Enhancing Well Integrity in the Hengelo Brine Field Continuous Improvement of Well Integrity in One of World’S Largest Brine Fields

SOLUTION MINING RESEARCH INSTITUTE Technical www.solutionmining.org Conference Paper 105 Apple Valley Circle Clarks Summit, PA 18411, USA

Telephone: +1 570-585-8092

Enhancing well integrity in the Hengelo brine field Continuous improvement of well integrity in one of world’s largest brine fields

T.P.F. Koopmans & A.M. den Hartogh AkzoNobel Industrial Chemicals, Hengelo, the Netherlands

SMRI Fall 2017 Technical Conference 25 - 26 September 2017 Münster, Germany

Solution Mining Research Institute Fall 2017 Technical Conference Münster, Germany, 25 - 26 September 2017

Enhancing well integrity in the Hengelo brine field Continuous improvement of well integrity in one of world’s largest brine fields

Tjeerd Koopmans1 & Marinus den Hartogh Mining Development & Compliance - AkzoNobel Industrial Chemicals Boortorenweg 27, 7554 RS Hengelo, the Netherlands

Abstract

AkzoNobel mines Röt salt from its brinefield between Hengelo and Enschede at depths of approximately 400 meters. Since 1933, over 250 relatively flat caverns with over 550 connecting wells have been developed in the bedded salt deposit.

Early 2016 oil was detected in the shallow groundwater near well 335. Well 335 was drilled in 1986 and gives access to a so-called multi completion cavern (MCC) with three access wells. During a video run two distinct leaks were encountered at depths of 54 and 97 meters (177 and 318 ft) . Subsequent investigations showed that a breach of integrity of the cemented casing had probably caused the leakage of oil to the environment. As the annular space only contained blanket oil in 1986-1987, this integrity breach probably dated back to those years. The volume of oil that potentially had leaked into the environment was assessed to be 12 m3 (424 cf).

Next to the necessary environmental investigations near well 335, AkzoNobel immediately started an extensive integrity program covering all its wells. Due to the relatively small amount of salt that can be produced from one cavern in this bedded salt, compared to the larger caverns in salt domes, the number of wells is large (over 550). Until the early 2000s approximately 140 wells were abandoned, leaving over 400 in open connection with the underlying caverns. The integrity program was split in three parts: 1. A desk study to find out which abandoned wells were potentially suspect to have had any historical integrity issues. Being abandoned, well specific investigations could then only consist of environmental investigations; 2. A desk study to find out which inactive wells were potentially suspect to have had any historical integrity issues. Although inactive, these wells are still accessible, and can be subjected to well investigations, such as a USIT/CBL/PMIT-run, plugging, pressure testing and, if necessary, running a video through the cemented casing; 3. Pressure testing of all active wells, most of which were recently drilled wells to single completion caverns (SCC-wells, with one access well). These pressure tests provide proof of integrity of the wells that are currently producing brine. This was done by placement of a retrievable plug during a normal, scheduled workover.

Since mid-2016 almost 70 active and inactive wells have been tested. All 37 tested active wells (until August 2017) were found tight providing confidence that the current brine production methods are sound. Besides well 335, eleven of the suspect inactive wells were found to be not tight. To prevent any ongoing leakage of fluids present in the wells these remain plugged. These wells divide in two groups: 1. Seven wells drilled in the late 80s which are very comparable to well 335; 2. Four wells drilled in the period 1963-1970 with characteristics quite different from well 335.

Subsequently to the well integrity investigations, a root cause analysis resulted in a number of indicators for wells to be not tight. More indicators means more suspicion and this was used to prioritize the further testing of other wells. Key words: well integrity, casing, oil blanket, soil pollution, root cause analysis

1 corresponding author; email: [email protected]

