The project intends to apply machine learning techniques to decrease maintenance costs and maximize up-time of medical LINACs for particle therapy by applying “personalized” preventive maintenance plans. This one-year pilot intends to evaluate whether this innovative approach would indeed be advantageous for commercial systems, in which case a follow-up project would be proposed. The project will use data from CERN LINAC 2 and 4 as well as data securely collected in real-time using a block-chain-based infrastructure. A set of “blueprint” reference models of an appropriate sub-section of a typical medical LINAC will be used as starting point. This overcomes the difficulty in accessing blueprints of existing commercial machines