A novel data management technique for renewable energy systems

Renewable energy systems like solar and wind farms require an extensive array of data regarding energy production, demand, weather, and systems performance. The complexity involved in modern data centers is a burden for the conventional data centres strategies. As a result of inefficiency, performance of the system will be degraded. In this paper we will explain an innovative data management system suitable for renewable systems such as solar, wind, and hydro. We at first use sensors in order to complete multiple types of data from renewable energy. With the aim of getting rid of duplicated data and equalizing performance, Z-score normalization is conducted on the data collected. Our method is fundamentally based on machine learning applications to conduct data-driven analysis and decision-making. We create a model for renewable energy generation prediction that observes changes in environment using a lion swarm optimized recurrent neural network with much higher rate (ELS-ORNN). Finally, the experimental results conducted under MAE (1.25 J/M²), MSE (5.60 J/M²), RMSE (1.40 J/M²), and R-squared (95.15%) are evaluated. As the experiments have proven, the ESL-ORNN method is superior to the rest of techniques. With respect to renewable energy prediction, the suggested ELS-ORNN method demonstrates improvement in performance.