JPT logo
SPE logo

Technology Update

Use of TEG-MeOH To Control Wellbore Hydrate Formation

The Southwest Kinsale (SWK) reservoir is part of the Kinsale Head reservoir in the Celtic Sea offshore Ireland. SWK was completed with three subsea wells tied back to the Kinsale Head platforms in 1999 (Well 3) and 2001 (Well 4 and Well 5). After initial depletion, SWK was converted to gas storage in ­2001–02 and has been operated by PSE Kinsale Energy each year since. The storage is seasonal with injection in May to September and withdrawal principally from December to March. The nominal capacity is 8 Bcf with a maximum flow rate of approximately 95 MMscf/D.

Background of the Problem

The operation has experienced gas hydrate formation during the latter part of each withdrawal season since 2006. In 2011, the company decided to expand its storage business. Hydrate formation was identified as the key technical risk to that expansion. While the company has an excellent record of meeting withdrawal nominations, isolated incidences of hydrate formation have proved challenging. In March 2011, the company set up a structured program to understand and resolve the problem.

The majority of the withdrawal uses “free flow,” i.e., reservoir pressure, to generate gas flow over the 7-km distance to the Bravo platform, which has a choke and flowmeter, and thereafter, the 5-km distance to the Alpha platform, where the withdrawn gas is processed by separation and triethylene glycol (TEG) dehydration. From there, the gas travels another 50 km to the shore terminal. The flow can be boosted during peak demand in the latter part of withdrawal by a boost compressor upstream of the TEG dehydration. Production gas from other reservoirs commingles upstream of the TEG process.

There is neither downhole pressure nor temperature indication in the wells, nor is there a downhole chemical dosing point. On each Christmas tree between the master and wing valves, there is a chemical injection point (fed by two umbilical cores), a pressure sensor, and a temperature sensor. The hydrate inhibitor historically used was methanol (MeOH). No corrosion inhibitor is injected. The wells are in 90 m of water and completed at approximately 900 m total vertical depth below mean sea level. The reservoir temperature is approximately 27°C. Wells 4 and 5 are deviated while Well 3 is vertical.

The Experience With Hydrate

The native gas is lean at >98% methane and 0.2 mol% CO2, with predicted hydrate formation temperature of 4.5°C at 30 bar at typical water saturations (Fig. 1). Operating wellhead flowing pressures (WHFP) range from approximately 45 bar to 80 bar. Hydrate incidences appear to be occurring more frequently. The tendency for hydrate to form is compounded by progressive enrichment by the comparatively rich network gas injected into the formation each year. Historically, identifying hydrate incidences was hampered by frequent changes in flow rate because of operational and commercial requirements.

Production operations work had identified several possible causes of hydrate. The first occurs on the Bravo topsides choke and flowmeter run. This was mostly resolved by removing the flow straightener. Any current incidences of topsides hydrate are detected by flowmeter noise or a decrease in flow accompanied by rising WHFP. Local MeOH dosing or small actuations of the choke are used to resolve the problem.

The second hydrate formation region appeared to be in the well areas. Some background on the issue follows.

Studying Hydrate Formation

The first task was to develop a tool to predict the withdrawal flow capacity. Company reservoir engineers used a simple material balance model and developed this to account for the particular reservoir behavior, including compartments and regional connectivity. The result was a spreadsheet tool that reliably predicted flow capacity (+/- a few percent) at given WHFP and for a particular reservoir condition of gas in place and flow history. While the tool enabled operations to identify incidences of reduced throughput (indicating hydrate formation), it did not indicate where the hydrate was forming or provide knowledge of why it was forming.

The operating procedures to remove the hydrate were not consistently effective and consisted of shutting in the wells by closing the wing valve and injecting MeOH into the wellbore. “Normal” shut-in pressure was achieved on several occasions. However, the flow remained impeded, indicating the difficulty of verifying removal of hydrate plugs. There were also several instances in which hydrate plugs had formed but porosity allowed some flow to continue.

Early flow assurance work suggested that the wells should be operated at low pressure, below the hydrate formation curve (Fig. 1). An attempt in March 2011 to operate at the lowest possible pressures (20 bar) by “cold starting” the compressor to draw down the pressure of two flowing wells, following several days of shut-in, was unsuccessful. Hydrate started to restrict the flow noticeably within 2 to 4 hours, despite injection of significant levels of MeOH into the wells beforehand. It was evident that the MeOH had evaporated too quickly to inhibit hydrate formation sufficiently.

