基于数据融合的回热系统故障诊断方法研究

作者在火电机组回热系统常见故障的故障集、征兆集及故障树理论基础之上,利用模糊数学知识和相关理论,针对回热系统故障征兆参数的变化特征,采用不同的变化等级和阈值,建立了回热系统典型故障样本模式知识库及实时故障模式集。同时利用模糊推理和RBF神经网络,分别建立了回热系统故障诊断模型,在诊断结果层面上用数据融合算法进行适当的融合,将模糊理论与人工神经网络有机地结合起来,进行了优势互补。通过仿真典型故障数据和部分实测数据的验证表明,该方法提高了诊断的可靠性。

Heat recovery based on data fusion fault diagnosis method

In thermal power regenerative system common fault fault set,set of symptoms and fault tree based on the theory,fuzzy mathematics andthe theory of knowledge,for signs of heat recovery system failure parameters variations,changes in levels and with different threshold,the establishment of a heat recovery system samples the typical failure mode of the knowledge base and real-time fault pattern set.At the same time the use offuzzy inference and RBF neural network,heat recovery systems were set up fault diagnosi...