An Empirical Study of Meteorological Data-driven Electric Energy Consumption Prediction for Smart Campuses

As the global energy crisis and environmental pollution intensify, energy conservation and emission reduction have become essential objectives for the development of smart campuses at universities and colleges worldwide. Effective management of energy consumption in these environments relies on accurate power consumption prediction models. This paper investigates the prediction of power consumption for smart campuses through a meteorological data-driven model. By analyzing key meteorological factors, such as temperature, humidity, wind speed, and solar radiation, which influence electricity consumption on campus, we propose a prediction model using a genetic algorithm-backpropagation (GA-BP) neural network combined with sample entropy value and assignment methods. This model utilizes actual data on campus electricity consumption, augmented with multi-source meteorological data for training and validation purposes. Experimental results indicate that the model achieves high prediction accuracy across various meteorological data-driven conditions, thus enhancing the optimization of energy consumption management on campuses. The findings of this research offer theoretical support for energy-saving strategies in smart campuses and serve as a valuable reference for predicting energy consumption in similar public buildings.