A New Technology for the Combined Production of Charcoal and Electricity Through Cogeneration

biomass and bioenergy 69 (2014) 222e240

Available online at www.sciencedirect.com ScienceDirect

http://www.elsevier.com/locate/biombioe

A new technology for the combined production of charcoal and electricity through cogeneration

* Adriana de Oliveira Vilela a, ,1, Electo Silva Lora b, Quelbis Roman Quintero b, Ricardo Antonio^ Vicintin a,1, Thalis Pacceli da Silva e Souza a,1 a Rima Industrial S/A, Departamento de Pesquisa e Desenvolvimento, Anel Rodoviario, km 4,5, Belo Horizonte 30622-910, MG, Brazil b NEST e Nućleo de Excelência em Geraçao~ Termeletrica e Distribuı´da, Instituto de Engenharia Mecanica,^ Universidade Federal de Itajuba, Av. BPS 1303, CP 50, Itajuba 37500-083, MG, Brazil article info abstract

Article history: This paper presents an historical approach on the development of the existing biomass Received 17 February 2014 carbonization technologies in industrial operation in Brazil, the biggest charcoal producing Received in revised form country in the world. The gravimetric yield of charcoal from wood does not usually surpass 16 June 2014 25%; the time of each operation cycle is more than seven days; and less than 50% of the Accepted 27 June 2014 energy contained in the feedstock is transformed into charcoal e the rest is discharged into Available online the environment. The electricity generation associated with charcoal production is nowadays inexistent in Brazil. This paper presents the development of an industrial Keywords: technology of semi-continuous pyrolysis process, characterized by using metallic kilns Charcoal with forced exhaust system: the Rima Container Kiln (RCK). The results of the test runs are Pyrolysis gas related to 5 m3 and 40 m3 kilns, with a thermal power of 200 kW (pilot scale: 5 m3) and Cogeneration 3000 kW (industrial scale: 40 m3). The low heating value of the pyrolysis gases is 670 and Electricity 1470 kJ/m3, respectively. Biomass energy The main results are: a 3 h carbonization time; an average productivity per kiln of 1 ton of charcoal per hour; and a gravimetric yield of 35%. In this paper, four scenarios for the conversion of exhaust gases and tar into electricity were evaluated: the Conventional Rankine Cycle (CRC) and the Organic Rankine Cycle (ORC), each one with and without forest residues utilization. It is shown that the best economic indicators correspond to the scenario where ORC technology is used. The electricity generation cost is around U$30/ MWhe for ORC and US$40/MWhe for CRC. © 2014 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: þ55 31 3329 4483. E-mail addresses: [email protected] (A. de Oliveira Vilela), [email protected] (E.S. Lora). 1 Tel.: þ55 31 3329 4000. http://dx.doi.org/10.1016/j.biombioe.2014.06.019 0961-9534/© 2014 Elsevier Ltd. All rights reserved. biomass and bioenergy 69 (2014) 222e240 223

1. Introduction Table 1 e Products of carbonization. Products of carbonization % Dry base 1.1. The relevance of charcoal production in brazil Charcoal (80% fixed carbon) 33.0 Pyroligneous acid 35.5 Some reports indicate that around the year 500 A.C. the (Acetic acid) (0.5) Macedonians used wood to produce charcoal and tar. Even (Methanol) (0.2) before that, the carbonization of wood was already known and (Soluble tar) (5.0) utilized by the Egyptians, the Persians and the Chinese. The (Water and others) (23.5) Insoluble tar 6.5 process used by these ancient civilizations remains almost Non condensable gases (NCG) 25.0 unchanged today, especially from the energy loss point of (Hydrogen e 0.63%) (0.16) view, which can reach more than 50% of the biomass energy (CO e 34%) (8.5) e content. (CO2 62%) (15.5) Fig. 1 shows the share of the total energy produced from (Methane e 2.43%) (0.61) e charcoal and the respective energy loss in the State of Minas (Ethane 0.13%) (0.03) (Others e 0.81%) (0.20) Gerais, Brazil from 1978 to 2010 [1]. Energy losses reduction, as Total 100 observed in recent years, was only possible through process improvement, leading to increase in charcoal yield and more efficient forest handling. The phenomena that occur during carbonization are According to the National Energy Balance [2], Brazil has grouped differently depending on the author. For example, 44% of its energy matrix supplied by renewable sources. From Refs. [3] and [6], divide them in four stages as follows: this total, around 10% correspond to wood and charcoal, 15% to hydraulic electric generation, 16% to sugarcane, and 3% A: Up to 200 C, there is production of gases, such as water corresponds to wind and solar energy based generation. vapor, CO2, formic and acetic acid. Around 4% of the total installed capacity for electricity gen- B: from 200 to 280 C, the same gases from zone A are eration in Brazil corresponds to thermal power stations, released; but the emission of CO begins and there is a which burn coal, gas, oil and biomass (such as bagasse and substantial decrease in water vapor emission. The rewood dust). Nevertheless, there is not a single thermal power actions in this zone are endothermic. installation that uses exhaust energy from carbonization C: from 280 to 500 C. Carbonization occurs through processes. exothermic reactions. The products obtained in this stage are influenced by secondary reactions, including formation

1.2. Wood carbonization of fuel gases, tar, CO and CH4. D: over 500 C. All wood has been converted into charcoal. Wood carbonization involves a complex phenomenon that Various secondary reactions take place, catalyzed by the allows the generation of a wide range of chemical compounds, carbonization layer. which can be grouped as: charcoal, tar, pyroligneous acid and gases [3]. According to Ref. [7], sugarcane bagasse and wood pyroly- Table 1, adapted from Refs. [4] and [5], shows the mass sis can be divided by stages in a similar way based on thermal fraction content, dry basis, of the main products derived from analysis results. Stage B corresponds to hemicelluloses wood pyrolysis. These results were obtained at laboratory destruction and stage C to cellulose and lignin conversion into scale without oxygen supply and by using external heating. charcoal. Table 2, shows the main products generated in each stage of carbonization, according to the temperature evolution of the process. The values found in Table 2 correspond to tests performed at laboratory scale as show in Ref. [8]. Fig. 2 presents the photos of wood pieces at different carbonization stages as previously described. From left to right: in the first stage, the wood is dried and the released gases contain only water vapor. In the second stage, the product is a partially carbonized wood, or toasted wood. This toasted wood has the highest energy content per weight and also a great content of volatile matters. In the third stage, hydrocarbons start to be released and carbonization pushes forward to the center of the wood piece, reducing its volume in the radial di- rection. Finally, in the last stage, when the temperature rea- ches 500 C in the center of the wood piece, carbonization may Fig. 1 e Share of the total energy produced from charcoal in be interrupted. The charcoal at this final temperature has a the Minas Gerais State, Brazil from 1978 to 2010, as fixed carbon content of around 75%. Above 500 C, the charcoal granulated charcoal, other products (dust, tar and losses structure and composition continues changing and the fixed (smoke). carbon content can reach more than 90%. 224 biomass and bioenergy 69 (2014) 222e240

