Our PhD Students

My PhD work belongs to the Prognostic and Health Management (PHM) research field aiming to fully leverage the information of an industrial component to perform fault detection, fault diagnostic, prognostic of time-to-failure thereupon proactively manage maintenance and failures. Deep Learning (DL) is a branch of artificial intelligence based on the use of neural networks with deep architecture of a large number hidden layers and neurons which has the better capability of mining hierarchical representations from large-scale raw data. With the exploded amount of data available in the industry 4.0 era, deep learning-based PHM approaches are becoming more and more attractive. However, in PHM applications involving complex and safety critical industrial systems, the cost of acquisition of data describing the equipment behaviour in abnormal conditions is typically prohibitive. This is due to the fact that equipment tends to operate in healthy conditions, therefore, the historical data rarely refer to degraded or faulty conditions. Also, pure data-driven approaches can produce results which are inconsistent with well-known physics laws and, therefore, are unacceptable, especially in critical industrial sections, such as power production and energy systems. Since the demand for interpretable and transparent PHM methods is rising, my PhD work will investigate the possibility of developing a physics-informed data-driven PHM framework built on the recent advancement of physics-informed neural network (PINN) which combines the physical knowledge/constraints thereby mitigating the requirement of massive historical data and improving the physics-consistency of maintenance decision.