During February and March 2011, the apparent hydrate order of the wells was observed. Flow assurance work was initiated using a standard commercial dynamic flow assurance software package to model the wellbore and heat transfer into the surroundings. The modeling corroborated the observed hydrate tendency. The longer deviated wells with nitrogen in the annulus were more prone to downhole hydrate than the vertical well with annular brine. The higher flowing deviated well was repeatedly the first to experience flow restriction.

An analysis of the Christmas tree temperature responses showed that the two functioning devices were measuring different variables. The relatively invariable device on Well 3 measured seabed conditions (verified by seasonal temperature variation), whereas the apparently unstable device on Well 4 measured flowing temperature (the temperature decreasing with increased flow and equalling Well 3 when there was no flow). Well 4 was also the first to form hydrate. This enabled an analysis of wellhead flowing temperature (WHFT) to detect hydrate formation in the upper wellbore. Decreases in WHFT during otherwise steady flowing conditions indicate Joule-Thomson expansion across a hydrate restriction and enable the operator to identify hydrate formation.

Further WHFT observations included a dynamic temperature response after shut-in, whereby the temperature decreased and “undershot” the seabed temperature by 1.5°C to 4°C before returning to a steady-state temperature. It was also noticed that the flowing temperature varied throughout the season. A simple radial reservoir model was simulated in a standard commercial software package with a thermal-compositional module. The work demonstrated the effect of injecting gas at seabed temperatures and how the temperature of the subsequent withdrawal gas varied at postulated bottomhole conditions. The bottomhole temperature progressively decreased over five storage cycles from 24°C to 21°C.

In combination, the observations and studies agreed that the hydrate tendency had increased over the storage operation cycles and suggested that an alternative hydrate inhibitor would be more effective. The choices considered were TEG and methylene glycol (MEG). TEG is the more viscous of the two and has a lower vapor pressure. Modeling of the inhibitor loss caused by evaporation indicated that TEG would form a thicker film than MEG and persist longer downhole. MEG is not compatible with the TEG dehydration system because it will acidify in the regeneration system. Thus, despite the advantages of MEG in cost and inhibition efficiency, TEG was selected.

The TEG-MeOH Trial

In early February 2012, the withdrawal flow was restricted because of hydrate in line, yielding all the observations cited above regarding well hydrate tendencies, the WHFT decrease, and the capacity tool. The MeOH dosing was unable to remove the hydrate. The operation was forced to revert to lower flow, which at the higher flowline pressures also meant warmer gas and, therefore, hydrate dissociation.

In mid-February, a quantity of 70% TEG (cut with MeOH to reduce its viscosity to suit the existing pump specification) was injected at 18- and 24-hour intervals. The dose quantity was 20 US gal injected over approximately 50 minutes. This dose was calculated to create a film coating of approximately the upper 100 m of each well (tubing diameter 5½ in.). This was the target region of the wellbore because the downhole safety valves on the three wells were at approximately 100 m subsurface. Performance was tracked by comparing the achieved flow with the predicted flow for the WHFP.

Two dose intervals were longer than 24 hours (as long as 30 hours). It was found that these extended dosing intervals reduced the flow performance afterward. Also, during the run, the compressor speed was used to increase the flow rate. The performance was always able to achieve >92% of the predicted flow rate (Fig. 2). Repeatable performance with TEG-MeOH was demonstrated. Before the switch to using TEG-MeOH, the high-flow operation had only lasted 5 days before a significant hydrate event.

More tests corroborated the repeatability of the TEG dosing operation. This included a “dummy dose” when the wells were shut in as if they were dosed with the TEG-MeOH but were not. When the wells were reopened, there was an increase in flow (compared with the flow before shut-in). However, the flow reverted to the rate before shut-in within 2 to 3 hours. This showed that a) the diurnal variation in flow rate was not caused by reservoir effects, b) a hydrate did form during the interval between doses, and c) the TEG-MeOH dosing was controlling the hydrate.

2013 Operations

In summer 2012, the TEG-MeOH injection system was reinforced with a larger capacity pump and the process piping was rearranged to allow a continuous flow during compressor startup. A new infrared moisture analyzer was installed. During the hydrate risk period in the 2012–13 winter withdrawal season, the TEG-MeOH injection was rigorously applied at 24-hour intervals. The above refinements minimized the shut-in durations, thereby minimising cooling of the upper wellbore.

