Electricity-consuming forecasting by using a self-tuned ANN-based adaptable predictor

Abstract

Accurate forecasting of power consumption is an essential and modern approach for planning smart infrastructural projects that are required to overcome the future challenges of power markets. In this context, a new design of a self-tuned ANN-based adaptable predictor is presented in this paper. At first, the main design of the original adaptable predictor is clarified including its new architecture that partially relies on the Hebbian law, its training process is also explained as well as the dataset which is used for training. Then, all the details of the self-tuning-based technique are explained with all the relevant results that prove its high capability to produce more accurate forecasting outcomes. The impact of our suggested approach is introduced by explaining two different practical examples that employ the K-means clustering algorithm, and the genetic algorithm when it is used to optimize the operating/outage schedule for a group of local solar units, respectively.