1.3. Traditional wood carbonization technologies

The “Hot Tail” kilns are the most widespread kilns within Brazil, due to its simplicity and low cost, especially for small

900 producers. They are recommended for ﬂat sites and, in gen- e eral, are built with baked bricks, clay and sand mortar. Nor- Hydrogen phase mally, more than one kiln is used and they are disposed as batteries or tandems. The operation of the kiln starts with the ﬁrewood loading, followed by carbonization and unloading of charcoal. The use 700 700

e of dry firewood is essential for good carbonization, because the firewood moisture directly influences the yield of the kiln of charcoal Dissociation as show in Ref. [9]. There is an ideal temperature, around 60 C, for unloading the kiln because the contact of air and charcoal at superior m temperatures can lead to fires. A standard kiln operation H n consists of three to six days for carbonization, five days for cooling and one day to unload/load the kiln [9]. 500 500

e In addition to the “Hot Tail” kiln there are other carbon- release C Hydrocarbons ization technologies in Brazil, with similar productivity, gravimetric yield and energy efﬁciency, such as: slope kiln, surface kiln, rectangular kiln, beehive kiln, JG kiln, all made of hewn stone and without forced exhaust system. The names presented on this paragraph are a free translation from Por- tuguese to English. Table 3 and Table 4 present the results of studies [10] about charcoal production from the main types of kilns in Brazil.

release To ensure the economic and operational viability of the

380 380 mechanized charcoaling process, it was necessary to build e rectangular kilns, which can reach a production capacity Acetic acid methyl alcohol Lots of heavy tar Tar Low condensation

Beginning of hydrocarbons equivalent to five (5) surface kilns [11]. Today it is possible to find rectangular kilns in operation with and without external combustion. Their firewood capacity is higher than 700 m3 and they possess equipment for tar recovering, which is usually released into the atmosphere in conventional hewing stone kilns [12]. Table 5 presents the main charcoal production technolo- 280 280 e

production gies, now in operation in the world, visited in 2007, with the

acetic acid indication of capacity. Vital et al. [13], made a comparative analysis of the pres- ently used carbonization technologies. The results are presented in Table 6. There is a technology, under development at laboratory scale, based on the conversion of wood into charcoal by mi- 200 200

e crowaves. According to Ref. [14], it is possible to attain a high

68.00.0 66.5 0.2productivity 35.5 6.5 conversion, with 31.5a speciﬁc 7.5 energy consumption 12.2 42.7 0.5 80.9 of around 1000 kWh/(ton of wood). The constraints of the traditional technologies for charcoal production include:

Difficulty in the mechanization of firewood loading and charcoal unloading. Fragmentation of charcoal during unloading. ] 4.61 5.07 16.41 20.01 15.20 13.23 3 Difficulty in automation due to the lack of instrumentation, C] 150

in particular: weight, ﬂow rates and temperature

Evolution of carbonization as a function of temperature. monitoring.

e Impossibility to use wood chips, which favors the gravimetric yield and productivity e an exception is the 2

2 continuous retort. Amount of gas Very small Small Important Important Small Very small Condensable constituents in the gas Water vapor Water vapor and Table 2 Temperature [ COH HydrocarbonsHeating value [kJ/m 2.0 30.0 3.3 30.0 37.5 20.5 48.7 12.3 20.5 24.6 8.9 9.7 Carbonization stage Water removal Oxygenated gases Carbon content (% of theNon charcoal) condensable gases (%) CO 60 68 78 84 89 91 biomass and bioenergy 69 (2014) 222e240 225

Fig. 2 e Wood pieces at different stages of the carbonization process. Source: Author.

Table 3 e Performance of the carbonization kilns in Brazil. Type of construction Operation cycle (hours) Capacity Volumetric yield (st/CMC) Firewood (st) Charcoal (CMC) Hewing stone with internal heat source Slope 240 20 8.7 2.3 Hot tail 144e168 20 8.0 2.5 JG surface 144 10e11 4e5 2.2 Rectangular V&M 264e312 180e240 95e130 1.8 Rectangular ACESITA 264 110 65 1.8 Metallic with internal heat source Semi-continuous JG NA NA NA NA Metallic with external heat source DPC semi-continuous 72 80 53.3 1.5

NA ¼ Not available. st ¼ cubic meter of stacked wood. CMC ¼ cubic meters of charcoal.

High heterogeneity in the drying, pre-pyrolysis and the substitution of oil. Recently, there has been an evident in- pyrolysis profiles inside the kiln. crease of interest in the use of gases from charcoal produc- High rate of partially carbonized wood formation due to tion in burners and kilns, thus reducing gas emissions and heterogeneous thermal distribution in the kiln. possibly obtaining thermal energy in a first stage (gases Large operation time, with low productivity. burners) and electricity in a more advanced phase of Difficulty in gas collection and in sealing the kiln. development. According to Ref. [15], it is also important to mention the 1.4. Constraints of by-products utilization for technological barriers linked to the quality of the gases, whose cogeneration in conventional carbonization kilns composition and heating values are not homogeneous throughout the phases of the carbonization process. Such According to Ref. [15], companies that produce charcoal have barriers are present in the highly variable heating values of been developing alternatives to use the energy from gases the gases; especially in the initial phase of wood drying, when generated during the carbonization process. Some trials, at the produced gases are difficult to combust, due to their high the beginning of the 80's decade of the 20th century, allowed water vapor content. obtaining tar, which was successfully used as fuel in Among the main barriers that limit the use of gases and tar from charcoal production for thermal power generation are:

e Table 4 Constructive and operational characteristics of a) Low heating value of the gases. the hewing stone kilns. b) Variable composition and temperature of the gases in Kilns Slope Hot tail Beehive JG different carbonization stages. (3 m) (3 m) (5 m) (3 m) c) Variable moisture content of the tar generated during the Diameter [m] 3.0e4.0 2.9e3.8 3.0e8.0 3.0 carbonization process. e e Maximum 2.5 2.8 2.3 max. 3.2 5.0 2.3 max. d) Difficulties in developing a project of an adequate burner height [m] for the combustion of the pyrolysis gases with high and Loaded firewood 20.0 st 8 st 5.0e200.0 t 14 st variable content of particulates, moisture and condens- Charcoal 24.0 16.0e20.0 50.0e60.0 22 [m3/months] ables, which would require the installation of pre-filters. Cycle [days] 7e85e78e95e6 e) The dilution and partial burning of the gases due to Useful life [years] 3 2e33e52e3 admission of undesired air. 226 biomass and bioenergy 69 (2014) 222e240

