Convolutional Neural Network Based Power Information Network Security Situational Awareness Model

At present, most of the external attacks against the power network are through the intrusion of the network layer to tamper with control instructions and sensor data, resulting in abnormal operation of the power system, thus causing serious damage to the power grid process. A variety of defensive measures for security have solved the security problems in the power grid in a variety of ways, but they still can not improve the comprehensive analysis of data in situational awareness. This paper focuses on the collected error or abnormal data, which will lead to abnormal large or small range, and can not normally extract the characteristics of the data. A power load forecasting algorithm based on attention mechanism is proposed by combining the convolutional neural network and GRU. Embedding and convolutional neural networks are applied to prioritize the description of vulnerabilities, which effectively improves forecasting accuracy. The experimental results verify that the model in this paper has a good practical application effect in the power network situation prediction. It can predict the possible attack range and the attack purpose, and establish an effective early warning means to protect the network from attack.