Remaining useful life prediction of a piping system using artificial neural networks: A case study