Data-Driven Analysis of Renewable Energy Generation Using Meteorological Indicators

Abstract

Renewable energy generation is affected by various reasons related to the weather, and the specific reasons and analysis methods for solar, wind and hydropower systems are different. This paper studies the influence of irradiance, temperature, cloud cover, humidity, wind speed, air density, precipitation, runoff and seasonal changes in hydrology on renewable electricity generation. First, it introduces the key generation and weather indicators; then, it presents data preprocessing, descriptive statistics, correlation analysis, regression, time-series modelling and machine-learning methods. Representative studies are compared according to data sources, time resolution, areas covered, model design and evaluation indicators. Based on the above evidence, solar irradiance is a main meteorological driver of photovoltaic power generation; wind speed affects turbine power nonlinearly; and hydropower generation is also subject to hydroclimatic conditions and operational decisions. The main deficiencies are missing or misaligned observations, non-meteorological curtailment, regional heterogeneity, inconsistent temporal scales, limited model interpretability, and weak cross-site generalisation. Future work will combine multi-source weather and equipment data, physical constraints, external validation and uncertainty quantification. Data-driven analysis can be employed to investigate the reasons for changes in renewable energy generation under different weather conditions and other technology or operating factors.