Pilot tests were conducted on steam flood FWKO process water, oil plus solids from Wemco underdrain and skimmings. The tests were conducted at 70 to 180 F at various pH's. Water wet and oil wet membranes were used to test the various process streams. Ceramic element pore sizes tested ranged from 500 angstroms to 1.2 microns. The FWKO water chemistry and solids content was very variable. It was noted that the microfiltration system performed better (i.e. had higher flux rates and longer run times between cleaning) when a continuous chemical precoat was applied to the membrane. The chemical precoat helped prevent the premature plugging of the membrane pores. Several chemical additives were tested. The pilot tests showed that pH is a factor in the run times achievable on steam flood FWKO process water. The effects of dissolved gases on membrane performance was also investigated. Ceramic membranes are very durable and can be used at high temperatures, and pressures. The membranes can also be cleaned with aggressive solvents, acids and caustics. The membranes have a relatively low flux rate and are expensive when compared to conventional treatment process. They do appear to be useful in the cleanup of concentrated sludge or process water where other conventional filtration does not perform.
This crossflow microfiltration pilot investigation was set up to determine the long-term potential of treating Union Pacific Steamflood Knockout water. More specifically, the pilot was to establish chemical feed rates, flux, and cleaning methods for the ceramic microfiltration system. The pilot system consisted of one P1940 - 500A membrane pilot (CXP-2), and four P1940 - 0.8 elements in series (XP-2).
The pilot units were set up to take feedwater from the water leg of Union Pacific Feed Water Knockout (FWKO) vessels 1 and 2. The feed was obtained at vessel temperature and pressure, 160 to 180 F, and 35 psi to 48 psi respectively. The feed hose ran 200 feet, this coupled with the heat loss in the pretreatment tanks, other hoses and membrane housings, resulted in CFM system temperature being 10 F to 20 F lower than the FWKO temperatures. At night, and, especially if it rained, CFM system temperatures would be at or below 130 F, approaching a 50 to 60 F drop in temperature from the FWKO vessels.
The initial tests were run with and without the addition of FeCl3. The chemical was added to a pretreatment tank, having a retention time in the order of 15 to 30 minutes, based on the product flow. Total fluid retention time, which includes the recycle, would be 8 to 15 minutes.
After March 12, 1991, the system was operated as per the schematic shown in Figure 1. The FWKO was fed into tank 1, where the oil was allowed to skim and/or settle, then through the stripper and into tank 2, where 20–30 mg/L FeCl3 was added, and further skimming of oil occurred. The ferric solids and other precipitates resulting from the stripper were allowed to settle in the bottom of tank 2.
CFM feedwater was obtained out of the same suction line for the four 0.8 micron elements in series, as well as for the 500 A unit.
The elements were operated until the transmembrane psi pressure differential was 25 to 35 psi, and then the systems were shut down and the elements cleaned.
The following data was collected for each unit every 3 minutes during the test runs:
Transmembrane differential pressure.
Product flow rate.
Temperature.
pH.
P. 137^