Table 5 e Characteristics of the conventional carbonization kilns. Technologies Photos Characteristics Hot tail kiln Kilns with a maximum diameter of 3.8 m and 2.3 m high, producing no more than 4 m3 of charcoal per load. It presents low yield; control is done by the color of the smoke. There are smoke and tar emissions. The process lasts from 10 to 12 days.

Surface kilns Kilns with 5.0 m diameter, producing a maximum of 20 m3 of charcoal per cycle from 36 st of ﬁrewood; the process lasts from 10 to 12 days.

Rectangular kilns 13 m long, 4 m wide and 3.5 m high, with a capacity to process 200 e700 m3 of ﬁrewood in stalks; operation cycles of 15 days and a capacity of 2000 t of charcoal/year.

Slope kilns Widely used in the State of Minas Gerais because of their low cost and operation very similar to the “Hot tail” kilns, with yields that do not surpass 25%.

DPC process The stalks of firewood are dried, then carbonized and the charcoal is cooled inside cages that are inserted in the kilns. In this technology, the gas flow is modified according to the stages of the carbonization process. The gas from the carbonization is used as thermal fluid. The cycles last 60 h. biomass and bioenergy 69 (2014) 222e240 227

Table 5 e (continued) Technologies Photos Characteristics CML France Batteries of 12 ﬁxed kilns; a charcoal production of 2500 t/year, with a yield of 20e25%. Each load of 5th residual ﬁrewood produces 2 m3 of charcoal per cycle. The cycle lasts 24 h.

Continuous retort The carbonization is carried out by the hot gases from the combustion, and part of the gases from the pyrolysis. The ﬁrewood is fed, in pieces, with 20% humidity, at the top of the retort, the cold charcoal is discharged at the bottom, with tar recovering. The capacity of the equipment is from 2000 to 10,000 t/year.

It should be emphasized that there is an emission of 50 kg of methane for each ton of produced charcoal. This is the 2. Material and methods main environmental impact of the charcoal industry: it is well known that methane has 25 times more global warming po- 2.1. Materials tential than CO2 in a 100 years timeframe [16]. As methane is not captured during photosynthesis, it is necessary to ﬁnd a In this work, three reactors with different design, sizes and way to mitigate its emissions. A possible alternative is the operation, were tested and used to improve the industrial combustion of the gases generated during the carbonization container reactor project, or Rima Container Kiln (RCK). In process for heat recovery and electricity production. Other order to combine the slow pyrolysis furnace with electric possibility is to increase the gravimetric yield of charcoal, power generation, the following reactors or kilns were used: because it would imply in less gas emissions. This article intends to present a technology that can solve laboratory scale reactor (5 L) 3 the above listed challenges, providing a possible true fruitful pilot scale reactor (5 m ) 3 association between the charcoal production and the elec- industrial scale reactor (40 m ) tricity cogeneration through by-products utilization, reducing altogether the environmental impact of wood All of those reactors were designed, manufactured, tested carbonization. and evaluated during this study. The lab scale reactor, with

Table 6 e Comparison of the performance of the kilns at use today in Brazil. Kiln Capacity per kiln Investment Gravimetric yield Bio-mass based carbochemicals production

Hot Tail 5 ton per month R$14/ton.year 25% Very low Circular 8 ton per month R$27/ton.year 25% Low Rectangular 42 ton per month R$237/ton.year 30% 70e120 kg/t of charcoal Continuous Retort 450 ton per month R$648/ton.year 33% 250 kg/t of charcoal DPC 57 ton per month R$250/ton.year 35% 250 kg/t of charcoal 228 biomass and bioenergy 69 (2014) 222e240

Fig. 3 e Scheme of gases analyzer and thermocouples in RCK.

electrical heating, was used to obtain the reference results. silicon and magnesium production from Rima Industrial S/A Next, the project, construction and operation of the pilot requires an average consumption of 50,000 m3 of charcoal per container were the basis for the definition of drying, carbon- month. ization and cooling behaviors. Finally, the industrial reactor The initial input values for energetic, economic, environ- was projected maintaining the same structural proportions mental and thermodynamic calculations were taken from from the pilot kiln, but with structural, mechanical and theoretical and real data, measured on a pilot plant. operational improvements. The designed system is able to provide real-time, accurate For modeling and geometric definition of the pyrolysis re- measurement of mass flow rate of the main gas stream on the actors (pilot and industrial ones), a finite volume software was furnace exhaust. This value is obtained indirectly by used (Ansys CFX, R.13). Pressure gauge, flow meter, thermo- measuring temperature, pressure and gas content. The mea- couples, infra-red temperature sensor, gas analyzer, wind surement and calculation are based on the stoichiometric speed meter, precise weighbridge scales, weighing load cell definitions, given by Fig. 4. system and anemometer were used as instrumentation. Fig. 3 A program was specifically developed in MS Excel Solver is a scheme of the rig and its instrumentation. for the calculation of physical and chemical properties. Fig. 5 The wood used in the experiment was urograndis, a hybrid provides a block diagram of this program with the specific between eucalyptus grandis and eucalyptus urophylla. The wood inputs and outputs: pieces tested in the experiments had the following di- The entries in the program are: mensions: (100 ± 20) cm in length; (60 ± 20) cm in width and (40 ± 10) cm in thickness. Weight loss of loaded firewood, internal temperature of the kiln, flow rate and composition of the gases generated in 2.2. Local and procedures carbonization.

