健美操腰部训练器故障自动化识别系统设计

Aerobics trainers waist fault automatic identification system design

常用系统在提取故障识别特征参数时,仅提取了机械信号时域特征,而忽略了频域特征,导致故障识别准确率较低。针对这一问题,设计一种健美操腰部训练器故障自动化识别系统。硬件方面,采用双核结构处理器,嵌入惯性速度传感器,利用A/D转换芯片,采集训练器振动信号,优化信号预处理健相电路;软件方面,提取振动信号的量纲时域特征、无量纲时域特征、频域特征,获得信号时频图谱,输入卷积神经网络,识别训练器故障模式。采集训练器10种不同运行状态信号,作为测试样本,设置对比实验,结果表明,设计系统提高了故障位置、故障类型、故障程度的识别准确率,自动化识别结果更加可靠。

When extracting the characteristic parameters of fault identification, common systems only extract the time domain characteristics of mechanical signals, while ignoring the frequency domain characteristics, which leads to low accuracy of fault identification. To solve this problem, an automatic fault identification system for aerobics waist trainer is designed. In terms of hardware, dual-core structure processor is adopted, inertial speed sensor is embedded, A/D conversion chip is used to collect vibration signal of trainer, and the signal preprocessing phase-strengthening circuit is optimized. On the software side, the dimensional time domain features, dimensionless time domain features and frequency domain features of vibration signals are extracted, and the time-frequency maps of the signals are obtained, which are input into the convolutional neural network to identify the failure modes of the trainers. Ten different running state signals of the trainer are collected as test samples, and comparative experiments are set up. The results show that the designed system improves the recognition accuracy of fault location, fault type and fault degree, and the automatic recognition results are more reliable.