Reverse osmosis process successfully converts oil field brine into freshwater

F.T. Tao, S. Curtice, R.D. Hobbs, J.L. Sides, J.D. Wieser
Texaco Inc.
Bellaire, Tex.
C.A. Dyke, D. Tuohey
Texaco Inc.
Beacon, N.Y.
P.F. Pilger
Texaco E&P Inc.
Denver

Fouling Rate Calculation [210,535 bytes]
Based on papers presented at the Chinese American Petroleum Association Technology Conference, Houston, Apr. 30-May 2 and SPE/EPA Exploration & Production Environmental Conference, San Antonio, Tex., Mar. 7-10.

A state-of-the-art process in the San Ardo oil field converted produced brine into freshwater. The conversion process used chemical clarification, softening, filtration, and reverse osmosis (RO).

After extensive testing resolved RO membrane fouling problems, the pilot plant successfully handled water with about 7,000 mg/I. of total dissolved solids, 250 mg/I. silica, and 170 mg/I. soluble oil.

The treated water complies with the stringent California drinking water standard.

Water reclamation

Oil field operators are finding that reclamation of produced water is increasingly difficult because of disposal limitations, discharge regulations, and conservation measures.

Low-salinity water can be reclaimed easily with processes that remove critical contaminants. High-salinity water, however, requires more complicated processes such as distillation or reverse osmosis (RO).

Distillation is energy intensive and expensive and therefore, is not cost effective for handling produced waters. Reverse osmosis has only recently been used for handling produced water.

The Mount Poso cogeneration plant, near Bakersfield, Calif., was the first large-scale RO plant for oil field produced water . ¹ The plant treats the produced water by oil separation, clarification, filtration, RO, and demineralization to produce a boiler feed water for the cogeneration power plant.

This is also the conventional method for handling sea water. To control scale, the process operates at a relatively low 5-7 pH.

Unfortunately, the Mount Poso process proved ineffective for the San Ardo water (Table 1 [83,281 bytes]). In a test, the filtered product turned cloudy after a short period. The San Ardo water was very unstable at a pH of 5-7 and quickly fouled the RO membrane.

Controlling the stability of produced water proved to be the key for successful RO operation in San Ardo.

San Ardo process

Prior to the produced water entering the San Ardo pilot plant (Fig. 1 [365,169 bytes]), an air flotation unit removed dispersed oil. Then, a warm lime unit treated the water at a rate of 50 gpm (1,700 b/d), primarily to clarify the water and remove silica (Table 1). The silica was removed by adding lime and magnesium chloride, and a proprietary polymer coagulant was added to help clarify the water before further treatment.

The water was then pumped through a heat exchanger to reduce the temperature to about 95° F.

After the warm lime unit, the system flow rate was kept at 5 gpm (170 b/d) to match the capacity of the RO test unit. Because the treated water had a pH above 11, hydrochloric acid was added to reduce the pH. A course and fine adjustment stabilized the pH at the desired level.

After the pH adjustment, the water entered a pumice filter. A coagulant enhanced the removal of the residual oil and suspended solids.

Zeolite softeners, followed by a weak acid softener removed completely the hardness ions from the water. The softened water was then polished by 5 and 0.45 ? cartridge filters in series. The water leaving the cartridge filters had a turbidity of 0.30 NTU (nephelometric turbidity units) or less.

A small heat exchanger ahead of the RO unit provided a constant temperature feed water for the RO) membranes. To prevent precipitation, the heat exchanger warmed the water by about 10° F. The main concern was soluble organics that could precipitate because of changes in solubility after the pH adjustment.

The operating temperature was kept at 105° F. This was 10° F. below the temperature limit of the RO membranes. The 2 x I array of eighteen 2 1/2-in. x 40-in. RO membrane elements (Fig. 2 [101,570 bytes]) allowed a 75% recovery of the water as permeate.

The turbidity and silt density index (SDI) limits for the feed water were 0.5 NTU and 3.0, respectively. Typical operating ranges were 0.2-0.4 NTU and a SDI between 0-1.0.

For operations below a pH 10.6, a boron-specific resin removed boron to meet the irrigation standard of 0.5 mg/l. However, at a very high pH, above 10.6, the boron unit was not required because the RO membranes rejected most of the boron and produced a permeate containing less than 0.5 Mg/1. ² Therefore, high pH operations would provide substantial investment and operating cost savings.

The post-RO treatment system polished the water to meet the irrigation standards. After oxygen was added to the water in the aerator, the water entered the final unit, the sodium adsorption ratio (SAR) unit.

The SAR unit was a 10 in. x 6 ft column filled with crushed oyster shells that added calcium and magnesium to the water. These additions modified the water's SAR and conductivity to conform with the irrigation water standard.

Stability

The pilot plant operations showed that the produced water was very unstable. The turbidity of the water increased with time even after being clarified and filtered. The increase was caused by precipitation of soluble oil.

