An Improved Neural Network and its Applicable Study

In this paper, a robust neural network-based on line learning and artificial immune algorithm is proposed for a boiler combustion optimization system. This method involves a model modification and parameter optimization to the normal use of boiler combustion optimization system neural network. Neural network consists of working sets and standby sets of implicit strata real-time adjusted set number. Standby sets changed into working sets when the need of neural network relearned arised. Parameters of neural network are optimized by artificial immune algorithm. Analyzed results and illustrative examples show that the proposed neural network has a fast convergence to the optimal solution and effectively applied to real-time boiler combustion optimization system.