Research on automatic generation and transmission of power emergency information based on the mechanism of Transformer and self-attention

With the rapid development of information technology, the efficiency and accuracy of emergency response mechanisms are of great significance for reducing disaster losses and ensuring the safety of people's lives and property. This article aims to explore the automatic generation and delivery of emergency information in power grids and power systems, and study how to realize the automatic generation and delivery of emergency information through the mining and analysis of emergency indicators and data, combined with business rules. On this basis, this paper proposes an emergency information generation model based on Transformer and joint attention mechanism and builds an emergency information reporting platform. Compared with previous work, this work has two innovations. One is to use the Pearson correlation coefficient to screen emergency indicators to reduce the amount of data required for information generation. The second is to introduce external knowledge and combine it with a joint attention mechanism to screen the knowledge, and at the same time add emergency degree recognition to generate power emergency information that is accurate and can intuitively display the degree of emergency. Through experimental verification, it can be seen that the generation method proposed in this article can accurately and concisely generate emergency information, and the reporting platform has good emergency information classification capabilities. This study can improve the efficiency of emergency information processing, ensure the timeliness and accuracy of information, and provide strong support for natural disaster emergency management decisions in power systems.