基于改进深度信念网络的燃气轮机故障诊断

针对传统智能诊断方法使用的浅层模型难以准确挖掘信号特征量与对应故障原因之间复杂的映射关系,导致故障诊断精度不高的问题,提出了一种基于深度信念网络(DBN)的燃气轮机故障诊断方法。采用Apriori关联度分析法分析燃气轮机故障模式与故障特征向量之间的关系,生成关联度矩阵;根据关联结果筛选出满足精度的特征向量作为输入,同时结合遗传算法(GA)对深度信念网络的结构参数进行优化,建立了基于GA-DBN的燃气轮机故障诊断模型。最后通过诊断实例验证了所提方法的有效性。;The shallow model used in the traditional intelligent diagnosis method is difficult to accurately mine the complex mapping relationship between the signal feature quantity and the corresponding fault cause,resulting in low fault diagnosis accuracy.This paper proposes a method based on deep belief network (DBN) gas turbine fault diagnosis.The Apriori correlation analysis method is used to analyze the relationship between the gas turbine failure mode and the fault feature vector,and the correlation degree matrix is generated.The feature vector satisfying the accuracy is selected as the input according to the correlation result,and the structure of the deep belief network is combined with the genetic algorithm (GA).The parameters are optimized to establish a GA-DBN based gas turbine fault diagnosis model.A diagnostic example verifies the effectiveness of the method