(7b) Achieving the Next Level of Asset Integrity with Predictive Maintenance

Exceptional organizations manage to deliver superior operational performance without compromising on safety and integrity. On the contrary, high performing players are at the forefront of equipment reliability for critical equipment and safety performance.

In this context, introducing Predictive Maintenance can dramatically help achieve the next level of asset integrity by leveraging data - structured and unstructured, machine and non-machine based - in order to generate insights with advanced analytics methods that would not be possible with conventional techniques

Implementing Predictive Maintenance generates three benefits that reinforce one another:

· Predict when equipment will fail and help reduce the safety and integrity threats
· Avoid failures and positively contribute to reduce the scale of the value at risk
· Enhance specific asset knowledge and learn more about the behavior of critical equipment

In a recent client project for an upstream oil and gas facility, the main gas compressors have been identified as critical equipment: they account for many failures and a large portion of production losses. A Weibull analysis showed that out of several known failure modes, some show a random failure pattern, leading to sudden breakdowns releasing high energy and thus representing a safety issue. As per the random failure behavior, classical time- or usage-based maintenance strategies are not fruitful.

Combining 100’s of gigabytes of loss data, maintenance work order data and sensor data from 1,200 sensors around the equipment, a data driven model has been created showing that out of the 1,200 sensor sources, 43 explanatory variables predict the time to failure with ~80% accuracy. The model has been tweaked further in order to reduce the number of false negatives while accepting by this a certain number of false positives (i.e. false alarms) due to safety reasons.

As a result of this model, an early warning of an imminent failure means repairs can be planned, spares can be gathered and the equipment can now be shut down in a safe and orderly manner. Overall, this represents an average reduction from 14 to 4 days of downtime per breakdown event.