Is installing a sensor for predictive maintenance always the right answer: How to make the hype work?
The use of sensors in predictive maintenance is a hot topic in the discussion about asset performance. Although sensors are used widely today, their effectiveness is a subject of debate. One thing is certain: sensors are not the be all and end all. They only provide value when used in support of people (experts who know the machines) making informed and better decisions. Jonas Vancauwenberghe, an expert in the field, sets the tone.
Jonas Vancauwenberghe breathes asset performance. After a career as a maintenance technician and maintenance manager at Alpro, and his role being responsible for Industry 4.0 at ABB, he now supports companies from his position at REPM Dexis. He offers solutions aimed at optimising machine performance and addressing maintenance challenges.
Sensors and maintenance
Sensors have received a lot of attention because predictive maintenance is often touted as a way to reduce downtime, extend equipment life, and lower costs. However, the reality is often more complex. “Sensors offer value, but they are not the be all and end all,” explains Mr. Vancauwenberghe. “Their effectiveness depends on how well they are integrated into a broader maintenance plan, and how well we interpret the data they generate.
Companies must first return to first principles when it comes to their maintenance needs, and then use strategic research to consider different strategies such as reactive maintenance, preventive maintenance, and condition-based maintenance.”
More than data capture
Should companies opt for predictive or condition-based maintenance, the interpretation of data is the most important consideration. “Sensors generate an abundance of data, but true success depends on how you use that data. Do you have the resources and knowledge to interpret the data? Prediction is impossible without insight into failure modes, operational knowledge and its reflection on machine components,” says Mr. Vancauwenberghe.
Condition-based maintenance means it is not always necessary to resolve every fault immediately. If a pump continues to work despite having a few faults detected by sensors, it can last for years without replacement. “The assumption that every problem must lead to an immediate maintenance intervention is a misconception. The number of faults a machine can endure before actually failing is often higher than expected. As long as the machine continues to perform as required and is safe, faults can be tolerated. Therefore, it is important to determine when a machine is actually failing in terms of performance, and only then should we intervene. Sensors are great tools to accurately determine that point.
Sensors provide support, they are not decision makers
Finding the right balance is crucial. “Sensors should be used as a support tool, not as the sole decision-makers. People must always be involved in the decision-making process and combine the sensor information with their own operational knowledge,” emphasises Mr. Vancauwenberghe. “ Sensors can offer support and help people make better decisions, but they cannot make decisions.”
“Predictive maintenance must therefore be approached carefully and with open eyes. It’s important not to confuse tactics with strategy, and not to allow the hype to obscure the reality. A well-thought-out and balanced approach will ensure predictive maintenance can become the valuable tool it is intended to be rather than the panacea it is often perceived to be.” Mr. Vancauwenberghe concludes.
Gold Sponsor Asset Performance 2024:
