Title: Smart grid stability prediction based on Bayesian-optimised LGBM for smarter energy management

Authors: V. Ashok Gajapathi Raju; Manohar Mishra; Jai Govind Singh; Janmenjoy Nayak; Pandit Byomakesha Dash

Addresses: Department of Computer Science and Engineering, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar, India; Department of Information Technology, Aditya Institute of Technology and Management (AITAM), Tekkali, 532201, Andhra Pradesh, India ' Department of Energy and Climate Change, School of Environment, Resources and Development, Asian Institute of Technology, Pathum Thani 12120, Thailand; Department of Electrical and Electronics Engineering, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar, India ' Department of Energy and Climate Change, School of Environment, Resources and Development, Asian Institute of Technology, Pathum Thani 12120, Thailand ' Maharaja Sriram Chndra Bhanja Deo (MSCB), UniversityBaripada, Odisha, India ' Department of Information Technology, Aditya Institute of Technology and Management (AITAM), Tekkali, 532201, Andhra Pradesh, India

Abstract: Smart grid (SG) stability relies on collecting consumer data, assessing it against power supply criteria, and sharing utilisation reports. Artificial intelligence (AI) methods predict Smart grid stability status (SG-SS) to expedite this. Despite a huge progress in AI, superior models with enhanced accuracy are sought after. This work proposes a light gradient boosting machine (LGBM) for predicting SG-SS in decentral smart grid control (DSGC) systems. The LGBM is trained on the dataset from simulations of a four-node network with stable and unstable grids. Using the Bayesian optimisation (BO) algorithm, we optimise the LGBM's performance through hyperparameter tuning. Evaluation with ROC curves and confusion matrices shows that the proposed model achieves nearby perfect accuracy of 99.95% with AUC 100%, offering highly reliable predictions for grid stability. The proposed approach is further compared with other AI algorithms and other recently published works, where it outperforms all comparative approaches in predicting grid stability in DSGC systems.

Keywords: artificial intelligence; AI; smart grid; SG; hyperparameters; light gradient boosting machine; LGBM; Bayesian optimisation; BO; Gaussian process; GP.

DOI: 10.1504/IJAMECHS.2024.143382

International Journal of Advanced Mechatronic Systems, 2024 Vol.11 No.4, pp.226 - 241

Received: 02 Sep 2024
Accepted: 18 Oct 2024
Published online: 16 Dec 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article