All experimental data were taken from the pilot plant in And the outputs are: operation in the Rima Group Forest Unit, located in Buritizeiro, Minas Gerais. Composition, temperature, volumetric and mass ﬂow of Rima Industrial S/A is a metallurgy enterprise founded in the outlet gas 1975, located in Belo Horizonte, with factories in the northern Heating value of the outlet gas part of the State of Minas Gerais, Brazil. The ferro-alloys, Thermal power of the process biomass and bioenergy 69 (2014) 222e240 229

Percentage of tar, pyroligneous acid, carbonization gas and 25% of non condensable gases, with a mean composition

charcoal which must be burned to sustain the carboniza- of: 8% CO2, 12% CO, 2% of CH4, 1.5% H2, 2.5% O2 e 74% N2. tion process. 35% of pyroligneous acid solution, with 88% water, 5% of Oxygen excess content acetic acid, 5% of tar and 2% of methanol. Air flow inlet 6% of tar. Percentage or rate of complete combustion Heat losses by gases The gases composition was determined by chemical anal- Heat loss from the kiln ysis carried out during the monitoring of the carbonization process in the Pilot RCK. The MS Excel Solver provides these responses when the Based on the results from the pilot reactor and with the difference between the mass content of non-condensable gas expectation of achieving better results in an industrial scale components measured and the theoretical values derived reactor, the company decided to project, install and operate from the reaction is less than 2%. The final results presented an industrial unit (Industrial RCK) for the production of in the next sections are derived from the conduction of 10 charcoal. A schematic picture of this technology can be seen tests in industrial model Rima Container Kiln (RCK). in Fig. 7, that also shows the firewood loading, the carbon- For the evaluation of the electricity generation potential, ization process and the unloading of charcoal. the following scenarios were considered: This project started with the operation of a single kiln, with sequential test runs. The tests in this kiln (Fig. 8) supplied 1A: a CRC (conventional Rankine Cycle) using only the necessary information to evaluate the viability of energy pyrolysis gas as fuel. cogeneration, based on the pyrolysis gases. The information 1B: an ORC (Organic Rankine Cycle) using only the pyrolysis included: gas as fuel. 2A: a CRC using the pyrolysis gas and fines (forest residues) Thermal and fluid dynamic simulations using a finite vol- as fuel. ume software to optimize the geometry of the industrial 2B: an ORC using the pyrolysis gas and (fine forest residues) kiln, as well as the exhaust system and the flow of the as fuel. gases during carbonization. Evaluation of the carbonization gases' circulation pattern General data for all the considered scenarios are as follows: inside the kiln, from the hottest to the coldest section, in order to homogenize temperature distribution. This would Gas flow from each kiln: 6500 m3/h avoid firewood burning in the hot section and the forma- Number of kilns: 6 tion of partially carbonized wood in the cold section, which Gas heating value: 1470 kJ/m3 result in a low gravimetric yield. Gas thermal power: 15.8 MW The implementation of a continuous and on-line data Boiler efficiency in CRC: 80% attainment system. Steam turbine efficiency in CRC: 75% Instrumentation and automation of the process. Boiler efficiency in ORC: 90% Improvement in the utilization factor of the kilns and the Turbine efficiency in ORC: 80% increase of their productive capacity. Pump efficiency: 75% Development of a project for energy recovery from exhausted carbonization gases for wood drying and electricity generation. 3. RCK tests results 3.2. RCK development through modeling 3.1. Development of a semi-continuous kiln: the industrial Rima container kiln (RCK) From the studies of existing technologies, it was possible to elaborate flow profiles, which identify the pathways of Ferreira [5] developed a container kiln and evaluated its generated gases according to the process conditions. Table 7 performance. The results indicated that the container kilns shows the possible configurations of the gases pathways. have some advantages such as a higher productivity, gravi- The ignition procedure consists in the introduction/injec- metric yield and durability; faster cooling, loading and tion of energy into the kiln through the addition of burning unloading operations, which can also be mechanized [12]. pieces of charcoal in the ignition valves. Next, the carbon- The first patent referred to this technology was granted by ization takes place by gradually opening and closing other INPI, the agency for industrial property in Brazil, to Rima valves according to local thermal needs. Industrial S/A in 2011. The technology developed by Rima When the 40 m3 RCK was built and the tests were run with consists in the use of firewood pieces in metallic cylindrical type A ignition, as it can be seen in Table 7, different problems kilns, with lateral fissures, connected to an exhaust system, aroused. Explosions were frequent due to accumulation of as shown in Fig. 6. gases inside the kiln, and also due to formation of reflux zones The carbonization process in the Pilot RCK has the and gas pockets. Different pathway effects were analyzed following parameters: with the help of finite elements software, as it can be seen in Fig. 9. This study was based in the following premises: the kiln Yield: 33% of charcoal, with 81% of fixed carbon. is unloaded, the valves are open, the pressure inside the kiln is 230 biomass and bioenergy 69 (2014) 222e240

Fig. 4 e Block diagram for the gas LHV calculation. biomass and bioenergy 69 (2014) 222e240 231

Fig. 5 e Block diagram for calculation of physical and chemical properties.

equal to the atmospheric pressure and the gas flow is 6.500 m3/hr. The problems were eliminated by varying the ignition procedure and the pathway of gases. A continuous carbonization cycle was ensured, with a mean duration of 3 h, without throttling and/or internal blows or explosions. The developed RCK technology took in consideration that the best configuration for the gases flow pattern was the one with ignition at the bottom and with central and bottom forced exhaustion (type E in Table 7), keeping a continuous gas flow along the whole height of the kiln, vertically and radially, as shown in Fig. 10 and Fig. 11. Based on the flow pathway analysis, the ignition was changed to type E, meaning a modification from top to bottom ignition. This change leads to:

Homogenization of the internal temperature inside the kiln. Substitution of a linear carbonization front for a volumetric one. Fig. 6 e Top and lateral view of the Pilot RCK.

Fig. 7 e Industrial RCK loading, carbonization and unloading. 232 biomass and bioenergy 69 (2014) 222e240

Fig. 8 e Industrial RCK kiln with a capacity of 40 m3 of ﬁrewood.

