steam cycle

Total power production MW 102.59 152.71

Power production per kg of total kWh/kg steam

The second TLE should not cool below the boiling point of the heaviest hydrocarbons present, because this will cause severe fouling in the TLE due to condensation of the hydrocarbons.

3.4.2 Primary fractionation and quenching

After this rapid cooling stage, the cracked gas is direct-quenched with oil from the bottom of the primary fractionator, and is passed along to a fractionator. In the primary fractionator, gasoline and fuel (with a boiling point around 200 °) are condensed and fractionated. While this liquid fraction is extracted, the gaseous fraction goes to the overhead.

Results primary fractionator

The bottom temperature of the primary fractionator is 200.4 °C and contains 84.9 % of all the C9+. Cooling by contacting the cracked gas with recirculating oil from the bottom, results in a temperature decrease of 13.4 °C to 216.6 °C. The temperature of the overhead condensate is 36.9 °C and contains

36 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

all the lighter components. The cracked gas enters the column typically at 230 °C together with the dilution steam. However cooling down to 100 °C and removing of the dilution steam was required for the primary fractionator to function properly. Oil firing in the radiant burners of the furnace offers the opportunity to use a fraction of the pyrolysis fuel oil from the primary fractionator (i.e. the liquid product of the pyrolysis boiling above 180 °C) as fuel. This means the higher the amount of methane in the residual fuel, which could be exported to other consumers or be used for extra power generation.

Water quenching

Final cooling is accomplished by a direct-water quench tower, cracked gas is rapidly quenched with 100,000 kg/hr of water in which the cracked gas is cooled to near ambient temperature (42.8 °C) to stop the degradation of the olefins and coke formation. After gravity separation due to condensation of steam and fuel gasoline, heat could be extracted from the bottom in a circulating loop. Instead of performing gravity separation at two moments, only flash is performed at this point to evaporate the cracked gas to the top and 93 % of all water going to the bottom. Most of the excess water stream is used to produce dilution steam, the heat of this low temperature stream is typically consumed by the reboiler of the propene-propane fractionation tower.

3.5 PFD section III

In the third section, compression, acid gas removal, drying cooling of the cracked gas and gas oil stripping takes place, see figure 15.

Figure 15: Aspen Plus PFD naphtha steam cracker section III.

37 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.5.1 Compression stages

Cracked gas is compressed and cooled using multistage centrifugal compressors with interstage cooling. The purpose of the multiple compression stages and condensation (which takes place during the compression and cooling), is preparing the cracked gas for the specific cryogenic fractionation of the liquefied gases.

Result multistage compression

The gaseous fraction is passed through five stages of gas compression with temperatures at 30 °C. Eventually compression to 31.36 bar takes place, while some condensation of water and hydrocarbons occurs during the compression [63]. Pressure drops are assumed to be 2%. After each compression stage liquefied components are removed to prevent freezing, and the formation of hydrates which blocks the pipes and/or damages the equipment. The results from the compressor can be found in the table 6 below, as can be seen total cooling duty requires 54.24 MW and total work requires 40.79 MW.

Table 6: Multistage centrifugal compressor with interstage cooling results.

Unit Stage 1 Stage 2 Stage 3 Stage 4 Stage 5

Temperature °C 30 30 30 30 30

Temperature profile °C 90.5 78.5 81.1 82.6 84.8 (before cooling)

Pressure bar 2 3.92 7.84 15.68 31.36

Vapor fraction - 0.9787 0.9865 0.9842 0.9834 0.9772

Liquid knockout kg/hr 23,980 13,302 13,582 11,983 -

Work duty MW 8.84 8.26 8.30 7.96 7.43

Cooling duty MW 13.44 9.65 10.04 10.13 10.98

3.5.3 Gravity separation and gasoline stripping

In liquid cracking most of the gasoline fraction containing the C6 to C8 aromatics is condensed in the interstage coolers of the compressor. After gravity separation from process water (performed by a shortcut model separation block), this gasoline stream is fractionated in a stripper to remove C4 and lighter components before routing it to a hydrotreating step. The stripper is a distillation column without condenser in which the liquid feed is fed at the top tray and the liquid’s own vapour can be used to do the stripping. By contacting the liquid with gas, C4 and lighter components are removed from the liquid because they will be transferred to the stripping gas at the top. 71.02 % of the water and lighter components than C5 condensate with a 74.6 % purity, are routed to the first compressor stage.

38 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

96.4 % of the raw pyrolysis gasoline and the C5+ with a purity of 95.9 % are mixed with the bottom of the debutanizer. This stream could be fractionated into a gasoline and BTX part, of which the first could be sold as motor gasoline component after hydrotreatment, or it could be used as feedstock for aromatics production. Results and operating conditions from the column can be found in table 7.

3.5.2 Acid gas removal and drying

After the fourth stage of compression, the removal of acid gases, carbon dioxide and hydrogen sulphide is performed. Dimethyl disulphide is added at the cracking stage to reduce coking, which the source of sulphur in the system. The sulphur is removed with a shortcut model. The sulphur is typically removed in the form of hydrogen sulphide via caustic wash. Caustic wash can be performed with low pressure wet air oxidation or washing and steam stripping. Caustic wash or the regenerative solvent (MEA) for scrubbing can be used to remove carbon dioxide and hydrogen. Carbon dioxide needs to be removed because it can freeze at low temperature in the separation part and could also be absorbed into ethylene, it is also required to be able to meet the product specification for polymer-grade ethylene which needs a purity of 99.9 wt-%.

Drying

After acid gases have been removed, drying of the product gases is required to prevent ice formation and blocking of the pipelines in cryogenic separation equipment. Drying is accomplished by chilling with propane or propylene refrigeration to remove most of the water, this could be performed by passing gases through a tower for drying by absorption of water on a solid absorbent (e.g. activated alumina). In this case the drier exists of two bed with pressure swing adsorption (PSA). However multiple adsorption beds are needed for continuous water removal [64]. Drying can also be performed by molecular sieve drying. Molecular sieve drying system exists of a vacuum and complicated valve system. An upstream temperature of 15 °C is preferred for optimal separation of water during drying. After the last compression stage there is a pressure of almost 32 bar, which allows smaller dryer volume with lower adsorbent cost and less water removal. The drying in this model is performed via a shortcut method, removing 100 % of the water.

3.6 PFD section IV

This section is essentially a cryogenic separation process through distillation and refrigeration Cryogenic which is the predominant method for cracked gas separation, see figure 16. After cryogenic cooling, products are passed to a de-methanizer, a de-ethanizer, de-propanizer, de-butanizer, splitters and hydrogenation steps to get the final product streams. The compressed, condensed gas from the cold section can be fed to one of the three following (mainly-used) process sequences:

39 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

1. De-methanizer first with tail-end hydrogenation 2. De-ethanizer first with front-end hydrogenation 3. De-propanizer with front-end hydrogenation

These processes differ mainly in hydrotreating steps downstream or upstream and sequence of separation. However a comparison of all routes has shown that the first option has the best energy efficiency and is thus used in the model [3].

Figure 16: Aspen Plus PFD naphtha steam cracker section IV.

3.6.1 Cooling and hydrogen removal

The processed gas from the drying enters the cryogenic distillation at 32 bar and is typically chilled in three stages using refrigerants and a hydrogen/tail-gas stream from the process. CH4, CO and H2 are tail gas components which are separated in the flash at very low temperature of -170 °C and the de- methanizer in the separation train. These components can be used for flaring, fuel gas or as tail gas. In naphtha cracking it is important that hydrogen is obtained at a high purity because it is used in hydrogenation of acetylene and hydrogenation of methyl acetylene. Any excess hydrogen is used as fuel in the plant. At this temperature C2+ components are entirely removed by partial condensation.

Results cooling and flash

In the model, the cooling from 70 to -170 °C requires 67.60 MW. In the flash 85.3 wt-% of the H2 is removed together with almost all the CO and some CH4, resulting in a purity of 98.6 % H2, which is even higher than typical in literature (80-95 %). Both scenarios of using imported hydrogen or the high purity hydrogen from the flash on site, and the consequences of these scenarios for the power production were modelled. These results can be found in paragraph 3.6.6 Acetylene hydrogenation.

40 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 7: Results and operating conditions columns.

Unit Gasoline De-methanizer De-ethanizer De-propanizer De-butanizer C2 Splitter C3 Splitter stripper Number of stages - 2 50 50 55 50 120 200 Pressure bar 1.96 32 25 14 2.7 18 17 Temperature condenser °C 56.5 -94.5 -16.8 36.7 34.1 -32.7 41.4 Temperature reboiler °C 95.5 10.3 80.5 109.3 107.8 -7.7 57.8 Cooling duty condenser MW - -26.91 -7.69 -4.44 -0.62 -30.00 -54.89 Heat duty reboiler MW 2.66 14.51 20.34 8.29 3.84 33.84 59.33 Net heating(+)/cooling (-) MW 2.66 -12.40 12.68 3.85 3.22 3.84 4.44 duty Distillate to feed ratio - 0.120 0.209 0.583 0.609 0.782 0.884 0.904 Distillate rate kg/hr 7,484 58,416 129,139 56,322 28,278 114,568 51,253 Reflux ratio* - 9.03 6 0.7 0.85 0.2 2.8 12.8 Boilup ratio‡ - 0.23 0.80 2.80 2.61 4.28 23.59 133.50 *Reflux ratio is defined as the ratio between the reflux rate and distillate rate, mass based. ‡Boilup is defined as the bottoms stage vapor flow, excluding any vapor product. The ratio is boilup rate divided by bottoms rate, mass based.

The estimations for each of the distillation columns were calculated by using the Fenske Underwood Gillian method, see Appendix B for the equations used in this calculation method. In order to use this method the Antoine coefficient for each component for different temperature ranges were collected together with the property data. Due to the large amount of data, it was decided to leave these result out.

41 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.6.2 De-methanizer

Results and operating conditions from all the distillation columns can be found in table 7. Liquefied gases from the hydrogen flash are then pumped and heated to the feed temperature necessary for the separation of methane and the residual hydrogen in the de-methanizer distillation column at a somewhat lower temperature of -94.5°C. 46.18 MW of cooling duty can be recovered from this stream. The overhead stream consists of 95.9 mole-% methane and is used as a fuel for the cracking furnace. Methane and hydrogen are first expanded in a turbine to recover cooling duty. After de-methanization, all the hydrogen and methane are removed, however also some (<1%) of the ethylene is condensed during this distillation step. The bottom of the de-methanizer is passed to the de-ethanizer.

A challenge is provision of the chilling duty for individual cooling steps. Temperatures below -100 are achieved by methane refrigerant cycle, besides an ethylene and propylene recycle. Or generation of a light recycle stream, i.e. vaporizing condensed cracked gas fractions and recycling the vapors to the cracked gas compressor.

3.6.3 De-ethanizer

The de-ethanizer distillation column separates ethane and lighter components as liquefied overhead products from C3+ bottom components (e.g. propane, propylene, butane, butylenes, etc.). In the model the overhead temperature is -16.8 °C at 25 bar in which 100 % of all the C2s are going to the top, the composition of the overhead is as follows: 87.1 wt-% C2H4, 3.1 wt-% C2H2 and 8.1 wt-% C2H6 and some contamination by propylene (1.7 wt-%).

3.6.4 De-propanizer

The feed to this distillation column originates from the bottom of the de-ethanizer column, and separates propane and lighter fractions as overhead from C4+ fractions as bottom components (e.g. butanes, butenes, butadienes, and heavier components). 99.87 % of the C3 fraction goes to the overhead, which contains 86.3 wt-% C3H6, 2.00 wt-% C3H8, 8.9 wt-% C3H4 and some contamination by

C4H8 (2.7 wt-%) and traces of other components.

3.6.5 De-butanizer

The components entering the de-butanizer which come from the de-propanizer column, are the heaviest hydrocarbons left of the stream entering the separation train. These heavy components are always separated at the end. The C4+ components are typically separated at 36.7 °C in the model and 14

42 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

bar as overhead and C5+ fractions as bottom. The heavy components from the bottom stream from the pyrolysis gasoline stripper is an unstable stream and cannot be stored, as the unsaturated products react further, forming polymers and gum. Depending on the downstream processes a part of pyrolysis gasoline is selectively hydrogenated to olefinic compounds and fractionated into a C5 cut, a C6-C8 heart cut and a C9+ cut. The C9+ cut can be sold or blended as a pyrolysis gasoline component. Aromatics of various fractions can be recovered or they remain in the hydro-treated pyrolysis gasoline. The C4 fraction can be refined for butadiene, butene, isobutene, or mixtures thereof. If no C4s are desired C4 dienes and olefins are typically hydrogenated and recycled to cracking. C3 and C4 hydrogenation units consists typically of adiabatic beds or tubular reactors, integrated in a system of coolers, recycle pumps and separation drums. 99.17 wt-% of the C4 cut is separated to the overhead and contains 2.1 wt-%

C4H10, 3.7 wt-% C4H4, 60 wt-% C4H6 and 34 wt-% C4H8.

3.6.6 Acetylene hydrogenation

Acetylene is removed due to purity requirements for PE production. Removal of acetylene is usually accomplished by selective, catalytic hydrogenation of acetylene over a noble-metal catalyst (e.g. palladium, nickel/cobalt/palladium) in a packed-bed reactor operating at around 25-100 °C. Acetylene separation can also be performed via extractive distillation in an amine absorber by using DEA. Pressure ranges typically from 20-35 bar. Acetylene hydrogenation takes place in the gas phase of a pure C2 fraction. Hydrogenation reactors consist of one or two adiabatic beds, hydrogen is added to the reactor at a molar H2:C2H2 ratio of ca. 1.5, resulting in an ethylene and ethane gain of up to 50 %. Hydrogen and small amounts of methane and carbon monoxide introduced after demethanization must be removed by additional fractionation. The following reaction is introduced in an RSTOIC block in the model, at a temperature of 60 °C and 20 bar. It is assumed that 99.9 % of the acetylene is converted. 2CHHCHCH 3 2 2 2 o 2 4  2 6 (3.19)

Hydrogen removal

In front-end hydrogenation, no hydrogen removal after acetylene hydrogenation is necessary because it will all be separated before or during the de-methanizer. On the other hand it has disadvantages because other olefins get hydrogenated which means loss of significant by-products. Sometimes purging the light ends from the overhead ethylene fractionator condensate is possible, but additional multistage fractionation is usually required in the ethylene fractionation or in a small separate de- methanizer. However this is not the case, because hydrogen is not added in excess in this model. With the help of a simple fortran script and a calculator the amount of hydrogen required for the reaction is calculated. This is done by importing the mole flow of acetylene and exporting the value of hydrogen with the help of reaction 3.19. Besides hydrogen, the product stream typically contains around 1000 ppm of CO. CO is a poison for all hydrogenation processes, however all the CO is removed during the flash.

43 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.6.7 Ethane- ethylene separation

The feed for the C2 splitter originates from the top of de-ethanizer and separated into ethylene as the top product and ethane as the bottom product. This distillation column requires a high reflux ratio 2.8 (can be up to 4 according to Ullmann et al.) and 120 separation stages. Two separation processes are possible; high pressure fractionation and heat pumped C2 fractionation. The first one operates at a pressure of 17-28 bar, while low pressure propylene refrigerant is vaporized in the top condenser providing chilling duty. In the second process, the system operates at 8 bar and part of the compressed ethylene vapour is cooled and routed to the reboiler providing the heat duty for the reboiler. The latter has a few advantages over the first one, with respect to investment for equipment and energy consumption. Less equipment is involved, lower reflux ratio and less trays required for ethylene and ethane separation due to the better relative volatility of ethylene and ethane at lower pressure. However this process is not feasible in combination with tail-end hydrogenation which is used in this model due to the presence of light ends in the feed stream. In this model a distillation column operating at 18 bar with 120 stages and a reflux ratio of 2.8 results in 99.93 wt-% of all the ethylene going to the top. Eventually an ethylene yield (wt/wt naphtha) of 32.74 % is reached with a loss of 0.84 wt-% of ethylene during the whole separation train.

Table 8: Typical ethylene specifications (in ppm vol unless otherwise noted) and specifications of the ethylene from the model.

Component Polymer grade Model

Ethylene 99.9 % min. 99.94 %

Methane 300 0.7

Ethane 500 507.7

Propylene 10-15 Trace*

Acetylene 2 34.3

Hydrogen 10 56.2

Carbon monoxide 2 0

Carbon dioxide 2 0

Oxygen 5 0

Water 2 0

MeOH 5 -

Sulphur as H2S 2 (wt-%) Trace

Other polar compounds - Trace

DMF 2 (wt-%) -

*Trace implies there is a mole flow but the amount is less than 0.1 ppb.

44 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Polymer grade ethylene

Ethylene could be sent to storage, however it is typically directly used on site in a polyethylene (PE) synthesis plant. In order to use ethylene for production of PE, polymer grade ethylene must be produced which requires a purity of 99.9 mole-% of ethylene and some specifications on the contaminations. See table 8 for typical ethylene specifications compared with the results in the model. Sulphur, oxygen, acetylene, hydrogen, carbon monoxide, carbon dioxide are typically the biggest impurities, and are difficult to control when ethylene from multiple sources is mixed in transportation. As can be seen in the results, polymer grade ethylene is produced. However it can also be seen that too much acetylene, hydrogen and methane is present. Build-up of these components, especially due to acetylene could result in problems for PE production. This could be solved by increasing the column specifications by increasing the reflux ratio, which is at the moment lower than typical for an ethane ethylene splitter.

3.6.8 Methyl acetylene hydrogenation

The C3 fraction coming from the top of the de-propanizer typically contains 2-6 % of methyl acetylene and propadiene (MAPD). For purification of propene for producing polymer grade propylene (PP), and for economic reasons MAPD is hydrogenated to propene and propane. Modern hydrogenation typically yields a 60 % gain of propene and ca. 40 % of propane, however 50 % of each was assumed. The same calculator used in the acetylene hydrogenation was used for this RSTOIC reactor according to reaction 3.20. A fractional conversion of 99.9 % of methyl acetylene was assumed. 2CHHCHCH 3 3 4 2 o 3 6  3 8 (3.20)

Hydrogen for hydrogenation vs hydrogen for fuel combustion

Two processes are in commercial use in ethylene plants to purify hydrogen: (1) methanation of CO in a catalytic process, converting CO to methane and water or (2) hydrogen purification by adsorption in a pressure-swing adsorption unit. In total 843 kg/hr of hydrogen is required for both hydrogenations. The overhead of the flash contains 1331 kg/hr of hydrogen with 98.6 vol-% and 457 ppm vol CO. This is high purity hydrogen which could be used for hydrogenation of acetylene and methyl acetylene. Due to the large amount of residual fuel, the hydrogen could be used for the hydrogenation in both reactors and there would still be enough additional fuel left. On the other hand, due to the high purity of the hydrogen it is probably even economically better to sell it. With implementing a splitter in the hydrogen flow to the fuel combustion, the reduction in power from additional fuel was calculated. The amount of power decrease due to hydrogen use for hydrogenation, amounts to 9.64 MW. However raw material costs from buying pure hydrogen are then not necessary.

45 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.6.9 Propane- Propylene Separation

After compressing the product stream coming from the hydrogenation of methylacetylene, the propylene is separated from propane in the C3 splitter. Propylene is recovered as top product and propane as bottom in this splitter. Typically 150-230 stages and a reflux ratio of 20 is needed in case of polymer grade propylene, which requires a purity of 99.5 % instead of 93-95 % purity which is used for a chemical-grade product. If no waste heat for reboiling is available, or when upstream cooling water temperature is higher than 35 °C a heat-pumped propylene fractionator is applied, which operates at 0.8-10 bar. In a heat-pumped fractionator the overhead gases are compressed and condensed in the reboiler. However in the model the first type of fractionator was used, operating at 17 bar with 200 stages and a reflux of 12.8. At these operating conditions polymer grade propylene is produced, 99.51 % of the propylene goes to the top which has a purity of 99.53 % and has 0.45 mole-% of propane contamination, 155 ppm vol of H2 and 4 ppm of C2H6. MAPD concentration at the reactor outlet is typically 500-1000 ppm, when the propene product has polymer grade quality, however there is only a trace present in the model. Eventually a propylene yield (wt/wt naphtha) of 14.57 % was reached with a product loss of 4.78 wt-% due to separation processes.

3.7 Recycle, combustion and power generation

Still a few parts of the PFD are not discussed, such as the combustion of the residual fuel and recycle and power generation of the turbines from the steam production. Also power generation from cooling duty, namely the expanding of the methane and hydrogen fuel gas. See figure 17 for the combustion of the residual fuel and the recycle stream from the separation train.

Figure 17: Combustion of the recycle and residual fuel.

46 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.7.1 Recycle

Normally ethane in the cracked gas is separated and recycled to an ethane-cracking furnace, in which the operating conditions are optimal for ethane cracking. In case the ethane cracker does not have sufficient capacity, co-cracking of ethane in the naphtha cracker takes place with ethane conversion of 40-50 %. Due to build-up of large recycle streams, co-cracking should be avoided wherever possible. Due to the small amount of the recycle; 20,482 kg/hr which is 5.85 wt-% of the feedstock, this recycle will not be sent to the cracker but this stream will be combusted. This recycle consists of 71.1 mole-%

C2H6, 13.9 mole-% C3H8 and 9.8 mole-% C3H6.

Figure 18: Expanding of hydrogen and methane for power production and residual fuel combustion.

3.7.2 Expanders

Methane (see figure 18) is not employed as a refrigeration loop, but as an expander and generates 2.42 MW of additional power. Expanding the top stream of the hydrogen flash results in 0.28 MW power. It is assumed that ethylene will be put in storage, since the conditions of ethylene from the overhead of the C2 splitter are at the right conditions for storage, this stream will not be expanded. If expanding would be performed, this which would result in an additional 4.68 MW of electricity.

3.7.3 Additional combustion

The amount of air is again calculated with a calculator by importing all the cracked compounds and exporting the amount of oxygen and nitrogen to combust them in the right stoichiometric amounts. The Fortran statement in the calculator is as follows:

EXC=1

O2ST1= 2*CH4+0.5*H2+3.5*C2H6+3*C2H4+2.5*C2H2

O2ST2= 4.5*C3H6+5*C3H8+5*C4H4+5.5*C4H6+6*C4H8+6.5*C4H10

47 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

O2=O2ST1*EXC+O2ST2*EXC

N2=O2/0.21*0.79 (3.21)

The air flow rate is calculated with taking an excess of air into account to produce around 5% of oxygen in the exhaust combustion gas, as explained before to have complete combustion and no fuel wasted. The results from the combustion can be found in table 5.

3.7.4 Refrigeration cycles

Typically an ethylene and propylene refrigeration cycle are employed. Each of which generates chilling duty at two to five different temperature levels. Ethylene and propylene refrigeration systems can be operated at low temperatures within -10 and -150°C for cooling and pressures with 15-30 bar for compression. The following papers could be reviewed for detailed operating conditions on the refrigeration cycles; Zdonik et al. (1970), Grantan et al. (1987) and van Geem et al. (2008).

