摘要
针对氧化锆氧量传感器易老化导致火电厂锅炉烟气氧含量随着时间的推移测量精度下降和日常维修繁琐成本高的问题。提出运用粒子群优化最小二乘支持向量机的软测量方法代替氧化锆氧量传感器,实现对烟气氧含量的准确测量。首先根据统计学原理建立烟气氧含量最小二乘软测量模型,然后利用粒子群优化最小二乘支持向量机模型的核参数和正则化参数来提高模型的预测精度。运用现场实测烟气氧含量数据对所建立的预测模型与传统最小二乘支持向量机预测模型进行了比较,结果表明粒子群优化模型预测精度更高,泛化性更好,可以较好地解决烟气氧含量难以精确预测的问题,对锅炉燃烧效率的提高具有指导意义。
In order to increase measurement accuracy of oxygen content and reduce maintaining cost caused by ag- ing of zirconia oxygen sensor in thermal power plants, a soft sensor method is proposed based on least squares support vector machine （ LSSVM ） optimized by Particle Swarm Optimization （ PSO ） instead of zirconia oxygen sensor, this method can accurately measure flue gas oxygen content. Firstly, a soft sensor model of LSSVM is built based on the principle of statistics, then the kernel and regularization parameters of LSSVM optimized by PSO are used to increase prediction accuracy of the model. Compared with traditional LSSVM, the proposed method can measure oxygen con- tent with higher accuracy and better generalization. This method has guiding significance for improving the efficiency of boiler combustion.
作者
李艳
任锦
LI Yan;REN Jin(Institute of Electrical and Information Engineering, Shaanxi University of Science and Technology, Xi＇ an Shanxi 710021, China;Shaanxi Research Institute of Agricultural Products Processing Technology, Xi＇ an Shanxi 710021, China)
出处
《计算机仿真》
北大核心
2017年第11期58-62,共5页
Computer Simulation
基金
陕西省科学技术研究发展计划项目(2013K07-28)