2 1. Introduction

In the eastern Netherlands, salt mining by AkzoNobel first started in 1919 near Boekelo. Since 1933, after the opening of the Twente Canal, AkzoNobel moved its salt mining operations to the Hengelo- Enschede brine field within the Twenthe-Rijn, Twenthe-Rijn Uitbreiding and Helmerzijde concession areas. The brine field is located between the cities of Hengelo and Enschede (Figure 1). Using solution mining, bedded Triassic Röt salt is mined from caverns at depths ranging from 350 to 550 meters (1.148 and 1.804 ft) below ground level. One cavern can only produce a relatively small amount of salt from the approximately 50 m thick layer of useable salt, compared to the much larger amount of salt that can be produced from one cavern in a domal salt. Consequently, there are a large number of caverns in this field. Over 250 caverns were developed since 1936. These caverns are typically disc-shaped with an average height (at end of production) between 20 and 40 meters (66 and 131 ft) and a diameter/span between 80 and 150 meters (262 and 492 ft). Over 550 wells connect the caverns to the surface network of pipelines transporting fresh water from the canal to the caverns and brine from the caverns to the salt plant. Until the beginning of this century most caverns were equipped with two or three wells (so called multi-completion caverns or MCCs), meaning that one well was used for injection of fresh water and another well for extraction of brine. Younger caverns only have one well with two internal tubes: one tube for injection and the other for extraction (so called single-completion caverns or SCCs). Figure 2 shows the completion of these two types of wells.

Approximately 140 wells were abandoned, leaving over 410 wells still in open connection with the underlying caverns. These open wells enable 5-yearly sonar measurements, which are mandatory for caverns that are determined to be potentially unstable. Caverns are categorized as potentially instable when leaching in the past has left a remaining salt roof of insufficient thickness to ensure long-term stability (den Hartogh et al., 2017). Furthermore, salt caverns may be used for storage of liquids. Two caverns in the Hengelo-Enschede brinefield are being used for the strategic storage of 250,000 m3 (8.8 MMcf )of gasoil for the Dutch state (Koopmans et al., 2014).

Of the still open 416 wells, 86 wells are actually producing brine, of which approximately 60 are wells to SCCs. The other wells connect to MCCs which are used for recycling of salt plant waste streams (like bromide concentrate from the evaporation process and brackish water from rinsing and filtering processes) or for stabilization purposes (injection of calcium slurry, solids from the brine purification

3 process). Six wells are part of the oil storage caverns. The other 322 wells are not in use. These may be used for stabilization or recycling later. Figure 3 presents the different types of wells in the brinefield.

Figure 1: AkzoNobel’s Hengelo-Enschede brine field in the east of the Netherlands close to the German border

Figure 2: General completion of the different types of caverns used for solution mining in Twente: multi-completion caverns (MCC) and single-completion caverns (SCC). Most MCC-wells have a 7’’ cemented casing with a 5 or 5½’’ 5 tubing, while SCC-wells have a 9 /8’’ cemented casing with 7’’ and 4½’’ tubings.

Figure 3: Overview of the different types of wells present in the Hengelo-Enschede brinefield at the moment of writing (August 2017)

4 2. Well 335

May 2016, well integrity issues came to surface with well 335, one of the three wells giving access to cavern 334. The wells of cavern 334 were drilled in 1986 and produced brine from 1986 until 2010. Between 2013 and 2015, cavern 334 was taken into production again to replace the raw brine in the cavern with bromide concentrate from the plant. AkzoNobel monitored the shallow groundwater quality at producing wells by taking 3-yearly groundwater samples. Groundwater samples showed high concentrations of mineral oil. The possible presence of a soil and groundwater pollution with oil led to further soil investigations. It became clear that there was a possibility of a leak in the cemented casing from which the oil had leaked into soil and groundwater.

In the period March-May 2016 a series of mining engineering investigations was started at well 335 together with a further environmental investigation in the vicinity of the well. Due to the possibility of a deep source of the pollution, the environmental investigation focused on the deep subsurface (~40 m / 131 ft below ground level). The mining engineering investigations consisted of: • cleaning of the inside of the cemented casing (being done by running in hole a high water pressure nozzle); • a combined USIT/CBL/PMIT wireline measurement run to check cementation and casing quality (by Schlumberger); • placement of a permanent (drillable) packer at a depth of ~390 m( 1.280 ft) below ground level (wireline run by Schlumberger); • a pressure test performed at the highest operational oil pressure in the annular space; • extraction of most of the fluid from the well; • a video run to check for any inflow of fluids from outside the casing and interpretation of the video images.

The USIT/CBL/PMIT measurement resulted in qualitative data on the cement quality and the depth of the top of the cement. The top of the cement was determined to be at ~109 m (358 ft) below ground level. The pressure test clearly showed the 7’’ casing not to be tight, the pressure dropped 2.5 bar/hr (36 psi/hr) at a pressure of 20-25 bar (290 -363 psi) (i.e. 10%; see Figure 4). A leak rate of ~0.7 liter/hr (25 oz/hr) was deduced.

