Use of an oil/water separator

• By Michael B. Meyer
• Oct 01, 1999

Many industries use oil or oil-based products. Storage and handling of these materials may result in the contamination of nearby waters, due to spills or leaks. The enactments of the Clean Water Act (CWA) and, subsequently, the Oil Pollution Act of 1990 (OPA) mandate facility owners and operators be prepared to respond to a worst-case discharge and to a substantial threat of such a discharge. The cleanup operations following such oil spills and releases can be painstaking, lengthy, weather-dependent, costly and sometimes inefficient.

The U.S. Environmental Protection Agency (EPA) suggests several measures in its Compliance Assistance Guides for containing, controlling and preventing spills, such as the use of sorbents, dikes, berms, booms, diversionary drainage structures, on-site treatment plants and oil/water separators. Whatever the technique(s) used, EPA's main concern is that drainage and/or containment systems be designed to prevent accidental oil spills from reaching navigable waterways.

Depending on the size, nature, frequency and probability of occurrence, either a single or a combination of measures may be adopted. One spill countermeasure is the use of an oil/water separator. In cases where sudden releases may occur, the oil/water separator should be capable of handling sudden and large spills at elevated flow rates. The device separates oil-based liquid from any rain, process or wash water, thus discharging clean water and resulting in the reclamation of otherwise lost oil that can now be recycled.

Treated wastewater exits the properly sized device, in many cases, without the need for further treatment or human spill-response intervention. In facilities where fuel transfers are somewhat frequent, oil leaks are continuous, human error is possible or the probability of a sizable spill is high, the use of an oil/water separator can prove to be the most cost-effective measure to clean up spills.

Apparatus and testing method
The American Society for Testing and Materials (ASTM) developed Standard Practices D 6157 and D 6104 for testing the performance of oil/water separators subjected to a sudden release and to surface runoff, respectively. Other standards for testing the performance of oil/water separators subjected to a pumped influent are currently under development. A testing program was developed based on the guidelines of these practices. The primary purpose of these tests was to determine how a given oil/water separator would react when subjected to a spill. A Hydrasep¨ HS1000 oil/water separator - a 1,000-gallon separator rated at 81 gallons per minute (gpm) with a 370-gallon oil storage capacity - was selected for these tests.

The testing apparatus consisted of an elevated structure with three 1,000-gallon water storage vessels manifolded together and draining through a 3-inch line into the separator. Orifice plates were used to control the flow rate. The height of the structure with respect to the diameter of the storage vessels was such that the difference in flow rate caused by the loss in static head as the vessels drained did not exceed 5 percent. A 120-gallon oil storage tank was mounted piggyback on the water tanks, so that the extra elevation helped compensate for the difference in the densities of water and oil. The oil vessel drained into a Y-branch in the piping placed upstream from the orifice plates.

Re-entrainment or "drag-out" tests were initially performed to determine how the vessel would react to a sudden inrush of clean water while it stored a sizable amount of oil within. These tests are usually performed to establish a datum for the performance of the separator. The separator was filled with 300 gallons of diesel #2, and water was introduced at different flow rates until three volume changes were obtained. Drag-out tests were repeated at several flow rates, ranging from 100 gpm to 185 gpm. A minimum of eight effluent samples per test run were collected at equal intervals.

A sudden spill followed by a wash-down scenario was also simulated. The initial stored diesel quantity was reduced to 250 gallons. One-hundred and twenty gallons of diesel were drained from the oil vessel, and were immediately followed by 3,000 gallons of water. The 250 gallons combined with the 120 constituted the rated oil storage capacity of the separator. The water volume causes three separator volume changes and helps determine the efficiency and rate at which the separator "absorbs" the shock of the sudden release. Tests were performed at flow rates ranging from 62 to 120 gpm. A minimum of eight effluent samples per test run were collected at equal intervals. All samples were sent to an independent laboratory for analysis.

Results and analysis
All non-detect results were given the value of the detection limit set forth by the analytical laboratory - 1.4 milligrams per liter (mg/L). An average and a standard deviation were obtained from the data of each test run. These, along with the peaks from each run, were plotted versus flow rate for each type of test.

Figure 1 shows the obtained peak, average and standard deviation for each re-entrainment run versus flow rate. All the flow rates were 1.23 to 2.28 times the separator rating. Trend curves were plotted through the data. All the obtained peaks were below 10 mg/L, and the averages were below 3 mg/L. The low standard deviations indicate that most of the results were in the neighborhood of the averages. The volume of fuel lost with the 3,000 gallons of water ranged from 0.803 fl. oz. and 0.96 fl. oz. at 100 gpm and 160 gpm, respectively, to 2.8 fl. oz. at 185 gpm depicting an inception in performance degradation.

Figure 2 shows the obtained peak, average and standard deviation for each spill simulation run versus flow rate. The flow rates ranged from 0.76 to 1.48 times the separator rating. Trend curves were plotted through the data. The peaks ranged from 6 mg/L to 54 mg/L. The averages ranged from 3.3 mg/L to 10.4 mg/L. The standard deviations indicate that most of the results were in the neighborhood of the averages, demonstrating that the separator quickly dampened the effect of the spill. The volume of fuel lost with the 3,000 gallons of water ranged from 1.48 fl. oz. at 62 gpm to 4.67 fl. oz. at 185 gpm out of the 120 gallons spilled per run.

Five hundred gallons of diesel #2 were available for running the above-referenced test. In the spill simulation phase, 120 gallons were pumped out and recycled to run the next test. At the completion of all the tests, a sample of diesel was analyzed; it showed only 0.29 percent moisture content.

Conclusions
Based on these results, it can be concluded that the tested separator is quite capable of handling sudden spills and can be an ally in minimizing their environmental impact without the use of personnel. More importantly, it is in place and always ready for the next spill.

This article originally appeared in the 10/01/1999 issue of Environmental Protection.