Hyperparameter Optimization of LSTM MPPT Controller for Solar Power Plants

Abstract

Efficient tracking of the maximum power point (MPP) is crucial for optimizing energy extraction in solar power plants. This paper focuses on hyperparameter optimization of Long Short-Term Memory (LSTM) neural networks for MPP tracking control. Traditional and emerging MPP tracking techniques are reviewed, and the potential of LSTM-based controllers is explored. The study rigorously investigates the impact of hyperparameters such as learning rate, layers, hidden units, and dropout rate on the controller's efficiency. By navigating the hyperparameter space, the research aims to uncover optimal configurations that enhance MPP tracking precision and performance in solar power plants.