Identification of nonlinear dynamics of nuclear power reactor using adaptive feedforward neural network

In this paper,the authors have attempted the identification of real world non-linear time-varying pressurized heavy water reactor (phwr) dynamics. the nonlinear dynamics is identified using adaptive feedforward neural network design. artificial neural network (ann) has been introduced to solve phwr nuclear power plant operation complexity in dynamics. four multi-input single-output (miso) neural systems are designed for the prediction of control rod reactivity,moderator reactivity,moderator level and actual reactor power based on control rod and moderator level dynamics. a severe transient has been imposed in training and prediction phases with proposed ann. the proposed ann is developed in matlab. in essence,the proposed adaptive feedforward neural network mimics and identifies the actual nonlinear time-varying dynamics of phwr system for an operating phwr-type nuclear power plant in pakistan. the proposed ann is found highly efficient and fast. the performance of the proposed ann is investigated and the predicted results are in good agreement with the measured results.