Descripció del projecte
The thesis must be an expert reference work at ALSTOM in the field of predictive maintenance for rolling stock, especially through the use of a particular Machine Learning technique known as the artificial neural networks (ANN). The following list describes some of the goals that need to be attained:
– Cost-effectiveness: Prove that the ANN is the most suitable approach (i.e., good enough, developed in a short time, with increased productivity) to tackle PHM problems at ALSTOM (railway industry)
– Proven technology: Prove that the ANN technology of use stands the test of time (i.e., applied research with more than 5-7 years, showing a stable trend). Moreover, ANN is in the academic engineering curriculum, which is to be considered regarding long-term team management objectives
– Industrialisability: Prove that the lead time between a research prototype and an industrial-proof solution on any platform is minimised with ANN
– Flexibility: Prove that the ANN technique can be applied to the whole value chain of PHM for solving different problems in terms of components and variables, and also to improve the performance of the related products
– Robustness: Prove that the ANN succeeds in solving real-world railway problems, characterised by a shortage of useful data, which is very different from standard research datasets