Reduction in the theoretical carbonization time in an in- Top ignition, lateral exhaustion by the top Ignition at the bottom, lateralby exhaustion the top dustrial kiln. Increase of pyroligneous vapor fraction. A partial process of gasiﬁcation, with the formation of

synthesis gas (Syngas) composed by CO and H2. The elimination of dead zones (gas stagnation), specially in the case of hydrogen. Safe runs, without the formation of static pockets, mini- mizing the possibility of combustion that induces blows and explosions. Optimization of the RCK productivity with individual ca- pacities similar to the Continuous Retort. A charcoal gravimetric yield equals to (34 ± 1) %. This value is similar to the theoretical gravimetric yield, which was determined to be 35%. Top ignition, central exhaustion by the top Ignition at the bottom, centralby exhaustion the top Table 8 shows the calculation of the theoretical gravimetric yield. The calculation was based on the average content of wood [17] and the conversion of each component into charcoal [18]. The sum of the individual yields leads to the estimated total gravimetric yield of carbonization.

3.3. Temperature vs. time relationship for ignition at the top of the kiln

By monitoring internal temperatures in the industrial 40 m3 RCK, it was possible to obtain a thermal proﬁle of the kiln. The temperature monitoring system includes 28 points of temperature measurement near the wall of the kiln and 42 points in its internal part, as it is shown in Fig. 3. Moreover, there are

4 thermocouples located on the conical section of the kiln (at Top ignition, lateral exhaustion by the bottom Ignition at the bottom, lateralby exhaustion the bottom the very bottom of the kiln). As results of a series of tests with top ignition, the behavior of the mean temperatures along the height of the kiln was obtained. Fig. 12 shows the mean temperature for each height level, calculated from the values of 10 thermocouples in each level: 6 of them located in the middle of the reactor and 4 near the wall (Fig. 3). The mean temperature for the conical part of the kiln was calculated from the values obtained with the four thermocouples located in this section. The level sequence is descendent, with Level A being the top of the kiln and Level H being the bottom (conical section). Fig. 13 shows the three

dimensional temperature proﬁle of the kiln in the beginning, Possible pathways of the gases from the carbonization. middle and end of carbonization. e It is possible to infer that when ignition is done at the top of the kiln, there is a highly heterogeneous temperature proﬁle by the bottom by the bottom Table 7 Top ignition, central exhaustion Ignition at the bottom, central exhaustion during the carbonization process. At certain time intervals, if biomass and bioenergy 69 (2014) 222e240 233

Fig. 11 e Flow of gases in a set of RCK.

above 200 C. It is possible to see that the hot regions are restricted to the superior part of the kiln, moving next to the center and then returning to the top (passing briefly through the bottom). These problems are aggravated due to the low thermal conductivities of wood, partially carbonized wood and charcoal. In the individual wood pieces, as carbonization proceeds Fig. 9 e Gas flow simulation in the empty RCK using from the outside to the center (Fig. 14), the very low thermal computational fluid dynamics program. conductivities are themselves responsible for blocking the carbonization front in the core of the piece [19], especially if this piece is impregnated with tar. comparing the top and the bottom of the kiln on Fig. 12, the By using the top ignition, there is also formation of gas temperature difference is as high as 500 C. A carbonization pockets, corresponding to dead zones of stagnant gases and/ kiln with this temperature profile has a carbonization time or reflux. Hydrogen, being lighter, concentrates in the angles superior to 8 h. In addition, the very humid gases generated and edges of the kiln, where it remains, not able to follow the during carbonization show a descending pattern, going from turbulent flow of the gases. These regions are potentially the hot region to the cold region; leading the vapors, which explosive and, with favorable conditions, may cause explo- also contain tar and pyroligneous acid, to condensate over the sions with pressure waves so powerful that could lift a 40 ton non carbonized wood, significantly limiting heat transfer. kiln. Another problem with the ignition at the top is the low temperature in the bottom part of the kiln, which prevents the complete carbonization of the wood. This is also observed in Table 8 e Calculation of theoretical gravimetric yield of Fig. 13, which shows the regions of the kiln with temperatures carbonization. Component Average content Conversion of Final in wood component yield into charcoal

Cellulose 50% 34% 17% Hemicelluloses 20% 10% 2% Lignin 30% 55% 16% Total yield 35%

Fig. 12 e Distribution of the temperature along the RCK for Fig. 10 e Gas ﬂow pattern in the RCK. top ignition procedure. 234 biomass and bioenergy 69 (2014) 222e240

Fig. 13 e Three dimensional temperature proﬁle for ignition at the top of the kiln.

3.4. Temperature vs. time relationship for ignition at the virtually at the same time. This thermal homogeneity of the bottom of the kiln kiln also allowed reducing the carbonization time to 3 h in a kiln with 40 m3 capacity. Another advantage of this system is By using 74 thermocouples, it was possible to obtain a tem- the temperature increase at the bottom of the kiln, which now perature distribution along the height of the kiln for the makes possible a complete carbonization of the entire load carbonization process with ignition at the bottom. The tem- inside the reactor. This is observable on Fig. 16, in which the perature profile is shown in Fig. 15, in which the height level is regions above 200 C extend throughout the entire volume of descendent, with Level A representing the top of the kiln and the kiln e a clear difference from Fig. 13. Level H, the bottom (conical section) of the kiln. Fig. 16 shows the Three Dimensional Temperature Profile for the kiln in 3.5. Curves of weight loss for bottom and top kiln different moments of carbonization. ignition When ignition of the kiln is done at the bottom, it is ' 3 possible to obtain improvements in kiln s performance in The industrial 40 m RCK is equipped with a weighing system, comparison to top ignition. The first improvement is the ho- which continuously monitors the weight loss during the mogeneity of temperature distribution, verified on Fig. 15 by carbonization process. This system allowed the comparison of temperature differences from top to bottom of the kiln not effects from bottom ignition and top ignition in carbonization. superior to 250 C. This guarantees that materials in different Fig. 17 shows the curves of average weight loss registered levels of the kiln are undergoing the same carbonization step for carbonization with bottom and top ignition. Each curve is represented by the mean of 10 runs and the standard deviation is represented by error bars. It is possible to observe in Fig. 17 that, when carbonization is conducted with bottom ignition, the weight loss proceeds more rapidly and the end of carbonization happens 5 h earlier.

3.6. Mass and energy balance

The main results obtained with carbonization in the 40 m3 RCK can be summarized as follows:

Fig. 15 e Temperature distribution from top to bottom of Fig. 14 e Wood pyrolysis mechanism. the container kiln with the ignition at the bottom. biomass and bioenergy 69 (2014) 222e240 235

Fig. 16 e Three dimensional temperature proﬁle for ignition at the bottom of the kiln.

