Grease filters in a restaurant kitchen are the part of the extract system nobody sees until they are removed, and once they are off the housing the questions come thick and fast: can hot steam be aimed at them, does the mesh need to come off the frame, and who decides when the result counts as clean. A test run answers those questions on one set of filters before the method is written into a weekly schedule. The steps below are written for a kitchen that stays open, which means the trial has to be small, isolated and timed.
1. Choose the filters that make the problem visible
Do not start with the cleanest filter in the building. Pick two: one with a thin, fresh film of grease and one that has been in service since the last deep clean, where the deposit has gone tacky and started to harden. Those two give different results with the same method, and the difference is what you are measuring. Note the filter type and frame size as recorded on the equipment register, along with the position in the extract run, because filters nearest the cooking line always carry more load than those further along the duct.
Record what is on them before anything is applied. A photograph on a phone, taken from the same distance each time, is more useful in a review meeting than a written description, and it costs nothing. If your site keeps a fire-risk assessment that references grease loading in the extract system, the same photographs support that record.
2. Isolate the extract fan before the filter comes out
A filter being washed a few metres from a live hood will pull the operator towards the fan, and loose grease flakes fall into a running duct. Switch the fan off at its isolator, and where the site runs a permit system, lock and tag it. Then let the hood and the filter cool. This is not a formality: steam onto a warm metal surface produces an immediate burst of vapour that scalds hands and fills the working area so the operator cannot see the mesh. State the cooling time in the test sheet and have one named person confirm it before the wand is opened.
Clear the area under the hood. Cover or move dry goods, packaging and open chemicals, because vapour from a steam trial travels and settles on whatever is nearby. Sign the line so nobody walks into a live wand while the trial is running.
3. Set up a wash point, not a sink
Grease-laden run-off from filters must not enter a prep-area sink or a floor drain. For the test, use a tray or shallow tank under the filter stand, and route the run-off to the site's waste route through the grease interceptor. Record where the interceptor is, who checks it, and when it was last emptied — a full interceptor turns a controlled trial into a flooded floor, and it is the most common way a first test is written off as unworkable.
Decide where the water comes from before the trial starts. A fixed machine plumbed to a tap supply, such as a fixed steam cleaning machine, keeps the crew from dragging hoses and trailing leads across an aisle that is still in service. A mobile unit from the Cleaning Equipment Series suits a kitchen where several servery areas share one cleaning crew and no single wash point exists. The trial is the moment to find out whether the chosen machine's hose actually reaches the filter stand without crossing a walkway.
4. Take the deposit off in stages, and write down the order
Steam is not a substitute for the earlier steps, and treating it as one is how a trial fails. The order that works is: scrape or wipe away the loose surface grease into the waste tray; apply a degreasing detergent and give it the dwell time the manufacturer states; then use heat and pressure to move what the detergent has released; then rinse; then dry. Applying steam to a thick deposit without the detergent step wastes machine output and produces a half-clean filter that looks finished but is not.
Where the machine offers two modes, use them deliberately. On a dry and wet steam cleaner such as the PowerDe HW02 — listed on its product page with 160° steam temperature, 0.8 MPa vapour pressure, a 6 kW heater and an 80 L SUS304 tank, on 220 V — the drier mode puts heat into a hardened film on the mesh, and the wetter mode acts as the rinse once that film has released. Write into the test sheet which mode was used on which face of the filter, the distance the nozzle was held from the mesh, how long it dwelt on the deposit, and how many passes were made. Those four numbers are what turns a demo into a method the second shift can repeat.
Work from one edge of the filter to the other rather than chasing the worst patches first. Grease moves; chasing it means the clean areas get soaked while the operator hunts the last streak.
5. Protect the frame, the seal and the electrical work
The mesh is not the fragile part. On many filters the frame joints, any gasket or seal between filter and housing, and the grease cups or collection trays at the bottom of the extract run are what fail after repeated high-pressure washing. Do not aim the wand into frame joints or at seals. Wipe those instead. If the hood's manufacturer has stated in writing that the filter frame or housing must not be water-washed, mark that task wipe-only and treat the warning as a limit on the method, not an inconvenience.
