Design of the control and risk automatic warning platform for power grid operation based on AI+video technology

In order to accurately identify the risk behavior in the power grid operation, ensure the life safety of operators and the development of power enterprises, the power grid operation behavior control and automatic risk early warning platform based on AI+video technology is designed. The platform is composed of infrastructure layer, data resource layer, application support layer, platform management, and AI+video monitoring layer. The infrastructure layer is mainly responsible for collecting video of grid operation behavior, and considering the actual situation of grid operation, and accurately extracting the prospect target in the video image (i. e. monitoring target) under the jitter condition of the video camera. The data resource layer and the application support layer are responsible for selecting different power grid operation risk behavior judgment algorithms under different application scenarios; the platform management and AI+video monitoring layer are responsible for showing the monitoring and early warning results of the operation behavior of the power grid to users. The experimental results show that, under the condition of monitoring camera jitter, the prospect of grid behavior control and automatic risk warning platform based on AI+video technology is closest to the benchmark map, and the risk behavior identification results are consistent with the relevant standards of power grid operation in China. The area under the ROC curve has remained above 0.6, which can effectively identify the risk behavior in power grid operation, reduce the risk behavior of network operation and improve the safety of grid operation.