14000 The evolution in time of heating value and total thermal 3 12000 power of the released gases in the 40 m kiln with bottom ignition are shown in Fig. 18 and Fig. 19. Each curve corre- 10000 sponds to one of the 10 tests done with the industrial kiln. 8000 When comparing the 10 runs, it is possible to infer that the curves show a stable behavior and a relatively low standard 6000

Weight [kg] deviation, especially when it is considered that the tests pre- 4000 sented variation in the initial properties. As already ± 2000 mentioned, the size of the wood pieces was (100 20) cm in length; (60 ± 20) cm in width and (40 ± 10) cm in thickness. The 0 ± 0 100 200 300 400 500 600 humidity of these wood pieces varied in the range of (19 5) %. Time [min] The oscillations in Figs. 18 and 19 are the result of the Ignition done by the Top Ignition done by the Base process time scale. The measurement or analysis of the gas is done at 5 Hz, recording a moving mean of 10 s, which makes Fig. 17 e Comparison of the mean weight loss curves when the sampling interval shorter than the plotting interval. The ignition for the cases of top and bottom ignition. oscillations are represented as plateaus in Fig. 19 because of the reduced time scale (many points and small area). The results obtained in the industrial kiln were different Average gravimetric yield of (34 ± 1)% (the gravimetric yield from the results previously obtained in the pilot kiln, mainly is defined as the ratio between the weight of produced the composition of the produced gas, as shown in Table 10. charcoal and the weight of dry firewood fed to the kiln). The mean values presented on Table 10 are the results from 3 Volumetric conversion index of around 1.5 (this index is replicates conducted with the pilot kiln and 10 replicates with defined as the ratio between the volume of firewood loaded the industrial one. into the kiln and the volume of charcoal, in m3). The performance of the industrial kiln was better than the Carbonization time: fluctuating from 3 to 5 h, depending on 5m3 pilot one, considering that the released gas was less the moisture of the firewood and on the final properties of diluted. Another significant difference is that tar and pyro- the charcoal i.e., the fixed carbon content, the density and ligneous acid remained in the gas flow of the industrial kiln, the mechanical resistance. without condensation, as occurred in the pilot kiln. Table 11 shows average gas composition results of test Table 9 provides information on the charcoal quality from runs carried out with the pilot and industrial kilns. Fig. 20 the industrial kiln (40 m3 RCK) with ignition at the bottom. shows the energy distribution in carbonization products

Table 9 e Charcoal quality. Spin testa Drop testb Chemical analysisc Density d 3 e 3 Mean Deviation Mean Deviation %H2O %VM %ash %FC Bulk (kg/m ) True (kg/m ) 53% 9% 34% 10% 6% 18% 2% 81% 280 1500

a Done according to ABNT MB 1375-80. b Done according to ABNT 7416-84. c Done according to NBR 8112 ABNT-D176264. d Done according to ABNT NBR 9165. e Done according to ASTM D 167-73. 236 biomass and bioenergy 69 (2014) 222e240

Table 11 e Gas composition tests results e mean values (%v/v). Gases Pilot Deviation Industrial Deviation kiln kiln

CO2 7.1% 0.1% 11.6% 0.2% CO 0.05% 0.01% 11.9% 0.4%

CH4 0.99% 0.05% 0.50% 0.01%

C3H8 0.15% 0.01% 0.30% 0.01% ee H2 0.10% 0.01%

O2 13% 2% 4.9% 0.1%

N2 77% 2% 71% 2% Total 100% 100% Fig. 18 e Low heating value LHV (kJ/kg) of the gas. Low heating 670 2% 1470 1% value [kJ/m3]

Fig. 19 e Thermal power (MW) of the gas. Fig. 20 e Energy distribution of the carbonization products in the Pilot and Industrial RCK. from pilot and industrial scale. The differences are due, mainly, to design improvements in the industrial kiln and in the process of carbonization. There are higher energy density and temperature homogeneity in the industrial kiln, which lead to the production of a gas with higher heating value and higher potential for energy recovery. Fig. 21 is a chart that includes the ﬁnal energy balance, as well as the energy available for cogeneration.

3.7. Electricity generation using carbonization by- products

Container kilns make it possible to keep carbonization gases composition and heating value constant. These kilns also allow an energy recovery without implementing a cluster of various kilns operating sequentially in different stages, which would be an expensive and operationally complex procedure. Based on a 6 container kiln unit, with a charcoal yield of Fig. 21 e The ﬁnal energy balance for the RCK. 35% and wood with 19% initial humidity, it is possible to generate around 6.0 MWe of power from the carbonization

Table 10 e Mass fractions of the byproducts from ﬁrewood carbonization (wet base). Comparative analysis %Mass e pilot 5 m3 Deviation %Mass -industrial 40 m3 Deviation Charcoal 27.0% 0.3% 28% 1% Firewood gas 37.0% 0.9% 49% 7% Tar þ pyroligneous 17.0% 0.6% 5.0% 0.5% Humidity 19.0% 0.4% 18.0% 0.4% biomass and bioenergy 69 (2014) 222e240 237

gas. This calculation assumes that the carbonization gas has a heating value of 1470 kJ/m3 and the efficiency of the electricity generation unit is 20%. The use of tar and forest residues as complementary fuel would make energy generation increase notably. Fig. 22 shows the productive process and energy recovery of the RCK plant. The calculation of the installed power, when using commercially available generation technologies, must be preceded by its selection. For the conversion of energy from the carbonization gas into electricity, external combustion technologies were selected, as the gas contains tar, particles and other compounds that may affect the operation of internal combustion devices. External combustion technologies also allow the use of forest residues as complementary fuel. Among external combustion technologies, the only one at Fig. 22 e Flow of the productive process and energy commercial stage is the Conventional Rankine Cycle (CRC), recovery of the Rima Container Kilns. which uses water as working fluid. The Organic Rankine Cycle, known as ORC, which uses an organic fluid instead of water, is at the early stages of commercialization, with few hundred units in operation, mainly in Europe. One option to increase the efficiency of such systems is the use of prime movers as alternative to axial turbines systems in Conven- tional Rankine Cycles, such as screw expanders, steam en- gines and radial turbines. However, all these technologies are in developing stage and demand a high investment cost. The available power in a charcoal unit is of a few MWs e in this power range, Conventional Rankine Cycles are techni- cally feasible, although their efficiency could be low and the investment cost high due to the low efficiency of axial turbines. ORC systems have a higher efficiency in this power range. For modeling studies, related to energy recovery in RCK, e Fig. 23 Scheme of a conventional Rankine cycle. both the Conventional Rankine and ORC technologies were considered. It was also assumed that eucalyptus biomass is

