Preventing asset failure in the refining industry using machine learning
THERE is a very real need to carry out equipment failure prevention using data rather than estimates. The combination of mechanical- and process-induced breakdowns costs up to 10% of a global US$1.4trn manufacturing market (2012 McKinsey Global Institute report).
While companies have spent millions trying to address the issue of unplanned downtime, until now they have only been able to address wear- and age-based failures. Current techniques cannot detect failures early enough and lack insight into the reasons behind the apparently random failures that cause more than 80% of unplanned downtime. This is where machine learning software can play a role and capture process-induced failures.
Companies need to identify and respond effectively to early indicators of impending failures to avoid this unplanned downtime. However, traditional maintenance practices do not predict process failures. That would require a unique technology approach combining assets and processes; especially for asset-intensive industries such as the energy industry. With the right technology in place, organisations can sense the patterns of looming degradation with sufficient warning to prevent failures.