Special Report: Reverse osmosis treatment of CBM produced water continues to evolve

Oct. 5, 2009
A comparison of two treatment plant designs shows the evolution of the reverse osmosis process for treating produced water from Powder River basin coalbed methane wells.

A comparison of two treatment plant designs shows the evolution of the reverse osmosis process for treating produced water from Powder River basin coalbed methane wells.

Both designs incorporate reverse osmosis (RO) and recovery reverse osmosis (RRO) because this configuration has proven effective for meeting produced water treatment objectives.

Each year US oil and gas onshore operations generate about 15-20 billion bbl of produced water. By comparison, the Wyoming, Powder River basin produces between 700 and 900 million bbl/year of water from natural gas wells.

CBM production

Coalbed methane recovery techniques are unique when compared with other production methods because hydrostatic pressure holds the methane in the coal seam so that gas production requires removal of formation water or dewatering.

Removing the formation water depressurizes the formation, thus releasing the gas. Initial water production is high but decreases rapidly to allow release of the methane.

Producers must manage these considerable volumes of water generated during the dewatering process. Much of the water can be disposed of by direct discharge given the high quality of the CBM produced water in the Powder River basin.

Operators must manage produced water of a lower quality, however, depending on environmental compliance and economic objectives. This would include volume of produced water, proximity to surface water, rights-of-way, influent chemistry, discharge quality requirements, land use provisions (public or private) and recycle objectives.

The reverse osmosis RO-RRO process has been permitted through the Wyoming Department of Environmental Quality/Water Quality Division (WDEQ) Chapter 3 process, requiring review and monitoring by the department's water and waste water division engineers.

The produced water stream enters a ballast pond before being pumped to the Wild Turkey treatment facility (Fig. 1).

Both plants minimize waste by maximizing system recovery, and use an aeration pond for evaporating and concentrating the brine (Figs. 1 and 2). The plants include a design with bypass and blend provisions so that the plants can blend the produced water to a wide range of discharge specifications.

Waste water from the Wild Turkey system goes to evaporation ponds (Fig2).

Both plants maximize membrane performance with filtration and scale control, but differ in the approach to controlling scale (Fig. 3). It is the nature of the scale control that is the primary focus of this article.

The Wild Turkey plant RO-based treatment system prepares the water for surface discharge (Fig. 3).

The article also will discuss selected components of the treatment process for each plant and lessons learned as the plant design matured with treatment experience and the producer's needs.

Influent, effluent criteria

One must clearly understand feedwater characteristics for proper treatment plant design. This includes seasonal variability that may identify influent extremes or complex chemistries. Waste and product stream characteristics must also be understood so that service factor, redundancy, and compliance can be addressed in the plant system design.

By its inherent nature, CBM water is high in sodium and bicarbonate and low in hardness, and may also include suspended solids, iron, silica, and barium.

Sodium is a closely monitored aspect of the treatment plant effluent. Soils with an excess of sodium ions, as compared to calcium and magnesium ions, can affect the way plants adsorb water. The ratio of the sodium to calcium and magnesium is referred to as the sodium adsorption ratio (SAR).

The plants also require Wyoming Pollutant Discharge Elimination System (WYPDES) permits issued by the Wyoming DEQ for construction, operation and discharge of the produced water. The plants can discharge greater than 95% of the influent water into the Powder River.

The state permitting authority defines effluent standards to protect aquatic life and downstream uses of the water. The treatment systems have sufficient flexibility to meet the defined effluent recipe as it changes on a monthly basis.

Application engineers use solubility indices to understand the relationship of the dissolved ions as they move through the treatment process. For instance, one technique for predicting calcium carbonate solubility considers the bicarbonate carbonate and calcium concentration to access the potential for hardness scale formation.

This is the concept behind the Langelier saturation index (LSI). A positive LSI denotes an increased potential for calcium carbonate scale formation while a negative LSI denotes that calcium carbonate may dissolve in the solution.

LSI is one of the many solubility indices that facilitate design engineers' understanding of ion interactions as water chemistries change through a process. This information helps designers control the severity of the process and applies appropriate equipment and chemistries to moderate the behavior of the water as it progresses through the process.

Another constituent common in CBM water is silica. Because of its unique chemistry, silica poses special treatment challenges to design engineers. While the silica concentration in Powder River basin produced water is moderate, the high recovery rate of the membrane system creates ideal conditions for silica to scale membrane surfaces. Silica precipitation control is complicated further because control techniques for other ions conflict with methods for controlling silica.

Geographical, environmental concerns

The Powder River basin is a sparsely populated region, and unlike water treatment plants in industrial or municipal applications, the plants have intermittent manpower coverage. This must be considered when designing the system to ensure sufficient redundancy to address uptime and reliability objectives. Key considerations include redundancy, call-out features, response time, and safety.

