The key phrase here is ‘ always on ’, because for any company – especially manufacturers – there is no time so costly as machine downtime .

Unplanned downtime can cost industrial manufacturers billions of dollars annually . Nearly half of such incidents are caused by equipment failure . Traditionally , such failure has been addressed retrospectively , through reactive maintenance programmes .

However , predictive maintenance can help control these costs by enhancing performance awareness and reducing downtime . Predictive maintenance is the practice of analysing collective data from sensorized equipment using advanced analytics , the aim being to predict when a part is likely to fail , and thus when maintenance should take place .

The real-time flow of information , and the ability to analyse it , uses physical data to develop digital insights and drive informed physical action in the form of maintenance and upkeep .

Algorithms are used to analyse historical data , coupled with data relating to new environments , as well as user and machine data .

To better predict performance thresholds , alerts can signal when an asset is performing outside of an established threshold , so a technician can be scheduled , and spare parts ordered in advance of the asset breakdown . In many respects , predictive maintenance is the point at which IT and operations technology ( OT ) meet .

supplychaindigital . com 119