Research And Application Of The Power Plant Boiler Combustion Process Optimization Based On Improved Fast Learning NET

With the continuous development of the economy,people's demand for electricity is also increasing.As a populous country,China has a greater demand for power resources.Energy consumption and environmental pollution are unavoidable problems for thermal power generation.Therefore,optimization of combustion system in thermal power plants has become an important research topic for scholars at home and abroad.Traditional modeling methods usually face the problems of non-linearity and strong coupling when dealing with the complex engineering modeling of boiler combustion.In view of the shortcomings of traditional methods,the rapid learning network has excellent learning ability,which makes the network have the foundation to deal with complex modeling.Aiming at the shortage of generalization in the practical engineering of fast learning network,this paper studies and improves on the basis of fast learning network.Through the way of quantum computing,the Dropout of the network is processed,and the improved algorithm is applied to UCI data set and practical engineering modeling.In order to realize the construction of boiler combustion system in power plant,efficient optimization algorithm is also needed to optimize the network model.As a complex operation system,the combustion system of utility boilers can actually be reduced to a multi-objective optimization problem.In this paper,an efficient multi-objective optimization algorithm(r-MOWAP)is proposed by improving the multi-objective wolf swarm algorithm.The effectiveness of the r-MOWAP algorithm is verified by testing seven multi-objective test functions,which supports the practical application of the boiler combustion optimization model.In order to verify the optimization model proposed in this paper,the combustion system of a 300 MW boiler in a thermal power plant is studied.Quantum Dropout Fast Learning Network(QDFLN)is used to accurately predict the combustion parameters.The r-MOWAP algorithm is used to optimize the adjustable parameters in the combustion process of the boiler,and to find the optimum combustion parameters ratio,so that the boiler system can operate in the optimum state.The experimental results show that the model of boiler combustion system based on QDFLN and multi-objective wolf swarm algorithm has high prediction accuracy,strong generalization ability and obvious optimization effect,and achieves the comprehensive optimization goal of reducing NO_x emission and improving combustion thermal efficiency.At the same time,the visualization of the boiler combustion optimization model is finally realized,which provides support for the realization of coal saving and emission reduction in thermal power plants.