Condition Monitoring: A Practical Starting Point for Manufacturers

What condition monitoring is, stripped back

At its core, condition monitoring is just paying attention to what your equipment is telling you. Most mechanical failures do not happen without warning. There is usually a window — days, weeks, sometimes months — where something measurable changes before the asset actually fails. The bearing starts running warmer. The vibration signature shifts. The motor is drawing more current than it should. Condition monitoring is the discipline of noticing those changes early enough to do something about them.

That is it. The technology varies from a handheld thermometer to a fully automated sensor network, but the principle does not change.

Why most condition monitoring programmes fail

We have been called in to look at condition monitoring programmes that cost serious money and delivered almost nothing. The pattern is usually the same — someone bought a platform, installed sensors on a handful of critical assets, generated a lot of data, and then nobody in the maintenance team knew what to do with it. The system ran for six months, the dashboards got ignored, and it quietly got written off as "not working for us."

The problem was never the technology. It was that nobody connected the monitoring output to the maintenance planning process. Data without a decision attached to it is just noise.

A technician doing a walk-round once a week and recording bearing temperatures on a clipboard is doing condition monitoring. If those temperatures are feeding into the planning meeting and influencing when jobs get scheduled, that simple clipboard process will outperform a £50,000 platform that nobody trusts.

Start with the right assets, not the best technology

Before you think about sensors or software, work out which assets actually justify the attention. The question is simple: what happens when this fails?

If a motor fails and it takes the whole production line down for 16 hours with no standby, that asset needs condition monitoring. If a pump fails and you switch to the duty-standby in 10 minutes and carry on, it probably does not — at least not yet.

When we do this exercise with sites, we usually end up with somewhere between 10 and 25 assets that genuinely justify closer monitoring. That is a number you can manage without a big programme. It is also where the return is highest, because those are the assets whose failures cost you the most.

The mistake we see most often is trying to monitor everything at once. You spread the effort thin, the signal-to-noise ratio gets terrible, and the maintenance team loses confidence in the data. Start narrow, do it well, expand from there.

The techniques that are worth starting with

Vibration monitoring on rotating plant

If there is one technique we recommend first for heavy manufacturing, it is periodic vibration checks on critical rotating assets — motors, pumps, fans, gearboxes. Changes in vibration signature can flag bearing wear, imbalance and misalignment weeks before they become failures.

You do not need a permanently installed system to start. A portable vibration analyser and a technician who has been trained to use it, checking your top 15 critical rotating assets once a fortnight, will catch a significant proportion of the failures that currently arrive without warning. The investment is modest. The return, if you act on what you find, is not.

Thermal imaging of electrical infrastructure

Hot spots in switchgear, distribution boards and motor control panels are a fire risk and a failure waiting to happen. A thermal imaging survey of your main electrical infrastructure twice a year is one of the highest-value, lowest-cost things a manufacturing site can do — and the majority of the sites we visit have never done one.

We have walked into sites and found connections running at temperatures that should have triggered an immediate shutdown. Nobody knew because nobody had looked. That is not a technology problem, it is a habits problem.

Oil analysis on gearboxes and hydraulics

If you have critical gearboxes or hydraulic systems and you are not sending oil samples for laboratory analysis, start. A sample costs between £30 and £80 and tells you about wear particle content, contamination and fluid degradation before any of it becomes visible externally. We have caught gearbox failures months in advance this way. It is unglamorous and it works.

The thing that makes or breaks it

None of the techniques above matter if the findings do not change what happens in the planning meeting. This is where most programmes fall down, and it is the part that has nothing to do with technology.

Condition monitoring data needs to flow into a decision. Is this asset okay, is it trending in a direction we should watch, or does it need intervention now? That decision process needs to be owned by someone, happen on a regular cadence, and actually result in work getting scheduled.

We usually recommend building this into an existing weekly or fortnightly planning review rather than creating a separate process. The maintenance team is already in the room. The work order system is already open. You are just adding a data check to a meeting that is already happening.

Get that loop working on 15 assets and you will have learned more about condition monitoring than six months of reading about it. Then you can decide what technology to invest in — with actual site experience to inform the decision.