| 基于数据驱动的设备电路板无图纸故障诊断 |
| |
| 引用本文： | 刘香君,种银保,肖晶晶,赵鹏,张诗慧.基于数据驱动的设备电路板无图纸故障诊断[J].中国医学物理学杂志,2020,37(8):1047-1052. |
| |
| 作者姓名： | 刘香君 种银保 肖晶晶 赵鹏 张诗慧 |
| |
| 作者单位： | 陆军军医大学第二附属医院医学工程科， 重庆 400037 |
| |
| 摘 要： | 【摘要】针对医疗设备电路板结构设计复杂，传统故障诊断方法过度依赖图纸等技术资料和维修专家个人技术经验，导致维修贵、维修难等问题，提出一种基于数据驱动的无图纸故障智能诊断方法。在未知电路图纸信息以及电路板工作原理的前提下，模拟电路板不同故障状态，采集各外部接口引脚电信号作为原始故障数据；对故障数据进行预处理，并划分为训练集及测试集；使用机器学习的方法构建基于单层长短时记忆网络的故障智能诊断系统，利用Python编程进行模型训练，系统输出训练过程准确率及损失曲线。结果表明，该方法能实现对电路板故障的诊断分类，准确率达89.99%，效率较高，可靠性强。
|
| 关 键 词： | 医疗设备 电路板 故障诊断 长短时记忆网络 |
Fault diagnosis for equipment circuit board based on data drive and no circuit drawing |
| |
| LIU Xiangjun,CHONG Yinbao,XIAO Jingjing,ZHAO Peng,ZHANG Shihui.Fault diagnosis for equipment circuit board based on data drive and no circuit drawing[J].Chinese Journal of Medical Physics,2020,37(8):1047-1052. |
| |
| Authors: | LIU Xiangjun CHONG Yinbao XIAO Jingjing ZHAO Peng ZHANG Shihui |
| |
| Affiliation: | Department of Medical Engineering, the Second Affiliated Hospital of Army Medical University, Chongqing 400037, China |
| |
| Abstract: | Abstract: Considering the complicated design of medical equipment and the problems of traditional fault diagnosis methods such as the high cost and difficulties of maintenance due to the over-reliance on technical data such as circuit drawings, as well as the personal technical experience of maintenance specialists, an intelligent fault diagnosis method based on data drive and no circuit drawing is proposed. Under the premise of unknown circuit drawing information and the working principle of the circuit board, different fault states of the circuit board were simulated, and the electrical signals of each external interface pin were collected as the original fault data which were then preprocessed and divided into training set and test set. Machine learning method was used to construct an intelligent fault diagnosis model based on a single-layer long short-term memory network. The model training was carried out by Python programming, and the accuracy curve and loss curve of training progress were output. The result showed that the proposed method realized the diagnosis and classification of circuit board faults with high efficiency and strong reliability, and the accuracy reached 89.99%. |
| |
| Keywords: | Keywords: medical equipment circuit board fault diagnosis long short-term memory |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《中国医学物理学杂志》浏览原始摘要信息 |
|
点击此处可从《中国医学物理学杂志》下载全文 |