3.8 Emissions and waste

The water on the plant result from quench water, dilution steam, decoking water, and flare water discharges. Streams from the caustic scrubbing section require chemical treatment (oxidation) before discharge to the wastewater unit. Examples of process wastes are coke, cleaning acids, waste oils, sulfuric acid from cooling towers, catalysts for hydrogenation, amine tars, etc. Flue gas emissions due to combustion exist of 8.5 ppm vol NO and 0.6 ppm NO2, even with the tight regulations on NOx emissions (<50 ppm) no NOx treatment procedures have to be performed.

In total 221,210 kg/hr of carbon dioxide is emitted from flue gas emissions due to both combustion processes. For naphtha feedstock a typical emission ratio of 1.8-2.0 for tonnes of CO2 produced per tonnes of ethylene. These emissions are based on Ren et al. (2006) and are the result of fuel combustion and utilities, both of which use fossil fuel. In the model a ratio of 1.92 is reached from combustion of all fuel produced on site and combustion of the recycle. Combustion of utilities is not taken into account, however the additional fuel combustion results in the energy generation for the different utilities.

3.9 Coke deposition and storage

An accurate coke deposition calculation is required for calculating the amount of coke deposited in the equipment and produced during cracking. However this needs a complete free-radical mechanism that would include all major coke producing steps. A mathematical model could be used such as the Masoumi model/Linde software or the Kumar and Kunzru coke depositing model. This is however not implementable in Aspen Plus.

48 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Coking and fouling results in poor flow distribution in the entry cone of the coils and TLE (it blocks tubes), at tube sheet it causes eddies and backmixing. Coke also increases the pressure drop along the radiant coil, cracking yield is thus lowered because of increasing hydrocarbon partial pressure. A steam- air mixture is introduced to burn out the coke at 880-900°C or even up to 1100°C, while the air concentration is increased carefully to avoid overheating the radiant coil. Total decoking of radiant coil and TLE takes circa 36 hours (or even 1-4 days) for a naphtha steam cracker. Before decoking the furnace has to be shut down. Thus, a cracker furnace operates cyclically between the cracking and decoking operations. Excess steam may also convert some of the hydrocarbons or naphtha components to carbon monoxide and hydrogen and reduce the yield of olefins. Depending on the feedstocks, coil configuration and severity, decoking for steam cracking furnaces is required every 14-100 days on an average. Naphtha furnaces are typically decoked every 15-40 days. The maximum cycle time is around 60-100 days. During the course of one on-stream cycle, deposited coke can reduce the heat-transfer efficiency by 1-2 %, resulting in a 5 % fuel consumption increase [3].

The tendency for coke build-up is greater with heavier feedstocks, in the case of light naphtha a ratio of 0.5 is more standard. 0.01 wt-% of the feed becomes coke deposition (reducing heat transfer plus increasing pressure drop), which would thus result in 35 kg/hr of coke. Different developments in cracking coils have been offered to the industry to reduce the coke deposit, resulting from catalytic coking (forms on the contacted surfaces) or thermal coking (forms within the gas phase). One approach is coating of the coil surface with inert materials to minimize the amount of catalytic coking. Another is chemical treatment of the coils during or prior to startup, a disadvantage is that the chemicals introduced end up in the cracker products and could influence the quality of the products. However these technologies are rarely used in large-scale plants, since no major breakthrough has been achieved with the use of the solutions. Another method is the introduction of sulphur in the feed inhibits coking, however a sulphur content above 400 ppm may increase coking. In the model 40 ppm DMS is added.

Ethylene Storage

Ethylene can be stored in two ways, semi-refrigerated storage or atmospheric storage tanks. The first one stores at 15 bar and -40 °C, using spherical containers of a size up to 3000 m3. And the second method used atmospheric pressure and -103 °C in tanks with a volume range of 10,000-20,000 m3. A few pipeline systems are available for transportation, which transports under a pressure of 4-100 MPa with a temperature above 4 °C to prevent liquefaction of ethylene. However most ethylene is consumed locally, requiring little storage and transportation.

49 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.10 Energy analysis

Currently, ethylene manufacture is a high efficiency process, in many cases all energy required is generated on site and often excess energy is exported for use in various operations. For instance, the energy necessary to heat up the feed in the furnace usually comes from burning the hydrogen and methane which are produced together with the ethylene. With ethane as feedstock, small amounts of energy need to be imported, on contrary of ethane for other feedstocks, particularly propane and naphtha, the process is energy-balanced or provides export energy [60]. New plants are designed for 93-95 % thermal efficiency, and revamping of older ones can increase energy efficiencies up to 89-92 % [3]. Radiant efficiencies are in the range of 38-42 % for side wall firing, 40-45 % for a combination of sidewall and floor firing and 42-47 % for only floor firing.

3.10.1 Work requirement and power production

Table 9: Results from power production or requirement per block in model.

Block Stream Power production (+) or required (-) in MW

Turbine 1 W1 +102.59

Multi compressor W2 -40.79

Pump 3 W4 -0.01

Compressor 2 W5 -0.50

Pump 2 W6 -0.01

Pump 1 W7 -1.06

Turbine 2 W10 +152.71

Expand 2 W13 +2.43

Expand 1 W14 +0.28

Pump 4 W18 -1.07

Expand 3 W19 +4.68

Total +219.17

Table 9 on the previous page shows the results from the model considering the work required and power produced. As can be seen the highest power requirement results from the multistage compressor, which is typically the case. In the end an additional amount of 219.17 MW of electricity could be exported or used for other equipment that are not taken into account. Table 10 shows the energy analysis results of the full range naphtha from the model to literature values. The last column is

50 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

compensated for the fact that the literature values are for a 1.09 Mt/y of ethylene plant while the capacity of the plant in the model is equal to a 1.00 Mt/y of ethylene. Therefore a linearized relation was assumed for simplification.

Table 10: Typical consumption figures of a standard-size ethylene plant (1.1 Mt ethylene per year) based on Ren et al. (2006) compared to model results.

Feedstock Unit Ethane FR FR FR (lit.) Naphtha Naphtha Naphtha (lit.) (model) (model) x 1.09

Ethylene/propylene product t/h 125/0 125/55 115/51 125/56 rate, [t/h]

Fired duty of cracking furnace MW 613 772 372 406

Fired duty residual fuel and MW - - 378 412 recycle

Fuel gas import (+) export (-) MW 28 -142 -124 -135

SHP steam production of the t/h 450 at 6.5 458 at 403 at 8 439 at 8 furnaces MPa 11.5 MPa MPa MPa

HP steam import (+) export (-), t/h -72 Balanced -532 at 8 -580 at 8 at 4.6 MPa MPa MPa

Electric power kWel 3.5 7.5 2150 2343

Circulating cooling water m3/h 28,000 47,000 28,000* 47,000* temperature rise 10 °C

Total compressor power MW 72 85 41 45

Specific energy consumption kJ/kgethylene 16,000 23,000 20,880 22,759

*Not calculated, same values as literature was assumed.

The fired duty from the cracking and recycle combined (818 MW) is a bit higher. Due to the combustion steam cycle the steam export is quite high in the model. No water from acid gas removal was taking into account or decoking water, flare water discharger or additional water during compression. If these amounts would be taken into account, the system export and import should probably be almost balanced. The lower compressor power value can be explained by the lack of the compressors which are usually present in the ethylene and propylene refrigerant cycle. These compressors are not present in this model, since it was decided to perform a pinch analysis on all the streams instead of a complete optimized heat exchanger network. Circulating cooling water temperature rise is not calculated in the model, so the same amount was taken into account for the cooling water as the electric power in order to calculate the specific energy consumption. This sums up to 22,800 kJ/kg ethylene which is according to literature values for the same size plant.

51 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Energy consumption different sections

According to these values (more general) observations about the energy analysis can be made. For example the fact that most of the energy consumed, is introduced into the ethylene process via firing the cracking furnaces. The main energy consumers in the recovery section are the cracked gas compressor and the refrigerant compressors, and steam produced in the cracking furnaces and steam consumed are typically balanced. Cooling is most commonly performed via a cooling-water cycle, discharging heat to ambient temperature via a cooling tower. The ethylene process requires a number of utilities, some of them like fuel gas and steam being produced in the plant itself.

3.10.2 Distillation columns

See figure 19 for the energy required for the various distillation columns according to literature.

Figure 19: Distribution of heat load and work requirement of the refrigeration system in a typical front-end de-methanizer naphtha steam cracker plant, from Ullmann’s Encyclopaedia [3].

A) Distribution of heat load and work requirement with a) net heat absorbed by the refrigeration system; and b) net work done by the refrigeration system.

B) Relative cost of heat absorbed by the refrigeration system at different temperatures.

From the results in Table 7 it can be seen that de-methanizer has the third highest heating and cooling duties, after the C3 and C2 splitter (high energy requirement due to high reflux). However, the de- methanizer in the model is without the duty required for the hydrogen flash and the cryogenic chilling. However, it can be deduced that the de-butanizer requires the lowest amount of energy together with the de-propanizer. Due to the cryogenic equipment, the de-methanizer and hydrogen flash also consume a big part of the capital cost.

52 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.10.3 Pinch Analysis

Figure 20 shows the pinch analysis performed on the model. From the cold and hot composite curves, it can be seen that only hot utility is insufficient to complete the heating and cooling demand of the system. A heat duty of 5.886x108 kJ/h and no cold duty is required. These results show that a heat duty of 5137 kJ/kg ethylene is required, which is mainly due to the heating of the preheating of the additional fuel combustion and the preheating of the HPS and MPS.

Figure 20: Pinch analysis for naphtha steam cracking.

3.11 Aspen Plus model

In order to produce the model some assumptions have been made, fluid property packages have been chosen for calculating the properties of the components and of course also some problems occurred during the modelling which will all be discussed in this paragraph.

Assumptions in model

In order to produce the model some assumptions have been made, for example:

x No pressure drop will take place in the reactor x Residence time in the reactor is 0.1 s and cooling in the TLEs is performed in 0.1 s x No further separation of BTX, C4 fraction is required x Fractional conversion of acetylene and methyl acetylene are 99.9 % x No excess of hydrogen is required for hydrogenation of acetylene and methyl acetylene Fluid property packages

Aspen fluid packages usable for the separation train, quench tower and light hydrocarbons are Peng- Robinson and RK-Soave. Primary fractionators typically use Chao- SEA, or Grayson. For acid gas absorption ELECNRTL is typically used. In this model, Peng-Robinson was used as base method.

53 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

3.12 Steam cracking of ethane

Due to the recent increase in shale gas and switch to ethane crackers, the difference between an ethane and naphtha cracking plant configuration is briefly discussed.

To begin with, ethane is a very stable paraffin, higher cracking temperature is necessary than for other hydrocarbons and mixtures, i.e. naphtha. In case of ethane steam cracking a 60-75 % conversion and ethylene yield around 50-55 % is obtained for commercial furnaces. Short residence time is however in most cases not feasible due to short run length. On the other hand short-residence-time ethylene yield improvement is moderate (1.55% higher yield for 0.1 s compared to 0.5 s) and are therefore not applied in most cases. Steam cracking of ethane has a similar process as naphtha steam cracking. It contains a pyrolysis section, a quenching section and a recovery/separation section. The individual processes however differ due to different feedstock, which means different physical properties and plant configuration. Different feedstock often influences the distillation and separation section. A higher capacity of the C2 splitter is required, however no de-butanizer, de-propanizer, gasoline stripper or primary fractionator is required. Storage tanks or recovery equipment for propylene, butadiene and BTX aromatics are not needed, however an ethane vaporizer and superheater are required [4].

Processing of cracked gas resulting from liquid feedstocks is more complex as higher amounts of heavy hydrocarbons are condensed and must be removed in this section. Also the coke build-up is greater with heavier feedstocks such as naphtha, compared to ethane. Heavier feedstock also provides less yield of ethylene, which means more feed must be cracked to provide a given ethylene yield, see figure 21.

Figure 21: Feedstock requirements for a 1,000,000 t/a ethylene plant from Ullmann’s Encyclopaedia.

It shows that 2.5 times more naphtha feedstock is required, compared to ethane, to produce the same amount of ethylene. It can also be seen that naphtha cracking shows some sensitivity to short residence times, whereas ethane crackers are less sensitive with respect to feedstock savings with respect to shorter residence times.

54 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4 Energy analysis and plant design for OCM

4.1 Introduction

The catalytic oxidative coupling of methane (OCM) reaction to higher hydrocarbons (C2+) has been investigated over the past 30 years. Various publications in this field of research have shown that the OCM is a very promising reaction as an alternative method in the production of ethylene. Two main obstacles have prevented OCM to be commercialized to an industrial application, the first is the relatively low ethylene concentration in output gases and the second is the huge amount of energy required to carry out the reaction [65].

For over 20 years, several studies have been published regarding the exploit of methane. The main component of natural gas as can be seen in figure 22 below.

Average Compositions Natural Gas

Netherlands 83.2

Western Canada 95

Russia 97.6

USA-Alaska 99.7

0 10 20 30 40 50 60 70 80 90 100 vol-% USA-Alaska Russia Western Canada Netherlands CH4 99.7 97.6 95 83.2 C2H6 0.09 0.99 3.2 4.05 C3H8 0.03 0.32 0.2 0.71 C4+ 0.01 0.12 0.09 0.33 N2 0.17 0.84 1 9.94 CO2 0 0.09 0.5 1.79

Figure 22: Average compositions natural gas top for top three natural gas producing countries and the Netherlands based on data from [66]–[68].

Figure 22 gives an overview of the average composition of natural gas of the top three natural gas producing countries and the Netherlands, who is at the 8th place out of world production of natural gas. As can be seen on average natural gas consists primarily out of methane (90%), other alkanes, helium, carbon dioxide, nitrogen and sulphur. The different compositions of natural gas per country are due to

55 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

a combination of the different production locations and to the different processing technologies used. The Netherlands has a higher nitrogen content, because the gas is produced from onshore fields, which results in a lower energy content.

As mentioned in chapter 2, due to the recent shale gas revolution and exploitation of large world reserves of this hydrocarbons source it is very popular to research methane-utilizing techniques. At the moment, the quantity of methane used as raw material for the chemical industry accounts only for 5-7 % of the total consumption [69]. Natural gas utilization for power generation, directly or via NGL, is a well-established process and only a small portion is used as chemical feedstock. Producing ethylene from methane in addition to or instead of ethane would greatly increase the available feedstock for ethylene production. On the other hand with the crude oil depletion, methane from natural gas may well become the main energy source and the primary raw material for many chemical products. The indirect utilization of methane by converting it to syngas first, and processing the syngas to methanol or other products has been discussed in chapter 2. Besides the fact that this is a complicated process, high production costs and high amounts of energy are consumed. The direct utilization of methane to valuable components is economically more valuable, especially partial oxidation or OCM [29].

4.2 Theory

Oxidative coupling of methane is a complex system of heterogeneously catalysed and non-catalytic gas phase reactions. The reaction kinetics of these reactions will be discussed in the next paragraph. One of the best results to date, is 30 % conversion of methane and about 80 % selectivity to C2+ hydrocarbon products, ethane and ethylene [29]. Determining how to convert methane to C2+ with high selectivity is difficult because complete oxidation of CH4, C2H4, and C2H6 to CO2 and H2O must be suppressed. Efforts in gaining a higher yield were mainly researched via studying new reactor designs, but also led to no noticeable gains until now [70].

4.2.1 Reactions and kinetics

Despite extensive research on the OCM reaction, still many fundamental aspects which determine the choice of catalytic components remain unknown. For instance, distribution between surface-to-gas phase reactions, essential features for an optimal catalyst composition, as well as the participation of the non-equilibrium sites in the OCM process. However, Stansch et al. [71] provides a reaction scheme with a 10- reaction step kinetic model of OCM to C2+ over a La2O3/CaO catalyst. These kinetics are based on a study which was performed covering a wide range of reaction conditions in a micro-catalytic fixed bed reactor. These kinetics appear to be superior to previous kinetics published, moreover due to the wide range of reaction conditions and the achieved accuracy of the kinetics, it can be applied for reaction engineering simulations in this thesis. The kinetic model is characterized by the following set of stoichiometric equation.

56 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Step 1: C 4HO2 2 o 2  CO2 2 HO (4.1)

Step 2: 2CHOCHHO4  0.5 2 o 2 6  2 (4.2)

Step 3: CHOCOHOH4 2 o 2  2 (4.3)

Step 4: COOCO0.5 2 o 2 (4.4)

Step 5: CHOCHHO2 6 0.5 2 o 2 4  2 (4.5)

Step 6: CHOCOHO2 42 2 o 2  2 2 (4.6)

Step 7: CHCHH2 6o 2 4  2 (4.7)

Step 8: CHHOCOH2 42 2 o 2  4 2 (4.8)

Step 9: COHOCOH2 o2  2 (4.9)

Step 10: COHCOHO2 2 o  2 (4.10)

The course of this 10-step reaction mechanism is depicted in the reaction scheme of figure 23 below.

Figure 23: Reaction scheme for OCM based on Stansch et al. (1997) [71].

The rate equations for this model can be found in equation (4.11-4.16) and the kinetic parameters estimated for the reaction scheme and respective rate equations are summarized in table 11. Since the kinetic parameters didn’t have the proper units, the parameters were adjusted as shown in the fifth and sixth column, in order to use them in the equations.

57 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 11: Kinetic parameters OCM reaction from Stransch et al. (1997), [71].

Step Value Value Unit Ea, j k0, j m j n j [kJ mol-1]

-1 -3 -1 1 E1 48 k0,1 3.60 [mol Pa m s ] 0.24 0.76

-13 -1 ȴ,CO2,1 -175 k0,CO2,1 2.50 x10 [Pa ]

7 -1.4 -3 -1 2 E2 182 k0,2 4.18 x10 [mol Pa m s ] 1 0.4

-14 -1 ȴ,CO2,2 -186 k0,CO2,2 8.30 x10 [Pa ]

-12 -1 ȴ,O2,2 -124 k0,O2,2 2.30 x10 [Pa ]

-1 -1.42 -3 -1 3 E3 68 k0,3 9.36 x10 [mol Pa m s ] 0.57 0.85

-14 -1 ȴ,CO2,3 -187 k0,CO2,3 3.60 x10 [Pa ]

2 -1.55 -3 -1 4 E4 104 k0,4 1.98 x10 [mol Pa m s ] 1 0.55

-13 -1 ȴ,CO2,4 -168 k0,CO2,4 4.00 x10 [Pa ]

5 -1.32 -3 -1 5 E5 157 k0,5 3.06 x10 [mol Pa m s ] 0.95 0.37

-13 -1 -166 k0,CO2,5 4.50 x10 [Pa ] ȴ,CO2,5 5 -1.96 -3 -1 6 E6 166 k0,6 1.08 x10 [mol Pa m s ] 1 0.96

-13 -1 ȴ,CO2,6 -211 k0,CO2,6 1.60 x10 [Pa ]

7 -1 -3 -1 7 E7 226 k0,7 1.20 x10 [mol Pa m s ] - -

10 -0.97 -3 -1 8 E8 300 k0,8 1.67 x10 [mol Pa m s ] 0.97 0

2 -2 -3 -1 9 E9 173 k0,9 3.42 x10 [mol Pa m s ] 1 1

4 -2 -3 -1 10 E10 220 k0,10 4.68 x10 [mol Pa m s ] 1 1

Ea k eRT pmj p n j 0 j CO2 rj 2 ; j=1, 3-6 (4.11) H ª ' ad, CO2, j º 1 § K eRT · p «  ¨ j, CO2 ¸ O2 » ¬ © ¹ ¼

H n2 Ea,2 ' ad, O2 k eRT § K eRT p· p 0,2 ¨ 0,O2 O2 ¸ CH4 © ¹ (4.12) r2 n H n H § ' ad, O2 ' ad, O2 · 1 § K eRT p· K eRT p ¨  0,O2 O2  j, CO2 O2 ¸ ¨ ¨ ¸ ¸ © © ¹ ¹

58 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Ea,7 r k eRT p (4.13) 7 0,7 CH2 6

Ea,8 m n r k eRT p8 p 8 (4.14) 8 0,8 CHHO2 4 2

Ea,9 m n r k eRT p9 p 9 (4.15) 9 0,9 CO2 H 2 O

Ea,10 m n r k eRT p10 p 10 (4.16) 10 0,10 CO2 H2

4.2.2 Various process schemes

Besides OCM for C2+ production, the process can also be used for example for the production of gasoline, LNG or as add-on unit to a naphtha cracker. The latter uses the methane produced during naphtha cracking as a feedstock for OCM. It was concluded that this concept was technically and economically feasible, at 1992 prices, only if the catalyst meets the assumed conversion of 30 % and selectivity of 80

% towards C2 products [72]. Due to the coupling of the two processes, the same cryogenic separation for the naphtha cracking could be used for purification of the OCM products.

In the case of OCM to olefins production, different researches have been conducted and also studies reviewing the different processes have been performed. In the study of Gradassi and Green (1995) [73], one process concerning OCM to olefins was reviewed in detail. Figure 24 shows the schematic overview of the OCM process to C2+ hydrocarbons.

Figure 24: Process overview for conventional catalytic OCM to ethylene based on a process from Gradassi and Green [73].

59 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

In this process fresh natural gas, is fed together with oxygen and a recycle gas stream consisting of mostly unreacted methane and oxygen. This reaction is performed at 4.83 bar, while heat exchange of the product gas stream preheats the feed gas from 21 °C to a reaction temperature of 789 °C and reduces the product gas temperature from 816 °C to about 90 °C. The temperature of the reactor was maintained at below 816 °C by generation of high pressure (103 bar) steam. After this process, the purification of the stream takes place which is a similar process as the separation train of naphtha steam cracking. First, the product gas is compressed to 10.3 bar and condensate is removed. The gas then passes through an acid gas removal unit based on a 30 % aqueous mixture of monoethanolamine (MEA) for the removal of CO2. Hereafter, the stream passes through a refrigeration unit for the removal of additional water and then finally a molecular sieve column for drying and thus removal of the remaining water vapour and CO2. A cold box system is then used to recover the C2+ products from the product gas stream, ethylene is recovered in a typical ethylene plant separation scheme using a de-methanizer, de- ethanizer and a C2 splitter. In this process electric power is generated to meet plant requirement using steam turbine generators. High pressure steam is generated from the OCM reactor heat, and from using fuel gases from the purge gas and from the C2+ recovery system.

No export power needs to be provided. A quick analysis of this scheme compared to the naphtha steam cracking, shows that the separation train for OCM is less complicated due to the simplicity of the product mixture. A difference with other researches, is for example that this thesis includes the air separation unit together with the OCM process.

