基于深度主元分析的电站锅炉NO_x质量浓度预测模型研究  

Research and application of prediction model for NO_x concentration in utility boilers based on deep component analysis

作　　者：王林[1] 王小成[1] 郭亦文 姜守义 高林[1] 侯玉婷[1] 周俊波[1] 姜肇雨 刘涛 王芸 

机构地区：[1]西安热工研究院有限公司,陕西西安710054 [2]华能国际电力股份有限公司日照电厂,山东日照276826 [3]华电江苏能源有限公司苟容发电厂,江苏苟容210029

出　　处：《热力发电》2017年第12期68-74,共7页Thermal Power Generation

基　　金：国家重点研发计划项目(2017YFB0902101)~~

摘　　要：目前,国内电站大部分选择性催化还原(SCR)脱硝系统喷氨装置喷氨调节阀自动控制均存在一定的问题。本文通过深度主元分析,阐述了机组SCR脱硝系统入口NO_x质量浓度非线性模型的构建方法,同时利用该方法构造的模型以及模糊调节和时滞补偿控制,发展了新型的喷氨控制策略。运行控制结果表明,在动态和稳态过程中,烟囱出口NO_x质量浓度均能长时间稳定地控制在设定值附近,有效解决了SCR脱硝系统喷氨自动投入的问题。At present, most of the automatic control of ammonia injection valve in domestic SCR ammonia spraying system exist certain problems. To solve this problems, this paper expounded the establishment method for a non- linear predictive model for NOx concentration at the SCR system inlet, on the basis of deep component analysis. Moreover, combined with the fuzzy control and time delay compensation method, this model was applied to develop a new control strategy for ammonia injection. The result shows that, in both steady and dynamic state of the unit, the NOx concentration at the stack outlet can be controlled within the range close to the setting value for a long time. So the problem of automatic operation of the ammonia injection in SCR system was effectively solved

关 键 词：超净排放 喷氨控制 主元分析 非线性模型 模糊调节 时滞补偿 NOx质量浓度 

分 类 号：TP272[自动化与计算机技术—检测技术与自动化装置]