What Happens After a Failed LEV Test

The first shift: protect the people before you fix the plant

The instinct after a failed LEV test is to reach for spanners. The duty is to reach for the people first, because the failed hood is still sitting over a live process, and the gap between failure and repair is a period of uncontrolled or under controlled exposure unless you close it deliberately.

Reading the remedial list like an engineer instead of a menu

The remedial list on a failed TExT is written in a deliberate order, and treating it as a pick list of suggestions is the most common way sites waste the repair budget. The examiner sequences it from cause to symptom: airflow restored first, then capture geometry, then housekeeping, because a hood rebalanced before its blocked branch is cleared will be rebalanced again afterwards.

Distinguish the three kinds of entry. Defects are the failures: the crushed duct, the seized damper, the fan running backwards after a motor swap, the filter blinded solid. Modifications are the report telling you the system no longer matches the process: a hood too far from a repositioned machine, a branch added without rebalancing, a booth handling twice the airflow of its design. Observations are the trend warnings that cost nothing today and prevent the next failure. Budget for defects immediately, engineer the modifications properly, and diarise the observations rather than discarding them.

Resist the temptation to fix by intuition. LEV systems are pressure balanced networks: closing one damper raises every other branch, clearing one blockage redistributes flow everywhere, and a well meaning fitter balancing hoods by ear can fail three hoods while fixing one. Every adjustment should be made against the commissioning data, measured, and recorded, which is also exactly what the retest examiner will ask to see.

Where the list is long, agree the sequence with the examiner before starting, and ask the one question that saves the most money: which of these entries caused the failure, and which are passengers. Repairing in cause order often turns a ten line list into three repairs and seven items that resolve themselves.

Retest, full TExT or recommissioning: what the examiner needs to see

Not every failure closes the same way, and matching the closure to the failure saves both money and argument. A single hood failed on a local defect closes with a focused retest of that hood against its benchmark once the repair is made. A failure rooted in the main plant, the fan, the main duct, the filter, generally needs the affected sections retested as a system, because restoring the main airflow changes every branch. And a system that failed because it no longer matches the process needs modification and then recommissioning of the changed parts, with new benchmark data recorded, because the old commissioning figures no longer describe the system that now exists.

Get the closure type agreed with the examiner when the repair is planned, not when it is finished. They will say what measurements will satisfy them, whether smoke or dust lamp observation is needed alongside velocities, and whether the repair changes anything the Written Scheme of the report, the logbook or the risk assessment relies on. A repair that quietly changes hood positions or duct runs without new benchmarks has converted a failed system into an unmeasurable one, which is worse.

Keep the failed report, the repair records and the passing retest stapled together in the logbook, in that order. The sequence is the compliance story: found, fixed, verified. The five year retention applies to the failure as much as the pass, and the examiner at the next TExT will start from exactly this file.

If the retest fails again, stop patching. Two failures on the same hood mean the diagnosis is wrong or the system is undersized for the process, and the next spend should be engineering assessment, not another connector.

When the answer is redesign, not repair

Some failed TExTs cannot be repaired closed, because the system is doing everything it was built to do and the process has outgrown it. The signs are consistent: multiple hoods marginal rather than one hood failed, transport velocities sagging across the network, a filter cleaned monthly that was designed for quarterly, and a remedial list that repeats from one report to the next like a chorus.

Recognise that pattern early, because money spent repairing an undersized system is rent, not investment. The honest sequence is an engineering assessment against the process as it now runs: what contaminant, generated at what rate, at how many points, needing what capture velocity and what airflow. Sometimes the answer is a larger fan and a rebalance; often it is redesign of the worst branch; occasionally it is the finding that the process itself should change, because substituting the material or enclosing the source is cheaper than doubling the extraction plant, and sits higher up the hierarchy of control where COSHH wants you looking anyway.

Fund the assessment with the numbers the failed reports have already given you. Three TExTs of declining velocities are a trend an accountant can read, and the cost of repeated remedial visits, retests and production interruptions usually exceeds the redesign within two cycles. The examiner is an ally here: their reports are the evidence, and most will gladly distinguish this system needs maintenance from this system needs replacing if asked directly.

While the redesign is planned, the interim control question returns and stays: the failing system still runs over a live process, and the bridging measures from the first shift, process limits, RPE, task rotation, remain in force and documented until the new benchmarks exist.

The logbook, the risk assessment and the next fourteen months

A failed TExT should leave permanent marks on three documents, and a site that repairs the plant but not the paperwork has fixed half the failure. The logbook gains the failure file: report, repairs, retest, cause. The COSHH risk assessment gets reviewed, because an assessment that assumed adequate control has just been contradicted by measurement, and the review is what proves the contradiction was noticed. And the weekly user checks gain whatever line would have caught this failure early: the pressure gauge reading that was drifting for months, the hood position mark, the look behind the machines.

The weekly check deserves particular attention after a failure, because almost every TExT failure was visible to a competent weekly check long before the examiner arrived. Airflow indicators falling, dust settling where it never used to, operators smelling what the hood should capture: these are the system reporting its own decline in real time. A logbook that shows fifty two identical ticked entries and then a failed TExT is not evidence of checking; it is evidence of ticking, and inspectors read it exactly that way.

Set the next fourteen months up deliberately. The statutory clock to the next TExT runs from the examination date, not the retest, so diarise both. If the failure revealed a Schedule 4 process on site, the interval may be shorter than you have been assuming, and the diary needs correcting today rather than at the next failure.

Close the loop with the people once more: what failed, what was fixed, what the weekly check now watches. The operators who report the next drift early are the cheapest monitoring equipment the system will ever have.