基于深度学习的输电线路危险源智能监控系统

为了掌握输电线路的实时状态,提出一种基于深度学习的输电线路危险源智能监控系统.该监控系统主要包括:信号采集端、内部光纤网络、服务器控制中心、显示终端,其中服务器控制中心的输电线路危险源智能辨识算法是整个系统的核心.该算法采用稀疏自编码从图像/视频信号中学习特征,完成深度神经网络的训练;然后用卷积和池化对特征进行降维;最后采用softmax回归来解决危险源的多分类问题.经实例验证,基于深度学习的危险源辨识算法,可以准确对危险源进行判别.监控系统将判别结果反馈到显示终端,可以全面掌握整个线路运行情况,确保电力系统的安全运行.

Intelligent Monitoring System for Hazards of Transmission Line Based on Deep Learning

As the important support of the national economy, the safe operation of the power grid provides an endless motivation for the economic development. At the same time, the safe operation of the transmission lines is facing increasing challenges. To master the real-time state of transmission line, an intelligent monitoring system for the recognition of the hazards of transmission line, which is based on the theory of depth learning, is proposed. The monitoring system mainly includes: signal-acquisition equipment, the optical network, server control center, as well as the display terminal. The transmission line hazards intelligent identification algorithm which is in the server and control center, is the core of the system. First, the recognition algorithm uses characteristics by the sparse self-coding to train the neural network. Second, it employs the operator of convolution and pooling for feature dimension reduction. Finally, the softmax regression is utilized to solve the problem of hazard classification. The experiments confirm that the hazards identification algorithm based on deep learning, hazards can be accurate to discriminate. Then, the results will be transferred to computer terminal to fully control the whole operation, and all kinds of emergencies can be handled in time to ensure the safe operation of the power system.