Nuclear power plant transient detection and classification using fuzzy pattern matching

For the safety of a nuclear power plant, it is needed to classify and count the transient state during the entire period of a plant operation. Related to this, most of suggested methods require the input and output process data, simultaneously. However, the input data given from the operator's control action usually cannot be accessed in a nuclear power plant. So, in this work we propose a nuclear power plant transient recognition and classification algorithm depending on the output data only using a fuzzy pattern recognition technique. This can easily learn a new transient pattern by storing the pattern. Moreover, adopting the fuzzy logic permits the robustness for measurement noise and disturbance. Finally we apply it for a steam generator and show its usefulness