Gas producer upgrades water facilities to comply with strict Colorado emission rules

PAUL WHITE, Williams Production

Williams Production operates gas wells producing primarily from the Williams Fork formation in the Piceance basin of western Colorado. The gas wells also yield 58°API condensate and produced water. Williams separates and sells the condensate at the well pad and then transports the produced water to a centralized facility for further treatment. The treated water is stored in lined ponds, and then transported back out to the field for use in completion operations. Some of the water is disposed of into injection wells.

Recently, the Williams centralized water treatment facility required an upgrade to provide additional water throughput and to reduce hydrocarbon emissions both from its process equipment and at the downstream storage ponds. Removing hydrocarbons from Williams Fork produced water and controlling air emissions is problematic due to soluble volatile organic compounds (VOCs) from the condensate, suspended solids and the methanol used in operations. The operator identified the issues, arrived at the required process design, upgraded the facility to reduce air emissions in compliance with Colorado’s stringent air regulations, and accomplished the operational goal of increased facility throughput and storage.

BACKGROUND

Williams’ daily water production of 40,000 bbl consists of about 50% produced formation water and 50% completion flowback. Until 2008, all water processed at the centralized facility was either recycled for completion operations or evaporated. When drilling decreased in late 2008 and 2009, Williams began injection disposal of excess recycled water.

Centralized water treatment was initiated over 20 years ago. The treatment facility in Rulison, Colorado, initially consisted of a simple tank that received incoming fluids. The water exited the bottom of the tank into a water storage pond, while oil was segregated into a separate tank and then sold. As throughput increased, the facility grew to four inlet tanks operating in parallel, two parallel polishing tanks, and parallel box-type oil-water separators, Fig. 1. In addition, a small skim pond captured any remaining free-phase hydrocarbons—consisting mainly of C5 to C10 (condensate)—before the water entered the storage ponds. The condensate was transferred to condensate sale tanks. Small emulsion tanks batch treated emulsified liquids, sediment and waste from tank bottoms and other production operations to separate condensate and water.

Facility throughput typically averages 12,000 bpd. However, up to 90% of throughput occurs during daylight hours, and 70% between 8 a.m. and 4 p.m. The facility receives water primarily via vacuum trucks, which unload at four bays. It also receives water intermittently via pipeline. The water quality coming into the facility varies from nearly 100% water to 100% oil, with the amount of solids varying up to 50%.

In 2007, working with the Air Pollution Control Division within the Colorado Department of Public Health and Environment (CDPHE) to amend the facility’s air permit, Williams undertook an effort to upgrade the facility. New Colorado guidelines require application of reasonable available control technology (RACT) to control and reduce air emissions at oil and gas production-related water treatment facilities.

APPLICATION OF RACT

Fig. 1. Initial configuration of the centralized water treatment facilities in Rulison, Colorado.

Produced water storage ponds are subject to the RACT requirement contained in Regulation 7, Section V(A), which states, “No person shall dispose of volatile organic compounds by evaporation or spillage unless RACT is utilized.” Therefore, a RACT analysis must be conducted for produced water holding ponds that require a permit. Storage ponds require an emissions permit if all emission points at the facility combined emit actual uncontrolled emissions of 5 tons or more of VOCs per year.

RACT is defined in the Colorado Common Provisions as “technology that will achieve the maximum degree of emission control that a particular source is capable of meeting and that is reasonably available considering technological and economic feasibility. It may require technology that has been applied to similar, but not necessarily identical, source categories. It is not intended that extensive research and development be conducted before a given control technology can be applied to the source. This does not preclude requiring a short-term evaluation program to permit the application of a given technology to a particular type of source.”

Colorado allows only one of two methods to estimate emissions from storage ponds: mass balance or the US Environmental Protection Agency’s AP-42 model, both of which overestimate emissions.

The term “reasonable” accounts for both technical feasibility and economic considerations. Operators have the obligation to meet this standard and should evaluate multiple RACT options with the goal to maximize emission reductions and minimize cost. A RACT analysis should be provided with any storage pond application for a permit. CDPHE interprets RACT to mean that, once any other facility that it permits uses a certain technology to reduce emission, and then all such facilities requiring permits should do so. Technical feasibility and maximum emission reduction are emphasized; economics are not a factor considered in RACT selection by the Air Pollution Control Division.

