How to Size and Service a Grease Interceptor Without Guessing

Why this matters

Two numbers decide whether an interceptor works: the flow it was sized for, and how often it gets emptied. Both are routinely set by habit rather than by measurement. A unit sized off a rule of thumb passes grease from the day it is commissioned, and a unit serviced on a calendar is either being pumped before it needs it or is out of compliance for weeks before anyone looks. Neither failure announces itself, because the customer's drains keep working the whole time - the consequence lands on the lateral downstream and on the sampling report. This card gives the arithmetic for both numbers and the measurement that turns a service schedule into a derived interval rather than a guess.

Before anything else

• An in-ground interceptor is a permit-required confined space. Nobody enters, for measurement or for cleaning; every step below is done from the surface. Where entry is genuinely unavoidable it is a permit entry under 29 CFR 1910 Part 1910.146 in general industry or 29 CFR 1926 Subpart AA in construction work, and it is not a service-call activity.
• Interceptors generate hydrogen sulfide. When the lid comes off, stand upwind and let the headspace clear before working over the opening. A sibling card covers why the smell fading is not reassurance.
• Cast lids pinch. Lift with a hook, hands off the rim, feet clear of the arc.
• Take influent temperature with a surface probe on the tailpiece or a clamp probe on the pipe, not by catching the stream; dish machine discharge runs hot enough to scald. If a hot sample is genuinely required, dip it on a pole in splash goggles and gauntlet gloves rated for the temperature, standing beside the opening rather than over it.
• Pumping is done by a licensed hauler with a vacuum truck. Nothing here involves a technician reaching into the tank.

Step 1: Establish which kind of unit you are working on

This decides which of the two arithmetic paths applies, and getting it wrong makes every subsequent number meaningless.

• A hydromechanical interceptor is rated in gallons per minute and depends on a flow control fitting upstream to hold the flow at or below that rating. Remove the fitting and the rating is fiction.
• A gravity interceptor is a tank rated in gallons of liquid capacity, and it works on retention time. A sibling card covers why retention time is the governing variable and what defeats it.

Skipping this step is how a shop ends up computing a retention time for a device that does not have one, or sizing a tank in gallons per minute.

Step 2: Compute the drainage load from the fixtures

The standard method works off fixture volume, not off a guess about how busy the kitchen is.

1. Measure each fixture's internal length, width and depth in inches and multiply for cubic inches.
2. Sum the compartments the interceptor serves.
3. Divide by 231 to convert cubic inches to gallons.
4. Multiply by a fill factor of 0.75, because nobody fills a sink to the rim.
5. Divide by the drainage period to get gallons per minute.

That last term is the most leveraged assumption in the entire method, and two values are in common use, one and two minutes. They differ by a factor of two, which means they produce interceptor selections one full size apart. Use the drainage period your jurisdiction's adopted plumbing code or FOG ordinance specifies, write which one you used on the calculation sheet, and expect a reviewer using the other one to disagree with your answer by exactly a factor of two.

Skip this step and size from the drain pipe diameter instead, and you have sized for a pipe's carrying capacity rather than for the flow this kitchen actually produces. Those are different numbers and only one of them is about grease.

Step 3: Verify the flow control fitting exists, on a hydromechanical unit

The fitting is a restriction with a vented inlet that holds the flow to the unit's rated gpm regardless of how fast the sink dumps. It is small, it is upstream, it is easily lost during a remodel, and its absence is invisible from the interceptor. Confirm it is installed, confirm its orifice matches the unit's rating, and confirm the vent is open. A hydromechanical unit running without one is a box the waste passes through.

Step 4: Measure the cap and the blanket from the surface

Two measurements, both taken with a corer or a sludge probe lowered on a line, never by entry:

• FOG cap thickness, the floating layer at the top.
• Settled solids blanket thickness, at the bottom.
• Total liquid depth of the tank, so the two can be expressed as a fraction of it.

The widely used service trigger, and the one most FOG ordinances write into a permit, is combined FOG plus solids reaching 25 percent of the liquid depth. Confirm both the figure and the measurement method against the ordinance your jurisdiction actually enacted, because some specify each layer separately rather than the sum.

Step 5: Turn two measurements into an accumulation rate

One measurement is a status. Two measurements at a known interval are a rate, and only a rate can produce a schedule. Measure immediately after a verified full pump-out, then again at a known number of days later, and divide the combined depth by the days.

The rate you get from a clean-tank interval is the fastest the tank will ever accumulate, and this is the part that gets misread. As the cap thickens the working volume falls, retention time falls, and capture efficiency falls with it. So accumulation slows as the trigger approaches while the effluent leaving the tank gets steadily worse. A tank that appears to have stabilized has not stabilized; it has stopped capturing.

Step 6: Derive the interval, then back off from it

Divide the trigger depth by the rate to get days to trigger. Because that figure assumes linear accumulation and the previous step says accumulation is not linear, do not schedule to it. Schedule at 75 percent of the computed days to trigger, rounded down to a whole number of weeks so it lands on a repeatable day. Re-measure at every service and re-derive the rate; a menu change, a seasonal swing or a new degreaser will move it, and the interval should move with it rather than sitting frozen in the customer's contract for three years.

