Modeling of Vibration Monitoring of Steam Turbine in Nuclear Power Plant using Modular Artificial Neural Network

This paper states a methodology for using a Modular Artificial Neural Network (ANN) in modeling the vibration monitoring of the Steam Turbine (ST) in Nuclear Power Plant (NPP). The input and the output signals of the vibration transducer are used as a source of the training data for the neural network model. The type of the network used in this methodology is the supervised Multilayer Feed-Forward Neural Networks with the Back-Propagation (BP) algorithm. The module architecture is according to the Human Factors (HF) Considerations in designing the Human-System Interface (HSI). The Vibration Severity limits are determined by the International Organization for Standardization (ISO) 10816. The model also contained 2out of 3 voting and dynamic trip limit value ANNs. The results show that the proposed Modular ANN has good generalization capability to monitor and protect the machine from the Vibration Severity, increasing the reliability of (ST), and good HSI. This modeling methodology can be used for the other non-redundant components in NPP such as Reactor Coolant Pump (RCP).