What is predictive maintenance?

Predictive maintenance means maintaining equipment when its measured condition says it is needed, before it fails, rather than on a fixed schedule or after a breakdown. Wikipedia describes the techniques as "designed to help determine the condition of in-service equipment in order to estimate when maintenance should be performed". It notes that the approach "claims more cost savings over routine or time-based preventive maintenance, because tasks are performed only when warranted".

How it works

Sensors measure what changes as a machine wears: vibration, temperature, current draw, noise, oil condition. IBM, which sells maintenance software, describes it as using "operational data and real-time condition monitoring (CM) to predict when assets are likely to fail". Maintenance is then scheduled before the failure, ideally at the next planned stop.

Where it pays

Not every machine is worth watching. In a mill, downtime on the bottleneck costs output directly, while a machine with spare capacity can often lose an hour at little cost.

From Quarri's own work with a sawmill: the trim line was the bottleneck, down 75%+ of the time. A figure like that shows where to look first. It doesn't show that the stops were failures a sensor could predict. Some downtime is starved or blocked flow, changeovers or planned stops, which predictive maintenance won't touch.

Two cautions

Bottlenecks move with product mix and log supply, so the machine that limits output this month may not next month. And a failure on a machine with a long repair lead time, such as a main chipper, can stop the whole line wherever the bottleneck sits. Ranking machines by the cost of their failure covers both cases.

What it isn't

Predictive maintenance isn't preventive maintenance, which works to a calendar. Nor is it only sensors: someone has to decide what happens when an alert fires.

What to check

Which machines, if they failed, would cost the most output or the longest wait for parts? Measure any programme by the output it recovers, not by alerts raised.

Quarri for sawmills is built around how a sawmill runs, from log intake to shipped order.

Quarri is an AI-native data platform for the timber supply chain. It connects buying, production, sales and inventory for forest management, sawmill, wood products and pulp, paper and packaging operators.