| Station boiler is important drive fountain in industrial production and human's life. It consumed sizeable proportion in energy sources. It was a significant question for discussion in technical alteration and retrenching energy sources that improving thermal efficiency of boiler burning, reducing consumption of energy. The key problem of boiler control is to implement effective control on combustion process. How to keep the main steam pressure having better dynamic characteristic and keep the fuel entered the boiler burning thoroughly is most important to combustion process control.Boiler combustion control system is a typical complex system with characteristic of non-linearity, large time lag, time variation, and etc. It is difficult to model the plant and is hardly to obtain perfect control performance with conventional control method; hence we should find new intelligent control strategy. At present, fuzzy theory and neural network is gradually showing enormous latent capacity in the control field, and generalized predictive control has have more and more application in the process of production. Based on the analysis and research of the previous three kinds of control technology and made use of the three synthetically, a new intelligent control strategy, that is fuzzy predictive control based on neural network, is brought forward. Firstly, aimed at the rolling optimization problem in predictive control, study how the rolling optimization be realized by fuzzy controller; Secondly, against the limitation of conventional predictive control that based on linear model, the predictive model is built with neural network to forecast the future output; Finally, study this algorithm applied for the boiler combustion system, and the control algorithm programs is worked out with Matlab, the simulation experimental curves show that the control effect is effective, and the control strategy can be applied for other systems that characterized by non-linearity and large time lag. |
PDF Full Text Request