CONDENSATE CLEANS UP OILY PRODUCED WATER

Nov. 25, 1991
C. Dean Cockshutt, Gerry Fode Alberta Energy Co. Ltd. Calgary Condensate and water treatment chemicals added to the produced water tank through a homogenizing centrifugal pump have been very effective for de-oiling produced water. Synergistic effects between condensate and the current chemicals produce water qualities that exceed the water quality from using either chemicals or condensate alone. A key feature of the technique is the ability to minimize treatment costs by recycling added
C. Dean Cockshutt, Gerry Fode
Alberta Energy Co. Ltd.
Calgary

Condensate and water treatment chemicals added to the produced water tank through a homogenizing centrifugal pump have been very effective for de-oiling produced water. Synergistic effects between condensate and the current chemicals produce water qualities that exceed the water quality from using either chemicals or condensate alone.

A key feature of the technique is the ability to minimize treatment costs by recycling added condensate. Condensate losses are minor due to their low natural solubility in water.

HEAVY OIL PROBLEM

A significant problem facing many heavy oil operators is de-oiling produced water prior to reinjection into disposal or pressure maintenance wells. Because the specific gravity of heavy oil is very close to that of water, unassisted gravity segregation of oil and water requires prohibitively long residence times. Thus, many operators use various combinations of chemicals and mechanical separation aids in an attempt to clean up the produced water.

In the Alberta Energy Co. Ltd. South Jenner battery, continuously recycling condensate results in a very economical water treatment process. Multivariate statistical analysis was used to isolate the effects of condensate additions from the effects of other plant operating variables.

TREATMENT SYSTEM

Fig. 1 is a flow schematic of the battery water treatment system.

A stable water continuous emulsion from heavy oil wells enters the free water knockout (FWKO) vessel at about 85% water cut. Following primary separation, water drawn off the FWKO contains about 1,000 ppm of heavy oil.

The water is treated with chemical agents and pumped to the produced water tank where secondary separation occurs. Separated oil is skimmed from the water surface and recycled to the FWKO.

Vapors are drawn off to a vapor recovery system where natural gas liquids are condensed out. Water that commonly contains over 200 ppm of oil is pumped off the bottom of the produced water tank to an injection tank.

Following solids filtration in a disposable cartridge-type filter, produced water is pumped at high pressure to a series of injection wells.

Several options other than gravity segregation are available for separating oil from the produced water:

  • Chemical demulsification

  • Froth flotation

  • Centrifugal (cyclone) treatment

  • Coalescer beds

Gravity segregation, the most common method of oil and water separation, requires a tranquil environment for separation. This necessitates a large volume vessel where long settling times must be obtained. The rate of settling is more rapid for large droplets, a large density difference between phases, and a low viscosity of the continuous phase.

Because the density of heavy oil approaches that of produced water, very small flotation forces exist. Moreover, the natural organic acid surfactants present in heavy oil allow high dispersion of very small oil droplets.

Stokes' law settling calculations for these conditions will yield settling residence times of many days. Obviously, vessels constructed to meet the Stokes' criteria would be prohibitively expensive.

A reduction in settling time can be achieved by adding chemicals that increase interfacial tension. Typical treatment chemicals are blends of polymers, alcohols, and petrochemicals.

The chemical mixture is slightly water soluble but oleophilic in nature. This gives the mixture the ability to diffuse through the bulk water phase to the surface of the oil droplet. At the surface, destabilization of the droplet's interfacial film occurs.

One theory on interface destabilization contends that the migrated surfactant molecule changes the orientation of the hydrophilic and hydrophobic groups of the natural emulsifiers. 1

Breaking the stabilizing film allows coalescence of individual droplets into larger droplets. Stokes' settling law predicts that large particles will settle or float at a greater rate than small particles. Thus, coalescence increases the flotation rate.

Often, the bulk phase of the treatment chemicals is low specific gravity, oil-soluble hydrocarbons. Mixing the high specific gravity heavy oil with low gravity hydrocarbons gives an overall decrease in density. This enhanced density difference with the produced water speeds flotation. The density reduction effect is secondary to the film destabilization mechanism unless very large quantities of chemical are being added.

