论文部分内容阅读
基于油液光谱分析界限值在油液检测中的重要性,探讨运用遗传算法增强BP神经网络界限值动态调整模型的泛化能力,建立了基于BP神经网络的动态界限值模型。结果表明:该模型不仅能完成基于光谱分析界限值的监测工作,而且能进行某型飞机发动机的故障诊断,并根据油液光谱监测数据实现故障定位。运用这种方法,可以将光谱分析界限值的学习固化于神经网络的连接权中,形成初步的诊断报告,与监测数据一起上交飞机发动机质量监测小组,形成更加科学的某型飞机发动机的监测机制。
Based on the importance of oil spectrum analysis in oil detection, this paper discusses the application of genetic algorithm to enhance the generalization ability of the dynamic adjustment model of BP neural network threshold, and establishes a dynamic threshold model based on BP neural network. The results show that this model can not only perform the monitoring work based on the spectral analysis threshold but also can diagnose the fault of a certain type of aircraft engine and locate the fault according to the oil spectrum monitoring data. Using this method, the learning of spectral analysis thresholds can be consolidated into neural network connection rights to form preliminary diagnostic reports, which can be submitted with the monitoring data to the aircraft engine quality monitoring team to form a more scientific monitoring of certain aircraft engines mechanism.