The first step in choosing a water treating system is to characterize the influent water streams. It is necessary to know both the oil concentration in this stream and the particle size distribution associated with this concentration. This is best determined from field samples and laboratory data.

Various attempts have been made to develop design procedures to determine oil concentration in water outlets from properly designed freewater knockouts and treaters. A conservative assumption would be that the water contains less than 1,000 to 2,000 mg/1 of dispersed oil.

It is possible to theoretically trace the particle size distribution up the tubing, through the choke, flowlines, manifolds and production equipment into the free-water knockout using equations presented in previous sections. However, many of the parameters needed to solve these equations, especially those involving coalescence, are unknown.

Because of the dispersion through the water dump valve, the oil size distribution at the outlet of a free-water knockout or heater treater is not a significant design parameter. From the dispersion theory it can be shown that after passing through the dump valve a maximum droplet diameter on the order of 10 to 50 microns will exist no matter what the droplet size distribution was upstream of this valve.

If there were sufficient time for coalescence to occur in the piping downstream of the dump valve, then the maximum droplet diameter would be defined by Equation 7-2 prior to the water entering the first

vessel in the water treating system.

The solution of this equation requires the determination of surface tension. The surface tension of an oil droplet in a water continuous phase is normally between 1 and 50 dynes/cm. It is not possible to predict the value without actual laboratory measurements in the produced water. Small amounts of impurities in the produced water can lower the surface tension significantly from what might be measured in synthetic water. In addition, as these impurities change with time, so will the surface tension. In the absence of data it is recommended that a maximum diameter of between 250 and 500 microns be used for design.

It is clear that there will be distribution of droplet sizes from zero to the maximum size, and this distribution will depend upon parameters unknown at the time of initial design. Experimental data indicate that a conservative assumption for design would be to characterize the distribution by a straight line as shown in Figure 7-19.