Study On Boiler Control Based On Artificial Neural Networks

Boilers are important components in power systems, Boiler is quite complex and huge in size. The process happened in it is multi-variables and its parameters are heavily coupled together. So its control and optimization are as one of the most key problems by researchers.In this paper, RBF Neural Network was selected as a tool for modeling according to the characteristics of boiler systems. The paper Based on the low-cost DCS platform and aiming at the characters of medium-sized and small complex process control system, The boiler system was modeled through its static operation data. 'Multi-Complex' Method was selected as the implementation of multi-goal optimization, and was used in optimization of the process based on the model gained, searching for the optimal input parameters.The pruning strategy of RAN algorithm is improved in this paper. The pruning strategy not only to delete the hidden neurons continuous contributing little to the network output, but also to combine the similar hidden neurons and thus to implement a more compact network structure. RAN algorithm based on radial basis function (RBF) neural network is a dynamical neural network and suitable for process online modeling. The online identification of power plant thermal process nonlinear model is carried out by the method, simulation study indicates the validity of the RAN algorithm based on radial basis function neural network and established model with higher precision, less calculation and can be used in model based control algorithm directly.