Electricity Consumption Prediction Method Based on Bi-LSTM and Edge Computing

With the development of the smart grid, electricity consumption forecasting has become increasingly important in power systems. Accurate forecasting not only helps power companies optimize their operation strategies but also improves resource efficiency and reduces energy waste. However, existing electricity consumption prediction systems face challenges in handling complex consumption behavior, large-scale data, and real-time responses. To address these challenges, this paper proposes an electricity consumption prediction system based on Bi-directional Long Short-Term Memory (Bi-LSTM). The system considers both forward and backward time-series data to capture the complex, time-dependent, and nonlinear features in electricity consumption data, thereby improving prediction accuracy. Additionally, edge computing devices are deployed at the user end to collect and process electricity consumption data in real time, enabling rapid generation of prediction results. This distributed computing model reduces data transmission requirements and network latency, allowing for timely adjustments in forecasts and grid strategies when electricity demand fluctuates. The local processing capability of edge devices effectively handles high-frequency and large-volume data streams, preventing the waste of computational resources and improving system energy efficiency. These innovations make the prediction system more efficient and flexible in power management.