4.2.3 Reactor design

The reaction system and kinetics have been investigated in various works [74]–[76], in the extent that knowledge about the sub-processes and their limitations are well-known. The most studied reactor types are: conventional catalytic fixed-bed reactor, plug flow reactor with distributed oxygen addition, counter current moving bed chromatography (which is a simulated reactor and only a theoretical concept), fluidized bed reactors and membrane reactors. However, all reactor process designs suffer from the high costs of separation of the products from unconverted methane and the required gas recycling. As soon as these challenges and the selectivity of the catalyst have been overcome, the conversion of methane will certainly become economically viable [65].

4.2.4 Catalysts

Also work on catalysts for OCM has been reviewed, the activity of the catalysts for this reaction are very dependent on the experimental conditions, such as pressure, CH4/O2 ratio, temperature, space velocity and so forth [77]. Challenging properties of the catalysts involve their ability to generate methyl radicals without their consecutive heterogeneous oxidation, and to perform this activation in the presence of other reactive reaction products (C2H4 and C2H6). However despite the knowledge of the fundamentals

60 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

of the OCM reaction and its catalysts, yields of C2 hydrocarbons higher than 30 % have not yet been achieved. Some researches do not expect that higher values will be achieved in the near future based on present catalyst design only [78]. Hydrogen and steam are sometimes added to reduce coking on catalysts.

4.2.5 Cryogenic air separation unit

The cryogenic air separation unit (ASU) in the OCM process can be described as a process which separates oxygen from air by using two or more distillation columns in which the air separation takes place in two fluid streams respectively rich of nitrogen and oxygen. The ASU is a large instillation, which works at very low temperatures (-170 °C) and could reach oxygen concentrations higher than 95 vol-%. The oxygen produced in the ASU can be used in different areas, such as combustion in oxycombustion systems, space propulsion or in this case for oxidation of methane to ethylene. ASU is very energy intensive, due to the cooling of the gases and the selective separation though cryogenic distillation. Other methods for oxygen separation such as membranes, pressure swing adsorption (PSA) and vacuum pressure swing adsorption (VPSA), are commercially used but on a smaller scale than cryogenic ASU.

For the large tonnage oxygen customers (more than 300 t/d, which would be the case if this OCM plant with ASU would be built), a cryogenic plant is usually the most economic compared to other oxygen production technologies. Cryogenic ASU is also the only economically available technology which is able to produce the oxygen amounts required and at high purity [79]. The reason to use pure oxygen in the OCM process, instead of air, is due to the large recycle stream with unreacted methane in which nitrogen would also be present and build up in the recycle. These impurities will affect the reactor designs, and make their size even larger, which affects the final equipment costs [80]. Besides the reason afore mentioned, the molar ratio of methane to oxygen (99% purity) should be controlled around 2:1 in order to reach the desired selectivity to ethylene in the OCM reactor.

4.3 Aspen Plus model

The complete model for the ASU and OCM reaction with purification, are for organizational purposes divided into two figures, see figure 25 and figure 26 for respectively the OCM + purification and ASU. In which stream 39B comes from the ASU containing the high purity oxygen.

61 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 25: PFD OCM and purification of the products to polymer grade ethylene.

62 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 26: Model for the cryogenic air separation unit in Aspen Plus.

63 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.4 ASU model

A cryogenic air separation unit was modelled in this model together with the OCM and its purification in order to get polymer grade ethylene, see figure 27 for the complete model of the ASU. The ASU will provide the high purity oxygen required for the OCM reaction. The cooling of the gases requires a large amount of energy to make this refrigeration cycle work and is partially delivered by an air compressor. Modern ASUs use expansion turbines for cooling, the output for the expander helps drive the air compressor, for improved efficiency.

4.4.1 Feedstock and inlet conditions

The composition of air used for the cryogenic ASU in the model in mole fractions, can be found in figure 6 28. For a 1 Mt/y of ethylene in an OCM plant with an inlet ratio CH4:O2 of 1.25, 5.4 x10 kg/h (51.79 kmol/sec) of air is required as feedstock for the ASU. This air stream is fed to the ASU at 25 °C and atmospheric pressure.

Composition air

20.95% 0.038 % 1% 0.93% 0.032 % 78.05%

Nitrogen Oxygen Argon Carbon dioxide Water

Figure 28: Pie chart of air composition used in the ASU model.

4.4.2 Compression and acid gas removal

Before compression air is typically pre-filtered of dust, however dust is not implemented in the model and this step will not be taken into account. Air is first compressed to typical pressure ranges of 5-10 bar. See figure 29 and table 12 for the configuration and operating conditions of the compressors and accompanying coolers. The process air is typically passed through a molecular sieve bed in order to remove the remaining water vapour and carbon dioxide, which would otherwise freeze the equipment. To compare both models, the molecular sieve drying is modelled as a shortcut method separator block similar to the first model. The net amount of work required for these compressors is 399.99 MW and the amount of cooling duty required is 397.67 MW.

64 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 12: Results and operating conditions compression stages and heaters in ASU unit.

Unit Comp1 Comp2 Comp3 Comp4 Comp5

Temperature in °C 25 20 20 20 20

Temperature out °C 90.5 84.7 67.6 21.5 20

Temperature after °C 20 20 20 20 20 cooling

Pressure inlet bar 1 1.84 3.36 5.28 5.36

Pressure outlet bar 1.84 3.36 5.28 5.36 74.59

Work duty MW 103.98 102.64 40.19 1.25 151.93

Cooling duty MW 106.69 98.17 38.62 1.21 152.98

Figure 29: Results and operating conditions compression stages and heaters in ASU unit.

Hereafter the product gas stream is split into two stream, of which one is further compressed with interstage cooling and the other stream is again split into two streams. The streams are split, so that each stream could be fed to one of the three distillation columns at different pressures. This results in the highest energy efficiency for separation of oxygen and results in a rich nitrogen stream and a rich oxygen stream [81].

4.4.3 Integrated heat exchanger I

Processed gas stream from molecular sieve drying, is split into three streams and individually further compressed before sending through an integrated heat exchanger (HX05in model). The integrated heat exchanger, which is usually a plate fin heat exchanger, cools the gas against the product and waste

65 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

cryogenic streams resulting from the distillation columns further on in the model. In this step part of the air is liquefied which will be sent to the medium and low pressure column (the stream which was compressed to almost 80 bar), after being cooled to -175 °C in the integrated heat exchanger. The remaining streams are also cooled to the same temperature, however they are not liquefied due to the lower pressure of these streams. The various splitting, which also takes place further on in the plant, is performed to send one or more streams to each distillation column.

4.4.4 Joule-Thomson effect in the expander

The refrigeration required for producing liquid products is obtained using the Joule-Thomson effect in an expander which feeds compressed air directly to the low pressure column. This Joule-Thomson effect describes the temperature change of a real gas or liquid when it is forced through a valve or porous plug, while kept insulated so that no heat is exchanged with the environment. In order words the change in temperature unit per unit change of pressure, while the enthalpy is constant [82], [83]. The power generated during this expansion to 2.31 bar is 2.70 MW stream.

4.4.5 Distillation columns

The liquefied stream is partly fed to all three columns see figure 30 which are a high pressure (COLHP), medium pressure (COLMP) and low pressure column (COLLP). The operating conditions and results from these distillation columns can be found in table 13.

After these cryogenic distillation steps, eventually a stream rich of oxygen and a stream rich of nitrogen will be produced. The high purity nitrogen stream comes from the overhead of the high pressure column. While the high purity oxygen stream will come from the bottom of the low pressure column. The condenser of the high pressure column requires refrigeration which can be obtained from expanding the oxygen rich stream further across a valve or through an expander. Alternatively the condenser may be cooled with heat from the reboiler in a low pressure distillation column. To minimize the compression cost, the combined condenser-reboiler operation between the high and low pressure columns should have a temperature difference of only 1-2 K. Since this is not the case in the model an alternative energy efficient method was performed, namely combined condenser-reboiler operation between the low pressure and medium pressure column which provides 119.57 MW cooling duty to the reboiler and afterwards the reboiler requires 15.18 MW in the reboiler. However combined condenser-reboiler operation between high and medium pressure would probably be better in terms of compression costs. A certain part of the air is not separated and leaves the low pressure column as a waste stream from the overhead.

66 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 13: Results and operating conditions columns ASU.

Unit HP Column MP Column LP Column Number of stages - 50 50 100 Pressure bar 5.16 3.1 1.38 Temperature condenser °C -178.5 -184.7 -192.7 Temperature reboiler °C -173.0 -177.8 -181.0 Heat duty condenser MW -95.71 -119.57 7.71 Heat duty reboiler MW 32.58 43.93 14.84 Distillate rate kg/h 1.10 x106 1.23 x106 4.11 x106 Reflux ratio* - 0.8 0.9 0.25 Boilup ratio‡ - 2.30 3.18 0.20 Nitrogen rich stream purity % 98.1 99.4 98.7 Oxygen rich stream purity % 50.1 64.5 95.5

Figure 30: Distillation columns, expander and the second integrated heat exchanger in ASU.

Argon

As mentioned before, air is mainly composed of nitrogen and oxygen, however argon is also present in air and needs to be separated to get the required purity of oxygen and nitrogen. The boiling point of argon (186 °C), lies exactly between the boiling points of nitrogen (-196 °C) and oxygen (-183 °C). Which results in argon building up in the bottom of the low pressure column. Typically a vapour side draw is taken from the low pressure column, where the argon concentration is the highest, in order to separate

67 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

the argon from the oxygen in a separate column. Though argon is only present for less than 1% of the feedstock, the column required for this separation requires a high amount of energy due to the high reflux ratio needed for this separation (around 30). Cooling of the argon column can be supplied from the cold expanded rich liquid or by liquid nitrogen. The use of modern structured packing which have very low pressure drops enable argon with less than 1 ppm impurities. However this column is not taken into account in this ASU plant to get a lower energy requirement and the assumption that 95.5 % oxygen purity is high enough. Typical oxygen purities recovered from the ASU lie between 97.5-99.5 % and influence the maximum recovery of the oxygen.

4.4.6 Integrated heat exchanger - part II

Finally the top from the low pressure column, which is rich in nitrogen, is fed to the second integrated heat exchanger, in order to recover some of the cooling duty from this stream before sending it back to the first heat exchanger in which the rest of the cooling duty is recovered by other streams. After this cooling duty exchange, the rich in nitrogen stream has an end temperature of 17.9 °C. As mentioned before the rich nitrogen and rich oxygen stream are used in the first integrated heat exchanger for cooling the feed stream. The rich oxygen stream from the bottom of the pressure column is first fed to a pump and pressurized to 41.24 bar which requires 2.25 MW and then circulated back to the first integrated heat exchanger. After leaving the first heat exchanger, the oxygen rich stream is warmed up to ambient temperature against the incoming feed and is gaseous again. All the exit streams from the medium and high pressure column are routed to the second integrated heat exchanger and cooled by a few degrees Celsius to around -190 °C.

4.4.7 Valves

In the ASU model, various valves were used for pressure release between the streams from column to column. In this case the disturbance from pressure differences is kept low.

4.4.8 Products and storage

Eventually 97.89 wt-% of the oxygen is recovered which amounts to 1.224 x106 kg/h (339.9 kg/sec or 10.6 kmol/sec), which has as purity of 95.4 %. This purity is lower than conventional oxygen separation with cryogenic ASU due to the argon contamination. However, the oxygen’s purity is still high enough for the OCM reaction. The question is if the energy gain due to not performing the argon separation through cryogenic distillation, and separate it later on in the de-methanizer (which also results in lower purity thus lower yield and recycling of argon) weighs up to the purity decrease in the de-methanizer and thus build-up of argon. 99.15 wt-% of nitrogen is recovered which amounts to 4.04 x106 kg/h, which has a purity of 98.2 mole-%. The high purity nitrogen with low oxygen contamination, may be used as

68 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

inert gas in storage tanks of ships and tanks for petroleum products. The separated products are sometimes supplied by pipeline to large industrial users near the production site, for longer distances transportation is done by shipping liquid product for large quantities. In this model, the air separation unit is on site.

4.4.9 Work requirement and power production

Table 14 below shows the results from the model considering the work required and power produced.

Table 14: Results from power production or requirement per block in model.

Block Stream Power production (+) or required (-) in MW

Comp1 W1 -103.98

Comp2 W2 -102.64

Comp3 W3 -40.19

Comp4 W4 -1.25

Comp5 W5 -151.93

Turbine W6 +2.70

Pump W7 -2.25

Total -399.54

The power required for the first two compressors is almost similar due to the same pressure ratio. The total power required for this ASU producing 1224 t/hr of oxygen, is 399.54 MW. This means 0.319 kWh of electricity required per kg of oxygen consumed (0.325 kWh electricity is required per kg of high purity oxygen produced). The result is in the same range compared to for example a conventional cryogenic ASU according to Spallina et al. (2014), however compared to a conventional cryogenic ASU from Linde Engineering with 95% oxygen purity and a specific energy consumption of 0.232 kWh/kg, the energy requirement for this ASU is higher. Moreover, advanced energy optimized process schemes for new applications, e.g. IGCC, oxyfuel and IGSC have a specific energy consumption of 0.167 kWh/kg for a 95% oxygen purity. Of which the latter has a reduction by >25 % of power consumption compared to the conventional cryogenic ASU [84].

69 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.5 OCM part I

The first part of the OCM model has been explained, which included the oxygen separation from air in a cryogenic ASU. With the high purity oxygen coming from the ASU together with imported natural gas, the OCM reaction can take place in a plug flow reactor. Hereafter the produced gas needs to be cooled, compressed and purified during the cryogenic separation. The following paragraphs will explain these parts, the overview of these parts modelled in Aspen Plus can be seen in figure 25.

4.5.1 Feedstock, catalyst and inlet conditions

A few average compositions of natural gas were mentioned in the introduction of this chapter, from this average composition the following composition of natural gas was used in the model. Multiple situations were modelled, e.g. with recycle to cool the reactor, with recycle of methane and ethane, different feedstock compositions, etc. The composition of the rich oxygen stream is however always the same. The natural gas is always fed at 25 °C, and preheated to 200 °C. The simulations explained in 4.5.3 were performed with composition 1 of natural gas, and all the remaining simulations were performed with composition 2 of natural gas. As can be seen in table 15, the compositions are only slightly different. The simulations are performed at 800 °C and 1 bar, unless mentioned differently. See figure 31 for the first part of the OCM model.

Table 15: Different compositions of natural gas used during simulations.

Composition 1 Composition 2

Component Mole fraction Mole fraction

CH4 0.956 0.92

C2H6 0.022 0.04

CO2 0.004 0.01

N2 0.018 0.03

Catalyst

3 3 This Excel model has a solid fraction of 0.6 mp /mr . The catalyst used in the model is the same as Stansch

3 et al. uses which is a La2O3 (27 % at.)/CaO catalyst with a density of 3600 kgp/mp and an active catalyst 3 weight fraction of 0.27 kgcat/kgp, which thus results in a catalyst weight in the reactor of 583.2 kg/mr . To give an idea, for the reactor design in this model, this results in 42,920 kg of catalyst for a 1.220 Mt of ethylene plant without recycle streams.

70 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 31: Model of mixing, OCM reactor and cooling recycle.

4.5.2 OCM reactor model

Preheated natural gas (up to 200 °C) is mixed with oxygen at atmospheric pressure and a possible recycle stream from the product exiting the OCM reactor. The recycle is not always used in the analysis of the OCM process, this will be mentioned when applicable. The OCM reactor has been modelled in two ways. With the first method, the kinetics from Stansch et al. were directly implemented in the Aspen software, see Appendix C. However the results from this method differed too much from the paper. In the second method, which will be used from now on in all the simulations, the plug flow model was made in Excel with the kinetics from Stansch et al. This Excel model was coupled to the reactor in Aspen Plus model (a similar method as the naphtha steam cracker).

OCM packed bed reactor model with Excel

First, the results from Stansch et al. were reproduced in a plug flow modelled in Excel. Then, Excel model including these kinetics was implemented into a calculator in Aspen. This Excel sheet is a bit more complex than the Excel model used for the naphtha steam cracker. Since in the latter case the in- and outlet did not change and no kinetics were coupled to the file, just components going in and out. Since kinetics implementation did not work out for this Aspen model, a set of five equivalent reactions is considered in order to achieve the same gas conversion from Excel and Aspen. Together with this set of reactions and the calculator, the results can be obtained with the kinetics as well in Aspen as in Excel. The components of the feed entering the reactor (natural gas, oxygen and possibly a recycle stream) and the pressure are imported from Aspen into the Excel file, which then calculates the conversion of all components along the length of the plug flow reactor. During the simulation, the conversion of methane, selectivity of C2+, and the yield of C2+ were calculated along the length of the plug flow reactor. Simply put, Aspen gives the inlet pressure and feedstock composition to Excel, the Excel file calculates the output according to a plug flow model with Stansch kinetics and hereafter Excel gives the results from the outlet composition back to Aspen.

71 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Conversion, selectivity and yield

During the first simulations while studying the optimum molar ratio and the cooled reactor, the yield, conversion and selectivity were defined the same according to Stansch et al., see the following definitions:

Methane conversion:

CC CH  CH X 4 L x 4 L x ' x CH (4.17) 4 L x C CH 4 L x

Selectivity of ethylene and ethane (C2+) combined to methane:

2CC CH  CH S 2 4 L x 2 6 L x C (4.18) 2 L x C C CH  CH 4 L x 4 L x ' x

Yield of ethylene and ethane (C2+) combined:

YXS C CH u C (4.19) 2 L x 4 L x2 L x

Further on in the project, when the recycle of remaining methane and ethane was implemented, definitions for the yield, conversion and selectivity were redefined. This was done for better comparison to the naphtha steam cracker, since ethylene is the desired product the focus was no longer on C2+ but on ethylene. See the following equations 4.20-4.22, the methane conversion will be the same as in equation 4.17.

Ethane conversion:

CC CH  CH X 2 6 L x 2 6 L x ' x CH (4.20) 2 6 L x C CH 2 6 L x

Selectivity ethylene to methane:

C CH S 2 4 L x CH (4.21) 2 4 L x CC / 2 CH  CH 4 L x 4 L x ' x Yield ethylene based on methane and ethane conversion:

CC CH  CH Y 2 4 L x 2 4 L 0 CH (4.22) 2 4 L x C X / 2 CX CH u CH CH u CH 4 L x 4 L x 2 6 L x 2 6 L x

72 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Simulations

Since Aspen and Excel now generate results according to the paper, further design of the OCM plant and generating results from the kinetics could be performed which is similar to the separation section in the naphtha steam cracker. Except, no oil quench, primary fractionator, de-ethanizer, de-propanizer, water quench, gasoline stripper and de-butanizer are required First simulations were performed at much lower flow rates, to get an idea of the conversion and yield and separation section. According to these results, the plant was scaled up to produce 1 Mt of ethylene per year (without taking any recycle into account).

4.5.3 Results from OCM + ASU without recycles

After scaling up, one of the parameters interesting for OCM reaction was studied, namely the ratio of methane to oxygen fed to the reactor. Therefore, various simulations have been performed to see what the influence of this ratio was on the mole fractions of the main components, the conversion of CH4, selectivity of C2+, and the yield of C2+. This was performed by calculating the aforementioned definitions values along the length of the plug flow reactor. Therefore by varying the molar ratio results were generated by varying the flow rate of the natural gas, in order to find the optimum molar ratio of CH4 to O2. These results were generated at 1 bar in such a way that oxygen completely reacted away, by varying reactor dimensions.

Results from varying molar ratio

The results of the conversion, selectivity and yield per molar ratio can be found in figure 32-35. The results also show that along these ranges of molar ratios, the C2+ yield varies between 12.45 and 22.65

%, while the methane conversion varies between 12.49 and 100 %. The C2+ selectivity varies between 12.59 and 99.66 % for these molar ratios. This shows that 100 % methane conversion at a ratio of 0.5 can be obtained, however this results in a low selectivity and as consequence low yield. Also C2+ selectivity of 100 % is obtained, however together with only 12.5 % methane conversion, this still results in a low yield. Unfortunately the results show that at high conversions there are no high values of C2+ selectivity and vice versa. From analysis of these results, the optimum molar ratio at the highest C2+ yield of 22.65 %, can be established at a molar ratio of 1.25 with a methane conversion of 60.20 % and a C2+ selectivity of 37.63 %.

73 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Influence of CH4:O2 ratio on X, S and Y 100 90 80 CH4 conversion C2+ selectivity 70 C2+ yield 60

% 50 40 30 20 10 0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

Molar ratio CH4:O2

Figure 32: Influence of CH4:O2 ratio on methane conversion, C2+ selectivity and C2+ yield.

The results from the reaction rates along the length of the plug flow reactor can be found in figure

33.The results from the conversion of methane, selectivity of C2+, and the yield of C2+ calculated along the length of the plug flow reactor can be found in figure 34. The results from the molar fractions of

CH4, C2H4, C2H6 and O2 along the length of the plug flow reactor can be found in figure 35.

Reaction rates at CH4:O2 ratio 1.25 700 dr1

600 dr2 dr3 s)] 3 500 dr4 dr5 400 dr6 300 dr7 dr8 200 dr9 Reaction rate [mole/(m 100 dr10

0 0 3 6 9 12 15 Length reactor [m]

Figure 33: Reaction rates along the plug flow reactor for the OCM reaction with a CH4:O2 molar ratio of 1.25.

74 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

X, Y and S at CH4:O2 ratio 1.25 100% CH4 Conversion 90% C2+ Selectivity 80% C2+ Yield 70%

60%

50%

40%

30%

20%

10%

0% 0 3 6 9 12 15 Length reactor [m]

Figure 34: Conversion, selectivity and yield results along the plug flow reactor for the OCM reaction with a CH4:O2 molar ratio of 1.25.

Molar fractions components at CH4:O2 ratio 1.25 60% C2H4 CH4 50% C2H6 O2 40%

30%

20%

10%

0% 0 3 6 9 12 15 Length reactor [m]

Figure 35: molar fractions of methane, ethylene, ethane and oxygen along the reactor.

75 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Results optimum molar ratio

Figure 35 shows that oxygen is totally reacted away after 12 m in the reactor, at this point figure 34 shows that the maximum achieved methane conversion is 58.36 % with a C2+ selectivity of 42.99 %, and

C2+ yield of 25.09 %. A closer look to the selectivity shows a C2H4 yield of 20.76 % and for C2H6 yield of

4.33 %, and a selectivity respectively of 35.56 % to C2H4 and 7.43 % to C2H6. The paper obtained results in wide ranges, e.g. methane conversions between 0.9-32.7 %, oxygen conversion between 15-100 %,

C2H6 selectivity between 1.2 -33 %, and C2+ yields above 20 %, due to the use of different catalysts. The definition of ethylene yield and C2+ selectivity from Stansch et al. do not make sense for the focus of this project, also no recycle is taken into account and if ethane is in the feed the definitions give a wrong representation of the yield and selectivity. At the beginning of the reactor nothing is yet produced, however these definitions show there already is a yield of 10 % due to the ethane which is present in the feed and thus at the beginning of the reactor. Because of the aforementioned reasons, the definitions for yield, conversion and selectivity were adjusted for the focus of this work which is ethylene. From now on the new definitions for yield, conversion and selectivity will be used.