The topsides moisture analyzer showed a dehydration effect because of the injected chemicals, pointing to further reduction of hydrate formation risk. Operationally, the TEG-MeOH injection resulted in additional liquid volumes in the inlet separator, with the volume separated amounting to about 80% of injected volume. Previous dosing with equivalent volumes of methanol had yielded practically no separator liquids.

Winter withdrawal has recently concluded without a single hydrate incident. The confidence in the operation helped the operator to maximize uptime, and extra gas was available for withdrawal from storage during the March cold weather period.


The operation concluded that batch dosing TEG-MeOH allows control of downhole hydrate in a subsea gas storage well that has no downhole dosing capabilities. The intervention has allowed the operator to reliably meet peak customer demand in the highest risk period of storage gas withdrawal.

Treatment Enables High-TDS Water Use as Base Fluid for Hydraulic Fracturing

Paul D. Lord, SPE, and Renee LeBas, SPE, Halliburton

The beneficial use of produced water is receiving great attention in the oil and gas industry. As oil and gas are produced, large quantities of water containing high amounts of total dissolved solids (TDS) are produced in the process. Most of the high-saline produced water is reinjected into disposal wells. Reclaiming produced water as the base fluid for hydraulic fracturing not only helps to alleviate the industry’s dependence on fresh water, but also lowers the overall cost of the fracturing treatment.

Contents of Produced Water

Produced water usually comprises the formation water and the injected fluids from previous treatments. It can contain hydrocarbons, high levels of TDS, suspended solids, and residual production chemicals. The use of produced water for hydraulic fracturing has many advantages, such as reduction in disposal of produced water, reduction of freshwater consumption during completion and production operations, and realization of economic benefits by the operator. If produced water is gathered at or near the production site and a minimal treatment is applied in hydraulic fracturing fluids, recycling and reuse programs may become economically and environmentally beneficial.

As hydraulic fracturing in shale plays continues to require large amounts of fresh water for oilfield operations (i.e., 4 to 6 million gallons per well in some cases), reusing produced water reduces freshwater consumption. With increased restrictions on water availability from subsurface or surface sources, fresh water is becoming more difficult to obtain.

In some basins, the use of produced water in oil and gas drilling and slickwater-based fracturing treatments has been explored. Typically, these applications use water with low TDS levels. Little work has been done on the use of produced water with high TDS levels (>200,000 ppm) in fracturing fluids designed with linear or crosslinked gel bases.

Treating High-TDS Water

To reuse high-TDS produced water effectively in crosslinked gel-based hydraulic fracturing fluids, the water must first be treated. The treatment need not generate potable or fresh water, as some oil and gas industry water treatments do. The goal of the treatment is to remove only minerals that hinder the development of the crosslinked fluid or that cause scale buildup in the well. One effective form of produced water treatment is electrocoagulation (EC).

In EC, an electric current is applied across metal plates to remove various contaminants from the water. Heavy metals (ions) and colloids (organics and inorganics) are primarily held in solution by electrical charges and particle size. The electrical charge applied to the solution of contaminated water destabilizes the charges on the particles and generates a coagulation reaction.

A produced water sample was treated using a bench-scale EC process. This step in the recycling program removes specific contaminants, including suspended solids, which allows the treated water to be used in future fracturing applications. Fig. 3 shows a visual comparison of produced water samples, one untreated and the other treated. In the EC treatment, the suspended solids are concentrated into a waste stream that accounts for 3% to 5% of the total volume. This waste stream can be reduced further by approximately 90% through the use of a filter press. This step enables the removal of the dry solids and the filtration of the water, which is returned to the EC treatment stream.

Scaling Tendencies

Produced water can contain high levels of scaling ions such as barite, calcite, and siderite. It is important to reduce these ions, when cost-effective, and control the scale, when ion removal is not economic. This requires predicting scaling tendencies and developing a scale inhibitor suitable for use in fracturing fluids.

The Pitzer theory of electrolytes calculates the scaling potential of water over a wide range of temperatures, pressures, TDS, and cosolvents. A scaling model based on the Pitzer theory can be used to assess the scaling tendency of produced brines under the temperatures and pressures of a potential fracturing treatment.

EC is very effective in removing total suspended solids, turbidity, oil and grease, and heavy metals, i.e., copper, zinc, lead, arsenic, chromium, and iron. The removal of heavy metals and the reduction of divalent ions typically results in EC-treated produced water having lower scaling tendencies than untreated produced water. When necessary, treated produced water can be tested by use of a dynamic scale loop to determine the minimum effective dose of scale inhibitor. It is recommended that scaling tendency tests be performed before development of a fracturing fluid so that the scale inhibitor is included in the fluid analysis.

