Data-driven decision-making for smart operation and maintenance

#degradation #data-driven #operation #maintenance

This talk presents dada-driven decision-making for smart operation and maintenance, focusing on failure prediction and maintenance optimisation. I will first discuss extending the applicability of deep learning (DL) algorithms for failure prediction with limited time series data. To broaden the application of DL algorithms in predicting failures using relatively limited datasets, we propose leveraging data augmentation methods. Different from existing approaches that indiscriminately blend synthetic and real data without considering the selection or weighting of synthetic samples, our novel method involves generating, selecting, and reweighting synthetic samples to enhance prediction accuracy. A case study of failure prediction in a wastewater treatment plant will be used to illustrate the effectiveness of the proposed method.

Subsequently, I will discuss finite-horizon maintenance strategies based on Markov decision processes, and also briefly discuss the application of online reinforcement learning in maintenance management, towards automation in operation and maintenance decisions.