微型电脑应用

目前，常规的激光除锈在线监测系统主要通过对电气设备实施除锈操作前的锈蚀层厚度与除锈后的剩余锈蚀层厚度进行测量来实现监测分析，由于缺乏对激光除锈信号的特征提取，导致监测精度较差。对此，提出基于卷积神经网络(CNN)的激光除锈在线监测系统。在硬件方面，对监测系统的数字电路以及模拟电路分别进行设计，并采用传感器对激光除锈表面信息进行收集。在软件方面，构建CNN结构，采用卷积操作对传感器采集到的信号进行信号分解与特征提取，通过分析传感器信号峰值对应的时间，计算电气设备表面剩余壁厚，并将上一时刻计算的剩余壁厚作为参考值，实现对激光除锈操作的在线监测。实验结果表明，采用所提出的系统对激光除锈操作进行监测时，除锈深度绝对误差值较低，具备较为理想的监测精度。

At present, the conventional laser derusting online monitoring system mainly measures the thickness of the rust layer before the derusting operation of the electrical equipment and the remaining thickness of the rust layer after derusting to achieve monitoring and analysis, which leads to poor monitoring accuracy due to the lack of feature extraction of laser derusting signals. In this regard, a laser derusting online monitoring system based on convolutional neural network(CNN) is proposed. In terms of hardware, the digital circuit and analog circuit of the monitoring system are designed, respectively, and sensors are used to collect the laser derusting surface information. In terms of software, the CNN structure is constructed, the signal decomposition and feature extraction of the signals collected by the sensors are performed by convolutional operation. By analyzing the time corresponding to the peak value of sensor signals, the remaining wall thickness on the surface of electrical equipment is calculated, and the corrosion wall thickness calculated at the previous moment is used as a reference value, to achieve online monitoring of laser derusting operation. The experimental results show that when the proposed system is used to monitor the laser derusting operation, the absolute error value of the derusting depth is low, and it has relatively ideal monitoring accuracy.