Predicting solar power potential via an enhanced ANN through the evolution of cub to predator (ECP) optimization technique

Forecasting plays a vital role in solar power generation and skillfully managing renewable energy resources. The traditional artificial neural network (ANN) has certain limitations with its pre-defined network structure in predicting solar power potential and not suitable because traditional ANN mus...

Full description

Saved in:
Bibliographic Details
Published in:Electrical engineering 2024, Vol.106 (5), p.6069-6080
Main Authors: Nasab, Morteza Azimi, Zand, Mohammad, Miri, Mohsen, Sanjeevikumar, P., Guerrero, Josep M., Saket, R. K., Khan, Bassem
Format: Article
Language:English
Subjects:
Artificial neural networks
Configuration management
Economics and Management
Electrical Engineering
Electrical Machines and Networks
Energy Policy
Energy sources
Engineering
Optimization
Performance evaluation
Power Electronics
Predators
Solar power generation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Forecasting plays a vital role in solar power generation and skillfully managing renewable energy resources. The traditional artificial neural network (ANN) has certain limitations with its pre-defined network structure in predicting solar power potential and not suitable because traditional ANN must be configured. Configuring the ANN model is time-consuming through manual process or trial and error. Therefore, this work evaluates the combination of different optimization techniques for ANN model configuration. In other techniques, the forecast performance is between 84.9 and 91.6% on average. Experience in handling the traditional methods helps develop a novel evolution of cub to predator (ECP) optimization technique. The investigation revealed that incorporating optimization results in superior performance over the traditional ANN, whereas the proposed ECP unveils 97.2%, which is an excellent result of contest optimization techniques. At the same time, the proposed model with the lowest MSE = 0.0029, and MRE = 0.0809.
ISSN:0948-7921
1432-0487
DOI:10.1007/s00202-024-02302-1