为进一步增强对新能源发电站的安全巡检能力,保障新能源发电站的安全稳定运转,提出了一种基于数字孪生的电力设备实时巡检监控方法.该方法在数字孪生技术的基础上对电力设备安全巡检系统进行了建模和仿真,使用5G网络切片技术确保电力设备数据在巡检系统中的传输速度.同时,还设计了一种轻量级残差密集卷积神经网络,并将其与能谱图结合,实现了对电力设备故障的快速、精准识别.实验结果表明,所提出的实时巡检监控技术方案具有良好的综合性能,识别准确率可达98.32%,相比于仿真性能最优的SENet仍高出5.11%,从采集数据到回传识别结果所需的时间仅为2.1～3.3 s.
Research on real-time inspection monitoring method of power equipment based on digital twin
To further enhance the safety inspection capability of new energy power stations and ensure their safe and stable operation,a real-time inspection and monitoring method for power equipment based on digital twins is proposed.This method models and simulates the power equipment safety inspection system based on digital twin technology,and uses 5G network slicing technology to ensure the transmission speed of power equipment data in the inspection system.At the same time,a lightweight residual dense convolutional neural network was designed and combined with energy spectrum to achieve rapid and accurate identification of power equipment faults.The experimental results show that the proposed real-time inspection and monitoring technology scheme has good comprehensive performance,with a recognition accuracy of 98.32%,which is still 5.11%higher than the SENet with the best simulation performance.The time required from collecting data to returning recognition results is only 2.1～3.3 seconds.
digital twinpower equipment inspection5G network slicedeep learning
万方数据
维普
