This paper focus on the Pellet-Cladding Interaction (PCI) evaluating by a self-organizing Radial Basis Function Neural Network (RBFNN) during the operation of Zircaloy clad fuel in the Water cooled Reactors. Using the neural network, which built through the analyzing of the existing data but the physical process of PCI, is a suitable way to reduce the calculation complexity. In this sense, the stage of PCI can be evaluated online. A self-organized RBFNN is used, which can vary its structure dynamically in order to maintain the prediction accuracy. The hidden neurons in the RBF neural network can be added or removed online based on the neuron activity and mutual information, to achieve the appropriate network complexity and maintain overall computational efficiency. The PCI experiment data from literatures is used to test this method, and the results demonstrate its effectiveness.
