LOGISTICS
“With predictive maintenance you should always be looking out for the exception ”
Predictive Maintenance and Service with SAP
ANDY HANCOCK GLOBAL VP OF SAP ’ S DIGITAL SUPPLY CHAIN CENTRE OF EXCELLENCE
Predictive maintenance is rooted in SCADA systems “ SCADA has been around for 40 years ,” says Hancock , “ but Industry 4.0 technologies such as artificial intelligence , machine learning and Internet of Things ( IoT ) have taken it to a whole new level of sophistication .”
He adds : “ SCADAs were all about condition-based monitoring . So , ‘ If this happens , then do that ’. But what people increasingly want is a predictive element , such as gaining insight on the useful remaining life of an asset somewhere out in the field . With big data , it ’ s now possible to carry out trend analysis to get at this kind of information .”
Using today ’ s technologies , it ’ s also possible to exert fine control over individual assets , wherever they might be . An electric motor on a piece of machinery in the desert , for example .
“ The amperage in a failing motor is different to one that ’ s working properly ,” says Hancock . “ So you can put thresholds into your algorithm to alert the operator about this before any damage happens .”
He adds : “ In this way , over time you get an understanding of the optimal way to set up servicing intervals for equipment , to make sure that nothing ever fails .”
Too much data can be as bad as too little But Hancock warns that datadriven solutions can also bring their own problems .
“ With predictive maintenance you should always be looking out for the exception ,” he says . “ Think of a temperature gauge on a piece of equipment , which is feeding back data . If everything is running fine , it will always be roughly the same temperature . You don ’ t need to keep feeding back data about that piece of equipment . The only data you want to capture is if something changes - say if the thermostat fails .
“ With data , less is usually more . Systems can soon get inefficient if overloaded with data . And then you end up chucking more tech at the problem , where what you really need to do is cut back on the data .”
Prescriptive maintenance also on the rise Predictive maintenance might feel like a new concept to many , but , says Hancock , some companies are already taking things
supplychaindigital . com 81