The Effect of Inlet Conditions on
Oil-Water Separators at
Sohio's Toledo Refinery
A. T. WALLACE, Assistant Professor of Civil Engineering
Clemson University
Clemson, South Carolina
G. A. ROHLICH, Professor of Sanitary Engineering
J. R. VILLEMONTE, Professor of Sanitary Engineering
University of Wisconsin
Madison, Wisconsin
Laboratory studies on oil-water separators at the University of Wisconsin's
Hydraulic and Sanitary Laboratory during the period from 1948 to 1950 (1) had
shown conclusively that for comparable rates of flow, the major factor in basin
hydraulic performance was the inlet design. A later field study at Hamburg,
Germany comparing reaction-jet baffles with vertical post baffles (2) for removal
of oil in parallel separators gave tacit support to the previous study.
During the summer of 1964 the Standard Oil Company of Ohio (SOHIO),
working in cooperation with the American Petroleum Institute (API) made two
oil-water separators at their Toledo refinery available for a comparative study of
hydraulic characteristics and oil removal. These separators were part of a bank
of ten parallel basins, each receiving feed from a common channel. Oily waters
enters an individual basin through three 18 in. by 18 in. slide gates which are
operated by hand valves located on the walkway over the inlet end of the separators. The effluent leaves over simple weirs at the end of each separator and is
conveyed away in a common channel. Each basin is 80 ft by 18 ft in plan and
has a water depth of seven ft, 8 ins. Basin No. 1 had a simple overflow weir inlet and Basin No. 2 had an inlet consisting of four 10-in. reaction-jets, baffles
11 ins. in diameter with a three-in. hole at the center. Figures 1 and 2 show
the basin and inlet details.
METHODS AND PROCEDURES
In general the experimental method was to obtain dispersion curves for both
basins at comparable flow rates and to simultaneously measure the removal of oil
by the basins. A sharp-crested weir with end contractions was installed in the effluent channel between the third and fourth basins to measure the combined discharge from basins 1 and 2. Basin No. 3 was not operated during the study so
that it would not contribute to the measured discharge. The head on the weir was
measured with both a point gage and a Steven's stage recorder located four ft upstream of the weir crest. Once the basin discharge had been regulated by means
of the inlet slide gates, the head on the weir did not vary by more than one-
hundredth of a ft from the mean.
Dispersion curves were obtained in the following manner:
1.   The basin slide gates at the inlet were adjusted until the appropriate point
fage reading was obtained and the basin allowed to operate for at lease one full
etention time to insure a stable flow condition before beginning a test.
- 618 -