Recycle and Reuse

An effective fracturing fluid must have these characteristics: easy preparation, low fluid loss, good proppant transport capacity (viscosity), low pipe friction, efficient recovery from the reservoir, and low gel residue. Through the use of gelling agents, fluids achieve excellent viscosity. For water-based fluids, guar gum and their derivatives, such as hydroxypropyl guar and carboxymethylhydroxypropyl guar (CMHPG), can be used. The derivative guars generally have less residues and faster hydration compared with non-high-yield guars. Adding borate crosslinkers or metal-based crosslinkers, such as zirconium, titanium, or aluminum, significantly increases the viscosity of guar-polymer gels.

Crosslinked fluids have a greater capacity to transport proppants than uncrosslinked fluids. This transport capacity is measured by a rheometer in which fluid is tested at specific temperatures and shear rates. In these tests, it is possible to observe the transporting elastic regime and proppant settling. A minimum transport capacity of 400 cp measured at 40 sec–1 with a B2 bob is commonly desired. This value was used as a reference for fracturing fluid tests developed using the rheometer.

The proprietary fluid system used in this study is designed to hydraulically fracture a sandstone formation (~6,800 ft in true vertical depth [TVD]) in the Permian Basin. The fluid consisted of a derivative polysaccharide guar gum, a metal-based crosslinker, oxidizing breaker, breaker catalyst, and a nonemulsifier surfactant. The bottomhole temperature (BHT) tested was 140°F. As seen in Fig. 4, the crosslinked fluid with 100% EC-­treated produced water retained more than the desired minimum viscosity of 400 cp for the required pump time of 60 min.

Field Trial

A field trial was performed in which seven wells with a total of 97 fracturing stages were tested using 100% ­EC-treated produced water. The well stimulations were performed to evaluate the validity of the laboratory test findings under field conditions in New Mexico. The sandstone formation depth was approximately 6,800 ft in TVD (11,300 ft in measured depth) with an approximate BHT of 140°F. The fracturing fluid system was a 30-lbm derivative crosslinked gel. The 30-lbm derivative crosslinked fluid was also used in the bench-scale testing.

Dynamic break tests of a 30-lbm derivative crosslinked gel with treated produced water as the base fluid were conducted in the laboratory using a Fann 35 viscometer to help ensure proper fluid rheology before stimulation. The derivative guar gum was hydrated in the treated produced water at a pH of less than 7, before an oxidizing breaker and a breaker activator were added.

The gel was crosslinked using a metal-based crosslinker at 140°F. To break the crosslinked gel on an optimal basis, the amounts of breaker and breaker activator used were adjusted. The tested fluid was designed to maintain greater than 400 cp—the minimum viscosity recommended for proppant transport capabilities—for 60 min. The fluid performed as specified in the test procedures.

One trial well was stimulated using 1.4 million gallons of treated produced fluid and approximately 1.7 million lbm of proppant placed in 17 treatment stages. The average treatment pressure was 1,911 psi and the average treatment rate was 47 bbl/min. The pump time for each stage was approximately 55 min and consisted of pad and proppant stages with an average maximum proppant concentration of 4.4 lbm/gal. No mechanical failure or screening out was encountered during the stimulation. The treating pressure trends observed were as modeled with 2% KCl water as the base fluid, indicating no adverse effects on the treatment pressure or ability to place proppant when using high-TDS treated produced water (>270,000 ppm).

Reducing Use of Fresh Water

Based on the results of the trial, it is feasible to use treated produced water with more than 270,000 ppm of TDS as the base fluid for hydraulic fracturing. The fluid performed comparably to fluid developed using fresh water as the base fluid under the same test conditions. Initial oil and gas production rates from the stimulated wells in the field trial proved comparable to, if not better than, than the production rates from an offset well.

By applying a minimal and cost-efficient treatment to the produced water and designing the appropriate fluid chemistry, the treated produced water was effectively used as a base fluid in a crosslinked application. More than 8 million gallons of fresh water were saved and more than 1,400 truckloads were eliminated in this multi-well trial, with the operator realizing a significant savings in water management costs. These results show a considerable potential for the industry to reduce its dependence on fresh water for hydraulic fracturing and, hence, reduce the cost of fracturing treatments.