From reactive to preventive with a small team

You have four people in maintenance, two shifts and a plant where the phone rings the moment you walk through the gate. Every day starts with one call-out and ends with another. The preventive maintenance plan exists on paper, but nobody has had time to carry it out for months. Two more people would help, but without a work order they too would get lost in emergencies. Below is a sequence that moves part of your time, step by step, from emergencies to planned work.

Rank the equipment

In the first two weeks, list all the equipment that matters and give each item a simple score. You do not need a criticality matrix with ten criteria. Four questions are enough, each scored from 1 to 3:

• If it stops, does delivery stop? Is there a spare machine or a way round it?
• Can it injure someone or cause an environmental problem?
• Does it damage the product or pass nonconforming parts downstream?
• How long does the repair take and what does it cost, including waiting for parts?

Multiply or add the scores, as you like, but the same way for everything. Do the exercise with the production manager and someone from quality, in an hour, not alone at your desk. The result does not have to be perfect. It has to be accepted by the people who will ask you tomorrow to drop everything and come to another machine.

Pick the few machines that hurt

From the list, take two or three items with the highest score and the most stops in recent months. You work on those for the next 90 days. The rest of the plant stays on the current regime, with repairs on request.

The temptation is to build a plan for the whole machine park at once. With four people, a plan for 80 machines stays an Excel file. A plan for three machines can be done, and the result shows on the line and earns you the trust you need to expand.

A stop log that gets filled in

Without data on stops you do not know where you lose time and you cannot show progress. The log must be simple enough that a tired technician at 3 a.m. can fill it in within two minutes. A few fields are enough: the machine, the date and time of the stop, the restart time, what was seen, what was done, which part was replaced.

A sheet stuck next to the machine or a form on a phone work equally well. What matters is that someone reads them weekly and asks when a stop that everybody knows about is missing. If nobody looks at what is written, people stop writing after a month.

Once a week, take the longest stop or the one that keeps repeating and keep asking why it happened until you reach something you can change: a tightening, a missing lubrication, a clogged filter, a wrongly mounted sensor.

Simple routines with operators

Operators stand next to the machine eight hours a day. They are the first to hear the new noise and the first to see the oil leak. Autonomous maintenance in TPM starts from here: the operator takes on the basic tasks of cleaning, inspection, lubrication and tightening, while technicians keep what needs tools and training.

For each of the chosen machines, build together a 10-minute checklist per shift. With photos, not long text. Lubrication points marked in colour, level indicators with a green and a red zone, places to clean and what to look for there. Do the first thorough cleaning with the technician present, because that is when the loose bolts, cracked hoses and leaks hidden under dust come to light.

The checklist also needs a clear step for what the operator does on finding something: where to note it, whom to tell. Otherwise the inspection becomes a tick box.

Spare parts for what matters

For the critical machines, look in the log and in the purchasing history for parts that were requested urgently and those with long lead times. Bearings, belts, contactors, sensors, seals, an old control board that is no longer made. Make a short list of what must be on the shelf, with a minimum stock, and keep them in a separate place, labelled by machine.

Do not buy parts “just in case” for all the equipment. The stock budget goes first to the parts without which a critical machine stands for days.

Protect PM time

The preventive plan usually dies like this: it is scheduled for Tuesday at 10, at 9:45 another machine stops, and the preventive job is pushed to “next week”. After three postponements it is no longer a plan.

A few rules that help:

• One person, or a fixed slot in the schedule, is allocated to PM and is not called to emergencies unless it is a safety matter or a total stop on a critical machine.
• The PM window is agreed with production in advance, for example the first hour of the Monday shift, and appears in the production plan.
• Any postponed PM is rescheduled within the same week and discussed at the production meeting.

The first weeks will be uncomfortable, because emergencies do not drop immediately. They usually drop only after the routines on the chosen machines have run for a while without interruption.

What to measure

A few indicators, tracked monthly, on the chosen machines:

• MTBF, mean time between failures: operating time divided by the number of unplanned stops. It should rise.
• MTTR, mean time to repair: total repair duration divided by the number of repairs. It shows how well prepared you are with parts, tools and documentation.
• PM compliance: how many planned jobs were done on the set date out of those scheduled.
• The share of planned work in the maintenance team’s total hours. This is where you see whether you are actually moving away from emergencies.

With low volumes the figures jump from month to month. Look at the three-month trend and read each number together with the stop log.

Common mistakes

Copying the intervals from the manufacturer’s manual for all components. The manual does not know how much dust you have in the hall, how many shifts you run or what material you process. Start from it, then adjust based on what you find at inspections and on the failure history.

Buying a CMMS before there is a routine. A maintenance system organises a process that already works. If people do not fill in a paper sheet, they will not fill in a screen either. Move to a CMMS when the log and the plan are used consistently and Excel has become hard to manage.

PM plans that nobody carries out. If the plan asks for more hours than the team has after emergencies, it is abandoned in the first month. Calculate the hours, remove what is not linked to a real failure mode and start with what can be done.

PM without feedback. If the technician finds a worn belt at inspection and nothing changes in the plan or the stock, the inspection served no purpose.

At the end of the 90 days, take the next two or three machines on the list and repeat.

If you want to compare how you organised this transition, the discussion continues in the Maintenance & Reliability group.