Health Monitoring of Complex Systems using Parallel Neural Networks

Complex machinery require monitoring numerous sensitive parameters. A system of parallel neural networks (PNN) can be designed to monitor the offset of any signal from its healthy operation. This paper presents the use of interprocessor communication mechanism (IPC) in PNNs. It describes integration of multi-neural networks cells in monitoring complex manufacturing machines. Each NN cell is trained with critical status of a subsystem of the machine. IPC signals activate an associated NN cell for identifying real-time status of each subsystem of the compound system. NNs cells have independent topologies. Experimental results indicate that the use of IPC in PNNs architecture achieves a high degree of precision in real-time monitoring.