Lighting fittings inside the hood, the fan motor housing, and any switch or terminal enclosure near the filter bay are wipe-only as well. Never direct steam or hot water at them, and check that covers and cable glands are still seated after the trial before power is restored.
6. Check for damage the deposit was hiding
A heavy grease film holds a mesh together visually. Once it is removed, stretched mesh, a split joint, or corrosion at the frame can appear. Inspect both faces against the light, and pull gently on the mesh at a corner. A filter with a damaged mesh goes out of service rather than back into the hood, and the replacement interval for that position in the run is probably too long. Record the finding against the equipment register entry so the interval gets reviewed.
7. Define "clean" in a way an operator can verify in seconds
Write the acceptance test before the trial, not after. Checks that hold up without argument: no visible film on either face of the mesh; no grease blocking the perforations or the drain path of the tray; the filter is dry to the touch at the frame joints before refitting; no detergent smell on a filter that will sit above food; and no pooled water in the hood sump. Compare the after-photograph with the before-photograph from the same distance. If a patch still needs hard scrubbing after the stated method, that is a finding about the method and the dwell time, not a fault in the operator.
8. Dry, refit and confirm the fan works
Filters go back dry. A wet filter in a running hood holds moisture against the frame and the seal, and water driven into the fan housing is a fault waiting to happen. Allow the stated drying time, refit with the correct orientation and seating, remove the lock and tag, and run the fan briefly to confirm normal noise and extraction before the line reopens. Where the hood feeds a duct, note whether the accessible end of the duct still shows heavy deposit — a clean filter behind a dirty duct does not solve the load, and the next task in the plan belongs to whoever cleans extract ducting.
9. Time the whole thing and turn the trial into a task
Record the total time from isolation to refit, how many people it took, and how much water the machine used. Then write the method into the schedule as a named task with a named owner, a frequency and a duration, alongside the daily and weekly tasks for the cooking line. A schedule line that assumes an unclaimed forty-minute job will not happen on a Friday night; a line that states two people, twenty-five minutes, after the last service, has a chance of surviving.
Keep the completed sheet with the interceptor emptying records and the fire-risk assessment paperwork. Grease loading on filters and in extract lines is a fire issue as much as a hygiene one, so the record of what was cleaned, when, and to what standard carries as much weight as the clean itself.
10. Machine upkeep, water and a working boundary
A trial that depends on a machine needs the machine looked after. Note the service interval, the consumables to hold — seals, nozzles, hoses, brushes — and who to call if the unit stops heating. Where the local supply is hard, scale collects inside a heater and output falls well before the machine fails, so a descaling routine needs an owner. The HW02 product page lists leakage, short-circuit, high-temperature, high-voltage, water-shortage and high-voltage-explosion protections; the crew should know they exist, and should also know that a machine tripping one of them is reported rather than reset repeatedly during a shift.
One boundary worth stating plainly: the PowerDe product page for the HW02 describes its current use as vehicle exterior cleaning at fuelling and parking sites, interior vacuuming and disinfection, and engine-compartment cleaning. The kitchen scenarios in this article describe a trial procedure to be run against published parameters, not a factory claim for filter cleaning. Before the method is adopted across a site, put the equipment register, photographs of the filters and the hood layout to the contact page and ask for confirmation of the machine, mode and pressure suited to your extract system, where customized cleaning equipment can be specified against your line layout and duty cycle.
What to take away from the first test
Run the trial on two filters, cool and isolate properly, control the run-off, and treat steam as the third step rather than the first. Write down the mode, distance, dwell and passes, the drying time, and the total labour. Those are the numbers a kitchen manager needs before the method goes into a weekly schedule — and the photographs that show whether the filters are actually clean or only look it.
