Until recent times, factory machinery just broke and was then fixed when the maintenance guy could get hold of the spare part and could slot in the time in between all his other pressing tasks. This meant long periods of unplanned downtime and low OEE (Overall Equipment Effectiveness) scores. The only alternative was to over maintain and that bring in its own costs and challenges. Things are better now, right?
Unfortunately, not. But they could be! Solutions of sorts have been deployed for a number of years – generally referred to as condition monitoring. These have centred around diagnostics – whether via an engineer sticking a vibration probe on a machine or adding a sensor and reviewing the measurements remotely (e.g. using a historian) to check the machine’s current health. These give you warnings of impending failure – you’re looking at no more than a few hours, perhaps a day or two for some equipment - which gives limited benefit. Any operations manager will remind you, lead times of spare parts and the availability of an appropriate time to shut down the line for repairs are things you can’t take for granted.
We recently visited a site where PROGNOSYS is being deployed in a pilot. An operator noticed increasing vibration levels in one of the machines. Within two days its operating tolerances were exceeded and the machine had to be shut down. They ordered a replacement bearing (not held in stock as “stock is bad”) which had a three-week lead time. The result – three weeks of unplanned downtime and lost production and orders. Sadly, this type of event is all too common.
This is why we’re developing PROGNOSYS to provide prognostics capability to the wider manufacturing community and forecast machine failures. Prognostics is all about forecasting the risk of failure long before there is operational impact, providing a time to failure (referred to as Remaining Useful Life, or RUL). In short, condition monitoring is about the now. Prognostics is about tomorrow, next week and next month. With this sort of information, you can properly tune your maintenance operations and drive up those KPIs, like On Time and In Full (OTIF).
True Condition Based Maintenance (CBM) regimes are still some way off and involve a lot more than just neat technology. Anyone who’s tried rolling out CBM will tell you that organisational change is often the biggest challenge. But the underlying tech pieces are coming together, with other advances like the Industrial Internet of Things and new low-cost sensors, making CBM more of a reality for more businesses.