Pump diagnostics using machine learning

The master thesis shows how transfer learning can be used to train machine-learning models to automatically recognize faults in a UHPLC pump. With the developed machine-learning models it is possible to recognize faults in real-time, estimate the seriousness of a fault and to differentiate between several faults. With the help of the 20-Sim model the UHPLC pump, all kind of faults can be replayed and log data can be generated. With this generated log data, machine-learning models can be trained. Using the trained machine-learning models it should be possible for untrained users of the pump to recognize several faults and to solve them.