基于证据理论的循环流化床锅炉煤质扰动识别

燃烧扰动是循环流化床锅炉常见扰动之一,基于人工判断很难将煤质变化引起的燃烧扰动与其他因素导致的燃烧扰动正确区分。针对以上问题,提出一种基于证据理论的煤质好坏在线识别方法。首先对锅炉燃烧过程进行机理分析,构造了机组负荷/给煤量、热量/给煤量、给水流量/给煤量这3个可以反映当前煤质的证据;然后用典型样本法建立置信函数;最后用D-S融合法则将这3个证据进行融合,得出最终的诊断结论。取电厂实际运行数据对该方法进行验证,诊断结果表明该方法可有效区分煤质正常和煤质变差两种燃烧状况;相比单一证据诊断结果,对目标模式的支持度增加,不确定性减小,提高了识别的准确性。

Identification of Coal Disturbance in Circulating Fluidized Bed Boiler Based on Evidence Theory

Combustion disturbance is one of the most common destabilizations in circulating fluidized bed boiler.It is not easy to distinguish the combustion disturbance caused by coal quality change from that caused by other factors based on artificial judgment.Aiming at solving the above problems,an online identification method of coal quality based on evidence theory is proposed.Firstly,the mechanism of the combustion process is analyzed,and the unit load/coal feed rate,heat/coal feed rate,and water supply flow/coal feed rate are constructed,which can reflect the current coal quality.Then the typical sample method is used to establish the confidence function.Finally,D-S fusion rules are used to fuse the three evidences to obtain the final diagnosis conclusion.Based on the actual operation data of power plant,it is verified the method is used to verify the results.The diagnosis results indicate that the method can distinguish the two combustion conditions of normal coal quality and bad coal quality.Compared with the single evidence diagnosis result,the proposed method’s support for the target model is increased,and the uncertainty is reduced.Therefore,this method improves the accuracy of recognition.