A video run that was performed after having partially emptied the casing revealed two distinct leaks of the buttress threaded connections (BTC). The defects were located at depths of ~54 and ~97 m (177 and 318 ft) below ground level (see Figure 5), above the top of the cementation. Subsequent investigations showed that a breach of integrity of the cemented casing had probably caused the leakage of oil to the environment. As the annular space only contained blanket oil in 1987-1989, the breach of integrity probably dates back to the start of production in 1987. The fact that the annular space only contained blanket oil during 20 months and taking into account the size of the defects and the physical properties of oil, the maximum volume of oil that potentially had leaked into the environment was assessed to be 12 m3. From 1989 onwards, leakage of brine via the same pathway cannot be ruled out.

After the initial groundwater samples, further soil investigation was performed in the vicinity of this well using Fugro’s ROST/UVOST (Rapid/Ultraviolet Optical Screening Tool), an ultraviolet induced fluorescence sensor deployed by Cone Penetration Testing (CPT) equipment. ROST measurements can detect the presence of oil. Electrical Conductivity (EC) measurements were also done using an EC- sensor deployed by the same CPT equipment. Difference in electrical conductivity indicates the presence of brine. In the typical Hengelo soil conditions, CPT investigations can be conducted to depths of ~50 m (164 ft) below ground level.

The ROST-measurements showed the presence of oil both in the shallow subsurface (0-3 m or 0- 10 ft below ground level) and in the deeper subsurface (38-43 m or 125-141 ft below ground level; see Figure 6). The EC measurements did not show any abnormal salinity of the groundwater.

Pressure(bar)

Time

Figure 4: Graph showing the results of the pressure test of well 335 performed at a starting pressure of 25 bar, being ~110% of the highest possible operational pressure (10 bar = 145 psi)

Both stratigraphy and soil quality results from the different types of CPTs were checked by installation of deep wells and drillings (again down to ~50 m (164 ft) below ground level). These confirmed the CPT results.

Based on all available information the following situation was deducted: - pollution of soil and groundwater with mineral oil (diesel) at groundwater level (0-3 m below ground level) in the close vicinity of the well (up to 20 meters downstream from the well); - groundwater pollution with mineral oil (diesel) in distinct sand layers / lenses within a clay formation at 38-43 m or 125-141 ft below ground level in the close vicinity of the well (up to 20 meters (66 ft) downstream from the well); - pollution with brine was not detected.

3. Well integrity program

Next to the necessary environmental investigations near well 335, AkzoNobel started an extensive integrity program. With over 410 wells still in open connection with the underlying caverns, the size of this program was considerable. The integrity program of checking these open wells was split in three parts: 1. A desk study to find out which inactive wells were suspect to have (had) integrity issues; 2. Well investigations wells suspect to have (had) integrity issues; the prioritization should be based on the differing level of suspicion; 3. Pressure testing of all active wells to provide proof of integrity of the wells that are currently producing brine.

Investigation of abandoned wells that were potentially suspect of any historical integrity issues was postponed until more information on the root cause of the breach of integrity was available. Being abandoned, well specific investigations could then only consist of environmental investigations.

Figure 5: Still from the video run in well 335 showing the leaking connection at a depth of ~97 m below ground level. Bubbles can be seen forming just above the leak.

Figure 6: Results of one of the CPT-ROST measurements in the vicinity of well 335, showing the presence of oil from surface down to ~3 m below ground level and from ~38 down to ~43 m below ground level. The rightmost two columns show the total fluorescence measured, the leftmost columns provide regular CPT information on the stratigraphy (cone resistance and friction ratio respectively).

4. Inventory of the integrity of all inactive wells

After the oil pollution at well 335 was detected and it was clear that the cemented casing was not tight, AkzoNobel decided to check the recorded history of all inactive wells. Over 410 well books, both analogue and digitally available, were studied by a team of mining engineers. This lead to the identification of 4 wells of more or less the same age as well 335, in which probably the same type of casings were used and that at some time within their history showed potential indications that may have

7 caused integrity issues (see Figure 7). These wells, together with well 335, received the highest priority for further well integrity investigations (priority 1). Another 14 wells were identified to also be of similar age as well 335, but the well books did not show any indications that could have led to integrity issues. Another 15 wells of very different ages were identified that, at some point in their history, showed indications that could have led to integrity issues, like a sudden decrease of the oil pressure. Sometimes these issues were (temporarily) solved or wells were repaired, sometimes wells were taken out of use. These two groups together formed a group of 29 wells with priority 2. The remaining inactive wells (approximately 290 in total), were suspect until proven tight. This proof may be provided by either integrity tests or ongoing insights from the test programme on Prio-1 and Prio-2 wells. This ongoing insight would be reported as part of the Root Cause Analysis and lead to the formation of a third group of wells to be investigated (priority 3).