In pre-RACT centralized water treatment operations, Williams was successful in its objective to remove and sell all free hydrocarbons using gravity separation and adequate retention time. While some solids settling occurred in the primary and polishing tanks, solids removal was not an objective. The company was not required to—and therefore did not—track separation, collect tank emissions and transport them to a combustor, or control hydrocarbon, VOC and methanol levels at the treatment outlet to the storage ponds.

CDPHE allows only one of two methods to estimate emissions from storage ponds. Their preferred method is for operators to use mass balance, which is extremely conservative and often overestimates actual emission by a factor of between 500% and 1,000%, and sometimes as much as 1,500%. The second choice is to use the US Environmental Protection Agency’s model found in AP-42 (Compilation of Air Pollutant Emission Factors) Chapter 4.3, “Waste water collection, treatment and storage.” The model was developed for estimating emissions from sewage collection and processing. AP-42 also overestimates emissions and drives reductions in emissions that are overestimated.

In addition, the RACT rule empowers CDPHE to peg limits of water going to the ponds at some combination of values that typically range 50–150 ppm for total recoverable petroleum hydrocarbons, 60–100 ppm for benzene, toluene, ethylbenzene and xylene (BTEX), and 100–700 ppm for methanol. This combination of ranges will depend on total petroleum hydrocarbon levels, the surface area of the pond and the biomass concentration, as well as the available technologies to remove each constituent. One of the control factors in the AP-42 model accounts for biological activity. In the Rulison facility’s recent air permit amendment, CDPHE required a minimum of 50 g/m3 of biomass, which is a default value in the AP-42 model.

As a result, Williams began evaluating the sparse historical water-testing data that was available in order to understand existing conditions versus the conditions that may be required under the new guidelines. In addition, the company enlisted the assistance of the Produced Water Society to plan and implement evaluation and testing programs and examine options to proceed.

Initial testing results showed that it would take considerable effort to bring the water entering the storage ponds within the new CDPHE limits. Hydrocarbon levels in the treated water ranged 200–600 ppm. Total suspended solids were 200–500 ppm, and biomass was less than 50 g/m3. These results were obtained in water samples that contained no visible free hydrocarbons. From these results, it was inferred that the water was contaminated with suspended and possibly dissolved hydrocarbons.

IDENTIFYING THE CULPRITS

Filtration tests indicated that dissolved hydrocarbons were generally less than 10 ppm, meaning that they contributed to the problem but were not the major issue. Filtration tests also revealed that the solids in the filter cake contained oily residue. Filtration testing in the lab did measure significant improvement in water quality; there was little doubt that the high levels of solids, combined with the intermittent flow and surges of throughput through the facility, contributed to the lack of separation at the facility. However, filtration was not considered a reasonable process for full-scale implementation.

To understand the effects of retention, a sample of the inlet water was allowed to sit in a test vessel up to 12 hours, while the water from the bottom of the tank was periodically tested. It was determined that gravity separation with additional static time provided only marginal additional benefit.

Testing for hydrocarbons between the various stages showed no significant improvement in water quality between the polishing tanks and box separators, or after the box separators. The failure of the treatment train’s later stages to remove hydrocarbons was attributed to instantaneous flowrates as high as 40,000 bpd when multiple vacuum trucks unloaded, which caused the disturbance and carryover of solids from the trucks that settled and built up inside the facility’s inlet and polishing tanks.

UPGRADE CRITERIA

The final revised air permit had to reflect all vessels, air emission points and controls. This required an iterative process to design the equipment that could meet the air emission standards required and still allow the company to recycle water and operate as needed. In addition to removing hydrocarbons, reducing methanol and increasing biomass, there were other considerations centering on air permit compliance.

Hydrocarbon separation. The operator wanted to add whatever retention time was possible using larger and quieter front-end tanks that were capable of holding a constant water/oil interface, as well as more retention time in quieter polishing tanks. The company also sought to improve gravity separation through some method of enhanced gravity separation, and considered additional polishing downstream of this enhanced separation to remove dissolved hydrocarbons, if required.

A flexible process configuration was considered, to allow for parallel or series flow through vessels as needed. Increased pipe diameters were considered a necessity to reduce bottlenecks and allow the treatment process to work at high instantaneous flowrates. In addition, the operator sought to upgrade the quality of field gas with better liquids scrubbing and the installation of a glycol unit to remove moisture from supply gas.

