Research on a dangerous point analysis model based on access profile of power grid dispatching automation

In the simple rule analysis mode of power grid dispatching, only a relatively rough judgment can be made on the dangerous points, which greatly reduces the accuracy of dangerous point identification. Therefore, based on the access portrait of power grid dispatching automation, a new dangerous point analysis model is proposed. Combining the actual network topology and application deployment topology, the static application topology of the power grid dispatching automation system is constructed by refining the device connection relationship layer by layer and generating a layout array. By capturing logs and parsing traffic data, business access information between the client and the server, as well as between the server processes, can be identified. The dynamic business access information is layered and superimposed on the static application topology according to the time dimension, and the mapping relationship is established by combining the device naming specification with the IP address management database. Based on the business access portrait model, the characteristics are analyzed and the key nodes are identified to complete the construction of the dangerous point analysis model. The experimental results show that the research method is optimized in terms of dangerous point recognition rate, loss value and confusion matrix performance, which significantly improves the accuracy of dangerous point identification in power system operation.