Development of a fault diagnosis system based on fuzzified neural networks in chemical processing plants

Failure of chemical plants is very dangerous. Although such failure is solvable with extension of a sensor, and an increase of the operator, a huge amount of funds are required for these initiatives, and it cannot be realized easily. A model has been built by neural networks, and a diagnosis method using that model was proposed. In this paper, we propose the alternative diagnosis method using fuzzified neural networks.