Predictive analytics is not quite magic, but it could make a world of difference for companies who want to save costs and improve efficiency.
Here, Leroy Spence, Manager for the Americas at industrial equipment supplier, EU Automation, discusses how manufacturers can use predictive analytics to manage assets and minimise downtime.
Most manufacturers have been using predictive analytics for years in the form of spreadsheets and manual data entry. Although useful, these methods were subject to operator assumptions and human error.
As the number of sensors on the factory floor increases, data gathering becomes automated and data sets are easier to analyse.
New predictive analytics techniques also significantly improve data accuracy. Engineers can use these data sets for predictive maintenance purposes to determine the condition of in-service equipment and identify when it will need to be repaired or replaced.
Predictive analytics is able to compare real-time machine data gathered from sensors to a history of machine failure. It uses complex algorithms to spot behavioural patterns before a breakdown.
By combining sensor technology and big data analytics, manufacturers can minimise equipment failure. Knowing that a motor or drive is likely to break soon means the manufacturer can repair it or order a replacement before a breakdown occurs.
It also allows manufacturers to schedule maintenance work for a convenient time, instead of shutting down production as and when breakdowns happen.
Predictive maintenance not only minimises downtime, it also gives maintenance engineers and plant managers peace of mind. It frees up the time of maintenance staff, so that they can deal with tasks that are more valuable to the business. In the near future, it is conceivable that a smart production line could order spare parts automatically when necessary, with minimal human intervention.
Beyond predictive maintenance
The real benefit of predictive maintenance lies in the hidden potential of data. By analysing historical data and identifying machine or part breakdown patterns a manufacturer could easily identify weak points in their production line and implement measures to address these flaws.
For example, if there is an electric motor that breaks regularly on a production line, the plant manager could look at the historical data gathered from the motor and compare it with environmental data to identify the most likely cause of failure, such as excessive heat, power supply issues, humidity or vibration.
A slight redesign could help to reduce the likelihood of these factors leading to failure. An experienced engineer would also be able to identify the causes of constant equipment failure and suggest changes, but predictive maintenance makes the process much easier.
Perhaps the most exciting application of predictive analytics lies outside the factory doors. By integrating predictive analytics into their products, manufacturers can turn the technology into an add-on service.
The self-diagnosing and self-maintaining production line is still a few years away, but it is becoming more of a reality as predictive analytics technology becomes more widely available.
It all comes down to Arthur C. Clarke’s second law of prediction, which states that “the only way of discovering the limits of the possible is to venture into the impossible.”