Figure 7: Location of well 335 and the four identified priority-1 wells (wells 330, 336, 340 and 342), indicated by arrows, all located in the rural area between Enschede and Hengelo.

5. Well integrity investigations

Well investigations at priority-1 and priority-2 wells Well integrity investigations at the priority-1 and priority-2 wells in general consisted of the following steps (see Figure 8): - run out of hole the inner tubing (if present); - cleaning of the inside of the cemented casing (being done by running in hole a high water pressure nozzle); - a combined USIT/CBL/PMIT wireline measurement run (by Schlumberger); - placement of a permanent (drillable) packer by wireline (Schlumberger) or placement of a Fangmann retrievable packer on a drill string (at wells that have to remain open for stability or production reasons); - a pressure test performed at the highest operational oil pressure in the annular space; - if the well is found to be not tight: - partial expulsion of the fluid from the well; - a video run to check for any inflow of fluids from outside the casing and interpretation of the video images (see Figure 9B); - soil and groundwater investigations guided by the results of the mining engineering investigations.

All four priority-1 wells were found to be not tight. The wells showed pressure drops between 1.5 and 7 bar per hour ( 25 and 102 psi/hr at an applied test pressure of 25 bar/363 psi). USIT/CBL/PMIT measurements showed large variations in cement quality over depth. The depths of the top of the cementation ranged from surface to 112 m or 367 ft below ground level. Video runs showed leaking threaded connections at different depths between ~10 m and ~218 m (33 and 715 ft) below ground level and damaged casings (potentially due to overtorque) at some depths as well. Subsequent environmental soil and groundwater investigations indicated the presence of oil pollution at all four wells, both near surface and at a depth of approximately 40 m (131ft) (i.e. shoe of the conductor casing).

Mining engineering If necessary: investigations If pressure test fails: Environmental remediation (borehole downhole camera investigation plans to be measurements; inspection developed pressure test)

Figure 8: General overview of the integrity investigations workflow

A. B.

Figure 9: A) drillable Fangmann plug just before being run in hole. B) configuration during the video run by Haitjema C) Close-up of the Fangmann drillable 7’’ plug.

The investigations of the 29 priority-2 wells are ongoing. Until August 2017, 19 wells have been investigated following the above described well integrity testing program. Twelve wells were found to be tight, 7 wells were not tight with recorded pressure drops between 0.5 bar/hr and 1 bar/min. This brings the total of not tight wells to 12 (August 2017). At the 7 not tight priority-2 wells, video runs only showed inflow at the 3 most recently drilled wells (329, 339, 341). These wells were drilled in the late 1980s and are comparable to well 335. At the other wells, which were drilled in the period 1963-1970 and have characteristics quite different from well 335, hardly any inflow was found, indicating very shallow leaks, probably more or less at groundwater level or only a few meters below. Environmental investigations near these wells are ongoing. As these wells remain permanently plugged, no further leakage of brine or oil can occur.

Integrity testing of active wells The active wells were not suspect but to increase confidence it was decided to test a sample of five active wells as soon as possible and to include integrity tests as part of the normal workovers. These wells are mostly recently drilled wells to single completion caverns (SCCs), with one access well. This was done in August-October 2016. The sample included five wells of very different ages due to undergo a necessary phase change workover. This includes pulling out of hole the production strings. The 5 workover was extended to include running in hole a drill pipe with a retrievable 9 /8’’ Fangmann packer (see Figures 9A and 9C). After placement it was pressurized according to AkzoNobel’s internal pressure testing procedure. If necessary due to temperature effects (cooling of the well results in a pressure drop as well), the well was re-pressurized and re-tested. All five wells in this random test were found tight.

Having integrity tests included as part of the normal workovers is seen as another proof of the integrity of wells that are currently in use for brine production (see Figure 10).

9 It guarantees that: - only wells pressure tested shortly after drilling are being taken into production; - wells are tested after approximately 3 years of production; - wells that were not tested before will ultimately be tested within approximately 4 years; - wells that are taken out of production are being subjected to an end-of-life-pressure test.