Expressed as total organic carbon, the soluble oil contained in the water was about 170 mg/I. This amount corresponds to an oil and grease content of 80 mg/I. (EPA 413.1 method).

A solubility curve of the RO reject water during the high pH operation (Fig. 3 [96,257 bytes]) showed that as the pH was lowered, soluble oil started to precipitate. The precipitated oil was then extracted with freon and measured at various pHs. The curve shows that the solubility of the oil (organics) decreases rapidly as pH decreases below 8.

RO membrane fouling

The San Ardo water had an initial pH of about 8.0. Based on the conventional operating conditions for sea water membranes, the pH should be reduced to about 6.0 for scale control. Unfortunately, as indicated in Fig. 3, at a pH of 6.0, the solubility of the oil was only about 50 mg/I. oil and grease.

Because the oil and grease in the produced water was 80 mg/I., about 30 mg/l. was precipitated on the pumice and cartridge filters. Some of the finely precipitated oil passed through the filters and quickly fouled the RO membrane.

At a slightly higher pH (7.5), the solubility of oil was 225 mg/l. This solubility is above the oil and grease content of 80 mg/I. for the produced water. Therefore, there was no precipitation at the pumice and cartridge filters. However, when this water was concentrated four times (75% recovery) in the RO system, the oil and grease in the reject increased from 80 to 320 mg/I. Thus, it exceeded the solubility limit of 225 mg/l. and precipitated in the second stage of the RO system (Fig. 2).

The solubility curve (Fig. 3) clearly shows that at a moderate pH (7-8), the concentration of the soluble oil is beyond its solubility limits. Therefore, it would be precipitated on the membrane before the reject water leaves the RO unit.

Texaco's experience indicated that fouling from soluble oil occurred even at a pH as high as 9.5. This fouling point is in agreement with the measured solubility of 320 mg/I. (four times 80 mg/I.) on the solubility curve.

Membranes at high pH

The pilot test confirmed that RO membrane fouling could be controlled by raising the pH to 10.6-11.0. The upper pH operating limit of most RO membranes is 11.

For the pilot test, the fouling rate was represented by the fouling factor, k_f (see box). Fig. 4 [93,725 bytes] shows k_f as a function of time for one of the tests conducted at 105° F. The feed water to the RO membrane was kept at 5 gpm, and at 75% recovery. This translates to 3.75 gpm of permeate.

Because the RO membrane gradually fouled, the permeate rate was held constant by increasing the operating pressure. The pressure increased from 455 psig at the beginning to 515 psig as the membrane accumulated foulants. The results show that the fouling rate was near zero after day 45 for both stages.

Water quality

The quality of treated water conforms with the stringent California Title 22 Drinking Water Maximum Contaminant Levels. These standards include inorganic, organic, bacteriological, radioactivity, and secondary drinking water standards.

Title 22 requires total dissolved solids (TDS) below 500 mg/I., chloride below 250 mg/I., and low levels of heavy metals and organic materials. Table 2 [114,873 bytes] compares the treated water with California's code.

Costs

The major operating cost is for chemicals. For a 50,000 b/d plant, chemicals cost about $0.03-0.05/bbl of water processed by the plant. Estimated operating cost, including chemicals, power, and membrane replacement is $0.06-0.08/bbl of water.

In high pH operations, the chemical and operating costs are lower because a boron unit is not required.

The investment cost for a full size 50,000 b/d plant is estimated to be $7-9 million.

References

VandeVenter, L.W., Ford, B.R., and Vera, M.W., "Innovative Processes Provide Cogeneration Power Plant with the Ability to Utilize Oil Field Water," 50th Annual Meeting International Water Conference, Pittsburgh, Pa., Oct. 23-25, 1989.
Dyke, C.A., Tao, F.T., Curtice, S., Tuohey, D., Hobbs, R.D., Sides, J.L., and Wieser, J.D., "Removal of Salt, Oil, and Boron from Oil Field Wastewater by High pH Reverse Osmosis Processing," AIChE Annual Meeting Membrane Separation Fundamentals and Applications to Pollution Control, Miami Beach, Fla., Nov. 1-6, 1992.

The Authors

F.T. Tao is a senior research consultant in Texaco Inc.'s environmental technology section of the E&P technology department (EPTD), Bellaire, Tex.
S. Curtice is a research associate at Texaco's EPTD environmental technology section, Bellaire, Tex.
C. A. Dyke is a research associate at Texaco Inc.'s R&D department, Beacon, N.Y.
R.D. Hobbs is an advanced lab technician at Texaco's EPTD environmental technology section, Bellaire, Tex.
J.L. Sides is a senior research associate at Texaco's EPTD in Bellaire, Tex.
P.F. Pilger is a senior production engineer with the production division of Texaco E&P Inc., Denver.
D. Tuohey is a senior engineer at Texaco's R&D department, Beacon, N.Y.
J.D. Wiser is a senior research scientist at Texaco's EPTD, Bellaire, Tex.