基于环境声识别的工业设备状态监测实验系统设计

Experimental System Design for Industrial Equipment Status Monitoring Based on Environmental Sound Recognition

为了实现低成本的工业设备故障检测,设计了一套基于环境声识别的设备状态监测实验系统。基于Arduino开源硬件开发低成本的智能传感器,并部署限制感受野的深度学习模型RFL-MobileNet。通过在边缘端进行智能数据处理,系统为工厂设备提供了实时且准确的状态监测,以减少传统方法中数据传输到云端所带来的延迟以及数据泄露风险。基于公开数据集和实际数据集的实验结果表明,系统运行良好,电动机故障检测准确率达95.4%。

A test system for industrial equipment status monitoring based on environmental sound recognition is designed to achieve low-cost equipment fault detection.The RFL-MobileNet model is developed,and deployed in Arduino open-source hardware to implement low-cost intelligent sensors.By performing intelligent data processing at the edge,the system provides real-time and accurate status monitoring for factory equipment,reduces the latency and data leakage risks brought by data transmission to the cloud in traditional methods.Experimental results based on public datasets and actual datasets demonstrate that the system’s functionality can work well,with the detection accuracy of the operating status of direct-current motors reaching 95.4%.