论文部分内容阅读
为研究工程机械液压系统故障的诊断,提出一种墨西哥草帽函数改进的ART神经网络学习算法,增强输入矢量距敏感特征近的区域对模式分类的影响。采用距离分区技术调整输入模式的特征矢量,提高诊断的精度和效率。用Amesim_HCD库建立液压系统液压泵、多位换向阀结构模型,用Amesim建立起重机液压系统模型。通过对液压系统各类型故障仿真,验证了改进ART方法对起重机液压系统故障诊断的可靠性。开发了基于改进ART神经网络的起重机液压系统故障诊断专家系统。
In order to study the fault diagnosis of hydraulic system of construction machinery, an improved ART neural network learning algorithm based on Mexican straw hat function was proposed to enhance the effect of the near-region sensitive pattern of input vector on pattern classification. The use of distance partitioning technology to adjust the input mode feature vector to improve the accuracy and efficiency of diagnosis. With Amesim_HCD library to build hydraulic system hydraulic pump, multi-directional valve structure model, with Amesim to establish a hydraulic system model of the crane. Through the simulation of various types of hydraulic system faults, the reliability of the improved ART method for fault diagnosis of crane hydraulic system is verified. A fault diagnosis expert system for crane hydraulic system based on improved ART neural network was developed.