When a restaurant kitchen receives a deep-clean audit, the grease filter is usually the item that decides the outcome. Hoods, ducts and fans attract attention, but the filter is the first barrier: it catches what the cooking process throws into the air, and it tells you within a week of service how well your cleaning routine is actually working. The difficulty for a facility manager is not knowing that filters need cleaning. The difficulty is deciding, in writing, what "clean enough" means so that the kitchen does not keep stripping filters earlier than necessary, and does not keep running them past the point where they cause trouble.
Acceptance criteria are the answer to that problem. Below is a practical framework for setting them, one that works for a single restaurant line as well as for a multi-site operation with a shared cleaning schedule.
Start With What a Grease Filter Is Actually Doing
Most kitchen filters are metal mesh or baffle panels hung in a hood. They work by changing the direction of the airflow quickly enough that grease droplets lose their momentum and hit a surface. They are not designed to absorb grease; they are designed to knock it out of the air and let it drain. That single fact drives every sensible criterion you can write.
Two failure modes follow from it. The first is airflow loss: as deposits build up on the mesh, the hood pulls less air at the same fan speed, and smoke and vapour drift out of the capture zone into the kitchen instead of into the duct. The second is grease carryover: once the mesh surface is saturated, the filter no longer holds new grease, so it passes further into the exhaust system, where it becomes a cleaning problem and a fire-loading problem that is much more expensive to address. Cleanliness programmes for commercial kitchen exhaust systems, including those run under NFPA 96 practice, exist because of this second mode.
A good acceptance criterion therefore targets both: capture performance at the hood, and visible grease condition on the panel. Criteria that watch only one of them will let the other fail quietly.
Write Three Criteria, Not One
Facilities that handle filters well usually run three separate tests with three separate triggers. They are cheap to define and they remove most of the argument at the end of a shift.
- A condition criterion. What must be visible on the panel at inspection: no dark, continuous film; no drips at the panel edge; mesh openings still distinguishable. This is the criterion you can train any team member to apply consistently.
- A performance criterion. Capture behaviour at the hood — where smoke travels when a burner is running at normal load — combined with whatever airflow reading your exhaust system already provides. If the readings and the visual test disagree, trust the visual test and investigate.
- A frequency criterion. The interval between cleanings, set from cooking load rather than from a calendar habit. A char-broiler line and a steaming line inside the same building do not share an interval.
Write the three criteria on one page, with the pass/fail wording and the action for each failure. Assign the check to a named role. If no one owns the check, the criteria exist on paper and nowhere else.
Choose a Cleaning Method Before You Set the Frequency
Frequency depends on how thoroughly the cleaning method restores the filter. A soak-and-scrub cycle on a busy line can take panels out of service, needs a sink or tank large enough for them, and leaves detergent residue on the mesh if rinsing is rushed. Detergent residue is worth taking seriously: a panel that re-passes the visual test but smells of detergent has not been rinsed, and anything left on the mesh will end up in the air stream and near food contact surfaces.
High-temperature steam has become the usual alternative where a facility wants to clean filters in place and shorten the cycle. Commercial kitchen steam cleaning works by combining heat with pressure: the deposit softens and lifts instead of needing to be dissolved and abraded, and the rinse step is the steam itself rather than a second chemical pass. The PowerDe HW02 steam cleaner, for example, is a fixed 220 V unit rated at 6 kW with steam at about 160 °C and 0.8 MPa, dry and wet steam modes, and an 80-litre SUS304 tank — the kind of specification that matters when you are cleaning mesh panels and hood faces that have to be usable the same day. For facilities that would rather start smaller, PowerDe cleaning equipment covers hand-pushed and mobile machines as well as fixed installations.
The practical point for your criteria is this: if cleaning can be done in place, on a short cycle, with hot water and no solvent, you can set a shorter frequency and a stricter visual standard. If cleaning is a disruptive operation, intervals get stretched and the visual standard quietly relaxes. Method and criteria are one decision, not two.
Make the Inspection Repeatable
An acceptance criterion that two people read differently is not a criterion. Two habits fix this.
First, photograph the panels at the point where you decide they have to be cleaned, and again after cleaning. Two images on the kitchen's shared drive settle most disagreements faster than any written description, and they give a new hire a reference better than a paragraph of instructions.
Second, inspect on a fixed rhythm tied to the kitchen, not to a month-end form. A short check at the start of a shift on the same day each week, with the panel pulled and looked at rather than viewed through the hood, produces consistent results. Record the date, the panel, the outcome, and the action taken. A log with a few failing entries is far more useful when you justify a shorter interval or a cleaning machine purchase than a log of nothing but passes.
For food-safety and insurance purposes, keep the cleaning contractor's records for the duct and fan sections alongside your own filter log. Your criteria cover what you clean; the contractor's records cover what you cannot reach, and an inspector will ask about both.
Review the Criteria When the Kitchen Changes
Grease loading follows the menu and the equipment, not the calendar. Adding a broiler or a wok range, extending service hours, or switching to an oil-heavy preparation changes the interval that a criterion supports. The same applies when a fan is serviced, a duct section is replaced, or a hood is moved. Treat any of those events as a prompt to re-test capture behaviour and re-check the interval within a fortnight, and write down what you changed and why.
Where a group of sites shares one standard, allow the interval to vary by site while keeping the visual and capture criteria identical. A single fixed interval imposed across a fried-food outlet and a salad-focused outlet will be wrong at one of them from the first week.
If you are writing the standard for more than one building, or you want a machine specified to your hood layout and utility supply rather than chosen from a catalogue, PowerDe works as an intelligent cleaning equipment manufacturer and supports customised configurations; you can send the requirement details through the contact page and get an equipment proposal against it.
A Short Checklist to Take Into Your Next Kitchen Meeting
- Pass/fail wording for filter condition written in one sentence, with a reference photo.
- A capture test defined for each hood, and a named person responsible for running it.
- Cleaning interval set by cooking load, reviewed whenever the menu or equipment changes.
- A cleaning method that can actually be executed on that interval without taking the line out of service.
- A dated log of checks, failures and actions, kept with the exhaust contractor's records.
Filters fail inspection most often not because a facility ignored them, but because the standard for "clean" was never written down and the cleaning method could not keep up with the interval the kitchen needed. Fix the wording, then fix the method, and the criteria hold themselves.
