Flink-based Streaming Data Processing for Power Wireless Private Networks

Le résumé fourni par la source

In order to break through the processing bottleneck of massive real-time streaming data in the electric power wireless private network, an innovative scheme integrating Flink distributed stream processing architecture and weathercoupled LSTM prediction is proposed. By constructing a threelayer stream processing framework, combining Kafka message middleware and flow arithmetic mechanism to realize dynamic data scheduling and anomaly cleaning; designing LSTM model based on sliding window, incorporating temperature, humidity and other meteorological factors to predict power load. Simulation shows that the scheme can effectively reduce the delay of billion data processing, improve the quality of data sets, and stabilize the error of power grid load prediction within $\mathbf{3. 8 \%}$. The research results significantly improve the real-time data processing efficiency and power load prediction accuracy of electric power wireless private network, provide highly reliable technical support for the dynamic scheduling decision-making of smart grid, and effectively guarantee the safe and stable operation of power grid under complex working conditions.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.