A steam trial on a restaurant kitchen grease filter answers one narrow question — does the heat release the deposit? — and leaves the three that actually decide a purchase untouched: whether the filter can be cleaned at all without dismantling the extract system, whether the result can be signed off by the person who insures and inspects the premises, and how the run-off is handled inside a working kitchen. Reviewing only the visible patch of clean metal is how a trial gets approved and the method still fails six months later. The checklist below is what an operator should take away from the test.
1. What the trial was actually allowed to reach
Most grease filters in a kitchen extract run are accessible only after panels, support brackets and the adjacent duct entry are removed. Write down what the wand reached during the trial: the filter mesh faces, the frame, the grease trough, the first section of duct beyond the filter, or all of them. A trial that cleaned a mesh laid flat on a floor in a yard has tested a different job from one that cleaned a mesh still sitting in a hood above an active line, and the two results are not comparable.
Record the dismantling and refitting time separately from the cleaning time. On many sites the labour around the filter costs more than the cleaning itself, and a method that halves cleaning time while doubling the handling time is not an improvement.
2. Deposit type before deposit thickness
Two filters of the same apparent thickness can behave completely differently. Freshly accumulated grease is soft and moves with hot water. Grease that has been repeatedly heated has polymerised into a hard varnish that behaves more like a coating than a soil, and this is the layer that responds to commercial kitchen steam cleaning, because the heat softens the film before the pressure lifts it. Note which class of deposit the trial was run on, and inspect the reverse face of the mesh — the side facing the duct usually carries the older, harder material, and it is the side that determines whether the method works.
3. Which mode was used, and what dwell time was given
A trial run without the mode and dwell recorded cannot be repeated by the second shift. For a baked-on film, the drier steam mode puts heat into the deposit; once the film has released, the wetter mode acts as a rinse. A dry and wet steam cleaner lets both steps be done with one machine, but only if the trial records the working distance, the dwell on the deposit and the number of passes used. Write those three numbers down on the test patch while the crew is still standing in front of it.
4. Mesh condition after cleaning, not just before
Cleaning can damage the thing it cleans. Inspect the mesh after the trial for bent or flattened layers, distorted seams, loose rivets, and any mesh that has been pushed out of its frame. Also check the filter's own condition record: a mesh with sagging layers, damaged edges or a failing seal may clean well and still fail a fire inspection. If the filters in your extract are past the point where they can be refitted with confidence, cleaning was never the right task, and replacement should be scheduled alongside the cleaning decision.
5. Run-off, drainage and the interceptor
Every kilogram of grease lifted off a filter has to go somewhere. In the trial, follow the run-off from the lance to the drain and confirm the route. Greasy water leaving a filter clean belongs in the site's waste route through the grease interceptor; a floor drain, a prep sink or a yard gully turns a compliant clean into a blockage and an environmental problem. Record who checked the interceptor before the trial and how much capacity was left, since a nearly full interceptor converts an acceptable clean into a flood within the same shift.
Water volume is a trial result in its own right. Note how much water the method consumed and how long the surrounding floor stayed wet. A fixed unit fed from a tap supply at a defined wash point keeps hoses out of walkways; a mobile or hand-pushed unit suits sites where one crew services several areas. Where the trial had no convenient services, that constraint belongs in the report, because it will decide the machine as surely as the cleaning result does.
6. Electrical, gas and lighting components in the hood
Extract hoods carry lamps, lamp holders, fan motors, damper actuators and, on some installations, fire-detection wiring above the filter line. A trial that got everything wet has not demonstrated a method, it has demonstrated a shutdown risk. List every component within reach of the wand, confirm with the equipment's documentation which parts must stay dry, and note which of them were covered or removed during the test. Any component that cannot be protected should be marked wipe-only in the method, not left to the operator's judgement.
7. Downtime and the window the kitchen can actually give
Record the clock. When the extract is down, the kitchen is usually down with it. Note how long the filter was out, how long cleaning took, how long drying took, and when the line returned to service. Less water on the surfaces means a shorter return-to-service window than a hose wash, which is the practical reason a kitchen might choose steam over a pressure washer — but only if the drying time was measured rather than assumed.
8. Machine specification against the duty you ran
Judging a machine on a small test patch flatters it. Read the published ratings against the duty you actually ran: heating power sets how quickly the unit reaches temperature and how well it holds that temperature through a long session, tank volume sets how long a session runs before a refill, working voltage sets what supply the site must provide. The Fixed Steam Cleaning Machine PowerDe HW02 product page lists rated values in this form, together with the protection items — leakage, short-circuit, high-temperature, high-voltage and water-shortage protection. Confirm each of those against your supply and your crew's habits, and check what the machine does when a protection trips mid-session.
Water supply belongs in the same review. In a hard-water area, scale collects inside a heater and output falls long before the machine fails. If the trial ran on mains water, ask the supplier what the descaling interval is and who owns it before the machine is bought, not after performance has dropped.
9. What the trial proves to an inspector or a fire risk assessor
Decide before the next trial what evidence the site needs afterwards. A date, the filter identification, who did the work, what method and mode were used, and what the acceptance check found is a record; a set of photographs is not. Deep grease deposits in filters and extract lines are a fire issue as much as a hygiene issue, so whoever signs your fire risk assessment will want documentation that matches your cleaning reality.
10. Acceptance criteria that settle the question on site
- Both mesh faces free of a continuous grease film, checked with a hand held behind the mesh and by looking through it against light.
- Frame, trough and the first duct section clean to a depth that is stated in the method, not "as far as practical".
- Mesh layers flat, seams intact, filter refitted with a proper seal and no visible gaps at the frame.
- No standing water in the trough, on the fan housing or around electrical components before the system restarts.
- Run-off accounted for, with the interceptor checked before and after the work.
- Airflow or extract performance checked after refitting, with the before and after readings recorded.
11. What to ask the supplier for after the trial
Take the trial record to the supplier rather than the brochure. The Cleaning Equipment Series covers fixed, hand-pushed and mobile units, and the right one depends on your hood type, how far the machine must travel, what supply is available and how many filters are in the rotation. If the filters or the extract layout are unusual, customized steam cleaning equipment can be specified against that configuration. Send the hood photographs, the filter type, the deposit description and your measured timings to the contact page and ask for a written method with the mode, dwell and run-off route set out, rather than a catalogue page.