Step 7: A pump-out means the tank is empty, then refilled

Removing only the floating cap and leaving the water and the blanket is a skim, not a service, and most ordinances require complete removal of all contents. Two things have to happen on every service and one of them is routinely missed:

• Everything comes out: cap, liquid and blanket. Then the walls and the tees are inspected while the tank is empty and visible from the surface, because this is the only moment a broken inlet tee or a collapsed baffle can be seen.
• The tank is refilled with clean water before it goes back into service. An empty interceptor separates nothing. Until it fills, everything the kitchen sends goes straight out the outlet, and the first hours after a service are the worst effluent the facility produces all cycle.

The hauler's manifest is the compliance record. File it against the account with the measured cap and blanket depths from step 4, because the two together are what proves the interval was derived rather than invented.

The filled-in artifact: one facility, both calculations

A restaurant with a three-compartment sink, each compartment measuring 24 inches by 24 inches by 12 inches deep, discharging to an existing 1,000 gallon gravity tank with a liquid depth of 48 inches. The operator wants to know whether the tank is right and how often it should be pumped.

Load calculation (step 2).

• Per compartment: 24 times 24 times 12 equals 6,912 cubic inches
• Three compartments: 20,736 cubic inches
• Converted: 20,736 divided by 231 equals 89.8 gallons
• Fill factor: 89.8 times 0.75 equals 67.3 gallons of actual fixture volume
• Drainage period, named explicitly: this jurisdiction's adopted code uses a one-minute period, so 67.3 gallons per minute governs. Had the two-minute period applied, the answer would be 33.7 gallons per minute, half as much and one selection size smaller. The one-minute figure is used for everything below.

Retention check on the existing tank (step 1 path for a gravity unit).

• Nominal: 1,000 gallons divided by 67.3 gallons per minute equals 14.9 minutes
• Working volume at the service trigger, with 25 percent of depth occupied by cap and blanket: 1,000 times 0.75 equals 750 gallons
• 750 divided by 67.3 equals 11.1 minutes
• Short-circuiting is unquantified because nobody has tracer-tested this tank, so the geometric figure is an upper bound: retention time at the trigger is at most about 11 minutes, written with one inequality and no interval

Eleven minutes is workable for a kitchen producing coalesced droplets and is not workable for an emulsion; which one this kitchen produces is a chemistry question the sibling card answers. The tank is not obviously undersized, which means the service interval is where the attention goes.

Accumulation measurement (steps 4 and 5).

• Day 0, immediately after a verified full pump-out and refill: cap 0 inches, blanket 0 inches
• Day 60: cap 5.0 inches, blanket 1.5 inches, combined 6.5 inches
• Rate: 6.5 inches divided by 60 days equals 0.108 inches per day

Interval (step 6).

• Trigger depth: 48 inches of liquid depth times 0.25 equals 12 inches combined
• Days to trigger at the measured rate: 12 divided by 0.108 equals 111 days
• The linearity correction, applied visibly: 111 days assumes the rate holds, and step 5 says it will not, because capture efficiency falls as the cap builds. Schedule at 75 percent: 111 times 0.75 equals 83 days
• Rounded down to a whole number of weeks: 11 weeks, or 77 days

Failure mode if this is done wrong. A shop that took the 111 days and sold a quarterly contract would be servicing at 91 days, inside the linear estimate and comfortably outside the real one, with the tank passing degraded effluent for the last several weeks of every cycle. Nothing in the customer's building would indicate it. The evidence would arrive as a failed sample from the sewer authority, months of non-compliance already banked, and a service record that says the schedule was met.

What changes the answer. Add a food waste disposer and the blanket rate rises sharply while the cap rate barely moves, so the combined trigger arrives sooner and the mix of what is pumped changes; a sibling card covers why. On a hydromechanical unit none of the retention arithmetic applies and the whole question becomes whether the flow control fitting is present and correctly orificed. And if the ordinance limits cap and blanket separately rather than as a sum, re-derive against whichever hits its own limit first, because 5.0 inches of cap may trigger long before the 6.5 inch combined figure does.

How to check your own numbers before you hand them over

Four checks, all cheap, each catching a specific error this method invites:

1. Restate the drainage period on the sheet. Without it, a reviewer cannot tell whether your gpm is right or off by a factor of two, and neither can you in a year.
2. Confirm the day-0 measurement was actually zero. A rate derived from a tank that was skimmed rather than fully pumped reads low and produces an interval that is too long.
3. Multiply the interval back out. 77 days times 0.108 equals 8.3 inches against a 12 inch trigger; that margin is what the linearity correction was buying.
4. Ask what changed since the last measurement. A rate is only valid for the operation that produced it. New menu, equipment, chemical or hours: re-measure.

References

• 29 CFR 1910 Part 1910.146, permit-required confined spaces, general industry; 29 CFR 1926 Subpart AA for construction work
• The plumbing code and the FOG or pretreatment ordinance as adopted and amended by the local jurisdiction, which set the sizing method, the drainage period, the service trigger and the manifest requirements
• Manufacturer documentation for the interceptor and, on hydromechanical units, for the matching flow control fitting
• See related: What a Grease Interceptor Is Actually Separating; Why a Food Waste Disposer Changes the Drainage Problem