Fig. 24 e A CRC for pyrolysis gases modeled using cycle Tempo software. 238 biomass and bioenergy 69 (2014) 222e240

Pyrolysis gas unit was assumed to be 18 years and the operation and Organic maintenance costs assumed to be 5% of the total investment. Thermal Oil Turbine Cooling Table 16 shows the results of the economic evaluation. The Electric Tower Boiler Condenser generator lowest generation cost corresponds to scenario 2B, which Condenser reached a value of 29.71 USD/MWh. The greatest NPV (Net Present Value) corresponds also to scenario 2A. The minimum commercialization price for attaining economic feasibility, for a corresponding IRR (internal rate of return) of 14%, was also Pump Pump Pump calculated for each scenario. In Brazil, renewable energy sources represent 44% of the e Fig. 25 Scheme of an ORC system using pyrolysis gases. total energy source, while in the world this ratio is only 14% and in developed economies it is 6%. From the renewable energy sources in Brazil, 33% correspond to hydraulic energy chopped in the field and fines, not feasible for charcoal pro- and 58% to biomass energy. Approximately 22% of renewable duction, are separately transported to the charcoal unit. The energy are forest based (firewood and charcoal). When it fraction of fines is considered to be 3e5% and the average comes to the generation of electricity, the most important transport distance 5e10 km. energy source is hydraulic, which represents 81% of the total Schemes and results for the four (4) evaluated scenarios generation (MME, 2012). are presented in Figs. 23e26 and Tables 12e15. Considering an average production of charcoal in Brazil Scenario 1A: CRC using only pyrolysis gases: equal to 10 million ton/year, and based on the results of the Scenario 1B: an ORC cycle using only the pyrolysis gas as present work, it is possible to generate more than 800 MWe of fuel. electric energy in the country from the use of pyrolysis gases. Scenario 2A: CRC using the pyrolysis gas and fines (forest Nevertheless, today, this contribution is null due to the lack of residues) as fuel. consolidated technology. For scenarios 2A and 2B, the available fuel's thermal power Therefore, the generation of energy from all charcoal units must include both the pyrolysis gases from the 6 kilns and the in Brazil could supply 5% of the electric demand of the coun- fine forest residues, achieving a final value of 21.42 MW. try, considering altogether the use of biomass residues, Scenario 2B: Pyrolysis gases and fines utilization in an ORC generated during firewood cutting. system. There is great heterogeneity of costs in the electric sector, which range from R$84.58/MWh for large hydroelectric plants 3.8. Investment and electricity generation costs and R$956.70/MWh for thermoelectric plants based on diesel calculations oil. It is then possible to infer that the estimated costs in this work for the generation of electric energy from pyrolysis gases An economic evaluation of each scenario was carried out to are highly competitive, especially if considering the current determine the generation cost and the minimum commer- stage of energy recession in Brazil, fundamentally dependent cialization cost. In this calculation, the life of the generation on the inconstant rainfall regime.

Fig. 26 e A conventional Rankine cycle for scenario 2A, modeled using the software cycle Tempo. biomass and bioenergy 69 (2014) 222e240 239

solutions as operation in synchronized clusters of kilns and Table 12 e Main results obtained from the modeling of the built-up of a system of ducts is required. Another solution scenario 1A. is the implementation of a technology of semi-continuous or Parameters Unit Value continuous kilns. Fuel thermal power MW 15.8 The project development and tests results of a continuous e Work fluid Water kiln e The Rima Container Kiln (RCM) were described and Net power MW 2.10 discussed. Also, a technical and economical evaluation of Efficiency % 13.36 cogeneration options, based on a charcoal unit using RCK, was performed. The pilot RCK was improved and it resulted in an industrial e Table 13 Main results obtained from scenario 1B. unit, with an increased capacity, an instrumentation Parameters Unit Value arrangement, a control system, a mechanized operation and Fuel thermal power MW 15.8 improved thermal capacity. Work fluid e Benzene During the tests, the load, the gas flow, the gas composi- Thermal oil e DyPhenyl C2 tion, the pressure and temperature values, the inlet air flow, Net power MW 3.04 the firewood temperature in 74 points and the volume of Efficiency % 18.1 generated pyroligneous acid were monitored parameters. All the data were continuously registered in real time. The tests lead to a stabilized operational regime with the following 3 Table 14 e Main results obtained from scenario 2A. characteristics: volumetric yield: 1.3 st of firewood per m of firewood; gravimetric yield: 34%; carbonization time: 3 h; gas Parameters Unit Value generation: 6500 m3/h in each kiln, with a heating value of Fuel thermal power MW 21.425 1470 kJ/m3; thermal power in each kiln: 3 MW; pyroligneous Working fluid e Water acid production: 150 kg in each cycle and approximately 1 ton Net power MW 2.9 Efficiency % 13.37 of charcoal per hour, per kiln. The RCK semi-continuous kiln allows solving the problems present in conventional kilns for cogeneration based on pyrolysis gases and other carbonization products. The gas Table 15 e Main results obtained from scenario 2B. composition and its heating value are constant in time. The

Parameters Unit Value average gas composition in the industrial prototype was: CO2 -

Fuel thermal power MW 21.45 11.2%, CO - 11.9%, CH4 - 0.50, C3H 8- 0.3, H2 - 0.1%, O2 -4.90, N2 - Working fluid e Benzene 70.7%. Thermal oil e DyPhenyl C2 The industrial kilns present higher productivity and po- Net power MW 4.1 tential for utilization of carbonization by-products because Efficiency % 18.2 the composition and heating value of the pyrolysis gases are constant. This avoids the necessity of building up ducts for gases transportation and synchronizing kilns operation in a 4. Conclusions pre-defined sequence. Conventional Rankine Cycle (CRC) and Organic Rankine There is an enormous potential in the State of Minas Gerais Cycle (ORC) are the most suitable technologies for cogenera- and in the entire Brazil for electricity generation using pyrol- tion, using RCK by-products. ysis gases and other byproducts from wood carbonization. Four scenarios were evaluated: CRC and ORC technologies The reduction of the environmental impact related to gases with and without utilization of forest fines residues. From the released into atmosphere must be also considered. For con- point of view of the economic indicators used in the feasibility ventional carbonization technologies, the energy recovery study (generation costs and NPV), the best scenarios are 1B from exhaust gases is complex because of the variation in its and 2A respectively. The best result is: ORC technology, which composition during the different carbonization stages. Some has an electricity generation cost of 29.71 and 29.50 US$/MWh,