Results reactor design and gas velocity for OCM without recycles

Thet total amount of feed fed to the reactor is (together with the recycle and oxygen) comes down to 863 m3/s. The diameter of the reactor is kept constant at 2.5 m during all simulations, while the length of the reactor was varied (with 2500 steps, and thus varying dx). This means the cross-sectional area of this reactor is 4.91 m2, according to the simple area circle equation 4.23.

2 A S r (4.23) According to this cross-sectional area, a gas velocity rate of 175.8 m/s applies to this feed, for a PBR this should be in the range of 0.5-1.0 m/s. This means multiple reactors would be necessary and or increase of the reactor diameter. Let say that eventually a velocity of 1 m/s would be accomplished, then one reactor with a radius of 13.3 m would be required. Since this is not realistic, for example 36 reactors with a diameter of 2.5 m, 9 reactors with a diameter of 5 m or 5 reactors with a diameter of 7 m have been considered for this plant with a feed of 2.16 x106 kg/h and produces 139,259 kg/h of ethylene. This results in a 1.220 Mt/y of ethylene in case of continuous operation without shut down. For a factor 1.220 smaller plant, assuming the same conversion and yield, 8 reactors with a diameter of 5 meter would be required. Complete conversion of oxygen takes place at 12-13 m in the reactor, with a gas velocity of 1 m/s and a residence time of 13 seconds.

4.5.4 Results adiabatic vs cooled reactor

Hereafter, simulations for these different molar ratio were performed in an adiabatic and a cooled reactor. The results are presented in table 16 and figure 36. The results in this table are obtained at

76 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

oxygen conversion of 100 %, at a pressure of 1 bar, a reaction temperature of 800 °C and a diameter of 2.5 m.

Also the temperature rise in the reactor in case of adiabatic operation and the cooling duty required in case of a cooled reactor were studied. The last two things are interesting to study, since the OCM reaction is exothermic and the heat of reaction is a complex issue in the OCM process. The cooled reactor was studied, to see if it is possible to provide such a cooling system required for cooling a reactor with a capacity of producing 1.220 Mt/y of ethylene.

Table 16: Results conversion, selectivity, yield, temperature rise adiabatic reactor and required cooling duty cooled reactor.

CH4:O2 CH4 C2+ C2+ Temp. Cooling duty required Conversion Selectivity Yield Adiabatic reactor (Treactor= 800 °C) [molar ratio] [%] [%] [%] [°C] [MW] 0.50 99.98 12.59 12.59 4292 3266 0.75 84.43 23.77 20.07 3697 3437 1.00 70.23 31.7 22.26 3089 3275 1.15 63.75 35.43 22.59 2814 3175 1.25 60.20 37.63 22.65 2665 3113 1.35 56.89 39.64 22.55 2527 3050 1.50 52.41 42.62 22.34 2343 2958 2.00 42.49 50.14 21.30 1942 2705 2.50 36.03 56.02 20.19 1682 2473 3.00 31.40 60.97 19.14 1496 2247 3.50 27.86 65.34 18.20 1354 2022 4.00 25.20 69.09 17.41 1246 1807 4.50 23.02 72.56 16.70 1156 1589 5.00 21.23 75.75 16.08 1083 1372 6.00 18.48 81.42 15.04 968 945 7.00 16.41 86.54 14.20 880 514 8.00 14.82 91.21 13.51 812 87 9.00 13.54 95.56 12.45 757 0 10.00 12.49 99.66 12.45 711 0

The results show that cooling the reactor produces a lot of energy, for instance 3.1 GW for the optimal molar ratio case, which means 5.26 MJ of heat released per kg feed (1.46 kWh/kg feed). Calculated for the amount of ethylene produced this means 98.26 MJ/kg (27.3 kWh of heat which needs to be removed per kg of ethylene produced). The range lies between 0 and 3.4 GW, however with 0 MW only 13.54 % of methane is converted and 12.45 % yield of C2+ is obtained. Around a yield between 15 %, the temperature increase is still acceptable with outlet temperatures between 711 and 968 °C, however with higher yields removing heat from the reactor becomes a serious problem. These outlet temperatures are found in the Aspen model, since kinetics in Excel are always calculated at a fixed temperature of 800 °C.

77 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Cooling duty required per molar ratio CH4:O2 3500 Cooling duty 3000

2500

2000

1500

Cooling duty[MW] 1000

500

0 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00

Molar ratio CH4:O2

Figure 36: Cooling duty required in case of cooled reactor to 800 °C for different CH4:O2 molar ratios.

The linear relation between the cooling duty required and the molar ratio is due to the varying of the methane conversion, which is lower at higher molar ratios, thus lower amounts of heat need to be removed from the reactor. The temperature increase in the reactor, for different molar ratios, is also due to the varying methane conversion. For this ratio and a total feed of 2.12 X106 kg/h (including oxygen) was fed for the optimum ratio, 124,380 kg/h of ethylene was produced resulting in a 1.09 Mt/y plant.

Results reactor cooling with cold recycle

Another option could be working with a cooled product gas recycle stream, which is cooled to 500 °C and fed to the reactor while keeping the reactor temperature at a maximum of 800 °C. This is calculated with a design specification in the model, which varies the split fraction and thus flow rate of the recycle stream while maintaining the reactor temperature at a maximum of 800 °C. The results from both of these simulations can be found in table

In this simulation initially 293 kg/s of natural gas (composition 2) at a CH4: O2 molar ratio of 1.25 at 1 bar was fed to the reactor. This led to a recycle stream, as big as 88.18 % of the product gas, which needed to be recycled to the reactor in order to keep the reactor temperature at 800 °C. This recycle stream was then cooled to 500 °C in order to release heat during the reaction. The cooling of the recycle to 500 °C requires 2.594 GW of cooling duty. Besides this huge amount of cooling duty required, other disadvantages due to the high recycle are present such as the big increase of reactor volume required. First, without the recycle was required which amounted to 44 reactors with a diameter of 5 m and reactor length of 5 m. With the recycle however, the number of reactors required for this gas flow

78 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

velocity is 700 with a reactor length of 20 m in order to obtain 100 % oxygen conversion. Thus upscaling of the reactor volume up to 16 times is necessary (which was already quite large to recall the results for the simulations without recycle).

Table 17: Results simulation without cold recycle and with cold recycle directly to the reactor.

Parameter Unit Without cold With cold recycle recycle

CH4 conversion % 58.37 16.91

C2H6 conversion % 57.85 14.46

C2H4 selectivity % 36.11 50.30

C2H4 yield % 33.24 46.53

O2 Mole-% 0.00 0.00

CH4 Mole-% 20.23 18.35

CO2 Mole-% 17.76 17.15

H2O Mole-% 40.12 45.02

CO Mole-% 3.20 1.65

H2 Mole-% 9.77 6.09

C2H4 Mole-% 5.05 7.94

C2H6 Mole-% 1.04 0.89

N2 Mole-% 2.83 2.90

Total flow to reactor kg/h 2.178 x106 1.842 x107

Total nr of reactor (D=5 m) 44 700

Reactor length m 5 20

CH4:O2 ratio - 1.25 4.68

Recycle flow kg/h - 1.624 x106

Ethylene going to separation section kg/h 141,019 213,425

Catalyst mass kg 57,227 228,906

Natural gas kg/h 896,400 896,400

Ethylene production per year Mt/y 1.235 1.870 (without recycle)

79 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.6 Energy analysis and plant design OCM part II

The methane from the overhead of the de-methanizer column containing methane, and the bottom stream of the C2 splitter containing ethane could be used as a recycling stream and fed directly to the reactor to increase yield of ethylene. However due to the high contamination of light components in the methane rich stream, it is better to separate this stream first which is performed by purging. This additional separation step should be taken into account when comparing these two model in terms of equipment cost and yield increase. The overhead stream contains 300,871 kg/h of methane (39.2 % of the methane present in the feedstock) and the bottom stream of the C2 splitter contains 62,506 kg/h of ethane with a stream purity of 99.76 %. Instead of feeding the ethane stream as recycle to the reactor, it could also be fed to a furnace to crack the ethane which results in a higher yield of ethylene (60-70 %) than with OCM. Both streams are big enough to use as recycle or use in an additional plant on site. Therefore this option should be further investigated. For both models, the reactor is considered cooled to 800 °C.

The two plant configurations are as follows:

1. OCM with ASU, no recycles. Methane and ethane exiting the de-methanizer and de-ethanizer are combusted to provide heat for pressure steam generation and eventually power production from high. This plant has a feed flow rate of 896,400 kg/h (249 kg/s) of imported natural gas and 1.286 x106 kg/h oxygen to the reactor. 2. OCM with ASU, with methane and ethane recycle after separation to increase ethylene yield. To prevent build-up, 10 % of this fuel recycle is purged and combusted to provide heat pressure steam generation and eventually power production from high. This plant started with a 644,953 kg/h (179 kg/s) feed of imported natural gas simulated recycle to try to get a 1 Mt/y plant with recycle. With the recycle, the feed flow rate resulted in 1.365 x106 kg/h (379 kg/s), which is a mixture of imported natural gas (565,470 kg/h) combined with the recycle, and 1.286 x106 kg/h oxygen to the reactor.

Results on X, Y, and S for the two OCM processes

The ASU plant configuration and energy consumption from the ASU are for both plants exactly the same. Because it turned out the optimum shifts due to ethane conversion, a higher molar ratio is used in the second process. Since ethane is recycled in this process and a new optimum molar ratio was established for the highest CH4:O2 ratio. The results of these simulations can be found in table 18.A quick study of these results show a lower methane conversion for the second process but higher selectivity of ethylene. The recycle flow contains 41.4 wt-% CH4, 0.28 wt-%-H2 and 4.0 wt% C2H6. A higher yield of ethylene is obtained and a similar ethane conversion is reached. From this preliminary results there could be concluded that the recycle is always better, since the yield is higher and eventually more ethylene is produced, respectively 45,781 kg/h for the first process and 156,226 kg/h for the second process. A recycle however results in disadvantages such as a bigger reactor volume, namely 10 additional reactors are required, bigger heat exchangers, higher reboiler duties, etc. These results will be showed further in this chapter together with the energy analysis of both of these plants.

80 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 18: Results OCM with ASU with and without recycle fuel.

Parameter Unit Without With recycle recycle

CH4 conversion % 58.36 54.38

C2H6 conversion % 50.14 49.99

C2H4 selectivity % 35.58 40.22

C2H4 yield % 33.1 37.0

Molar ratio CH4:O2 - 1.25 1.33

Outlet composition

O2 Mole-% 0.00 0.00

CH4 Mole-% 20.25 20.94

CO2 Mole-% 17.81 16.88

H2O Mole-% 40.07 35.25

CO Mole-% 3.13 3.15

H2 Mole-% 9.78 8.93

C2H4 Mole-% 5.05 5.02

C2H6 Mole-% 1.05 1.08

N2 Mole-% 2.85 8.75

Total flow to reactor kg/h 2.178x106 2.647x106

Total nr of reactor (D=5 m) - 44 54

Reactor length m 4 5

Recycle flow kg/h - 799,673

Ethylene going to separation section kg/h 45,781 156,226

Catalyst mass kg 42,920 57,227

Natural gas imported kg/h 896,400 565,470

Ethylene production per year Mt/y 0.401 1.369

81 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.6.1 Transfer line heat exchangers and turbines for power production

After the OCM reaction, the cooling, compressing, purification, etc. have to take place in order to obtain polymer grade ethylene. The temperature at the outlet of the reactor should be around 800 °C, to immediately stop the reaction and prevent degradation of ethylene formed. Therefore is necessary to cool the reactor output gases, hereby preventing the secondary reactions from happening which would otherwise reduce the value of the OCM gas. This can be done via a similar system as used in the naphtha steam cracking process, see figure 37.

Figure 37: TLEs together with pumps, heaters and turbines in the OCM model.

In this figure you can see two transfer line heat exchanger (shell and tube heat exchangers) in which the inlet will be cooled by pressurized water just below saturation temperature, which will become steam after contacting with the heat from the product gas at the outside of the tubes. In this model, the same multistep cooling with two TLEs in series, syngas cooling is obtained by producing HPS (80 bar) and MPS (10 bar) and expanded in two turbines to 10 bar in the first turbine and eventually to 0.08 bar in the second turbine. The outlet from the first turbine, is mixed with the medium pressure steam at 10 bar in order to expand both streams together to 0.08 bar without pressure disturbances. Since no high hydrocarbons are present in the gas product, there is no problem of fouling in the TLES due to condensation of the higher boiling hydrocarbons.

Results TLEs: Steam generation and power production

The TLEs recover the heat contained in the gas, when the gas is already cooled to 800 °C. The results from the steam and power generation at these temperatures can be found in Table 19 below.

82 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 19: Transfer line heat exchanger and turbine results.

Without recycle With recycle

Parameter Unit TLE 1 TLE 2 TLE 1 TLE 2

Aspen component - EXC5-B EXC6-B EXC5-B EXC6-B

Temperature hot stream in °C 800 320 800 320

Temperature hot stream out °C 320 190 320 190

Temperature cold stream in °C 290 175 290 175

Temperature cold stream out °C 496.1 306.4 497.0 296.39

Pressure steam bar 80 10 80 10

Heat duty exchanged MW 632.60 144.79 721.15 165.89

Preheating heat duty MW 412.17 7.58 469.35 49.03

Pump duty MW 2.94 0.08 3.35 0.09

Unit Turbine 1 Turbine 2 Turbine 1 Turbine 2

Steam flow rate combustion kg/h 5.616 x106 5.616 x106 1.213 x106 1.213 x106 steam cycle

Total steam flow rate kg/h 6.732 x106 6.952 x106 2.484 x106 2.738 x106

Power production without MW 141.84 243.37 161.78 277.56 combustion steam cycle

Total power production MW 849.42 1250.33 316.87 496.48

Power production per kg of kWh/kg 0.127 0.182 0.128 0.181 total steam

Without recycle

If the results show amongst others that 200.88 MW of the power production is due to the expanding of the high pressure steam flow from 10 bar to 0.08 bar. As can be seen from the results, 1.116 x106 kg/h (310 kg/sec) of steam is produced at 80 bar due to the heat of the produced gas from the reactor, which produces 342.72 MW of electricity when expanded to 0.08 bar in two different turbines. While the medium pressure steam with an amount of 219,000 kg/h generates 42.49 MW of electricity. This results in a total of 385.21 MW electricity produced without the expansion of the high pressure steam generated from the combustion of all the remaining fuel from the separation process. The additional amount of power production of 1714.54 MW is due to combustion of the unconverted fuel. Which is a huge amount, since also a huge amount of fuel is combusted (562,555 kg/h of methane, ethane and hydrogen). This results in a total amount of 2.1 GW of electricity produced, while 139,259 kg/h of ethylene is produced.

83 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

The difference in kWh/kg of steam produced is higher for the second TLE since for expanding steam the pressure ratio of the in- and outlet important is for the power produced, the higher the more electricity can be produced. This ratio is in the first TLE 8, while it is 125 in the second TLE.

With recycle

In total 813.35 MW of electricity is produced in the second process, of which 374.01 MW is due to steam cycle production from the combustion of the purge stream of methane, ethane and the hydrogen resulting from the flash. The total steam rate is lower in the second process due to the fact that only 10 % of the fuel is combusted in this case.

Process 1 vs. process 2

In the first process, a total amount of 562,555 kg/h of fuel (methane, ethane and hydrogen) is combusted and in in the second process 163,455 kg/h of fuel. This is a factor, 3.44 while the ratio in steam production from combustion is a factor 3.90. This can be explained by the different composition of the fuel streams, 56 % CH4, 19.6 % C2H6 and 6.1 % H2 for the first process and 16.7 % CH4, 0.8 % C2H6 and 59.4 % H2. Methane combustion is higher exothermic than hydrogen combustion, and therefore more heat is produced due to different composition in the first process. The increase in preheating duty of the steam to the TLEs is 98.63 MW due to the higher flow rates, the increase in pump duty is 1.23 MW which is negligible compared with the other values. The total flow rates of the high pressure and low pressure pumps (before the TLEs, not the combustion) and the accompanying heaters do not differ much, which means the same size pumps and heaters will be necessary. The product flow rates through the shell and tube heat exchangers differ a factor 1.2 (2.648 x106 to 2.178 x106). In comparison, due to the much higher steam cycle flow rate for the combustion of the first process, the total flow rate in the turbines is much higher. 6.732 x106 kg/h compared to 2.484 x106 kg/h in the first turbine, which is a factor 2.7. For the second turbine the increase in volume is a factor 2.5.

4.6.2 Heater, flash and compression

See figure 38 for an overview of the flash, compression, AGR and chilling.

Results heater

After the second TLE the product stream is first further cooled to a bit above ambient temperature (40 °C) in order to flash most of the product gas from the water of which the water will go to the bottom. The heat duty released from cooling this stream from 190 °C to 40 °C, is 622.60 MW for the first process and 600.14 Mw for the second process. The heat required for the second process heat was expected to be bigger than the first process due to the higher flow rate. However not only the mass but the composition of the stream is also different due to the recycle and therefore also the heat capacity is different and in this case lower.

84 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 38: Chilling, hydrogen removal and compression.

Results flash

90.4 % is flashed for the first process which amounts to 37,602 kg/h, and 86.9 % for the second process of the water which amounts to 33,992 kg/h are flashed at 40 °C, both with 99.999 % purity. These amounts should be considered waste streams with some traces in the ppb range of the product gases, so purification will not be necessary.

Results compression with interstage cooling

Hereafter the gas is compressed and cooled in between via interstage cooling. This multistage centrifugal compressor with interstage cooling to 30 °C and has five stages with a pressure ratio of two at each stage, and four liquid knockout streams. The purpose of the multistage compression with interstage cooling is condensation of the remaining water while compressing the gas to prepare the product for purification by cryogenic distillation at high pressures. Table 20 shows the results from this multistage compressor with interstage cooling.

For the first process a total amount of 250.02 MW is required for the compressors, and 198.00 MW of cooling duty. For the second process a total amount of 341.40 MW is required for the compressors, and 273.09 MW of cooling duty. The common problem of condensation of the aromatics and C5-C9 hydrocarbons in the steam cracking process of naphtha is not the case with the OCM reaction since these components are not present in the product stream.

4.6.3 Acid gas removal and drying

After the fourth stage of compression, CO2 is removed. CO2 is present due to the composition of the natural gas, but also remaining carbon dioxide from the rich oxygen stream for oxidative coupling, and it is also produced during the process.

85 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 20: Multistage centrifugal compressor with interstage cooling results.

Without recycle With recycle

Parameter Unit Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 1 Stage 2 Stage 3 Stage 4 Stage 5

Temperature °C 30 30 30 30 30 30 30 30 30 30

Temperature profile °C 105.1 93.6 95.6 95.3 96.5 108.9 97.3 99.5 99.7 100.2 (before cooling)

Pressure bar 2 3.92 7.84 15.68 31.36 2 3.92 7.84 15.68 31.36

Vapor fraction - 0.9540 0.9913 0.9955 0.9978 0.9989 0.9540 0.9913 0.9955 0.9978 0.9989

Liquid knockout kg/h 49,074 8,835 4,534 2,236 - 67,133 12,091 6,206 3,064 -

Work duty MW 42.17 38.88 39.58 39.06 38.31 57.98 53.49 54.53 53.94 53.15

Cooling duty MW 80.86 43.76 42.03 41.17 42.20 110.43 60.02 57.57 56.20 57.14

Table 21: Multistage centrifugal compressor with interstage cooling of CO2 results.

Without recycle With recycle

Parameter Unit Stage 1 Stage 2 Stage 1 Stage 2

Pressure bar 52.31 87.33 52.31 87.33

Work duty MW 6.67 5.51 7.51 6.21

Cooling duty MW 12.59 37.55 14.19 42.28

86 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

In the first process 731,910 kg/h (203.3kg/sec) of carbon dioxide is present in the product gas fed to the acid gas removal unit of which 3.13 % was already present in the feed to the reactor. In the second process 824,054 kg/h (228.9 kg/s) of carbon dioxide is present in the product stream, of which 1.75 % was already present in the feed due to the natural gas composition. Most of the CO2 is produced during the OCM reactions, namely 96.87 % and 98.25 %. The ratio of CO2: C2H4 produced during OCM in the output stream are respectively 5.09 and 5.18. Which is significantly higher than naphtha steam cracker, however this is not an emission value. Since 100 % CO2 removal during acid gas removal is assumed and thus no recirculation or emissions of CO2 will happen in both processes, the emission would be due to flashing the CO2 during the flash of hydrogen and would thus be otherwise sent to the combustion reactor. The remaining water flow rate in the first process is 2624 kg/h and 3,603 kg/h in the second process, which is removed during drying. The acid gas removal and drying is performed by a separation block. This stream will thus be compressed and send to the final storage.

4.6.4 Multistage compression of CO2

Since there is no liquid left and the pressure inlet is at 32 bar, only two stages with a pressure ratio of

1.7 are necessary to compress the CO2 to 87.3 bar and 30 °C in order to liquefy the stream, hereafter the stream should be further compressed to 115 bar for transport and eventually storage.

Results compression CO2

See Table 21 for the results of the multistage compressors for both processes. The costs of CO2 compression are 12.18 and 13.72 MW for process 1 and 2.The total cooling duty required is respectively 50.14 and 56.47 MW for process 1 and 2. Both systems are assumed to have a 2 % pressure drop at each stage. At the outlet the CO2 has become liquid at 87.33 bar and 30 °C and could thus be compressed further with a pump to 115 bar, which requires much less energy than compressing a gas.

4.6.5 Cryogenic cooling and hydrogen removal

After the acid gas removal, cryogenic cooling from 30 °C to -170 °C at 32 bar takes place to remove most of the hydrogen in the product gas by flashing the hydrogen gas from the liquefied stream.

Results cooling and flashing

This cooling requires 148.15 MW of refrigeration duty in the first process and 198.18 MW in the second. During the flash respectively 89.2 % (11,941 kg/h) and 87.4 % (17,467 kg/h) of the hydrogen are removed. The stream leaving the overhead flash in the first process mainly contains hydrogen, namely

87.6 mole-% H2, 1.8 mole-% CH4, 3.6 mole-% CO and 6.6 mole-% N2. The stream leaving the overhead flash in the second process is almost similar and contains 80.99 mole-% H2, 1.6 mole-% CH4, 2.8 mole-%

87 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

CO and 12.8 mole-% N2. These overhead flash streams have a much lower purity than hydrogen flash with naphtha steam cracking, which is quite logical due to the high amounts of light components present in the product gas from OCM reaction compared with cracked gas from naphtha steam cracking. A large amount of carbon monoxide, argon and nitrogen are still present after flashing the gas at -170 °C which will be separated at the de-methanizer and will leave at the top stream.

