当前位置：首页 > 过刊浏览->2023年45卷9期
深度神经网络在船用齿轮箱故障诊断中的应用
Application of deep neural networks in fault diagnosis of marine gearboxes
- DOI:
- 作者:
- 顾钦平
GU Qin-ping
- 作者单位:
- 江苏航运职业技术学院, 江苏 南通 226010
Jiangsu Shipping College, Nantong 226010, China
- 关键词:
- 深度神经网络;船用齿轮;故障诊断
deep neural network; marine gears; fault diagnosis
- 摘要:
- 船用齿轮箱故障诊断系统，通常采用的是基于模型的故障诊断方法，需要依赖专家对采集到的振动信号进行分析和判断，且设备的运行状态会随着时间的推移发生变化。传统的故障诊断方法受专家知识和经验的影响较大，难以获得全局最优解，导致其准确率较低。针对该问题，本文基于深度神经网络（DNN）的故障诊断方法，通过大量实验研究，确定了DNN模型中最佳参数和超参数。实验结果表明，在船用齿轮箱故障诊断领域，DNN模型不仅能够有效地对齿轮箱进行故障诊断，而且具有较高的准确率和较快的收敛速度。
The fault diagnosis system for marine gearbox usually adopts a model-based fault diagnosis method, which relies on experts to analyze and judge the collected vibration signals, and the operating status of the equipment will change over time. Traditional fault diagnosis methods are greatly influenced by expert knowledge and experience, making it difficult to obtain global optimal solutions, resulting in low accuracy. To solve this problem, based on the fault diagnosis method of deep neural network (DNN), this paper determines the best parameters and hyperparameter in the DNN model through a large number of experimental studies. The experimental results show that in the field of marine gearbox fault diagnosis, the DNN model can not only effectively diagnose gearbox faults, but also has high accuracy and fast convergence speed.
2023,45(9): 176-179 收稿日期：2022-11-22
DOI：10.3404/j.issn.1672-7649.2023.09.039
分类号：U667.65
作者简介：顾钦平(1984-),男,硕士,工程师,主要从事计算机网络安全及应用研究
