基于优化BP神经网络的快速起竖装置液压驱动系统故障诊断

快速起竖装置在自卸车等工程机械以及导弹发射车等军事领域均有较为广泛的应用。液压驱动系统是快速起竖装置的核心,准确诊断其故障具有重要意义。传统BP神经网络故障诊断准确率随着故障类型的增加急剧下降,难以满足工程需求。以快速起竖装置液压驱动系统为研究对象,提出一种基于BP神经网络和AdaBoost算法的故障诊断方法,将BP神经网络与逐步叠加建模算法结合,构造多分类BP-AdaBoost算法,建立故障诊断模型,搭建故障诊断实验台并设置8种典型工况。分析实验数据表明,该BP-AdaBoost算法与传统的BP神经网络方法相比具有更优的分类性能。

Fault Diagnosis of Hydraulic Drive System of Rapid-erection Device Based on Optimized BP Neural Network

The rapid-erection device is widely used in the engineering machinery and military field, such as dump truck and missile launching vehicle respectively. The hydraulic drive system is the core of the quick erecting device, so it is of great significance to diagnose its faults accurately. With the increase of the number of fault types, the accuracy of fault diagnosis of traditional BP neural network drops sharply, which is difficult to meet the engineering requirements. A fault diagnosis method based on BP neural network and AdaBoost algorithm is proposed for the hydraulic drive system of quick erecting device. Through the combination of BP neural network and step-by-step superposition modeling algorithm, a multi classification BP AdaBoost fault diagnosis model is established. Design the experimental system and set up 8 typical working conditions. The results show that the BP AdaBoost algorithm used in this paper has better classification performance than the traditional BP neural network method.