摘要
建立了基于统计分析的锅炉热力系统性能辨识和优化模型,采用稳态运行的有效数据样本和多维变量二次型进行系统建模,并利用统计分析判别模型的质量。模型的质量满足要求后,根据最优化原理在可调整输入参数的运行区间内进行系统性能寻优,获得了可调整输入参数的最佳控制数据;与人工神经网络法相比,该方法通过对数据样本的筛选,减少了计算量,能够满足实时需要。在1台电站锅炉上的应用表明:它是一种实用、有效的优化方法。
Performance characteristics identification and optimization model is established for boiler thermodynamic system on foundation of statistical analysis, using static, effective sample data and quadratic function of multiple variables, system model is obtained, statistical analysis is used to check the model, in operation region of adjustable parameters, system performance is optimized resorting to optimization methods after the model is proved acceptable, and the optimal control data for adjustable parameters are achieved. Comparing with artificial neural network method, this model needs less sample data and reduced calculation, fulfilling on-line demands. Application on a utility boiler shows that it is an effective, practical method.Figs 3,tables 3 and refs 3.
出处
《动力工程》
CSCD
北大核心
2004年第4期477-480,494,共5页
Power Engineering
关键词
动力机械工程
电站锅炉
统计分析
性能优化
节能
power and mechanical engineering
station boiler
statistical analysis
performance optimization
energy-saving