Gas injected at the bottom of the treatment tank can effectively de-oil produced water if the oil viscosity is low enough. This technique is used effectively at some bitumen extraction plants.

However, efficient gas froth flotation is dependent on having oil droplets that are of low enough viscosity to spread on the low-density gas bubble. These conditions prevail at bitumen extraction plants because of the high solvent addition rate and elevated temperature. Operating conditions at most heavy oil batteries do not favor froth flotation.

Rather than increasing droplet size or reducing droplet density, fluid cyclones increase the gravity force acting on the emulsion. For some heavy crudes, even doubling the gravity force will not appreciably enhance separation.

HIGH RESIDUAL OIL

Alberta Energy Co. Ltd.'s South Jenner oil battery produces daily about 700 cu m (4,400 bbl) of 14 API heavy oil, 4 cu m (25 bbl) of condensate and 4,000 cu m (25,000 bbl) of water.

At the South Jenner battery, produced water is treated by chemically enhanced gravity separation. This method has proven to be superior to either froth flotation or cyclone treatment. However, water qualities were consistently poor, typically 200 ppm oil and greater.

This high level of oil plugs the solids filtration media as well as the sand face in the injection well. The plugging results in high operating costs and excessive injection pressures.

Increased concentrations of many types of demulsifiers and chemical clarifiers failed to economically reduce oil concentrations to less than 100 ppm. Moreover, high chemical concentrations caused operating problems. Solids which normally sank to the bottom of the treatment vessel became concentrated at the water/oil interface due to wettability changes at high chemical concentrations.

Gas condensates and other hydrocarbons are commonly added to heavy oil treaters to de-water crude by specific gravity enhancement. A variation of this same technique was used to successfully reduce the oil concentration in the South Jenner produced water to approximately 50 ppm.

DE-OILING PROCESS

The primary function of treatment chemicals is to destabilize the interfacial film between the oil and water phases. Once this effect has been achieved, additional chemical volumes have little beneficial effect except at very high concentrations. At high concentrations the oil gravity reduction mechanism will begin to dominate and separation will be enhanced.

Because treatment chemicals are added on a once-through basis, operations using high chemical concentrations are uneconomical.

The South Jenner oil battery is equipped with a vapor recovery system which draws hydrocarbon and water vapors off the FWKO tank, water treatment vessels, sales oil tank, and evaporator. Compressed vapors are condensed yielding 3-4 cu m/day of C5 to C10+ natural gas condensates.

These recovered condensates are typically used to reduce the specific gravity and viscosity of the sales oil to specifications agreed upon by the pipeline operators.

For years, the produced water/oil concentrations were measured by solubilizing a sample of produced water in a colorless hydrocarbon solvent, followed by calorimetric comparison with known standards.

As an experiment, condensate was substituted for solvent. The positive results on oil separation shown in Fig. 2 resulted in applying the technique to the plant. Condensates were added to the produced water tank along with the usual treatment chemical. A noticeable reduction in produced water oil concentration was achieved.

However because natural gas condensate has a low specific gravity and a low water solubility of about 100 to 200 ppm depending on the composition 2 gas condensate does not easily disperse in the aqueous phase and cannot easily contact the oil droplets.

Thus, the beneficial effects were hindered by a lack of mixing. Under these conditions, the low specific gravity condensate floated in large globules to the surface of the treatment vessel without contacting appreciable amounts of oil.

A significant advancement in the technique resulted from adding the condensate to the water stream prior to the centrifugal pump that pumped the stream to the produced water tank. The high degree of agitation provided by the centrifugal pump caused the condensate to be well dispersed within the dilute oil-in-water emulsion. Thus the enhanced surface contact between the oil droplets and condensate was achieved.

The resulting droplet of oil and condensate had a reduced viscosity and a higher interfacial tension with water. Thus the oil and condensate has a greater tendency to coalesce into larger droplets after being pumped into the undisturbed treatment vessel. Moreover, the droplet has a reduced specific gravity which enhances gravity separation.

