摘要
基于某116 MW燃气热水锅炉的实际运行数据,采用多层感知器(MLP)神经网络和支持向量回归机(ε-SVR)数据辨识方法对其运行特性进行仿真建模,分析燃气消耗量及NOx排放量与锅炉运行工况之间的关系,并将两种方法的精确度和泛化能力进行比较。对比发现:MLP模型预测的燃料消耗量与实际数据的误差在-2%~3%之间,NOx的排放量误差在±5%以内;而ε-SVR模型预测的燃料消耗量误差在±2%以内,NOx的排放量误差在±3%以内,ε-SVR方法具有更高的准确性和泛化能力。
Simulation modeling on the performance characteristic of boiler was built through multilayer perceptron neutral network(MLP)and support vector regression(ε-SVR)data recognition algorithm based on the operating data of a 116 MW gas-fired water boiler.The relations among gas consumption,NOx discharge and operating conditions of boiler were analyzed and the accuracy and generalization ability were compared between MLP andε-SVR algorithms.It indicates that the error of gas consumption predicted by MLP algorithm is between-2%and 3%,and about±5%for the NOx discharge,while the accuracy of gas consumption and NOx discharge predicted byε-SVR algorithm is about±2%and±3%,respectively.The results demonstrate that theε-SVR algorithm has high accuracy and good generalization ability.
作者
张立申
李仲博
李淼
姜业正
ZHANG Li-shen;LI Zhong-bo;LI Miao;JIANG Ye-zheng(Beijing District Heating Group,Beijing,China,100000;Beijing Huare Technology Development Co.,Ltd.,Beijing,China,100028;Engipower Technology Co.,Ltd.,Changzhou,China,213000)
出处
《热能动力工程》
CAS
CSCD
北大核心
2020年第2期219-223,共5页
Journal of Engineering for Thermal Energy and Power
基金
北京市重点研发计划(D171100001217001)。
关键词
燃气锅炉
数据辨识
神经网络
支持向量机
误差
gas-fired boiler
data recognition
neutral network
support vector machine
error