Transforming the grid: AI, ML, renewable, storage, EVs, and prosumers

It is estimated that fifty percent of the world’s energy consumption will be in the form of electrical energy by 2050. The major shift of combustible vehicles to green electrical engines powered through batteries and fuel cells will be seen in the coming future, which will be used as sources when idle. These, and the conventional renewable energy sources like solar-PV, wind energy, and energy storage systems like batteries, will contribute as distributed energy resources (DERs). Application of data-driven technologies will be required to make the modern grid resilient and secure by self-corrective measures toward the source and load intermittent behavior. Machine learning (ML) and optimization techniques are extensively used in the forecasting and operation of these systems. In this chapter, the recent trends followed in applying the optimization techniques based on natural evolution and swarm behavior, like genetic algorithms, differential evolution, particle swarm optimization, ant colony optimization, etc., in the modern grid influenced by various DERs, will be presented. In a grid operating with multiple renewable sources and loads, these methods are fruitful in deciding the optimal source selection for economic load dispatch. These techniques are also being employed on the component level, for example, in the control of power electronics control, which helps in performance enhancement by selecting the states that will produce the least harmonics in the system. Further, the ANN can be trained to perform according to the datasets obtained from the optimization techniques. This chapter will show a case study where a metaheuristic technique and artificial neural network integrate the prosumer-based solar-PV integration through a multilevel inverter, enabling constant fundamental voltage with minimum total harmonic distortion.