As a result of decreased pressure and increased temperature in the produced water tank, some condensate vapors flash and are picked up by the vapor recovery system.

The bulk of the added condensates remain in the oil layer that is skimmed off and pumped back to the FWKO. Because condensate is a valuable product, the amount lost in the disposed water phase was of concern.

Table 1 compares the theoretical condensate solubility with the amount available. The theoretical solubility of 146 ppm compares favorably with the actual solubility of 130 ppm measured in laboratory analysis of produced water. Condensate does not plug the water filtration media or cause known undesirable effects in the injection well.

PROCESS EVALUATION

Improvements in produced water quality as a result of adding condensate were difficult to quantify due to the effects of many other operational variables. Fig. 3 plots produced water quality vs. time. Note that the condensate concentration was intentionally dropped to zero during a portion of the trial period.

Because it is not necessary to have exceedingly tight process control at an oil battery, fluctuations in operating data are normal. However, during the condensate addition trials, market conditions forced plant throughput to exceed 100% of capacity. This high capacity caused larger than normal process departure.

Multivariate statistical techniques were used to filter poorly fixed plant operating variables from treatment variables. Table 2 presents a dimensionless normalized correlation matrix developed from 124 daily samples for seven out of ten analyzed variables.

The degree of association of one variable independent of all other variables is represented by a value from -1 to 1. A positive value indicates a positive correlation, i.e. increasing one of the variables causes an increase in the correlated variable. A negative value indicates an inverse correlation.

An arbitrarily selected correlation coefficient cutoff of less than 0.5 (both positive and negative) was used to indicate no significant correlation.

As shown in the diagonal matrix a significant negative correlation coefficient exists between condensate addition (COND) and the produced water/oil content (PW). This value indicates that increased condensate concentrations caused decreased produced water/oil concentration.

In addition, a water quality variable defined as the percent change in oil concentration between the produced water tank inlet (no condensate) and outlet (after condensate addition), shows a positive correlation coefficient of 0.53.

Interestingly, correlation coefficients for both the clarifier and demulsifier concentrations were of little statistical significance. These chemicals are known to have a beneficial effect on water quality and reasons for this unexpected result are currently being considered.

A scatter plot, Fig. 4, was generated for produced water/oil concentration vs. condensate concentration. The curve drawn through the data was not generated using best fit algorithms, but was drawn arbitrarily to show the apparent relationship between condensate concentration and water quality (entrained oil concentration).

One interpretation of the data, Fig. 4, is that upon initially adding condensate a synergistic effect with the treatment chemicals enhances droplet coalescence through interfacial tension effects. Because the settling rate is a function of the droplet size squared, any coalescence causes a rapid drop in produced water oil concentration. This corresponds to Region I on Fig. 4. Once an equilibrium droplet size is reached due to decreasingly enhanced interfacial tension, the specific gravity reduction mechanism dominates. This results in oil concentration decreasing linearly with condensate concentration as would be described by Stokes' law for a density reduction. Region 11 on Fig. 4 illustrates the gravity reduction effect.

No further analytical work was done to support this hypothesis. However, recent field data tend to support the functionality of the curve drawn on Fig. 4.

Condensate and water treatment chemicals added to the produced water tank through a homogenizing centrifugal pump have resulted in a very effective treatment program. Synergistic effects between condensate and the current chemicals produces water qualities that exceed those using either chemicals or condensate alone.

A key feature of the technique is the ability to minimize treatment costs by recycling added condensate. Condensate losses are minor due to their low natural solubility in water.

ACKNOWLEDGMENT

The authors would like to thank Alberta Energy Co. Ltd. for permission to publish this work.

REFERENCES

  1. Parichay, K.D., and Hartland, S., "Effect of Demulsifiers on the Separation of Water in Oil Emulsions," Chemical Engineering Communications, Vol. 92, 1990, pp. 169-181.

  2. Yaws, C.L., Yang, H., Hopper, J.R., and Hansen, K.C., "232 Hydrocarbons: Water Solubility Data," Chemical Engineering, April 1990, pp. 177-182.

Copyright 1991 Oil & Gas Journal. All Rights Reserved.