Unobservable fault-propagation control of serial production systems based on discrete-event models

The paper formulates the multiagent fault-tolerant supervisory control of serial production systems under partial observation based on unobservable fault identification utilising the difference information of the controlled results. This allows the productivity of the serial production systems to be improved and the facilities to be protected in a certain abnormal situation. A case study of the fault-propagation control of polypropylene pre-polymerisation and polymerisation reactors is provided to illustrate the proposed control scheme