一种电站锅炉过热蒸汽温度数据解析模型  

A Superheated Steam Temperature Data Analytics Prediction Model of Power Station Boiler

作　　者：唐振浩[1] 杨名璇 曹生现[1] TANG Zhen-hao;YANG Ming-xuan;CAO Sheng-xian(School of Automation Engineering,Northeast Electric Power University,Jilin 132012,China)

机构地区：[1]东北电力大学自动化工程学院,吉林吉林132012

出　　处：《控制工程》2020年第8期1351-1357,共7页Control Engineering of China

基　　金：国家自然科学基金(61503072,51376042);吉林市科技创新发展规划基金(20166009)。

摘　　要：火电厂锅炉生产过程具有强非线性、强噪声等特点,这些特点导致过热蒸汽温度难以准确建模。针对这一问题,提出一种锅炉过热蒸汽温度数据解析建模方法。首先采用分类回归树(Classification and Regression Tree,CART)算法计算得到相关变量重要性,选择重要性大于0.5的变量作为数据驱动建模的输入。然后采用粒子群优化(Particle Swarm Optimization,PSO)算法优化最小二乘支持向量机(Least Squares Support Vector Machine,LSSVM)的参数,建立过热蒸汽温度的预测模型。最后基于实际生产数据的实验结果表明,所提出的建模方法取得比常用建模算法具有更好的预测精度,能够实现对锅炉过热蒸汽温度的准确预测。The production process of the thermal power plant boiler has strong non-linearity and strong noise,which leads to the difficulties in establishing an accurate model of superheated steam temperature.To solve this problem,a data analyze modeling method of superheated steam temperature is proposed in this paper.Firstly,a classification and regression tree(CART)algorithm is used to compute the importance of the relevant variables,and the variables with weight greater than 0.5 are selected as input of the prediction model.Secondly,the prediction model is constructed by a least squares support vector machine(LSSVM)model whose parameters are optimized by particle swarm optimization(PSO).Finally,the experimental results based on the practical data illustrate that compared with common prediction model,the proposed model obtains better prediction accuracy to predict the superheated steam temperature.

关 键 词：数据解析 过热蒸汽温度 特征选择 模型 算法 预测 

分 类 号：TH3[机械工程—机械制造及自动化]