Estimating Influential Parameters Affecting Exploitation of Bushehr's Nuclear Power Plant Using Neural Network

The present study has been conducted with the aim of estimating influential parameters affecting exploitation of Bushehr's nuclear power plant using a neural network. Safe exploitation of a nuclear power plant needs permanent monitoring of reactor core using reliable methods and careful analysis of the process of changes in safety margins such as DNBR (Departure Nucleate Boiling Ratio). For online monitoring of reactor core, calculation and prediction of some parameters with high speed and high accuracy are required. Inaccuracy and carelessness in estimating these parameters is high in traditional methods. According to features of data-based methods such as neural network, it can be performed adequately and with high speed and high accuracy with these methods. As a result, increased accuracy in calculations can improve safety margins and increase output power. Obtained results based on different error criteria show that parameter prediction by the developed neural network with a maximum error level of 10% has higher accuracy than other predicting systems based on the current model with error level about 30-40%.