AI-Based Energy Consumptions Predictions in Smart Grids

Smart Grid is an electrical network technology that can regulate electricity supply and maintain its operations. Smart Grid can help Indonesia overcome the excess energy supply that occurred as a result of the creation of mega-electricity projects aimed at meeting Indonesia's electricity needs. However, due to COVID-19, electricity demand in Indonesia has decreased, resulting in an excess supply of electrical energy. Therefore, the aim of this research is to optimize existing electrical energy. To be able to eliminate wasted energy and eliminate losses that occur. The methods we use are artificial neural networks, convolutional neural networks, recurrent neural networks, and long short-term memory. The samples we got from the weather dataset from Indonesia in 2018–2020 from Badan Meteorologi, Klimatologi, and Geofisika (BMKG) Indonesia. We use this dataset to be able to consider the dependency between energy use and weather data. Comparison of (ANN), (CNN), (RNN), and (LSTM). The experimental results of the data set we obtained were that the CNN model outperformed any model. Through experiments and research that the method (CNN) reduces overfitting better than other methods. This approach can be proposed to help predict peak demand periods and automatically control smart equipment to reduce consumption and reduce energy consumption costs for users.