4.6.6 De-methanizer

See figure 39 for the two distillation columns, the expanders and the combustion of the fuel.

Figure 39: Chilling, hydrogen removal and compression.

After hydrogen flashing, the liquefied bottom products from the hydrogen separator is pumped to the de-methanizer to 32 bar since some pressure drop had taken place during compressing and cooling. Hereafter the liquefied product is heated to the optimal feed temperature of the column, providing part of the refrigeration duty of the system. This column is used for the separation of all the remaining light components from the ethane and ethylene in the product gas with a purity of 99.99 %. The light components, of which methane is the main component present in the stream, emerges from the overhead of this column and could be used as a recycle stream to the reactor.

Results recovering cooling duty and de-methanizer

The operating condition from both processes for the two distillations columns can be found in table 22. From the bottom stream resulting from the flash, 106.52 MW for the first process and 140.61 MW of cooling duty for the second process can be recovered. This results in a feeding temperature of -50 °C. The overhead temperatures of the de-methanizers has a final purity of 58.0 mole-%.

88 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 22: Results and operating conditions columns OCM purification.

Without recycle With recycle

Parameter Unit De-methanizer C2 Splitter De-methanizer C2 Splitter

Number of stages - 50 120 50 120

Pressure bar 32 18 32 18

Temperature condenser °C -96.1 -32.7 -111.3 -32.7

Temperature reboiler °C -3.2 -11.5 -7.4 -11.4

Heat duty condenser MW -60.27 -45.43 -86.61 -50.76

Heat duty reboiler MW 30.34 59.06 46.99 65.45

Distillate to feed ratio - 0.698 0.704 0.817 0.812

Distillate rate kg/h 465,565 138,792 852,622 155,075

Reflux ratio - 1.5 3.5 1.5 3.5

Boilup ratio - 1.979 9.989 3.607 19.316

4.6.7 Ethane ethylene separation

The C2 fractionator requires a high reflux ratio and many separation stages and operates at high pressures.

Results C2 splitter and PE grade ethylene

There is a valve between the two distillation columns in order to release some pressure of the liquid stream, since the C2 splitter operates at 18 bar. For the first process this generates 0.16 MW and for the second process 0.15 MW. The heat exchanger thereafter, which cools to the feed temperature for the distillation column down to -30 °C, requires 5.54 MW cooling duty for the first process and 4.31 MW for the second process. As can be seen from these results, the first process now requires more cooling duty. This is due to the higher ethane amount in the stream resulting in a total flow rate of 201,440 kg/h in the first process and 190,979 kg/h in the second process. Respectively 99.90 % and 99.97 % of the ethylene are recovered from the feed stream to the column of both processes, which amounts to 138,768 kg/h of ethylene and 155,109 kg/h. These results show that in the first process a loss of 0.35 % took place due to separation and purification steps after the OCM reaction and in the second process a product loss of 0.7 %, of which the latter is almost 100 % in the recycle. Based on the feed flow rate and the resulting ethylene yield, the capacity is higher than the benchmark technology, 1.216 Mt of ethylene produced per year for the first process and 1.358 Mt of ethylene produced per year for the second process. The recovered ethylene purity from the model compared to literature can be found in table 23. The ratio of the scales (around 1.2) should be taken into account when comparing the two models.

89 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 23: Typical ethylene specifications (in ppm vol unless otherwise noted) and specifications of the ethylene from model.

Component Polymer grade Model without recycle Model with recycle

Ethylene 99.9 % min. 99.984 % 99.995 %

Methane 300 Trace* 2 ppb

Ethane 500 158.4 49.2

Propylene 10-15 - -

Acetylene 2 - -

Hydrogen 10 0 0

Carbon monoxide 2 0 0

Carbon dioxide 2 0 0

Oxygen 5 0 0

Water 2 0 0

*Trace implies there is a mole flow but the amount is less than 0.1 ppb.

As can be seen in the previous table, both the ethylene product streams have polymer grade purity and just two small contaminations of which none are higher than the specifications. Some assumptions made such as 100% CO2 separation during AGR are normally not allowed, however in this plant configuration the remaining CO2 would be flashed and water also separated due to the cryogenic separation. Another note on the purity is the natural gas configuration, which in this case does not contain any C3 or C4 components. Which means no propylene formation, which is a possible reaction via oxidative coupling of propane. Also no sulphur or acetylene is present or formed during the OCM.

4.6.8 Expanders

To generate more power for increasing the plants efficiency, the overhead stream of the flash, de- methanizer and C2 splitter could be connected to expanders. The power generated from these streams when expanded to 1 bar, is 3.06 MW for the expanding of the hydrogen stream, 22.1 MW for expanding methane, and 4.38 MW for expanding the ethylene stream. For the second process, these values are respectively 4.89 MW, 27.923 MW and 4.10 MW. However, ethylene is produced at the same conditions at which it could be stored, around -30 °C and 18 bar. Therefore this streams, will not be expanded but stored at these conditions. As for naphtha, the low temperature of gases after the expander will be recovered as refrigeration duty for the system.

90 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.6.9 Combustion

For the first case with no recycle, all the fuel is combusted. For the second process, 90 % of the methane and the ethane were first expanded, then mixed together with combusted together with the hydrogen, the rest is sent to the burners. Then combustion in both processes, is performed the same as the combustion in the naphtha steam cracker which combusts the recycle streams, producing HPS for the steam turbine. The exhaust leaves the furnace at a temperature of 450 °C, additional heat recovery includes the air preheating up to 300 °C and water economizer.

Results combustion

The results from the combustion outlet, fuel inlet, pump duty and heat released from the combustion to the steam cycle, can be found in Table 24. The results from the steam cycle power production after expansion in turbines can be found in Table 19.

The energy balance of the combustion section, in which it is important to notice that higher CO2 is emitted in the case without recycle as well as more heat is available for steam production. Fuel preheating from -166 °C to 300 °C requires 182.40 MW. In the Gibbs combustion reactor, the possible products are specified, and during combustion 5238.99 MW of heat is released to the steam cycle.

Table 24: Results from combustion and steam cycle.

Parameter Unit Without With recycle recycle

HPS (80 bar) kg/h 5.616 x106 1.213 x106

Temperature exhaust gas °C 221 228

Pump duty MW 14.80 3.01

Air flow rate kg/h 9 x106 2 x106

Fuel composition

CH4 kg/h 302,818 39,797

C2H6 kg/h 62,506 3,587

H2 kg/h 13,381 17,718

Total flow rate combusted kg/h 562,555 163,455 stream

Duty fuel preheating MW 182.40 57.76

6 CO2 produced kg/h 1.143 x10 147,116

H2O produced kg/h 912,655 254,333

Combustion heat MW 5238.99 1138.67

91 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.6.10 Total work requirement and power production

The results from the work requirement and power production of both models can be found in Table 25. It shows that the power produced during the OCM in the second model is sufficient for the power requirement of the ASU which is very energy intensive and requires 399.54 MW, and a residual amount of 51.27 MW is produced. In the first model, a huge amount of power is produced due to recycling all the fuel which results in an additional amount of 1498.93 MW of power which could provide electricity to multiple plants on site or be sold.

As can be seen, the exact same blocks are required for both models, and the main power requirement and production are all in the same blocks. They only differ in energy size due to different flow rates. In both models, the compressing and pumping electricity for both models required is quite high, mainly 627.04 and 692.44 MW of which 397.99 MW in both comes from the ASU.

4.7 Emissions and waste

Due to less water quenching, a feedstock with less components and specifically lower hydrocarbons, the amount of waste from the OCM is much less. On the other hand, less high value components are produced since the overall outlet composition of the reactor is not very diverse. However, there are CO2 contaminations in the feedstock and CO2 is produced during the OCM reaction scheme, therefore a lot of CO2 is present in the product gas. These amounts of CO2 are however no waste or emission since this stream is treated and stored. The CO2 produced during the combustion is however an emission stream. 6 For the first process this stream has a flow rate of 1.143 x10 kg/h while the second process has a CO2 emission of 147,116 kg/h. For the amount of C2H4 produced in these systems, this results in ratios of kg

CO2 per kg C2H4 produced of respectively 7.77 and 0.95. The first ratio is quite out of proportion due to the extreme large flow rate of the combustion fuel, the one from the second model is however lower than naphtha steam cracker.

Water

In both processes, water comes from separation steps, combustion, and steam cycles. The first process separates 2,664 kg/h of water during the acid gas removal and has 646,777 kg/h of liquid knockout due to compression at ambient temperature. Waste water from the acid gas removal is 3,606 kg/h, water from the liquid knockout due to compressing at ambient temperature results in 88,494 kg/h. These streams have very few contaminations therefore no purification will be required except for the water stream. This is a small stream compared to the other streams and will not result in high purification costs. The water formed during the combustion together with the CO2, nitrogen and remaining oxygen needs to be cooled to near ambient temperature and then purged to separate water from the gas. This is a quite large stream for both models and should be taken into account when providing and extensive techno-economic evaluation. For the first process and the combustion need to be considered.

92 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 25: Results from power production or requirement per block in model. Without recycle With recycle

Block Stream Power production Block Stream Power production (+) or required (-) (+) or required (-) [MW] [MW]

W7 +2.25 W7 +2.25

Turbine1 W8 +849.42 Turbine1 W8 +161.78

Pump2 W10 -0.02 Pump2 W10 -0.03

Pump3 W11 -2.94 Pump3 W11 -3.35

Pump4 W12 -0.08 Pump4 W12 -0.09

Comp6 W15 -197.99 Comp6 W15 -273.09

Expand W16 +3.06 Expand W16 +4.89

Expand2 W17 +22.10 Expand2 W17 +27.93

Pump5 W19 -0.16 Pump5 W19 -0.15

Turbine2 W20 +1250.33 Turbine2 W20 +553.69

Pump15 W31 -1.00 Comp10 W30 -13.71

Comp10 W34 -12.18 Pump15 W31 -1.13

Pump11 W70 -0.04 Pump8 W32 -3.01

Pump12 W154 -14.80 Pump7 W70 -0.04

Subtotal +1898.47 Subtotal +451.49 OCM OCM Subtotal -399.54 Subtotal -399.54 ASU ASU Total ASU + +1498.93 Total ASU + +51.27 OCM OCM

4.8 Sensitivity analysis recycling methane and ethane

When methane and ethane are recycled, the imported natural gas fed to the reactor could be decreased to reduce feedstock costs. As previous results established, a molar CH4:O2 ratio of 1.25 was the optimum. However, previous results also show that the amount of ethane at the reactor, see the results from the different compositions, inlet also positively increases the yield of ethylene and could thus result in a shift of optimum conditions at the same CH4:O2 ratio. In order to model the recycles, and keep the molar ratio the same due to change from methane recycle but also due to oxygen consumption

93 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

from ethane oxidation, a fortran script was used. This script calculated the new amount of imported natural gas while keeping the molar ratio at 1.35 and subtracting the amount of oxygen necessary for oxidation of the ethane and methane recycle. In order to calculate this script, the amount of oxygen from the ASU, the amount of methane and ethane in recycle and the molar fractions of oxygen and methane are imported to this calculator. Hereafter, the new amount of methane (in the imported natural gas feedstock) is calculated and exported to the model. Due to the ethane recycle the set of equations previously used to couple the Excel output to Aspen, required an adjustment to correct the exported values into Aspen. Due to the promising results from the second model, various sensitivity analyses will be performed on this model.

4.8.1 Influence of recycle on X, Y and S at p=1 bar and T=800 °C

The results of the reaction rates of the second model with recycle at standard operating conditions, can be found in the figures below, figure 40 to figure 45 on the next pages. Figure 46 and 47 show the influence of the recycle compares with the first model without recycle. Figure 40 below shows an overall overview of all the components in the reaction, H2O, CH4 and CO2 are the main components present at the outlet of the reactor.

Influence of recycle on mole fractions 60% O2 CH4 C2H6 50% C2H4 CO2 CO H2 H2O N2 40%

30%

20%

10%

0% 0 1 2 3 4 5 Length reactor [m]

Figure 40: Mole fraction for OCM reaction with recycle of fuel at standard operating conditions.

Results reaction rates

These figures show the rate with which certain components such as ethylene are converted and produced. In figure 42 it can be seen that from the reactions producing ethylene (5 and 7) only 5 is quite high at 200 mole/(m3s), compared to 600 mole/(m3s) as a maximum reaction rate, however this

94 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

reaction rate decreases 75% after 2.7 m in the reactor. This reaction consumes ethane and produces ethylene and water. Reaction 7 in which ethane is converted into ethylene and hydrogen, is almost negligible. Results however already showed, 50 % conversion of ethane is achieved and more ethylene is produced. Reaction which consume ethylene (6 and 8), of which 6 is negligible and 8 is at half the speed of reaction 5.

Influence of recycle on reaction rate 600 dr1 dr2 500 dr3 dr4

s)] 400 3 dr5

300 dr6 dr7 200 dr8 dr9 100

Reaction rate [mole/(m dr10

0 0 3 6 9 12 15 Length reactor [m]

Figure 41: Reaction rates for OCM reaction with recycle of fuel.

Figure 43 shows that reaction 4, the production of CO2 is very fast. All reaction rates stagnate decrease around the same point, and most of them become zero around 15 m in the reactor length at which oxygen is fully consumed. Except for reaction 9 and 10, the water gas shift reactions (WGS) which can result in more CO2 production, optimizing the reactor length and residence time can restrict this increase. Figure 44 shows that the reaction for methane consumption are all high, and especially the rate in which methane is oxidative coupled. The last figure are the reaction rates involving ethane (2,5 and 7) which are all involved in the production of ethylene and consumption of methane and are all high except for the direct dehydrogenation of ethane into ethylene as earlier mentioned.

95 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Reaction rates involving C H 2 4 Reaction rates involving WGS + CO2 250 600

/s) dr4

dr5 3 /s)]

3 200 500 dr6 dr9 150 400 dr7 300 dr10 100 dr8 200 50 100 0 0

0 3 6 9 12 15 Reaction rate [mole(m 0 3 6 9 12 15 Reaction rate [mole/(m Length reactor [m] Length reactor [m]

Figure 3: Reaction rates involving conversion of C2H4. Figure 4: Reaction rates involving the conversion of CO2, CO, H2 and H2O.

Reaction rates involving CH4 Reaction rates involving C2H6 600 250 dr1

/s)] dr2

3 500 /s)] dr2 3 200 400 dr5 dr3 150 300 dr7 100 200 100 50 0 0 0 3 6 9 12 15 0 3 6 9 12 15 Reaction rate [mole/(m Length reactor [m] Reaction rate [mole/(m Length reactor [m]

Figure 5: Reaction rates involving the conversion of CH4. Figure 45: Reaction rates involving the conversion of C2H6.

96 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Influence of recycle on X, S and Y 0.6

0.5

0.4

0.3

CH4 conversion 0.2 C2H6 conversion C2H4 selectivity C2H4 yield 0.1 CH4 conversion recycle C2H6 conversion recycle C2H4 yield recycle C2H4 selectivity recycle 0 0 1 2 3 4 5

-0.1 Length reactor [m]

Figure 46: Influence of recycle on X, S and Y.

Influence recycle on mole fractions 60% O2 CH4 C2H6 C2H4 50% O2 recycle CH4 recycle C2H6 recycle C2H4 recycle 40%

30%

20%

10%

0% 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Length reactor [m]

Figure 47: Influence of recycle on mole fractions.

Figure 46 shows that the fuel recycle results in a higher ethylene yield, a lower methane and ethane conversion but a higher ethylene selectivity. This will thus result, as can be seen in figure 47, in a higher methane mole fraction at the outlet, however the rest is almost the same.

97 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Reactor design and gas velocity

The 54 reactors were calculated with a diameter of 5 m accounts for a 1.366 Mt/y ethylene plant. For a factor 1.366 smaller plant, assuming the same conversion and yield, 40 reactors with a diameter of 5 meter would be required and a length of 5 m. Since complete conversion of oxygen takes place at 4-5 m in the reactor, with a gas velocity of 1 m/s, there is a residence time of 4 seconds.

4.9 Pressure sensitivity analysis

In previous simulations, different parameters have been investigated for instance the molar ratio of methane and oxygen, also for different process model namely with and without a recycle. However other parameters, such as the pressure, can also be of great influence for reaction kinetics. In this paragraph the influence of the pressure together with the most promising process model of the OCM, with the recycle of methane and ethane, will be investigated.

4.9.1 Pressure influence on reactor design

Since the model turned out to be unstable at high pressures, a relaxation factor was implemented in the Excel model for the WGS reactions. The use of this relaxation factor was validated after changing it for different simulations and no difference in the outlet molar fractions was observed.

Results pressure sensitivity analysis

Figure 48 shows that increasing the pressure, the methane conversion, ethylene selectivity and ethylene yield decrease. However the conversion of ethane increases significantly. Unfortunately, due to the decrease of methane conversion and selectivity decrease of ethylene, this still results in a decreased yield. This result however shows potential, perhaps changing the methane ethane ratio in the reactor would be a good idea to research. So, if the ethane recycle is used and maybe also an additional feed of ethane is fed to the reactor, ethylene yield could be higher. The overall optimum is at 2 bar of these four streams.

98 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Influence of pressure on XCH4, XC2H6, YC2H4 and SC2H4 80%

70%

60%

50%

40%

30% CH4 conversion 20% C2H6 conversion 10% C2H4 Selectivity

0% C2H4 Yield 0 4 8 12 16 20 24 28 32 Pressure [bar]

Figure 48: Influence of pressure on the X, Y and S.

Figure 49 shows a slight decrease in mole fractions of ethylene and ethane at the outlet of the reactor while changing the pressure. As the previous figure already suggested due to the lower ethylene yield and higher ethane conversion while varying the pressure. Both these figures also show that methane conversion decreases at higher pressures and thus more methane is present at the outlet of the reactor. Another interesting result is that at 2 bar, the ethylene yield is the same as at 1 bar but the ethane mole fraction is lower and the methane mole fraction is higher than at 1 bar.

Influence of pressure on outlet composition 30%

25%

20% O2 CH4 15% C2H4

Mole Mole fraction 10% C2H6

5%

0% 0 4 8 12 16 20 24 28 32 Pressure [bar]

Figure 49: Influence of pressure on the mole fractions at the outlet.

99 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 50-54 on the next page show the pressure influence on the yield of ethylene, conversion of methane and ethane and the ethylene selectivity at different pressures along the reactor length. These figures give some more detail on the advantages and disadvantages of operating at 1 or 2 bar in terms of X, S and Y. However this part of the reactor, ethane conversion and methane conversion are the highest at 4 bar and ethylene yield is not very different from the optimum.

When looking at the second half of the reactor the methane conversion, ethylene yield and ethylene selectivity are all the highest at a pressure of 1 bar. Except again for the ethane conversion which reaction rate is almost linear when operating at a high pressure of 32 bar and almost reaches 75 % conversion at the end and could even be higher since it did not reach its equilibrium just yet. Which suggests that at these higher pressures the seventh reaction, which is the dehydrogenation of ethane, dominates in ethylene conversion over the fifth reaction, which is the oxidation of ethane, because oxygen is almost all reacted away at this point. Also the following observation can be made:

- Increasing the pressure, the methane conversion decreases because of the pressure dependence in the reaction rates 1, 2 and 3 involving methane conversion.

- In case of ethylene yield, the decrease is due to decrease of methane conversion and ethylene selectivity.

- Due to increasing the pressure, the ethane conversion increases due to the pressure dependence in the reaction rates involving ethane conversion.

- Due to increasing the pressure, the ethylene selectivity decreases because the methane conversion decreases.

100 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Pressure influence on CH4 conversion Pressure influence on C2H4 yield 60% 60% 50% 50%

40% CH4 conversion 32 bar 40% CH4 conversion 16 bar 30% 30% CH4 conversion 8 bar 20% CH4 conversion 4 bar 20% C2H4 yield 32 bar C2H4 yield 16 bar 10% CH4 conversion 2 bar 10% C2H4 yield 8 bar C2H4 yield 4 bar CH4 conversion 1 bar C2H4 yield 2 bar C2H4 yield 1 bar 0% 0% 0 5 10 15 20 0 5 10 15 20 Length reactor [m] Length reactor [m]

Figure 50: Pressure influence on CH4 conversion. Figure 51: Pressure influence on CH4 conversion.

Pressure influence on C H selectivity 2 4 Pressure infuence on C2H6 conversion 60% 80% 50% 60% 40% C2H6 conversion 32 bar 30% 40% C2H6 conversion 16 bar 20% C2H4 selectivity 32 bar C2H4 selectivity 16 bar C2H6 conversion 8 bar 20% 10% C2H4 selectivity 8 bar C2H4 selectivity 4 bar C2H6 conversion 4 bar C2H6 conversion 2 bar 0% C2H4 selectivity 2 bar C2H4 selectivity 1 bar 0% C2H6 conversion 1 bar 0 5 10 15 20 0 5 10 15 20 Length reactor [m] Length reactor [m]

Figure 52: Pressure influence on C2H4 selectivity. Figure 53: Pressure influence on C2H6 conversion.

101 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.9.2 Pressure influence on plant configuration

Hereafter also the plan configuration in the Aspen model was adjusted to these pressures, such as the pressure outlet of the valve from the oxygen rich stream, the pressure outlet from the expanders of the top of the de-methanizer and the pump pressure from the bottom of the C2 splitter and of course the inlet pressure of the imported natural gas was changed so a compressor was added to the system. For instance due to operation at 32 bar, the expander coupled to the top of the de-methanizer becomes obsolete. So the pressure influence on the plant configuration can be quite high in terms of work electricity production, reactor design (residence time, number of reactors required for the reaction, gas velocity), etc. In order to this six variations were made in terms of plant configuration on the OCM with ASU and recycle, each plant configuration operating at a different pressure. The two distillation columns were optimized for each pressure in terms of providing polymer grade ethylene. The multistage compressor with inter stage cooling is adjusted in terms of stages (1 up 5) for the different pressure models, with a pressure ratio of 2 and a. intermediate cooling to 25 °C.

Results of pressure influence on plant configuration

The results from this analysis of pressure on the plant configuration can be found in table 26. The following things can be concluded from these results:

- Higher pressure results in lower volumetric flow rates, which results in lower gas velocities with constant reactor dimensions. Beside higher conversion of ethane, operation at 32 bar also result in a significant smaller reactor design. Only two reactors are required for the OCM with recycle at 32 bar due to the volumetric flow rate being a factor 31.7 smaller. For the other interesting case of 4 bar, still 14 reactors are necessary for this volume. However this is four times less than the 54 reactors required when operating at 1 bar. The number of reactors, mentioned in this analysis is an indication of the plant size and investment cost but the reactor geometry will result from a techno-economical optimization.

