基于LSTM算法的门座式起重机减速箱故障诊断研究

为实现门座式起重机减速箱机械故障的智能诊断和分类,运用长短期记忆网络构建了门座式起重机减速箱机械故障的自动诊断分类模型;首先设计并使用了基于labview的数据采集系统对门座式起重机的复合故障数据进行了采集,结合东南大学公开的齿轮箱故障数据建立了数据集;然后用数据增强的方法对数据进行预处理,接着采用长短期记忆神经网络,构建门座式起重机减速箱机械故障诊断模型;最后使用测试数据集对模型的诊断分类准确性进行了验证实验,结果表明该诊断模型能快速准确的对门座式起重机减速箱的机械故障进行自动诊断和分类,实现了96.8％的诊断分类准确率,与传统的基于CNN的诊断分类模型相比,准确率提高了4.1％,为下一步便携式智能诊断仪器的开发和应用奠定了一定的理论基础.

Research On Fault Diagnosis of Gearbox of Portal Crane Based on LSTM Algorithm

In order to realize the intelligent diagnosis and classification of mechanical failure of portal crane gearbox, the automatic diagnosis and classification model of mechanical failure of portal crane gearbox is constructed by using long-term and short-term memory network. Firstly, a data acquisition system based on labview is designed and used to collect the composite fault data of portal crane, and a data set is established based on the gearbox fault data published by Southeast University. Then the data is preprocessed by data enhancement method, and then the mechanical fault diagnosis model of portal crane gearbox is constructed by using long-term and short-term memory neural network. Finally, the diagnostic classification accuracy of the model is verified by the test data set. The results show that the diagnostic model can automatically diagnose and classify the mechanical faults of the portal crane gearbox quickly and accurately, and the diagnostic classification accuracy is 96.8%. Compared with the traditional diagnostic classification model based on CNN, the accuracy is improved by 4.1%, which lays a theoretical foundation for the development and application of portable intelligent diagnostic instruments in the next step.