Featured

A DIY toolbox for predictive maintenance

Assets are like fingerprints, says digitalisation expert Freddie Coertze. Each one has a unique signature – which is why certain condition monitoring techniques might not be the solution when it comes to improving equipment health.

“I think there is a misconception about monitoring assets, that perhaps you can only monitor just one value such as vibration or temperature, and that will be effective in determining failures,” says the IoT business development manager for ifm Australia.

“When actually, to get a holistic picture of an entire asset, you should be monitoring different values and patterns.”

According to Freddie, a holistic approach to asset health is the fundamental difference between condition monitoring and predictive maintenance. Condition monitoring is reactive, while predictive maintenance is proactive.

“Condition monitoring focuses on identifying problems as they occur, while predictive maintenance uses data analysis from sensors, values and process information to detect anomalies quickly and ‘predict’ health issues,” he said.

“Both approaches can provide benefits, but predictive maintenance is of course defined by the word ‘predict’.”

“It has the potential to provide even greater benefits by giving advance warning to act before the equipment is already starting to fail, which in turn reduces downtime and maintenance costs,” Freddie said.

Circling back to the fingerprint, Freddie says that each asset behaves differently. Which is why a solution needs to be able to interpret those unique characteristics or signatures.

To make this simple for businesses – regardless of their size – the ifm Industrial Internet of Things (IIoT) platform has an in-built DataScience Toolbox that has this capability.

“This is why moneo is so powerful. It makes it easy to get the right information about your assets, because it includes this DataScience Toolbox – you don’t need a data scientist involved to get access to machine health insights, it’s completely DIY,” he explains.

 “The DataScience Toolbox will establish what good behaviour is for each machine based on historical and current data. It will keep comparing the machine’s behaviour, and if there are any anomalies, it will bring that to immediate attention.”

The moneo DataScience Toolbox incorporates artificial intelligence (AI) algorithms that assess data recorded by sensors.

It uses data with advanced machine learning technology to set parameters and make sense of detected anomalies and patterns.

It does this with two specific tools – the moneo SmartLImitWatcher and moneo PatternMonitor.

“The SmartLimitWatcher will generate a model based on the status of a process that is monitored. If the data shows an anomaly outside the parameters in the model it will set an early notification,” says Freddie.

“Whilst the PatternMonitor is looking at structural changes in critical processes. Like its name, it will identify patterns in the process, and report on any changes in volativity or levels. It alerts you to any undesirable process changes.”

Send this to a friend