基于神经网络与遗传算法的锅炉系统的优化

Optimization of Boiler System Based on Neural Network and Genetic Algorithm

传统热工控制系统DCS在锅炉运行中存在多变量之间相关关系难以全面协调以及期望性能指标难以达到的问题。针对参数难以全面协调问题,采用BP神经网络来解决多变量间的非线性关系。针对性能指标的优化,在神经网络预测功能的基础上,采用遗传算法优化锅炉参数间的比值,并将优化后的风煤比在线监测数据作为反馈信号引入锅炉性能优化系统中,改进后的方法优化处理了热工控制系统问题。以青岛市某循环流化床锅炉为例,通过Matlab与C#语言编写的可视化界面曲线得出,预测与优化的综合方法能够将锅炉效率提高0.5%~3%。

In the process of boiler operation,traditional thermal control system（DCS）has two difficult issues,one is the relationship between multiple variables which is difficult to coordinate,and the other is hard to achieve the desired performance indicators. For the coordination of parameters relations,BP neural networks can solve the problem of multi-variable nonlinear relationship. For the optimization of performance indicators,the paper adopts genetic algo- rithms and uses BP neural network to predict the value of efficiency optimized. The improved method can optimize the boiler efficiency. In this paper,a circulating fluidized bed boiler of Qingdao is taken as an example,and analysis is done through Matlab software and C# language. The optimization results show that boiler efficiency is improved by 0.5% to 3%.