Furthermore, older MCC wells that will be taken into production again after a stand-still period, for example for recycling of salt plant waste streams or stabilization of caverns, will always be tested prior to re-use.

approximate age (in years) 1 2 3 4 5 6 7 8 9 10

MLS3 Sump 1st main leaching 2nd main leaching stage (MLS2) (if phase stage (MLS1)

apl)

(afterdrilling)

pressure tests pressure

pressure testpressure

end of production endproduction of

pressure test @ @ test pressure

phase change WOphase change

tested after MLS1) after tested

not yet pressure test (only if was (only test pressurewell Figure 10: Workflow for pressure testing of active wells

Until August 2017, 37 wells have been pressure tested in this way and all were found to be tight. Continuing at this pace will result in 50% of the active wells being tested at the end of 2017.

6. Root cause analysis A root cause analysis (RCA) comprising well book investigations was started in June 2016 and was extended in December 2016 with the ongoing insights of the integrity test program until that time, i.e. on five priority-1 wells (all found to be not tight), eight priority-2 wells (of which four were found tight and four were found not tight) and 15 active SCC wells (all tight). From this RCA the following conclusions were drawn. The most probable root cause for the largest group of defect wells is the use of (possibly re-used) casings of unequal length during the time period 1986-1989. The casings at these specific wells were possibly not tightened with the correct torque (signs of both under- and over torque were observed), causing some, but not all wells of this group to be not tight. The cementation is often not all the way to surface level and shows differences in quality. For this reason, leakage paths up to the conductor shoe may have developed, causing blanket oil to leak from the annular space. At the conductor shoe, it may have been collected until finally making its way to the shallow groundwater. The other group of not tight wells consisted of older wells, drilled between 1960 and 1975. No single root cause was found, but several causes or indicators for integrity issues were found, including: - the first wells drilled with the H20 drill rig, that was taken into use in 1960, are suspect. Possibly casings were used that were less suitable for being tightened with a hydraulic power tong and experience with the use of this new piece of equipment still had to be built up; - wells drilled by an external drilling contractor between 1965 and 1971. Possibly tightening of the casings was not always done properly, although the number of investigated wells in this RCA was still too small to certify this conclusion; - wells of which the conductor was removed shortly after the well was drilled. Between 1960 and 1970, this was more or less common practice because the conductor could then be re- used at the next well. In case of suboptimal cementation quality, groundwater may have caused corrosion of the last cemented casing, especially at groundwater level.

Following the results of this RCA, another six wells were added to the priority-2 group of wells. These wells were also drilled in 1986-1989, but do not have casings with unequal lengths.

Furthermore the results of the RCA were used to examine the level of suspicion of the remaining inactive wells and a preliminary group of priority-3 wells was formed. Besides these priority-3 wells based on the results of the RCA the other 284 open wells were labelled “no priority” meaning that they were no longer suspect. At the end of 2017 a further RCA update is planned to examine which of the priority-3 wells should still be tested in 2018.

7. Further investigations and integrity enhancements After completion of all priority-2 wells, the well integrity program will continue with the priority-3 wells in 2018. Next to the research on open wells, also wells that were abandoned at some time during the last decades will be subject to investigation. Based on the results of the RCA also these will be prioritized. These investigations will focus on the environmental (soil and groundwater) situation first, because the boreholes are no longer accessible for investigation.

The integrity tests of wells currently in use for brine production will continue and it is expected that in the course of 2019 all active wells have undergone a pressure test.

In the meantime the procedures for new drillings have been checked and further enhancements have been made to ensure the integrity of new wells and to minimize the chance of the development of any breach of integrity in the future.

8. References Den Hartogh, A.M., Steveninck, R. van, Leusink, H.J., Schicht, T. & Pinkse, T.M., 2017. Preventing subsidence caused by cavern migration in Hengelo and Enschede, The Netherlands. A risk based approach to monitoring and backfilling potentially instable caverns in the Hengelo brinefield, The Netherlands. SMRI Fall 2017 Technical Conference 25- 26 September 2017.

Koopmans, T.P.F., Groenenberg, R. & Pinkse, T.M., 2014. 21st century gasoil storage in Twente, the Netherlands. State-of-the-art, multiple-barrier design based on a novel risk management approach. SMRI Fall 2014 Technical Conference 29 - 30 September 2014.