A Hydraulic System Fault Diagnosis Method Based on FTA and FNN
-
摘要: 针对液压系统故障的复杂性和不确定性等特点,传统的故障推理方法难以满足液压系统故障诊断的要求,提出了基于故障树分析和专家经验知识的模糊神经网络故障诊断方法。以起重设备液压系统为研究对象,建立故障树模型,基于故障树信息和专家经验知识,建立模糊神经网络诊断模型及并提取训练数据,在此基础上,运用统计参数法确定模糊预处理所需的模糊隶属函数。将训练好的网络模型应用于实例诊断,实验结果验证了该方法的实用性和有效性。Abstract: The hydraulic system fault diagnosis method based on fuzzy neural networks( FNN),fault tree analysis( FTA) and expertise knowledge is proposed to overcome the shortcomings of the traditional fault diagnosis methods because of the complexity coupling and uncertainty of hydraulic system fault. Taking lifting equipment's hydraulic system as study object,a fault tree model is set up,and the fault tree information and expertise knowledge are employed to establish the FNN structure and extract training data. The fuzzy membership functions,which are needed in fuzzy pre-processing,are confirmed by using the statistical parameter method. Finally,the trained FNN model is applied to examples; the application results show that the hydraulic system fault diagnosis method based on FTA and FNN is effective and practicable.
-
Key words:
- efficiency /
- failure analysis /
- fault tolerance /
- fuzzy neural networks ( FNN)
-
[1] 周汝胜,焦宗夏,王少萍. 液压系统故障诊断技术研究现状与发展趋势[J].机械工程学报,2006,(09):6-14. [2] Ma C,Wang K S,Sun J H. Fault diagnosis to mid-size excavator hydraulic system based on fault tree analysis[A].2010. [3] 向长城. 基于可拓学的智能故障诊断与状态监测的理论及应用研究[D].重庆:重庆大学,2008. [4] He X Y,Shao Y M,He S H. Fault diagnosis of hydraulic system using dynamic GRNN[J].ICIC Express Letters,2011,(8A):2617-2623. [5] 陈炜,吴志良. 故障诊断专家系统结合故障树技术的结构研究[J].航海技术,2005,(06):43-44. [6] Pisu P. Hierarchical model-based fault diagnosis with application to vehicle systems[A].Barcelora,spain,2002. [7] 张敬芬,孟光,赵德有. 基于模糊神经网络的薄板不同指标裂纹诊断[J].机械工程学报,2006,(03):145-149. [8] Yznaga B I,Sun D,He Y K. Neural-fuzzy technique for inverter faults diagnosis in PMSM DTC system[J].Electric Machines and Control,2008,(03):132-139. [9] Pltkeisheim F. The three sigma nule[J].The American Statistician,1994,(02):88-91.
点击查看大图
计量
- 文章访问数: 195
- HTML全文浏览量: 21
- PDF下载量: 6
- 被引次数: 0