The two treatment plants discussed in this article are in the Powder River basin. Given the potential for inclement weather, inventory controls must incorporate the possibility for restricted site access during seasonal extremes.

One must carefully design acid feed systems to minimize risks to personnel and facility. The volume of hydrochloric acid needed to neutralize the alkalinity inherent in the CBM water is considerable.

Tanker trucks deliver the acid, often down lease roads and potentially during severe weather. The facility should:

• Store the acid outdoors in double-contained tanks.

• Have double-contained feed lines and valves.

• Locate tanks as close as possible to injection points to minimize the feed line length.

Another key criterion in system design is meeting discharge specifications to comply with WDEQ specifications for protecting aquatic life from toxicity. The test commonly used to confirm compliance is the Whole Effluent Toxicity Test, or WET test.

The test involves collecting effluent samples at appropriate outfalls and analyzing them to determine the effect of the discharged water on aquatic life in the receiving waters.

The acute WET test is a 48-hr static test using daphnia magna (water flea) and an acute 96-hr static test using pimephales promelas (fathead minnow), as collected from designated outfalls.

Toxicity occurs if mortality exceeds 50% for either species at the effluent concentrations. Chronic WET testing is a 7-day test using pimephales promelas. The test requires collecting a series of composite samples during several days.

The test dilutes the subject water with synthetic lab water to evaluate the degree of toxicity as compared to the lab control sample.

First plant

The first CBM treatment plant installed by Siemens for Petro-Canada Resources (USA) was at the company's Wild Turkey operations near Gillette, Wyo. The plant, commissioned in 2006, has a design for processing 120,000 b/d of produced water at peak production and discharging treated water to a Powder River tributary.

The plant's design allows for the discharge to have a blended sodium level and to meet WET standards. Fig. 4 shows a schematic of the plant.

Because Petro-Canada was eager to start processing the produced water as soon as possible, the project involved placing a temporary mobile treatment system online during the installation of the permanent system.

The mobile system components included media filtration and RO skids. The skids were contained in fully automated trailers that included instrumentation and climate control. Siemens personnel located onsite operated and maintained the application with support and critical spares sourced from the Siemens Colorado Springs, Colo., branch office.

Second plant

In 2008, Petro-Canada awarded Siemens a second operating contract for the treatment of Powder River basin CBM water at Mitchell Draw, also near Gillette. As of the date of article, the company has not commissioned the plant. The plant has a design for treating 72,000 b/d.

The Siemens engineers wanted to advance the Wild Turkey plant design by focusing on hardness and silica scale formation and acid feed.

Borrowing on capabilities introduced by DOW in 1983,1 Siemens added ion exchange softening into the process flow as a key innovation over the Wild Turkey design. Ion exchange removes polyvalent cations from the feedwater. The process removes constituents such as calcium, magnesium, barium, and soluble iron to very low levels by exchanging them for sodium on the ion exchange resin.

On first review, adding a strong acid cation sodium-form softener may not be an obvious addition because sodium is a strictly controlled effluent contaminant. The amount of calcium and magnesium in the CBM water relative to the amount of sodium, however, is low, so that the percentage increase in the amount of sodium is low.

The softener provides several advantages. First, it reduces the potential for scale formation by removing dissolved cations such as calcium, magnesium, and barium. This reduces the antiscalant and acid chemical requirement typically used for controlling solubility when influent concentration or system design affects the solubility limits.

The Mitchell Draw plant will operate at a higher pH than the Wild Turkey plant. As stated previously, the process typically feeds acid to a neutral or slightly acidic pH range to control hardness scale. Without acid feed, the bicarbonate alkalinity concentration increases, resulting in an alkaline feedwater condition.

The higher pH offers preferred operating parameters that increase the solubility of residual organics, thus reducing the potential for organic fouling on the membrane surface. The higher pH shifts the boric acid to borate equilibrium so that the membrane more easily rejects the boron, resulting in lower boron concentrations in the effluent water. The higher pH increases silica solubility, thus lowering the potential for silica fouling.

Fig. 5 shows a flow schematic of Mitchell Draw.

Lessons learned

The Wild Turkey treatment system was a success because it provided the intended recovery rate, was reliable, and handled changing effluent water standards. The system design followed a conventional approach including influent settling, media filtration-iron removal, acid feed, and RO.

Because disposal cost is a primary driver for produced water projects, the process met this objective by maximizing system recovery to minimize brine disposal costs.

Managing the soluble iron was one of the challenges addressed by the design engineers. Iron and other soluble metals can oxidize within the membrane modules and foul the membranes, resulting in reduced performance.

Wild Turkey produced water averaged greater than 10 ppm; however, it is not uncommon for Powder River basin CBM water to contain 20-30 ppm of dissolved iron.

