Application of Federated Learning in Predictive Maintenance to Predict Remaining Useful Life

Application of Federated Learning in Predictive Maintenance to Predict Remaining Useful Life

Data is at the center of training efficient ML models to predict remaining useful life of equipment. For example, remaining time to degradation in turbofan engines can be predicted such that proactive predictive maintenance can minimise downtime. However, data available to train models, though large in quantity, are privately sensitive in nature and often exist in silos. Access to data remains a hurdle to improving operational efficiency and reducing resource wastage using AI. To tackle this, the concept of federated learning was introduced in recent years to use the vast pool of data to train models whilst preserving privacy. We show how federated learning can be applied to predict the RUL of turbofan engines. We compare the results with baseline centralised models to show that despite the privacy preserving nature of the federated models, they offer comparable predictive performance.

Speaker