- The net heating-cooling duty required of the natural gas compressor is the highest for the 32 bar model, which differs a factor 3.36 with the 1 bar model. So almost an increase of 100 MW cooling duty is required because of the multistage compressor with inter stage cooling used at 25 °C. This cooling duty is not required for the 1 bar model, since the whole multistage compressor is not required.

- In all of the six different pressure models, power is generated at the plant due to HPS and MPS generation, but also from the expander at the top of the de-methanizer. However for the last case, with a pressure of 32 bar, this expander has become obsolete since the pressure of the top is 32 bar and this streams is recycled to the reactor. The loss of the turbine power generation and the additional power required for the compression, result in the lowest electricity production for the highest pressure. The trend is when the pressure is higher, the net electricity produced is lower.

102 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 26: Results on the plant performance and reactor design from pressure analysis.

Pressure [bar] Parameter Unit 1 2 4 8 16 32 NG imported kg/h 565,470 558,099 547,445 536,455 525,120 510,965 Volumetric flow rate reactor m3/s 1052.8 527.8 264.5 132.5 66.4 33.2 Length reactor m 20.0 20.0 20.0 20.0 20.0 20.0 Gas velocity m/s 214.2 107.4 53.83 26.97 13.50 6.763 Nr of reactors 54 27 14 7 4 2 (in case of d=5m, v=1m/s) Turbine top de-methanizer MW +27.93 +16.42 +13.53 +9.97 +5.55 -

Pump bottom C2 splitter kW +35.49 +31.58 +25.49 +16.50 +2.91 -21.53 Work compressing 1 kg NG W - 35.51 70.92 106.20 141.13 175.50 Compressor work duty NG MW - -39.60 -37.26 -56.99 -74.15 -89.71 Compressor cooling duty NG MW - -39.72 -37.38 -57.58 -75.84 -93.71 Work turbine HPS MW +313.41 +314.74 +311.77 +313.30 +315.81 +326.46 Work tubine MPS MW +499.92 +484.74 +485.14 +490.31 +498.26 +510.09 Condenser duty demethanizer MW -86.61 -87.30 -88.37 -90.55 -92.61 -96.00 Reboiler duty demethanizer MW +46.99 +47.44 +48.36 +50.53 +52.72 +56.65 Net heating-cooling duty MW -39.62 -79.58 -77.39 -97.60 -114.73 -133.06 Net electricity produced (+) MW +841.30 +776.33 +773.21 +756.61 +745.47 +725.31

103 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

4.9 Temperature sensitivity analysis

The last sensitivity analysis on the OCM with ASU and recycle performed, is the temperature sensitivity analysis. This analysis was performed to see the influence of the temperature on the conversion, selectivity and the yield. The kinetics in the Excel are calculated at fixed temperature and the Aspen model also uses a fixed temperature outlet in the reactor. During the simulation the temperature in the Excel was not changed, only in the first simulations the effect of the cooling of the reactor and the adiabatic reactor was studied. Since the reaction is highly exothermic and the Excel does not allow any temperature profile at the moment and neither does the Aspen model, the temperature influence was imitated by changing both values at the same time between 700 and 950 °C with steps of 50 °C.

Results of temperature sensitivity analysis

As can be seen in figure 54, there seems to be an optimum of ethylene yield at 810 °C. At this point also

CH4 and C2H6 conversion are high, the same C2H4 selectivity. However methane and ethane conversion appear to be the highest at 700 °C, unfortunately this again results in a lower ethylene selectivity and thus ethylene yield.

Temperature influence on X, Y and S 100% Ch4 conv 90% 80% C2H6 conv 70% C2H4 Select 60% C2H4 yield 50% 40% 30% 20% 10% 0% 700 750 800 850 900 950 Temperature [±C]

Figure 54: Temperature influence on X, S and Y.

The results of the exact values and the mole fractions can be found in Table 27. Each simulation was performed with oxygen completely reacted away. Because of WGS equilibrium CO production is higher at 950 °C, due to the reverse CO2, H2 and H2O production is much higher at 700 °C. The same accounts for ethane, due to the lower conversion at higher temperatures.

104 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 27: Temperature influence on X, S and Y and mole fraction at the outlet of the reactor.

Temperature [°C] 700 750 800 850 900 950

CH4 conversion 65.06% 52.12% 54.36% 52.00% 45.10% 40.93%

C2H6 conversion 80.60% 53.68% 51.34% 40.89% 44.89% 43.31%

C2H4 selectivity 31.05% 33.98% 40.14% 34.65% 24.48% 14.44%

C2H4 yield 26.65% 31.20% 36.94% 32.69% 22.21% 12.97%

O2 0.00% 0.00% 0.00% 0.00% 0.00% 0.00%

C4 12.31% 21.98% 20.96% 21.97% 24.37% 25.74%

CO2 20.16% 17.85% 16.99% 16.80% 16.94% 17.01%

H2O 35.12% 33.08% 35.02% 34.80% 33.05% 31.71% CO 2.72% 3.26% 3.18% 3.40% 3.69% 4.27%

H2 11.55% 11.08% 9.30% 7.89% 6.52% 5.93%

C2H4 3.56% 4.07% 5.01% 4.12% 2.45% 1.29%

C2H6 0.46% 0.92% 1.02% 1.03% 1.26% 1.32%

N2 14.11% 7.76% 8.52% 10.00% 11.72% 12.73% Qreactor [MW] 3010.836 3111.327 3017.626 2974.841 2947.648 2893.845

4.10 Pinch analysis

Also for the OCM with ASU model a pinch analysis was performed. From the composite curves, it can be seen that both hot and cold utilities are insufficient to complete the heating and cooling demand of the system. The key concept of pinch analysis is setting energy targets, targets for energy reduction. Targets obtained by pinch analysis are absolute targets, showing what the process is inherently capable of achieving if the heat recovery, heating and cooling systems are correctly designed [85]. The following formula is used for this concept:

T1 Q CPdT CP T T H (4.24) ³ 1  0 ' T0

With the CP as the heat capacity flow rate [kW/K]:

uCP m Cp (4.25)

With Cp as the heat capacity of the mixture in [kJ/(kgK)]. With CP assumed to be constant for a stream, required heat and total heat (Q) will be equal to the stream enthalpy change (̴H). So the pinch analysis is actually an overview of enthalpy changes of streams. Streams are considered flows which require heating or cooling but do not change composition. Consequently, the reaction process in the naphtha steam cracker and in the OCM reactor are not streams, because it changes composition.

After calculating the heat capacity flow rate for the streams, the heat load, and combining them together with the flow rate, inlet and outlet temperature of the streams, a pinch analysis could be

105 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

constructed. A special Excel sheet was made, which with these streams and results which could be exported to a different software programme called Aspen Energy Analyzer, which is a bit outdated and thus not have a lot of options. However, when all the results are properly imported, it gives the pinch analysis with the cold and hot composite curve, the pinch temperatures and the amount of cooling and heating duty required if necessary.

Results pinch analysis

The resulting T/H plot has two curves, see figures 55-59, one is the hot composite curve and the other one is the cold composite curve. Of which the cold streams represents the cold composite curve which require heating, and the hot streams represent the hot composite curve which require cooling. The overshoot at the bottom of the hot composite curve represents the minimum amount of external cooling required and the overshoot at the top of the cold composite curve represents the minimum amount of external heating required.

As can be seen from the result, the second model with recycle required the highest heating target and highest cooling target due to the higher flow rates. The ASU does not require any heating duty, but requires however a big amount of cooling duty.

Table 28: Heating and cooling targets for the two main OCM models and the ASU.

Without recycle With recycle

Unit OCM + ASU OCM ASU OCM + ASU OCM

Heating target x108 [kJ/h] 8.281 8.281 Sufficient 8.406 8.406

Cooling target x109 [kJ/h] 3.746 1.783 1.963 3.846 1.883

Pinch T hot [°C] 190 190 - 190 190

Pinch T cold [°C] 180 180 - 180 180

4.11 Aspen Plus model

In order to produce the model some assumptions have been made, for example:

x No pressure drop will take place in the reactor x Cooling in the TLEs is performed in 0.1 s x No C3 or C4 components are present in the feedstock and thus no oxidative coupling of propane takes place x Argon separation in the cryogenic air separation unit is not necessary for the OCM Aspen fluid packages usable for the separation train, and light hydrocarbons are Peng-Robinson and RK- Soave. Peng-Robinson is eventually used for the ASU and OCM model.

106 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Figure 55: Pinch analysis OCM + ASU without recycle. Figure 56: Pinch analysis for OCM without recycle. Figure 57: Pinch analysis ASU.

Figure 58: Pinch analysis OCM with recycle. Figure 59: Pinch analysis OCM without recycle.

107 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

5 Techno-economic analysis

In this paragraph an analysis on the economics of the three main processes, namely the naphtha steam cracker, the OCM with ASU without and with recycle of the remaining fuel. For this economic evaluation, the feedstock, product and energy prices were researched and some other data such as the natural gas density in this composition, and the price factors in case of the BTX and C4 mixture. In case of the feedstock and the product, it is simply a matter of the price per unit feedstock or credit per unit product times the flow rate. For the power production the following equation (5.1) was used based on an article from Ulrich et al. [86]:

C a CEPCI b C (5.1) S, u  S, f

Where CS, u is the price of the utility [$/kWh], a and b are utility cost coefficient, CS, f is the price of fuel [$/GJ] and CE PCI (Chemical Engineering Plant Cost Index) is an inflation parameter for projects in the U.S. Since this formula uses an inflation factor in or the U.S., also U.S. feedstock prices are used (since U.S., Indonesian and Russian natural gas prices can differ a factor 3). The a and b for onsite power charged to process module (power generation for onsite use), is used which are 1.4 x10-4 and 0.011. There are also a and b factors for exporting to another completely new plant (grass root plant), the calculations for both situations were made, however they did not differ much and will thus not further discussed. Different factors are provided by the paper for the CS, f , for instance for power production from coal.

Since the first plant combusts the recycle, mostly containing ethane and some higher hydrocarbons naphtha as fuel factor is used and natural gas for the other two models, which are respectively 11.5 for naphtha and 11 for natural gas. The density of the natural gas used is 0.6972 kg/m3. The flow rates of the models are 350, 736 and 416 t/h. During the OCM no other HVCs were produced, with naphtha steam cracking also propylene a mixture of BTX and C4 are produced.

Since price data from November 2015 were used, also the CE PCI for that year was used which is 579.8. The purification factors, because further purification is still required for these streams, are already included in the BTX and C4 mixture prices. Since the feedstock prices of natural gas, naphtha and ethylene fluctuate very often and also due to the different energy prices, a sensitivity analysis was performed with different pricing scenarios. For the natural gas, naphtha feedstock, ethylene prices an increase and decrease of -20 % and +20 % are considered, for the electricity prices an increase and decrease of 50 % is considered. Since power generation from coal can be 6.4 cents per kWh, while it could be 20.2 cents per kWh for naphtha.

108 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 29: Prices components and energy, with increase and decrease of 20 or 50 %.

Component Unit Value after Value base Value after decrease case increase

Naphtha $/t 440 550 660

Natural gas $/m3 0.1082 0.1352 0.1622

Ethylene $/t 960 1200 1440

Propylene $/t 800 1000 1200

Mixed BTX $/t 540 674 809

Mixed C4’s $/t 824 1030 1236

Electricity price naphtha model $/kWh 0.104 0.208 0.312

Electricity price natural gas models $/kWh 0.101 0.202 0.303

5.1 Base case economic analysis

For the base case the following results are obtained, which can be found in table 30. It shows that the OCM without recycle has the highest credit due to the high power production. From the ethylene production credit minus the feedstock cost there is even a negative value meaning that without recycle of fuel or fuel combustion for power production the first OCM model would not result in a profit. It can also be concluded from this table that the naphtha cracker is performing better in terms of ethylene and also mainly due to the HVCs credit, otherwise the naphtha cracker would not be profitable. The OCM with ASU and recycle does not require HVCs to compensate for the feedstock cost. The additional power production adds to the total credit in this economic analysis. However this second OCM model has still a factor 1.7 lower profit compared to naphtha due to no HVCs credit and less power production.

Table 30: Results from the economic analysis of the base case.

Component Unit Naphtha OCM + ASU OCM + ASU cracker without recycle with recycle

Cost feedstock X103 $/h -192.50 -142.70 -80.67

Credit ethylene X103 $/h 138.00 138.00 138.00

Credit HVCs X103 $/h 122.30 - -

Credit power production X103 $/h 45.55 249.08 8.94

Raw materials and products X103 $/h 67.80 -4.72 57.33

Total X103 $/h 113.35 251.53 66.27

Total X106 $/y 992.35 2203.42 580.56

109 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

5.2 Sensitivity analysis on economics

When ethylene prices increase or decrease, it is assumed that the other HVC prices increase and decrease at the same time with the same factor and that the electricity price is simultaneously decreasing and increasing in each scenario. The latter is due to the fact that the electricity price for different fuels, when selling, is always the same (the product, namely electricity is the same). For this reason and the fact that these prices do not differ that much, the electricity price will be the same price (0.208 kw/h, for the base case). So three variables, increasing, staying at the base case or decreasing leads to 81 scenarios for each model (since both natural gas prices will go up simultaneously). In order to keep the overview, first the electricity credit with the feedstock cost was varied and the product credit was kept constant at the base case which reduced the number of scenarios to 27. Hereafter, the product credit with the feedstock cost was varied and the fixed electricity price was kept fixed at the base case value. Since the latter will not vary much compared to the other two variables, since electricity is bought and sold with quite steady year contracts.

Results economics

The results from the sensitivity analysis described above can be found on the next page in table 31 in which the scenarios are numbered and in table 32 the results can be found of these scenarios. In the second table the course from low to high values is coloured from white to dark green, the yellow results in the middle are the base case values. Some of these scenarios are highlighted in the first table and will be discussed. A lot discussion can be done on this table, however most of them are quite logical:

- Lower feedstock prices, higher electricity prices and higher credit for ethylene (and other HVCs), result in a higher net credit. - The first OCM model is in all scenarios higher than the second OCM model and the naphtha steam cracker, except for scenario 4 compared to scenario 1 (orange scenarios). This scenario has a low electricity price and low feedstock price, so even though almost six times more power is produced is the benchmark technology better in terms of this economic analysis. - Two other interesting results are scenario 9 and 36 from the OCM with recycle compared with scenario 3 and 30 from the naphtha steam cracking (the grey and blue scenarios). Both of the coupled scenarios differ almost nothing, creating a potential for a scenario where the OCM with recycle is more profitable compared to naphtha steam cracking. Both scenario pairs have a high feedstock price, however the first pair (30 and 36) has a lower ethylene and HVC price and the second pair (3 & 9) has a decreased electricity price. - Since OCM with recycle not have any HVCs except for ethylene and has only a small additional power production compare to the naphtha steam cracking, thus lower electricity and HVC prices would have a much bigger impact on the naphtha steam cracking than the OCM with recycle. - Which should also be mentioned is the small amount of feedstock required, 350 t/h for naphtha and 416 t/h of natural gas in the OCM to achieve the same ethylene production. The methane and ethane yield decreases the imported amount of natural gas in such a way that the feed stock flow rate does not differ much from naphtha steam cracker. Next to that, if the same mass flow rated would be considered then the feedstock cost would differ a factor two exactly under these conditions.

110 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

Table 31: Numbers of the different scenarios for the sensitivity analysis on economics.

Model 1: Naphtha steam cracker Model 2: OCM + ASU without recycle Model 3: OCM with ASU with recycle F = 0.1082 F = 0.1352 F = 0.1622 F = 0.1082 F = 0.1352 F = 0.1622

F = 440 $/t F = 550 $/t F = 660 $/t $/m3 $/m3 $/m3 $/m3 $/m3 $/m3 E= 0.104 $/kWh 1 2 3 4 5 6 7 8 9 E= 0.208 $/kWh 10 11 12 13 14 15 16 17 18 E= 0.312 $/kWh 19 20 21 22 23 24 25 26 27

C2H4= 960 & E = 0.208 $/kWh 28 29 30 31 32 33 34 35 36

C2H4= 1200 & E = 0.208 $/kWh 37 38 39 40 41 42 43 44 45

C2H4= 1440 & E = 0.208 $/kWh 46 47 48 49 50 51 52 53 54

Table 32: Results sensitivity analysis on economics for three models.

Model 1: Naphtha steam cracker Model 2: OCM + ASU without recycle Model 3: OCM with ASU with recycle F = 0.1082 F = 0.1352 F = 0.1622 F = 0.1082 F = 0.1352 F = 0.1622

F = 440 $/t F = 550 $/t F = 660 $/t $/m3 $/m3 $/m3 $/m3 $/m3 $/m3

E= 0.104 $/kWh 1130.72 793.46 456.20 1330.70 1081.02 831.34 682.51 541.38 400.26 E= 0.208 $/kWh 1330.24 992.35 655.72 2453.10 2203.42 1953.74 721.68 580.56 439.43 E= 0.312 $/kWh 1529.76 1192.5 855.24 3575.50 3325.82 3076.14 760.86 619.73 478.61

C2H4= 960 & E = 0.208 $/kWh 874.63 537.37 200.11 2211.33 1961.64 1711.96 479.91 377.96 197.66

C2H4= 1200 & E = 0.208 $/kWh 1330.24 992.35 655.72 2453.10 2203.42 1953.74 721.68 580.56 439.43

C2H4= 1440 & E = 0.208 $/kWh 1786.4 1449.14 1111.88 2694.88 2445.20 2195.51 963.46 861.51 681.21

111 Energy analysis and plant design for ethylene production from naphtha and natural gas Technische Universiteit Eindhoven University of Technology

6 Summary and Conclusions

From these green routes only bio-ethanol from biomass, e.g. sugarcane, to ethylene (BETE) is momentarily commercially available, profitable and capable of producing high quantities of ethylene. The highest capacity BETE plant in the world produces 200,000 t/y, which is still only 20 % of the capacity of an average naphtha steam cracker plant. Another disadvantage is the huge agricultural land requirement to grow the crops. Fossil fuel based routes deliver multiple commercially available processes, such as coal to olefins (CTO) via syngas to methanol production. From an economical point of view, together with coal abundance this is a good option. However, from an environmental point of view, this is a step back due to the higher CO2 emissions. That leaves natural gas and crude oil, of which the latter is refined into naphtha. Shale gas revolution and consequently low prices of natural gas, and the large natural gas reserves result in several studies regarding the direct utilization of methane. One of the promising process technologies, is the oxidative coupling of methane (OCM) to ethylene as an alternative.

In this thesis the following plant configurations are modelled in Aspen Plus with success:

- Model 1: The benchmark technology of the naphtha steam cracker - Model 2: The oxidative coupling of methane with air separation unit (ASU), without fuel recycle - Model 3: The oxidative coupling of methane with air separation unit, with fuel recycle

Naphtha steam cracking

A naphtha steam cracker plant with de-methanizer first and tail end hydrogenation with a capacity of 1 Mt/y of ethylene was modelled. From feeding the liquid naphtha with a 350,000 kg/h feed up to the cryogenic separation and purification of the product gas. The energy, engineering and environmental analysis of this modelled plant configuration conclude that this model was successfully modelled according to industrial operation, in terms of energy efficiency, compressor power, waste, CO2 emissions, ethylene and HVCs yield, polymer grade ethylene production, HPS generation, fired duty of the cracking furnace and fuel gas import. This process was optimized for ethylene production and resulted in similar specific energy consumption of 22,759 kJ/kgethylene, according to literature values

(23,000 kJ/kgethylene) for a similar capacity plant configuration. An amount of 115 t/h with a purity of 99.94 % polymer grade ethylene is produced, with contaminations of 507.7 ppm ethane, 34.3 ppm acetylene and 56.2 ppm hydrogen. These values are a bit higher than typical polymer grade ethylene specifications. All product streams are considered, some of the products are sold (BTX, the C4 mixture and propylene), others are burnt for additional power production (the recycle, methane and hydrogen) and others are recycled (e.g. water and fuel oil).

112 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

Oxidative coupling of methane

For the second and third model, a plant configuration for the OCM reaction with a cryogenic ASU was modelled with capacities of 1 Mt/y of ethylene and higher, from feeding the natural gas and pure oxygen resulting from the ASU up to the separation and purification of the product gas. For the reactor design, a packed bed reactor was modelled in Excel with an OCM reaction mechanism including ten reactions and kinetics from Stansch et al., similar results according to the paper were reproduced. After this validation, the Excel reactor model with kinetics were coupled to the Aspen Plus model in order to reproduce the results in the reactor design of the Aspen Plus model.

Cryogenic ASU

The cryogenic ASU was modelled, including an optimized heat exchanger network and combined condenser-reboiler operation to increase energy efficiency. The energy analysis of the ASU concludes that 97.89 wt-% of the oxygen is recovered with a purity of 95.4 mole-%. The electricity required per kg of oxygen consumed is 0.310 kWh/h which is lower according to Spallina et al. (2014), but higher compared to a conventional cryogenic ASU from Linde Engineering (0.232 kWh/kg of oxygen consumed). A total amount of 379.42 MW is required in order to operate this plant.

Together with the ASU and kinetic model, the OCM plant configuration was modelled. The standard conditions used in the model are 800 °C, 1 bar, complete conversion of oxygen and a La2O3 (27 % at.)/CaO catalyst. First, simulations with low flow rates were performed to get preliminary results and work out the separation and purification section, hereafter the plant was scaled up and different variations on model 2 and model 3 were done to study potential process schemes and reactor designs.

OCM - without recycles

First the OCM without any recycles to perform a sensitivity analysis on one of the parameters important to OCM reaction, namely the methane to oxygen ratio. This parameter was varied, while analysing the methane conversion, C2+ selectivity, and C2+ yield. The optimum methane to oxygen ratio was found at 1.25 in terms of methane conversion, ethylene yield and ethylene selectivity. The selectivity shows a

C2H4 yield of 20.76 % and for C2H6 yield of 4.33 %, and a selectivity respectively of 35.56 % to C2H4 and

7.43 % to C2H6 at these operating conditions, according to Stansch et al. ranges. This OCM process model used a 2.16 Mkg/h natural gas feed to produce 139.3 t/h of ethylene of polymer grade purity, resulting in a 1.220 Mt/y ethylene plant. The flow rates in this model are too high (175.8 m/s) to operate this flow in one reactor, for a 1 Mt plant configuration with these flow rates, 9 reactors with a diameter of 5 m and a reactor length of 5 m are required.

113 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

OCM – adiabatic vs cooled reactor

Hereafter the reactor design was adjusted in order to study the heat management in the reactor. Since heat management is a serious issue in the OCM reaction. A reactor design studied in case of a completely cooled reactor and in case of an adiabatic reactor, the latter was performed to study the temperature rise in the reactor. For the completely cooled reactor the cooling duty required would be 3175 MW. For the heat management in the reactor also a third option was studied, with a recycle of the product gas implemented in the plant configuration. In order to keep the temperature outlet of the reactor at 800 °C, via a cooled (500 °C) product gas recycle, this resulted in a recycle containing 86.2 % of the product gas. The recycle results in a catalyst mass of 214,599 kg, with only 1,786 mole/s going to the purification section. The cooling of the recycle requires 2594 MW.