Table 16 e Results of the economical evaluation of scenarios 1A, 1B, 2A and 2B. Parameters CRC e gas ORC e gas CRC e gas þ ﬁnes ORC e gas þ ﬁnes

Fuel cost, USD$/(5e10 km) 0 0 1.36 1.36 Electric power, MW 2.1 3.0 2.9 4.1 Investment, USD$ 5,813,234 4,602,433 6,214,207 6,288,781 Levelized cost, USD$/MWh electric 51.65 29.50 39.74 29.71 Speciﬁc investment, USD$/MWe 2768.2 1534.2 2142.83 1533.9 NPV, USD$ 353115.6 184121.7 392895.9 249167.9 TIR, % 14.0 14.0 14.0 14.0 Minimum commercialization price, USD$ 108.6 61.2 82.5 61.4 MWh 1 240 biomass and bioenergy 69 (2014) 222e240

for the cases with and without use of the forest residues carvao~ vegetal. Belo Horizonte: Fundaçao~ Centro Tecnologico respectively. de Minas Gerais e CETEC; 1982. p. 77e89. The results obtained in this study lead to the conclusion [7] Silva E, Nogueira L. Dendroenergia, Fundamentos e Aplicaçoes.~ 2nd ed. Interciencia;^ 2003. that the RCK presents not only a better carbonization perfor- [8] Brito JO, e Barrichelo LEG. Consideraçoes~ sobre a produçao~ de mance but it may also be an efficient and easier way to use carvao~ vegetal com madeiras da Amazonia. IPEF-ESALQ. exhaust gases, tars and forest residues for electricity genera- Serie Tecnica, vol. 2; 1981. p. 1e25. #5. tion in a cogeneration arrangement. [9] Fundaçao~ Centro Tecnologico de Minas Gerais e CETEC. Produçao~ e utilizaçao~ de Carvao~ Vegetal, 1 v. Belo Horizonte: CETEC; 1982 [Serie de Publicaçoes~ Tecnicas] . [10] Pinheiros PCC, Sampaio RS, Rezende MEA. A produçao~ de Acknowledgment carvao~ vegetal: teoria e pratica. Belo Horizonte; 2006. [11] Nogueira CP, França GAC, Souza Ju´ nior L. Otimizaçao~ da The authors would especially like to thank the company Rima produçao~ de carvao~ vegetal em escala industrial. In: Industrial S/A for believing, investing and providing human, Seminario De Balanços Energeticos Globais E Utilidades, 21., 1999, Vitoria. Anais. Vitoria: ABM; 1999. p. 1e10 technical and financial resources for development, contin- [Mimeografado]. uous improvement and completion of this project. Special [12] Baer RF. Avaliaçao~ Economica,^ Ambiental e Tecnica de thanks are also given to FINEP (a Brazilian Public Agency for Quatro Fornos para produçao~ de Carvao~ Vegetal. Monografia. the promotion of Science, Technology and Innovation in UFV; 2008. companies, universities and research institutions) for [13] Vital MHF, Pinto MAC. Condiçoes~ para a Sustentabilidade da ~ ~ ~ believing and investing in this Project. Produçao de Carvao Vegetal para Fabricaçao de Ferro-Gusa no Brasil. BNDES Setorial 2009;30:237e97. [14] Leal TE. Produçao~ de Carvao~ Vegetal atraves de micro-ondas. references Monografia. Universidade Federal de Ouro Preto; 2012. [15] Moura LF, Brito JO, Silva Ju´ nior FG. Effect of thermal treatment on the chemical characteristics of wood from Eucalyptus grandis W. Hill ex Maiden under different [1] BEEMG. 26o Balanço Energetico do Estado de Minas Gerais, atmospheric conditions. CERNE (UFLA) 2012;18:449e55. Ano base 2011. CEMIG; 2011. [16] United States Environmental Protection Agency. Methane [2] EPE. Balanço Energetico Nacional. E. d. P. Energetica. Rio de and nitrous oxide emissions from natural sources; April e Janeiro: Ministerio de Minas e Energia MME; 2011. 267, 2010. from, https://ben.epe.gov.br/BENRelatorioFinal2011.aspx. [17] Santos ID. Influencia^ dos teores de lignina, holocelulose e ~ [3] Medeiros CA, Rezende MEA. Alcatrao Vegetal: Perspectivas extrativos na densidade basica e contraçao~ da madeira e nos ~ ~ ~ de uso e produçao. Fundaçao Joao Pinheiro 1983;13(9 rendimentos e densidade do carvao~ vegetal de cinco especies e a10):42 8. lenhosas do cerrado. Departamento de Engenharia Florestal ~ [4] Gomes PA, Oliveira JB. Teoria da carbonizaçao da madeira. e Universidade de Brası´lia; Fevereiro 2008. In: Penedo WR, editor. Uso da madeira para fins energeticos. [18] Sinha S, Jhalani A. Modeling of pyrolysis in wood: a review. ~ Belo Horizonte: CETEC; 1980. p. 158 [Serie Publicaçoes New Delhi e 110016-India: Department of Mechanical Tecnicas, n. 1]. Engineering. Indian Institute of Technology. ~ [5] Ferreira OC. O futuro do carvao vegetal na siderurgia: [19] Di Blasi C. Analysis of convection and secondary reaction ~ ~ emissao de gases de efeito estufa na produçao e consumo do effects within porous solid fuels undergoing pyrolysis. ~ & carvao vegetal. Revista Economia Energia 2000;4(21). Combust Sci Technol 1993;90:315e39. [6] Mendes MG, Gomes PA, Oliveira JB. Propriedades e controle da qualidade do carvao~ vegetal. In: Produçao~ e utilizaçao~ de