On initial inspection, iron-laden produced water may appear clear, but as the water is exposed to air and the iron oxidizes, the water takes on a rust color and becomes more turbid, representing increased loading for media filters and potential fouling of the membrane systems.

Petro-Canada addressed the soluble iron problem by installing a "riprap" system, which is commonly installed to control bank erosion. This application allows the influent to cascade over coarse stones to oxygenate the water and thus oxidize the iron prior to entering the influent equalization basin. As a precautionary step, the process adds chlorine to oxidize further the soluble iron. This process successfully converted the iron to an insoluble form that the iron removal filtration media could eliminate.

The Wild Turkey plant configuration addressed the dynamic operating conditions because produced water systems do not operate at steady-state conditions. As defined by the discharge permit, effluent standards change monthly so as to limit the sodium loading into the receiving water.

A second challenge was the changing influent silica concentration. Silica is a primary factor affecting system recovery for membrane systems. In the Wild Turkey plant, influent silica fluctuated between 8 and 12 ppm, which resulted in a corresponding fluctuation in RO recovery, between 92% and 96% (Fig. 6).

The following factors affected system productivity:

• When influent silica concentration was above design peak, RO system recovery was reduced.

• Silica concentration in brine needs to be limited for sustainable RO membrane performance.

• Antiscalant may not protect against silica scaling above a certain level.

• Silica levels appear to vary over time, either seasonally and/or with extended production from the CBM wells.

The Wild Turkey plant design addressed dynamic influent and effluent conditions. Key to this capability was the excess capacity on the RO system and control strategies that optimized the system configuration.

Acid feed control was vital for reliable operations of the Wild Turkey plant. Failure to control properly acid addition can result in scale formation on membrane systems. For instance, there was a period in which changing influent conditions resulted in scale formation on piping and the valves between the primary RO and the recovery RO. Increasing the acid dosage to reduce the pH reversed the scaling process.

Mitchell Draw improvements

Siemens expects the following benefits from the Mitchell plant, after it is commissioned.

CBM water is characteristically high in sodium and low in calcium and magnesium hardness. The low hardness concentration makes it an ideal application for sodium-form ion exchange. Removal of the hardness ions reduces the risk of hardness scale formation and the need for acid.

CBM water contains relatively high concentrations of bicarbonate alkalinity. Given the concentration effect across a membrane system, the pH of the feedwater will increase to increase the solubility of silica and residual organics and improve the rejection of boron.

The process eliminates acid for scale control, resulting in improved plant safety and system reliability. The plant does not need trucks hauling acid, thereby removing them from public highways and reducing risks to the plant and personnel.

The new design removes acid feed equipment and controls from the system because pH control is no longer a primary concern to system operation. In addition, it improves system reliability because inventory control is less vulnerable to delivery interruptions.

Water analysis comparison

The table above illustrates the changes to feedwater quality while applying each scale control technique. The feedwater data are approximated from actual Powder River basin produced water analyses. The analyses are then modeled using the Dow Chemical Co.'s Reverse Osmosis System Analysis (ROSA) program to illustrate the changing conditions.

In the pH-adjusted feedwater case (Wild Turkey plant), the process reduces the pH to 7.0 from 8.2 with hydrochloric acid as compared with sulfuric acid, since the sulfate in sulfuric acid can cause scaling as it bonds with divalent cations present in the feedwater.

Acid addition changes the characteristics of the water as hydrogen neutralizes bicarbonate to form carbon dioxide and chloride. This results in a marked increase in chloride concentration from the contribution of hydrochloric acid and a reduction in raw water LSI to –0.1 from 1.2.

In the softened feedwater case (Mitchell Draw plant), the plant reduces the divalent cations calcium, magnesium, strontium and barium to less than 0.3 ppm in the adjusted feedwater. There is a corresponding increase in sodium content as an equivalent amount of sodium is exchanged into the water.

Unlike the acid feed case (Wild Turkey plant), the pH and total dissolved solids (TDS) stay about the same. The LSI scaling calculation shows –1.3 in the adjusted feedwater. Furthermore, the barium sulfate and calcium fluoride scaling calculations also show reduced scaling potential with the softened feedwater case (Mitchell Draw).

Reference

1. Reyes, R.B., "Softening of Oilfield Produced Water by Ion Exchange for Alkaline Flooding and Steamflooding," Paper No. SPE 11706, SPE California Regional Meeting, Ventura, Calif., Mar. 23-25, 1983.

The author

James Welch (James.P.Welch @siemens.com) is the business development manager for onshore produce water for Siemens Water Technologies, Houston. His primary focus is on promoting innovation while working with clients to develop produced water treatment solutions. Welch has a BS in chemistry for Stephen F. Austin, a BA in marketing from the University of Texas, and an MBA from the University of Houston.

More Oil & Gas Journal Current Issue Articles
More Oil & Gas Journal Archives Issue Articles