OCM - without recycle and combustion

The second model 896,400 kg/h of imported natural gas and 1.286 x106 kg/h oxygen to the reactor, which results in a 139,259 kg/h of polymer grade ethylene production with a purity of 99.984 %. This model does not recycle any product gas, and combusts the separated methane, hydrogen and ethane which has a flow rate of 562,555 kg/h, for additional power production. The additional amount of power due to this total combustion is 1714.54 MW, together with the power production from HPS and MPS a total power production of 2.1 GW is produced. A methane conversion of 58.36 % was achieved with a ethane conversion of 50.14 %, an ethylene selectivity of 35.57 % and ethylene yield of 33.1 %.

OCM - with recycle and combustion

In the third model, the recycle of separated methane and ethane is implemented. In order to keep the molar ratio of methane to oxygen the same, a calculator was used to calculate the amount of imported natural gas based on the recycle flow rate. This resulted in a feed flow rate fed to the reacted of 1.365 x106 kg/h, which is a mixture of imported natural gas (565,470 kg/h) combined with the recycle, and 1.286 x106 kg/h oxygen. After the OCM reaction, 1555,109 kg.h of ethylene was produced with a purity of 99.995 %. 10 % of the recycled stream is purged to prevent build-up of contaminations present in these streams. The purge and hydrogen are combusted for additional fuel, which results in an additional of 813.35 MW together with the 374.01 MW due to HPS and MPS. A methane conversion of 54.38 % was achieved with an ethane conversion of 49.99 %, an ethylene selectivity of 40.22 % and ethylene yield of 37.0 %. Due to the recycle 3.9 % more ethylene was produced with a higher selectivity but lower methane conversion, compared to the simulation without the recycle. The recycle, and thus total higher flow rate, results in 91.40 MW additional power required for the compressors.

Carbondioxide compression for storage requires a total compressor duty of respectively 12.18 and 13.72 MW for process 1 and 2.The total cooling duty required is respectively 50.14 and 56.47 MW for process 1 and 2. This results in 0.017 kWh/kg of CO2 compressed for both processes. The cyrogenic

114 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

cooling requires 148.15 MW of refrigeration duty in the second process and 198.18 MW in the third. During the flash respectively 89.2 % (11,941 kg/h) and 87.4 % (17,467 kg/h) of the hydrogen are removed. From the bottom stream resulting from the flash, 106.52 MW for the second process and 140.61 MW of cooling duty for the third process can be recovered. The power generated from the expander connected to the top of the de-methanizer when expanded to 1 bar, is 3.06 MW for the expanding of the hydrogen stream, and 22.1 MW for expanding methane. For the second process, these values are respectively 4.89 MW, 27.923 MW and 4.10 MW. Ethylene is produced at the same conditions at which it will be stored, around -30 °C and 18 bar. The CO2 produced during the combustion is however an emission stream. For the first process this stream has a flow rate of 1.143 x106 kg/h while the second process has a CO2 emission of 147,116 kg/h. For the amount of C2H4 produced in these systems, this results in ratios of kg CO2 per kg C2H4 produced of respectively 7.77 and 0.95. The net power production for the second model is 1501.02 MW and for the third model this is 54.05 MW.

The analysis on reactor design for the third model concludes 40 reactors with a diameter of 2.5 m and length of 5 m are required for these flow rates (for a 1 Mt/y ethylene plant). The pressure sensitivity analysis for the third model shows that the pressure overall has a decreasing factor on the methane conversion, ethylene selectivity and ethylene yield. However the conversion of ethane increases significantly. The overall optimum is at 2 bar. The results of the pressure sensitivity analysis on plant configurations concludes that higher pressures lead to lower volumetric flow rates and therefore less reactors necessary. Only two reactors are required with these flow rates at a pressure of 32 bar. The net heating-cooling duty required is the higher for higher pressure, an increase of 100 MW cooling duty is required because of the multistage compressor with inter stage cooling used at 25 °C for the 32 bar simulation. This cooling duty is not required for the 1 bar model, since the whole multistage compressor is not required. In the 32 bar simulation, the expander at the top of the de-methanizer becomes obsolete. The trend is when the pressure is higher, the net electricity produced is lower.

Economics

The economic analysis concludes that the second model has the highest credit due to the high power production. It also shows that lower feedstock prices, higher electricity prices and higher credit for ethylene (and other HVCs), result in a higher net credit. The second model is in all scenarios higher than the first and third model, except for a low electricity price and low feedstock price. At a high feedstock price in combination with a lower ethylene and HVC price, or in combination with a decreased electricity price, there is almost no difference between the first and third model.

Final conclusion

Thus it can be concluded that the third model has potential in comparison with the first model in terms of energy and economic analysis based on the achieved yield and conversions for a 1 Mt/y ethylene plant. However this conclusions is based on the assumption that the produced heat in the reactor can

115 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

be removed. This is not achievable with a completely adiabatic reactor. The second model is not optimal or logical to perform, because you completely rely on the cost of electricity since the model is not profitable without the power production.

To conclude, the OCM with recycle of fuel and 10 % fuel combustion is an alternative for ethylene production when performed at higher pressure and thus decreasing the amount of reactors necessary to 2. The yield decrease is probably compensated by the fact that less equipment is required. In terms of CAPEX, has the separation section of the OCM far less components than naphtha steam cracker. No primary fractionator, oil quench, gasoline stripper, de-ethanizer, de-propanizer, or de-butanizer are required. However the three distillation columns and other equipment from the cryogenic ASU are required. Further techno-economic analysis on these two plant configurations should thus be performed. Other advantages of OCM over naphtha steam cracking are the lack of high hydrocarbon an thus less fouling in compression equipment, also a higher ratio of kg CO2 per kg of C2H4 is produced. From the raw materials economic analysis, it shows that with current feedstock prices OCM has potential. However compared to naphtha steam cracking overall yield (wt ethylene/wt feedstock) is much lower for OCM with natural gas. Naphtha steam cracking yield in not very high, compared to ethane cracking, however it also yields other HVCs in contrast to OCM.

116 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

7 Recommendations

As mentioned before, this thesis project is part of a four year European project, with its main focus on the air separation through novel Mixed Ionic Electronic Ceramic (MIEC) membranes integrated within a multifunctional unit and hopefully achieving higher yields compared with the conventional reactor for the OCM and benchmark technology modelled in this thesis. MEMERE will develop new O2 selective (supported) membranes which will be tested in this department.

The following two plant configurations are suggested:

1. As mentioned above, the study will eventually focus on novel porous membranes. For comparison, it would be good to model a base case of a plant configuration together with a porous membrane reactor instead of a packed bed reactor modelled in this study. This plant configuration could be studied considering feeding O2 from a cryogenic air separation unit to the membrane reactor. In summary, the plant configuration could be entirely the same as the third model however instead of a different reactor model with distributive feeding of oxygen.

2. The second plant configuration will include the complete OCM part again and a membrane reactor, only this membrane reactor will be a dense MIEC membrane reactor capable of air separation into high purity oxygen. While separating oxygen, the MIEC membranes will distributive feed the oxygen to the process of the OCM reaction. In this model, the cryogenic air separation becomes obsolete, allowing integration of different process steps in a multifunctional unit.

In this case, the distributive feeding of oxygen can be studied, and the advantages and disadvantages of a (porous) membrane reactor. Which could be later on compared to the model with (dense) MIEC membranes with distributive oxygen feeding. These MIEC membranes, could however make the cryogenic ASU in the first model obsolete if they turn out to be efficient enough and can produce high purity oxygen.

Research on dense and porous membrane reactors combined with the OCM reaction have been performed, however not with MIEC membranes. From the previous research results, some preliminary conclusions and or expectations can be mentioned. The oxygen membranes are expected to have an immediate result together with the high temperature reactor module on the increased ethylene yield because of the distributive oxygen feeding and temperature control. The conversion and selectivity generally increases, as the partial pressures of the reagents decreases. Therefore could oxygen distributive feeding and consequently lower partial pressure of oxygen, result in increased ethylene selectivity and yield and decreased the generated heat. This could thus result in overall improved plant efficiency and avoidance of costly cryogenic air separation unit which are used in competing technologies. Meanwhile downstream separation units could be simplified and or reduced in volume or operating costs, together with improved energy efficiency.

However, under certain operating conditions, a similar problem as observed in oxidative dehydrogenation of ethane could be present in OCM, e.g. undesirable oxygen accumulation phenomenon inside the catalyst bed may appear which can be minimized by carefully adjusting the

117 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

operating conditions. The presence of O2 at the reactor inlet significantly improves the ethylene production rates. However, if higher amounts of O2 are fed to the reactor inlet, the selectivity could drop, which may lead to hot spots near the inlet.

Some of the challenges for this future work will be scaling up the plant configuration with the membrane reactors, and improving the selectivity and methane conversion at the same time. Since the result from this thesis show that the number of reactors required for handling these huge flow rates, is high. Another challenge, which this study also showed, will be handling the heat management.

These models could be coupled together with the results from experimental work on the novel technology of air separation in MIEC membranes, such as the performance of the membranes, the oxygen permeation, the oxygen purity, etc. The latter will probably not be a problem since research on ceramic membranes shows results of oxygen purity’s above 90 % oxygen, however they require higher temperature to operate than for example polymeric membranes which can operate at ambient temperatures but can only produce 25-50 % oxygen enriched air. Next to this implemented model, which could be again made in Excel and coupled to the reactor design in Aspen Plus. This Excel file would then use the kinetics in combination with a membrane reactor mechanism together all the characteristics of the membranes. A similar study on economic, energy and environmental analysis could be performed for these two models in order to compare them to the three main plant configurations modelled in this study.

A third suggestion for a plant configuration could be a fluidized bed reactor, since previous research also show, that the heat management in these kind of reactors is much better than in a PBR. However, using another reactor would again involve a cryogenic ASU. The latter is perhaps not entirely necessary, since other non-cryogenic ASUs could be modelled for instance the oxygen MIIEC membranes mentioned before or pressure swing adsorption which operates at ambient temperature with a zeolite exposed to high pressure air. The latter however has the disadvantage of its smaller size. Beside fluidize bed reactors, and different non-cryogenic ASUs. Different plant configuration, also considering heat management at the same time, could be modelled. For instance multiple reactor in series with intermediate cooling by cold streams, perhaps even integrated with another plant with cryogenic cooling and thus using their cold streams and its cooling duty.

Next to these recommendations, a more elaborate OPEX and CAPEX economic analysis on the different components of the plant configurations should be performed. This however, is a study on its own and thus out of the scope of this thesis.

Finally, a recommendation towards other project and process technologies in case the MIEC membranes turn out to be very advantageous. For instance the MEMERE project can be extended to other partial oxidation processes such as methane auto thermal reforming, as the process faces the same challenges and the advantages of the novel approach are similar.

118 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

8 Acknowledgements

Before starting the master thesis, I did not know what to expect from the people, the project and the fact that you would be busy for nine months on one subject. I couldn’t be more happily surprised about all of it. The SMR group is great and time flies when you have fun and work hard. Moreover I thought it was exciting to be working on one subject and becoming an expert on your own subject. In the end, I can say that I learned a lot of things of which I didn’t know anything before and can be proud of my achievements and delivered work. Since a project is never performed alone, I would like to thank the people who had a contribution one way or the other during my thesis project.

First of all I would like to thank Martin van Sint Annaland for giving me the opportunity to do my master thesis within the SMR group and for his valuable and creative suggestions during the meetings. I also would like to thank Fausto Gallucci for pointing out pitfalls and providing and bringing me along to the kick-off meeting and to Brussel, together with Aitor, for the second MEMERE meeting. It was an amazing experience to be part of such a huge and innovative European project. I also would like to thank Timothy Noël for being my external committee member and Ildefonso Campos Velarde for coming on board. A special thanks to Vincenzo Spallina for supervising me on this project and providing me ideas from an energy engineering point of view.

I also would like to thank all the (PhD-) students from the SMR group for providing the awesome atmosphere and the lovely times together during coffee breaks, dinners, the Meerkamp, borrels, etc. Special thanks to Annelies, Beatrice, Fabio, Floris, Francesca, Gianmaria (my one and only neighbour), James, Jan-Willem, Jeroen, Martin, Michel, Tom and even Rik. And of course Judith and Ada, our wonderful secretaries. I will never forget these nine months, and all the stories.

I would like to express my deep gratitude to Michel for all the amazing times together! I always thought people were talking nonsense when they said you make friends for life during your time at the University. Never say never. You are “mijn mattie”, already for seven years and definitely more to come.

Last but definitely not least, I want to thank my family, my family in law and my boyfriend. They have always been there for me and were always cheering for me on the side lines. Thank you mom and dad, for the constant support and love and for always bringing me back to reality when I was completely stressed out. Thanks to Bas and Leia for bringing healthy competition into my life but especially the family craziness that I appreciate so much. Of course my four sisters in law and my lovely second families (also Helma and Nol). Oma, I wish you could have seen me graduate, I miss you! My final thanks and love go to Philip, who understands me (I still don’t know how) and for always supporting me (with chocolate, love and moral support) and also especially for teaching me how to relax.

119 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

9 Bibliography

[1] GlobalData, “US and China Driving Global Ethylene Capacity to Record 208 Million Tons per Year by 2017, says GlobalData,” 2014. [Online]. Available: http://energy.globaldata.com/media- center/press-releases/oil-and-gas/us-and-china-driving-global-ethylene-capacity-to-record- 208-million-tons-per-year-by-2017-says-globaldata. [Accessed: 15-Sep-2015]. [2] European Commission, “Reference Document on Best Available Techniques in the Large Volume Organic Chemical Industry February 2003,” Integr. Pollut. Prev. Control, vol. 1, no. February, pp. 267–289, 2003. [3] H. Zimmerman and R. Walzl, “Ethylene,” Ullmann’s Encyclopedia of Industrial Chemistry, vol. 13. Wiley-VCH Verlag GmbH & Co, pp. 465–529, 2012. [4] T. Ren, M. Patel, and K. Blok, “Olefins from conventional and heavy feedstocks: Energy use in steam cracking and alternative processes,” Energy, vol. 31, no. 4, pp. 425–451, 2006. [5] MEMERE, “MEthane activation via integrated MEmbrane REactors (MEMERE) Project,” 2015. [Online]. Available: http://www.spire2030.eu/memere/. [Accessed: 18-Jan-2016]. [6] R. Brelsford, “Rising demand, low-cost feed spur ethylene capacity growth,” Oil Gas J., vol. 112, no. 7, 2014. [7] W. R. True, “Special Report: Global ethylene capacity poised for major expansion,” Oil Gas J., vol. 111, no. 7. [8] METI, “Forecast of Global Supply and Demand Trends for Petrochemical Products (for the period 2006 to 2019),” 2015. [Online]. Available: http://www.meti.go.jp/english/press/2015/0612_03.html. [Accessed: 30-Sep-2015]. [9] C. Kelly, “Global ethylene market to hit $234.2bn by 2020,” 2015. [Online]. Available: http://www.arabianindustry.com/petrochemicals/news/2015/jul/22/global-ethylene-market- to-hit-2342bn-by-2020-5103787/#.Vop4jvnhC71. [Accessed: 15-Sep-2015]. [10] Global Processing, “Global petrochemical market to grow by 6.7% through 2018,” 2013. [Online]. Available: http://www.globalprocessingmag.com/global-petrochemical-market-to- grow-by-6-7-through-2018/. [Accessed: 15-Sep-2015]. [11] J. Richardson, “Market outlook: China growth declines on economic rebalancing,” ICIS Chemical Business. [Online]. Available: http://www.icis.com/resources/news/2013/05/13/9667046/market-outlook-china-growth- declines-on-economic-rebalancing/. [Accessed: 01-Jan-2015]. [12] B. Thinnes, “Bio-based polymers could be next big thing,” Hydrocarb. Process., vol. 91, no. 2, pp. 19–20. [13] CHEMSystems, “Plants to Plastics Can Nature Compete in Commodity Plastics? (Prospectus),” 2011. [14] J. Ungerer, L. Tao, M. Davis, M. Ghirardi, P.-C. Maness, and J. Yu, “Sustained photosynthetic conversion of CO2 to ethylene in recombinant cyanobacterium Synechocystis 6803,” Energy Environ. Sci., vol. 5, no. 10, p. 8998, 2012. [15] K. Ogura, “Electrochemical reduction of carbon dioxide to ethylene: Mechanistic approach,” J. CO2 Util., vol. 1, pp. 43–49, 2013. [16] A. Morschbacker, “Bio-Ethanol Based Ethylene,” Polym. Rev., vol. 49, no. March 2015, pp. 79– 84, 2009. [17] P. Haro, P. Ollero, and F. Trippe, “Technoeconomic assessment of potential processes for bio- ethylene production,” Fuel Process. Technol., vol. 114, pp. 35–48, 2013. [18] R. LEVANMAO, T. NGUYEN, and G. MCLAUGHLIN, “The bioethanol-to-ethylene (B.E.T.E.) process,” Appl. Catal., vol. 48, no. 2, pp. 265–277, 1989. [19] I. Munoz, K. Flury, N. Jungbluth, G. Rigarlsford, L. M. I Canals, and H. King, “Life cycle assessment of bio-based ethanol produced from different agricultural feedstocks,” Int. J. Life Cycle Assess., vol. 19, no. 1, pp. 109–119, 2014. [20] M. Steinberg, P. T. Fallon, and M. S. Sundaram, “The flash pyrolysis and methanolysis of biomass for production of ethylene, benzene and methanol,” in Novel Production Methods for Ethylene,

120 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

Light Hydrocarbons, and Aromatics, L. F. Albright, B. L. Crynes, and S. Nowak, Eds. New York: Marcel Dekker, Inc., 1992, pp. 481–508. [21] International Energy Agency, World Energy Outlook 2013. 2013. [22] MiningTechnology, “Coal giants: the world’s biggest coal producing countries.” [Online]. Available: http://www.mining-technology.com/features/featurecoal-giants-the-worlds- biggest-coal-producing-countries-4186363/. [Accessed: 06-Jan-2016]. [23] UOP, “Methanol to Olefins: A Pillar of Capacity Additions in China, but Does MTO have a Role in India?,” Des Plaines, Illinois. [24] BREF-ENE, “Reference Document on Best Available Techniques for Energy Efficiency,” Eur. Com., no. February, 2009. [25] C. M. Starks, “Two-stage hydropyrolysis-cracking process for producing ethylene,” 1975. [26] J. Q. Chen, A. Bozzano, B. Glover, T. Fuglerud, and S. Kvisle, “Recent advancements in ethylene and propylene production using the UOP/Hydro MTO process,” Catal. Today, vol. 106, no. 1–4, pp. 103–107, 2005. [27] J. H. Lunsford, “The Catalytic Oxidative Coupling of Methane,” Angew. Chemie Int. Ed. English, vol. 34, no. 9, pp. 970–980, 1995. [28] W. Hinsen, W. Bytyn, and M. Boerns, “Proceedings 8th International Congress on Catalysis,” 1984, pp. 581–592. [29] B. Vora, J. Q. Chen, A. Bozzano, B. Glover, and P. Barger, “Various routes to methane utilization- SAPO-34 catalysis offers the best option,” Catal. Today, vol. 141, no. 1–2, pp. 77–83, 2009. [30] L. Guczi, R. A. Santen, van, and K. V. Sharma, “Low-temperature coupling of methane.,” Catal. Rev. Sci. Eng., vol. 38, no. 2, pp. 249–295, 1996. [31] N. A. Baronskaya, A. A. Davydov, L. S. Woldman, and O. V. Buyevskaya, “Ethylene recovery from the gas product of methane oxidative coupling by temperature swing adsorption,” Gas Sep. Purif., vol. 1, no. 86, pp. 85–86, 1996. [32] A. H. Tullo, “Siluria’s Oxidative Coupling Nears Reality,” Chem. Eng. News, vol. 92, no. 27, pp. 20–21, 2014. [33] A. Gunter, W. W. Irion, and O. S. Neuwirth, “Oil Refining,” in ULLMANN’S ENCYCLOPEDIA OF INDUSTRIAL CHEMISTRY, Vol. 25, Wiley-VCH Verlag GmbH & Co, 2012, pp. 543–600. [34] J. G. Speight, “Ch. 5 Naphtha and solvents,” in Handbook of Petroleum Product Analysis, 2nd ed., John Wiley & Sons, 2015, pp. 88–99. [35] U. R. Chaudhuri, Fundamentals of Petroleum and Petrochemical Engineering. CRC Press, 2011. [36] K. M. Watson and E. F. Nelson, “Symposium on Physical Properties of Hydrocarbon Mixtures,” Ind. Eng. Chem., vol. 25, pp. 880–887, 1933. [37] G. Centi, F. Cavani, and F. Trifiro, Selective Oxidation by Heterogeneous Catalysis., 1st ed. New York, USA: Kluwer Academic Publishers/Plenum Press, 2001. [38] M. L. Rodriguez, D. E. Ardissone, E. Heracleous, A. A. Lemonidou, E. López, M. N. Pedernera, and D. O. Borio, “Oxidative dehydrogenation of ethane to ethylene in a membrane reactor: A theoretical study,” Catal. Today, vol. 157, no. 1–4, pp. 303–309, 2010. [39] F. Cavani, N. Ballarini, and A. Cericola, “Oxidative dehydrogenation of ethane and propane: How far from commercial implementation?,” Catal. Today, vol. 127, no. 1–4, pp. 113–131, 2007. [40] Nasdaq, “Crude oil prices,” 2016. [Online]. Available: http://www.nasdaq.com/markets/crude- oil.aspx. [Accessed: 30-Jan-2016]. [41] “Platts Global Petrochemical Index,” A Platts.com News & Analysis Feature. [Online]. Available: http://www.platts.com/news-feature/2015/petrochemicals/pgpi/index. [Accessed: 29-Dec- 2015]. [42] S. Andre, T. Hundertmark, and R. Rao, “McKinsey on Chemicals,” no. 4, pp. 46–58, 2012. [43] EIA, “U.S. Energy Information Administration: Henry Hub Natural Gas Spot Price.” [Online]. Available: https://www.eia.gov/dnav/ng/hist/rngwhhdm.htm. [Accessed: 01-Feb-2016]. [44] S. Azzarello, “Energy Price Spread: Natural Gas vs. Crude Oil in the US,” C. Gr. Mark. Insights, pp. 1–4, 2014. [45] InvestmentMine, “Prices of coal for the past 5 years,” 2016. [Online]. Available: http://www.infomine.com/investment/metal-prices/coal/5-year/. [Accessed: 06-Jan-2016]. [46] D. Hurd, S. Park, and J. Kan, “China ’ s Coal to Olefins Industry F . I . T . T . for investors Coal , to Syngas , to Methanol , to Olefins China ’ s Coal to Olefins Industry Coal , to Syngas , to Methanol

