In response to the increasing demands for economically efficient and environmentally sustainable operation in industrial sectors, industrial processes are becoming increasingly more complicated, and the degree of automation is growing rapidly. In this context, the cyber-physical systems (CPS) enhance the overall efficiency and performance of integrated systems by combining physical processes with computational capabilities and communication networks. This integration enables remote monitoring and control. However, physical processes within CPS are vulnerable to faults. Therefore, process monitoring and fault-tolerant control methods focus on triggering alarms, and compensating for the faults through controller accommodation. The first objective of this research is to develop a process monitoring scheme, based on a predefined performance index, to detect multiplicative faults. This approach is aligned with control performance monitoring (CPM) frameworks. To this end, two distinct performance indices are investigated. The index based on the loop performance degradation model is selected, as it remains independent of operating points and demonstrates better fault detectability. Following successful fault detection, the second objective of this work is to develop a controller accommodation method to recover the performance index. For this purpose, the plug and play control methodology is applied. The implementation and optimization of the plug and play controller are conducted on a cloud-computing platform to ensure reliable and efficient computation. Considering the cyber security and data privacy requirements in the context of reliable system operation, the edge-cloud computing configuration is adopted. Specifically, residual signals, which serve as fault information carrier are transmitted to the cloud computing platform instead of the system input and output data. This, in turn, complicates the design sophisticated attacks such as stealthy attacks, which traditional observer-based detection schemes may fail to detect. Finally, a case study on a three-wheel robot validates the effectiveness of the proposed process monitoring and cloud-computing-based fault-tolerant control schemes.
Abd Alghafer Hani Salah (Mon,) studied this question.