Solids and methanol. Solids buildup in the tanks was considered responsible for the lack of significant hydrocarbon removal in the pre-RACT treatment train. Therefore, it was considered advantageous to prevent the vacuum trucks from injecting air into the inlet tanks while unloading, which caused solids settled at the bottom of the truck tanks to carry over into the inlet tanks. Also considered were process tank designs that would allow removal of solids from the tank bottoms without having to open tanks for cleanout, as well as capture and removal of free solids at the inlet.

Options for reducing methanol content included reduced methanol use and methanol recovery.

Measurement. The operator sought to install instrumentation that would allow personnel to view the water/oil, water/emulsion and emulsion/oil interfaces, along with level of solids in the tanks, without opening the tanks. In addition, flow-metering in numerous locations throughout the facility would allow documentation of throughput and results. Other measurement options considered were an automation system to control, track, record and timestamp all processes, in order to demonstrate regulatory compliance in the future, and a system to continuously measure hydrocarbon content in the water.

Fig. 2. New water treatment configuration.

Fig. 3. The new emulsion tanks treat waste materials from the dissolved air flotation vessel and front-end tank bottoms with heat and microbes to produce condensate, water and sediments.

Operating conditions and safety. The equipment would have to operate at an average elevation on 6,200 ft with annual temperature ranges from −25°F to 115°F. Thus, the operator sought to improve the overall freeze protection of the facility. In particular, replacing the direct-fired tank heaters with electrical or hot oil heating systems would add freeze protection, as well as improve safety and reduce air emissions. Another safety consideration was to increase the distance between the combustors and any vessels that produced vapors.

IMPLEMENTATION

Williams met all the upgrade criteria in its final design. They employed three parallel front-end treatment trains, each having a 1,000-bbl-capacity high-efficiency skim tank, and two 500-bbl-capacity polishing tanks in series, to replace the old skim tanks and box separators. Solids control, instrumentation, supervisory control and data acquisition (SCADA), required measurement and programmable logic controller (PLC) automation are now employed in all processes. Enhanced gravity separation using dissolved air flotation was employed downstream of the front-end trains to remove any remaining solids and suspended hydrocarbons, Fig. 2.

In addition, new, larger and more advanced emulsion tanks were tied into the treatment system by piping. (The old emulsion tanks were separate and required fluids to be trucked to them.) The emulsion tanks batch-treat emulsified liquids, sediment and waste from the dissolved air flotation vessel, tank bottoms and other production operations, Fig. 3. The tanks are heated to break the emulsion and separate condensate and water from the emulsion liquids. The remaining materials, consisting of higher-molecular-weight hydrocarbons and sediment, are transferred to a land farm for biodegradation.

The design was finalized and equipment was ordered beginning in August 2009. Construction began shortly thereafter. In order to retrofit the required equipment into the existing facility, substantial civil work was required to improve containment and to relocate lines and utilities.

Work continued into what turned out to be the area’s worst winter in many years, with record snowfall and cold. It was necessary to construct 3,000-ft2 hutches of framing lumber with polyethylene walls, heated by high-rate kerosene heaters, so that welding and fitting could continue. The hutches were disassembled and moved several times as required while work progressed.

Williams originally planned on using dissolved gas flotation for enhanced gravity separation. However, the field gas was found to be too wet to achieve the company’s water quality objectives, so air was injected instead of gas in the flotation tanks.

Since the facility has low throughput at certain times, a fluid recycle circuit was installed from the clean storage pond to the front end. This assists with freeze protection, allows the company to maintain levels, and keeps the system ready for surges of throughput.

Williams also devised a method to cultivate microbes using condensate as feedstock. Nutrients and the condensate-degrading microbes are added to the front end, emulsion treating system and storage ponds on a weekly basis to enhance biological activity and hydrocarbon degradation, thereby reducing hydrocarbon emissions. In addition, the ponds contain several aeration units to enhance degradation and retard sulfur-reducing bacteria.

In May 2010, the facility was commissioned. Normal startup issues were ironed out, and official startup occurred in June. Since that time, all monthly targets for operation, water quality and air emissions have been achieved, with 100% run time. The storage ponds, which were formerly cloudy and had a dark red color, now have a light green tint and a clear appearance.