Planet Optimization with Machine Learning Enabled Power Usage Forecasting Modeling in Smart Grid Environment

The smart grid (SG) significantly enhances the conventional power grids with information and communication technologies, control decision-making systems, simulation analysis, advanced measurement and sensing. In contrast to conventional power grids, the SG has further benefits in stability, situational awareness, self-healing, information interaction, and renewable energy consumption. But the precise prediction of long-term electric energy consumption remains a challenge. Recently, machine learning (ML), particularly deep learning, has rapidly advanced and has demonstrated outstanding performance in various tasks of the SG fields. The representation capability of ML approaches is considerably enhanced. The SG is a curial framework area, hence ML algorithms including it should be interpretable for improving reliability of the system and increasing user trust. This study develops a new Planet Optimization with Machine Learning Enabled Power Usage Forecasting Modeling (POML-PUFM) in SG environment. The presented POML-PUFM technique forecast the utilization of power in the SGs for smart city applications. In the presented POML-PUFM technique, data pre-processing takes place to transform them into compatible format. For forecasting purposes, the POML-PUFM technique employs twin-support vector machine (TWSVM) model. In addition, the study used PO model to adjust the parameter related to the TWSVM method. The prediction performance of the presented POML-PUFM algorithm is examined and the outcomes are examined under several measures. The study verified the improvements of the POML-PUFM technique over other ML models.