基于改进朴素贝叶斯的脱硫循环泵启停优化  

Optimization of Start and Stop of Desulfurization Circulation Pump Based or Improved Naive Bayesian

作　　者：郑茗友 王伟[1] 赵文杰[1] ZHENG Ming-you;WANG Wei;ZHAO Wen-jie(School of Control and Computer Engineering,North China Electric Power University,Baoding 071003,China)

机构地区：[1]华北电力大学控制与计算机工程学院,河北保定071003

出　　处：《控制工程》2023年第4期746-752,共7页Control Engineering of China

摘　　要：石灰石-石膏湿法烟气脱硫(WFGD)装置是燃煤电厂的主要耗能设备,而浆液循环泵又是脱硫系统的主要耗电设备。在实际运行中,浆液循环泵投运台数主要依赖于运行人员的人为判断,从而影响了整个系统的脱硫效率与经济性。针对朴素贝叶斯算法中“属性条件独立性假设”的局限性,提出一种基于属性相关度的加权朴素贝叶斯算法。在对浆液循环泵的启停进行判断时引入相关性权重因子,并对训练集的选取与模型的输出进行优化,以提高浆液循环泵启停判断的准确性,实现脱硫系统浆液循环泵的经济运行。测试结果表明,改进的模型能够提升判断的准确度,减少误判的发生,对脱硫系统的智慧运行具有指导意义。Limestone-gypsum wet flue gas desulfurization(WFGD)device is a big energy consuming unit in coal-fired power plants,and the slurry circulation pump is the main power consuming equipment of the desulfurization system.In actual operation,the operating quantity of slurry circulation pumps mainly depends on the experience-based judgment of operators,thus affecting the desulfurization efficiency and economy of the whole system.Aiming at the limitation of"attribute conditional independence assumption"in naive Bayesian algorithm,a weighted naive Bayesian algorithm based on attribute correlation is proposed,the relevance weighting factor is introduced for the start-stop judgment of slurry circulation pump,and the selection of training set and the model output is optimized,which improves the accuracy of the start-stop judgment of slurry circulation pump and realizes the economic operation of slurry circulation pump for the desulfurization system.The test results show that the improved model can improve the accuracy of judgment and reduce the occurrence of misjudgment,which has guiding significance for the intelligent operation of the desulfurization system.

关 键 词：WFGD 朴素贝叶斯分类器 浆液循环泵启停决策模型 智慧决策 

分 类 号：TK39[动力工程及工程热物理—热能工程]