121 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

, to Olefins,” Dtsch. Bank Mark. Res., pp. 1–126, 2014. [47] “Increased coal-to-olefins processes in China,” Hydrocarbon Processing. [Online]. Available: http://www.hydrocarbonprocessing.com/Article/3096211/Increased-coal-to-olefins- processes-in-China.html. [Accessed: 30-Oct-2015]. [48] AspenTechnology, Aspen Plus ® Aspen Plus User Guide, vol. 10.2. Cambridge, 2000. [49] R. Schefflan, Teach yourself the basics of Aspen Plus. New Jersey: John Wiley & Sons, 2011. [50] AspenTech, Aspen Plus Steady state simulation Version 10 - Getting started modelling petroleum processes. 1999. [51] P. K. Niccum, C. R. Santner, and W. S. Letzsch, “Catalytic Cracking,” Houston, Texas: Access Engineering, pp. 1–121. [52] L. Tian, J. Wang, B. Shen, and J. Liu, “Composition-based Modeling for Predicting Olefin Yields of Naphtha Steam Cracking by the Monte Carlo Method,” Energy Sources, Part A Recover. Util. Environ. Eff., vol. 33, no. 19, pp. 1770–1781, 2011. [53] S. Seifzadeh Haghighi, M. R. Rahimpour, S. Raeissi, and O. Dehghani, “Investigation of ethylene production in naphtha thermal cracking plant in presence of steam and carbon dioxide,” Chem. Eng. J., vol. 228, pp. 1158–1167, 2013. [54] R. J. Zou, Q. K. Lou, S. H. Mo, and S. B. Feng, “Study on A Kinetic-Model of Atmospheric Gas Oil Pyrolysis and Coke Deposition,” Ind. Eng. Chem. Res., vol. 32, no. 5, pp. 843–847, 1993. [55] P. Kumar and D. Kunzru, “Modeling of naphtha pyrolysis,” Ind. Eng. Chem. Process Des. Dev., vol. 24, no. 3, pp. 774–782, 1985. [56] K. M. Sundaram and G. F. Froment, “Modeling of thermal cracking kinetics--/ථ͗dŚĞƌŵĂůĐƌĂĐŬŝŶŐ of ethane, propane and their mixtures,” Chem. Eng. Sci., vol. 32, no. 6, pp. 601–608, 1977. [57] K. M. Sundaram and G. F. Froment, “Modeling of thermal cracking kinetics--//ථ͗ƌĂĐŬŝŶŐŽĨŝƐŽ- butane, of n-butane and of mixtures ethane--propane--n-butane,” Chem. Eng. Sci., vol. 32, no. 6, pp. 609–617, 1977. [58] Z. Renjun, L. Qiangkun, and L. Zhiyong, “Study of kinetic models for the pyrolysis reaction of propane,” J. Anal. Appl. Pyrolysis, vol. 13, no. 3, pp. 183–190, 1988. [59] A. Chauvel and G. Lefebvre, Petrochemical Processes: Technical and Economic Characteristics, Volume I. Gulf Publishing Company, 1989. [60] Energetics Inc., “Energy and Environmental Profile of the U.S. Chemical Industry: The Ethylene Chain,” Columbia, 2000. [61] R. L. Grantom, “Combustion Air Should Be More Than Simply Recycling Energy,” in Industrial Energy Conservation, 1980, pp. 559–564. [62] R. L. Grantom, “Combustion Air Preheat and Radiant Heat Transfer in Fired Heaters–A Graphical Method for Design and Operating Analysis,” in Industrial Energy Conservation Technology Conference and Exhibition, 1981, pp. 151–159. [63] Metso Automation Inc., “Application report: Charge Gas Compression and Caustic Wash,” 2011. [64] Fisher, “Ethylene Production,” in Emerson Process Management (Chemical Sourcebook), 2010. [65] D. Salerno-paredes, “Optimal Synthesis of Downstream Processes using the Oxidative Coupling of Methane Reaction,” 2013. [66] UnionGas, “Data on chemical composition of natural gas by country.” [Online]. Available: https://www.uniongas.com/about-us/about-natural-gas/Chemical-Composition-of-Natural- Gas. [Accessed: 15-Nov-2015]. [67] RVO (Ministerie van Economische Zaken), “Gas Composition Transition Agency Report 2013: data on natural gas.” [Online]. Available: http://www.rvo.nl/sites/default/files/2013/12/New Natural Gas Report 2013 PNA.PDF. [Accessed: 15-Nov-2015]. [68] A. Marius, M. Hoppe, M. Montero, and G. Peureux, “Report on gas composition range in Europe,” INGAS Integrated Project, 2010. [Online]. Available: http://www.ingas- eu.org/docs/DB0.1.pdf. [Accessed: 01-Dec-2015]. [69] M. Thiemann, E. Scheibler, and K. W. Wiegand, “Natural Gas,” in Ullmann’s Encyclopedia of Industrial Chemistry, Vol. 24, vol. 2003, no. Ieo 2003, 2012, p. Vol 24. 177–223. [70] J. Wang, L. Chou, B. Zhang, H. Song, J. Zhao, J. Yang, and S. Li, “Comparative study on oxidation of methane to ethane and ethylene over Na2WO4-Mn/SiO2 catalysts prepared by different methods,” J. Mol. Catal. A Chem., vol. 245, no. 1–2, pp. 272–277, 2006. [71] Z. Stansch, L. Mleczko, and M. Baerns, “Comprehensive Kinetics of Oxidative Coupling of

122 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

Methane over the La 2 O 3 /CaO Catalyst,” Ind. Eng. Chem. Res., vol. 36, no. 7, pp. 2568–2579, 1997. [72] J. H. B. J. Hoebink, H. M. Venderbosch, P. C. van Geem, P. F. Van den Oosterkamp, and G. B. Marin, “Economics of the oxidative coupling of methane as an add-on unit for naphtha cracking,” Chem. Eng. Technol., vol. 18, no. 1, pp. 12–16, 1995. [73] M. J. Gradassi, N. W. Green, and N. Wayne Green, “Economics of natural gas conversion processes,” Fuel Process. Technol., vol. 42, no. 2–3, pp. 65–83, 1995. [74] W. Kiatkittipong, T. Tagawa, S. Goto, S. Assabumrungrat, K. Silpasup, and P. Praserthdam, “Comparative study of oxidative coupling of methane modeling in various types of reactor,” Chem. Eng. J., vol. 115, no. 1–2, pp. 63–71, 2005. [75] N. Yaghobi and M. H. R. Ghoreishy, “Oxidative coupling of methane in a fixed bed reactor over perovskite catalyst: A simulation study using experimental kinetic model,” J. Nat. Gas Chem., vol. 17, no. 1, pp. 8–16, 2008. [76] S. Manundawee, S. Assabumrungrat, and W. Wiyaratn, “Two-dimensional mathematical modeling of oxidative coupling of methane in a membrane reactor,” TiCHE Int. Conf., pp. 1–5, 2011. [77] T. V. Choudhary, E. Aksoylu, and D. Wayne Goodman, “Nonoxidative Activation of Methane,” Catal. Rev., vol. 45, no. 1, pp. 151–203, 2003. [78] S. Jašo, “Modeling and Design of the Fluidized Bed Reactor for the Oxidative Coupling of Methane,” Berlin University of Technology, 2011. [79] Linde Engineering, “Cryogenic air separation. History and technological progress.,” Pullach, Germany, 2014. [80] R. Smith, Chemical Process Design and Integration. 2005. [81] G. G. Crotti, “Analisi di unita ’ criogeniche per il frazionamento dell ’ aria,” Politecnico di Milano, 2010. [82] B. N. Roy, Fundamentals of classical and statistical thermodynamics. John Wiley & Sons, 2002. [83] D. W. Green and R. H. Perry, Perry´s Chemical Engineer´s Handbook, 8th ed. The McGraw Hill Companies, 2012. [84] B. Gerhard, “Enhanced Cryogenic Air Separation A proven Process applied to Oxyfuel - Future Prospects,” in 1st Oxyfuel Combustion Conference, 2009, pp. 1–38. [85] I. Kemp, “Pinch analysis and process integration: A user guide on process integration for the efficient use of energy,” p. 416, 2007. [86] G. D. Ulrich and P. T. Vasudevan, “How to estimate utility costs,” Chem. Eng., vol. 113, no. 4, pp. 66–69, 2006.

123 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

10 Appendix

A. Naphtha Composition

A.1 Naphtha composition resulting from comparison literature

Component Mol. Formula Density Mol. Weight Weight % Flow rate

Paraffins [g/mL] [g/mol] [kg/kg naphtha] [kmol/hr] Propane C3H8 0.493 44.10 4.9 388.92 n-Butane C4H10 0.248 58.12 16.85 1014.66 iso-Butane C4H10 0.251 58.12 1.5 90.33 2,2-Dimethylpropane C5H12 0.627 72.15 4 194.04 n-Pentane C5H12 0.626 72.15 8.5 412.34 Iio-Pentane C5H12 0.616 72.15 1 48.51 2,2- Dimethylbutane C6H14 0.791 86.18 2.5 101.54 n-Hexane C6H14 0.655 86.18 6.5 263.99 Methylpentanes C6H14 0.653 86.18 4 162.46 Methylhexanes C7H16 0.670 100.20 2.5 87.32 Dimethylpentanes C7H16 0.673 100.20 2 69.86 n-Heptane C7H16 0.710 100.20 5 174.64 Dimethylhexanes C8H18 0.710 114.23 1.5 45.96 Methylheptanes C8H18 0.698 114.23 2 61.28 n-Octane C8H18 0.703 114.23 3 91.92 Dimethylheptanes C9H20 0.720 128.26 - 0.00 Methyloctanes C9H20 0.720 128.26 - 0.00 n-Nonane C9H20 0.718 128.26 1.5 40.93 C10+ paraffins C10H22 0.730 142.28 - 0.00 n-Decane C10H22 0.730 142.28 1.5 36.90 Higher bp compounds C11+ 0.740 156.31 4 89.57 Total 72.75 3375.17

Cycloparaffins/napthenes

Cyclopentane C5H10 0.751 70.13 5 249.52 Cyclohexanes C6H12 0.779 84.16 6.06 252.02 Cycloheptanes C7H14 0.811 98.19 2.5 89.12 Dimethylcyclopentane C7H14 0.780 98.19 2.5 89.12 Methylcyclopentanes C6H12 0.749 84.16 - - Other C6-cycloparaffins C6H12 0.779 84.16 - - Methylcyclohexane C7H14 0.770 98.19 - - Ethylcyclopentane C7H14 0.763 98.19 - - Ethylcyclohexane C8H16 0.788 112.21 - - Dimethylcyclohexane C8H16 0.773 112.21 - -

124 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

n-Propylcyclopentane C8H16 0.780 112.21 - - 1,2,4-Trimethylcyclopentane C8H16 0.796 112.21 - - Other C8 cycloparaffins C8H16 0.834 112.21 2 62.38 C9 cycloparaffins C9H18 0.850 126.24 1.5 41.59 C10 cycloparaffins C10H20 0.871 140.27 1 24.95 Higher bp compounds C11+ 0.892 154.29 0.5 11.34 Total 21.06 820.03

Aliphatic olefins

Pentenes C5H10 0.641 70.13 0.025 1.25 Hexenes C6H12 0.678 84.16 0.025 1.04 n-Heptenes C7H14 0.710 98.19 0.025 0.89 n-Octenes C8H16 0.720 112.21 0.025 0.78 Branched nonenes C9H18 0.743 126.24 - - n-Nonenes C9H18 0.743 126.24 0.025 0.69 Brached decenes C10H20 0.741 140.27 - - n-Decenes C10H20 0.741 140.27 0.025 0.62 Higher bp compounds C11+ 0.750 154.29 0.06 1.36 Total 0.21 6.64

Cyclo olefins

C5 C5H8 0.771 68.12 - - C6 C6H10 0.811 82.15 - - C7 C7H12 0.824 96.17 - - C8 C8H14 0.848 110.20 - - C9 C9H16 0.861 124.23 - - C10 C10H18 0.873 138.25 - - Higher bp compounds C11+ 0.880 152.28 - - Total - -

Aromatics

Benzene C6H6 0.874 78.11 3.68 164.89 Toluene C7H8 0.866 92.14 1.85 70.27 Ethylbenzene C8H10 0.867 106.17 0.1 3.30 p-Xylene C8H10 0.861 106.17 0.1 3.30 m-Xylene C8H10 0.868 106.17 0.1 3.30 o-Xylene C8H10 0.879 106.17 0.1 3.30 Styrene C8H8 0.909 104.15 0.05 1.68 Methylethylbenzene C9H12 0.864 120.19 - - Trimethylbenzene C9H12 0.880 120.19 - - Methylisopropylbenzene C10H14 0.860 134.22 - - Higher bp compounds C11+ 0.880 148.25 - - Total 5.98 250.03

Total 100 4451.86

*Notes: bp is short for boiling point. Calculations made are based on a feed flow rate of 350,000 kg/h. - means component is not present in feed.

125 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

A.2 Naphtha Composition implemented in Aspen Plus

Component Mol. Formula Weight percentage Flow rate

[kg x/kg naphtha] [kmol/hr] Paraffins Propane C3H8 4.9 55.56 n-and iso- Butane C4H10 18.35 157.85 n- and iso Pentane (+ 2,2-dimethylpropane) C5H12 13.5 93.55 n-Hexane (+ methylpentanes) C6H14 13 75.43 n-Heptane (+ methylhexanes, dimethylpentanes) C7H16 9.5 47.40 n-Octane (+ methylheptanes, dimethylhexanes) C8H18 6.5 28.45 n-Nonane (+ dimethylheptanes, methyloctanes) C9H20 1.5 5.85 n-Decane (+ C10+ paraffins) C10H22 5.5 19.33 Total 72.75 Cycloparaffins/napthenes Cyclopentane C5H10 5 35.65 Cyclohexane (+methylcyclopentanes C6H12 6.06 36.00 Cycloheptane (+ dimethylcyclopentanes) C7H14 5 25.46 C8 (ethylcyclohexane, trimethylcyclopentane) C8H16 2 8.91 C9 cycloparaffins C9H18 1.5 5.94 C10 cycloparaffins C10H20 1 3.56 C11+ cycloparaffins C11H22 0.5 1.62 Total 21.06 Aliphatic olefins n-Pentenes C5H10 0.025 0.18 n-Hexenes C6H12 0.025 0.15 n-Heptenes C7H14 0.025 0.13 n-Octenes C8H16 0.025 0.11 n-Nonenes C9H18 0.025 0.10 n-Decenes C10H20 0.025 0.09 C11+ alkenes C11H22 0.06 0.19 Total 0.21 Cyclo olefins Benzene C6H6 3.68 23.56 Toluene C7H8 1.85 10.04 Ethylbenzene C8H10 0.1 0.47 Xylene (m-, o-, and p-) C8H10 0.3 1.41 Styrene C8H8 0.05 0.24 Total 5.98 Total 100

*Notes: bp is short for boiling point, Calculations made based on a feed flow rate of 350,000 kg/h. - means component is not present in feed

126 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

A.3 Light naphtha and heavy naphtha

#1 #2 Mol. Formula. #1 #2 Component Formula Weight % Flowrate Weight % Flowrate

[kg x/kg [kg x/kg Paraffins naphtha] [kmol/hr] naphtha] [kmol/hr] Propane C3H8 4.9 388.92 1 79.37 N-butane C4H10 16.85 1014.66 2.31 138.86 Isobutane C4H10 1.5 90.33 1.5 90.33 2,2-Dimethylpropane C5H12 4 194.04 4 194.04 N-pentane C5H12 8.5 412.34 3.94 191.07 Isopentane C5H12 1 48.51 1 48.51 2,2- Dimethylbutane C6H14 2.5 101.54 2.5 101.54 N-hexane C6H14 6.5 263.99 14.76 599.28 Methylpentanes C6H14 4 162.46 4 162.46 Methylhexanes C7H16 2.5 87.32 2.5 87.32 Dimethylpentanes C7H16 2 69.86 2 69.86 N-heptane C7H16 5 174.64 8 279.43 Dimethylhexanes C8H18 1.5 45.96 4 122.56 Methylheptanes C8H18 2 61.28 2 61.28 N-octane C8H18 3 91.92 6.25 191.50 Dimethylheptanes C9H20 0 0.00 0 0.00 Methyloctanes C9H20 0 0.00 0 0.00 N-nonane C9H20 1.5 40.93 3.5 95.51 C10+ paraffins C10H22 0 0.00 0 0.00 N-decane C10H22 1.5 36.90 2.5 61.50 Higher bp compounds C11+ 4 89.57 7 156.74 Total 72.75 3375.17 72.75 2731.14

Cycloparaffins Cyclopentane C5H10 5 249.52 5 249.52 Cyclohexanes C6H12 6.06 252.02 6.06 252.02 Cycloheptanes C7H14 2.5 89.12 2.5 89.12 Dimethylcyclopentane C7H14 2.5 89.12 2.5 89.12 Methylcyclopentanes C6H12 - - - - Other C6-cycloparaffins C6H12 - - - - Methylcyclohexane C7H14 - - - - Ethylcyclopentane C7H14 - - - - Ethylcyclohexane C8H16 - - - - Dimethylcyclohexane C8H16 - - - - N-propylcyclopentane C8H16 - - - -

1,2,4- - - - - Trimethylcyclopentane C8H16

127 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

Other C8 cycloparaffins C8H16 2 62.38 2 62.38 C9 cycloparaffins C9H18 1.5 41.59 1.5 41.59 C10 cycloparaffins C10H20 1 24.95 1 24.95 Higher bp compounds C11+ 0.5 11.34 0.5 11.34 Total 21.06 820.03 21.06 820.03 Aliphatic olefins Pentenes C5H10 0.025 1.25 0.025 1.25 Hexenes C6H12 0.025 1.04 0.025 1.04 N-heptenes C7H14 0.025 0.89 0.025 0.89 N-octenes C8H16 0.025 0.78 0.025 0.78 Branched nonenes C9H18 - - - - N-nonenes C9H18 0.025 0.69 0.025 0.69 Nrached decenes C10H20 - - - - N-decenes C10H20 0.025 0.62 0.025 0.62 Higher bp compounds C11+ 0.06 1.36 0.06 1.36 Total 0.21 6.64 0.21 6.64 Cyclo olefins C5 C5H8 - - - - C6 C6H10 - - - - C7 C7H12 - - - - C8 C8H14 - - - - C9 C9H16 - - - - C10 C10H18 - - - - Higher bp compounds C11+ - - - - Total - - - - Aromatics Benzene C6H6 3.68 164.89 3.68 164.89 Toluene C7H8 1.85 70.27 1.85 70.27 Ethylbenzene C8H10 0.1 3.30 0.1 3.30 P-xylene C8H10 0.1 3.30 0.1 3.30 M-xylene C8H10 0.1 3.30 0.1 3.30 O-xylene C8H10 0.1 3.30 0.1 3.30 Styrene C8H8 0.05 1.68 0.05 1.68 Methylethylbenzene C9H12 - - - - Trimethylbenzene C9H12 - - - - eMethylisopropylbenzen C10H14 - - - - Higher bp compounds C11+ - - - - Total 5.98 250.03 5.98 250.03 Total 100 4451.86 100 3807.84

*Notes: bp is short for boiling point calculations made based on a feed flow rate of 350,000 kg/h - means component is not present in feed

128 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

B. FUG method for distillation columns

B Partial pressures: logPsat [ bar ] A  (B.1) CTK []

sat Total pressure: P xi P i (B.2) ¦i

Bubble point: K x 1 (B.3) ¦ i i y Dew point: i 1 (B.4) ¦ Ki

Ki Volatility: Di (B.5) KHK

ª§xD · §xB ·º ª§xLK · § xHK · º ln log « » «¨1x ¸ ¨ 1 x ¸» ¨x ¸ ¨ x ¸ ¬©D ¹ ©  B ¹¼« ¬©HK ¹D ©LK ¹B »¼ Fenske: Nmin (B.6) ln log D AB D LK

c §Di * xi, F · Liquid feed at b.p.: ¦¨ ¸ 1q 0 (B.7) i 1 ©DTi  ¹

Average volatility: DDDaverage bottom* top (B.8)

xD 1 xD D AB x1 x c * x f  f §Di i, D · Underwood: Rmin 1 (B.9) 1 ¦¨ ¸ D AB  i 1 ©DTi  ¹

Gilliland: RR 1.2* min (B.10)

R  Rmin G (B.11) R 1

NN 1 54.4GG 1  min 1 exp ª§  ·§ · º (B.12)  «¨ ¸¨0.5 ¸» N 1 ¬©11 117.2GG ¹© ¹¼

2 ª x x º §NR · §B ·§F, HK ·§ B, LK · Feed stage location: log 0.206log « » (B.13) ¨ ¸ ¨ ¸¨ ¸¨ ¸ NS «©D ¹ xF, LK x D, HK » © ¹ ¬ © ¹© ¹ ¼

Column height: HNHcolumn total* spacing B.14)

129 Energy Analysis and Reactor Design for Ethylene Production from Methane Technische Universiteit Eindhoven University of Technology

C. OCM kinetics in Aspen

Aspen does not allow for simply putting the kinetics into a reactor design, but requires a more complicated and tedious method to implement the reaction kinetics. As can be seen below, reaction rates from 4.11-4.16 can only be implemented in the software via a Langmuir-Hinshelwood-Hougen- Watson mechanism. This reaction mechanism required rewriting all reaction rates to be able to insert them. Hereafter they were implemented together with the kinetic parameters.

kinetic_ factor driving_ force r (C.1) adsorption

n (ERTT / )> 1/  1/ 0 @ kinetic_(/) factor k T T0 e (C.2)

N N driving_ force k CD1 k C E1 (C.3) 1–i  2 – j i 1 j 1

m M N ª §n ·º adsorption K C 1 (C.4) «¦ i¨– j ¸» ¬« i 1 ©j 1 ¹¼» In order to study the reaction kinetics and reaction rates, and to reproduce the results from Stansch et al. a plug flow model was modelled in Excel. The results together with the implemented kinetics differed too much from this plug flow model in Excel and was therefore not further used. At first the kinetics directly implemented in the software resulted in higher yields and errors of 60 to 85% in outlet mole fraction, which were however according to the paper results. However the plug flow model did not lead to similar results according to the paper. Since rewriting the reaction rates and the complicated method of implementing the kinetics were assumed to be more prone to errors, the second method for the OCM reactor was used. However later on it turned out, there was an error in the Excel model which resulted indeed in higher yields. However reusing the first method could give more freedom in developing the reactor design in the PFD and working with pressure profiles and temperature profiles. Besides these advantages, other disadvantages are also overcome such as problems with running due to the Excel file in Aspen, longer run times and higher chance of crashing. Therefore it would be better to use kinetics directly in the model. Unfortunately this error was discovered later on in the project.

130 Energy Analysis and Reactor Design for Ethylene Production from Methane