基于D-S证据理论的燃烧稳定性融合诊断

针对电厂锅炉燃烧火焰监控信号判断问题，结合现有的FSSS系统燃烧稳定性的判断逻辑，提出了一种基于D-S(Dempster-Shafer)证据理论的燃烧稳定性判别方法。考虑不同工况下火焰的燃烧情况，以单层燃烧器的四个火检信号结合风煤比的大小以及相邻层火检信号来建立典型样本。以典型样本为参照依据，计算各证据在着火模式和灭火模式下的信度密度值，然后对其进行归一化处理得到信度函数分配。选取某时间段内的一层燃烧器的4个火检信号、风煤比、相邻层火焰作为证据，结合典型样本库计算出证据在各目标模式下的信度函数分配及不确定度，利用D-S联合规则进行数据融合，对燃烧稳定做出判断。经实验分析证明该种方法的有效性，并且提高了判断结果的准确性。在信度函数构造的过程中减少了主观性，便于工程应用。

Fusion Diagnosis of the Combustion Stability Based on D-S Evidence Theory

Combined with the judgment logic of combustion stability of existing FSSS system, a combustion stability discrimination method based on D-S (Dempster-Shafer) evidential theory was proposed for boiler combustion instability in power plant. This paper discussed the combustion conditions under different conditions of the flame, used the four flame monitoring signals of the single-layer burner combined with the size of the coal-to-air ratio and the fire signals of adjacent layers to establish a typical sample. It took typical sample as comparable data and computed the confidence density of ignition goal pattern and fire extinguishing goal pattern, then carried on normalized processing to the typical sample to obtain the confidence function distribution. In this paper, in a period of time within a layer of burner 4 fire monitoring signals, air-coal ratios and the adjacent layer of flames were selected as evidence, and combined with the typical sample database, the reliability function distribution and uncertainty of the evidence in each goal pattern were calculated. D-S joint rules were used to fuse the data to judge the combustion stability. The experimental result shows that the method is effective, and can not only improve the accuracy of the judgment result, but also reduce the subjectivity in the process of construction of reliability function, which is convenient for engineering application.