Regression learner machine learning approach to predict wind speed considering various parameters and integration of DG in mesh distribution system through GWO

ABSTRACT

The escalating global demand for electricity is driving a significant expansion in the size and complexity of electric power systems. Explosive growth strains power distribution, creating challenges for operators. Among these challenges, the imperative to minimise power losses is crucial. To enhance power efficiency and quality, Distributed Generation (DG) technology, particularly wind energy, is being integrated into distribution systems. However, the variable nature of wind speed poses a significant challenge to the seamless integration of wind energy into grids. To address this challenge, wind speed prediction methods are explored. In this work, a Machine Learning framework is employed for multivariate wind speed forecasting in Tamil Nadu, India $\left[{8.0883}^{\circ }\mathit{N},{77.5385}^{\circ }\mathit{E}\right]$. Key performance metrics, such as Root Mean Square Error (RMSE) and Mean Absolute Percentage Error values (MAPE), are used to assess the accuracy of the wind speed predictions. The predicted wind speed data is then utilised to estimate wind farm power output which is seamlessly integrated into the distribution system. Load Flow analysis is conducted using the robust and straightforward current injection method, focusing on IEEE-33 and IEEE-69 bus-balanced Mesh distribution systems. The aim is to determine the impact of wind-powered Distributed Generation on power losses within these systems.

KEYWORDS:

• Distribution System Load Flow (DSLF)
• Gaussian process regression learner (GPRL)
• Loss minimisation
• grey wolf optimization (GWO)
• machine learning algorithm (MLA)

Disclosure statement

The authors state that they have no prior financial conflicts that may seem to have influenced the work described in this study.

Data availability statement

All the data sources have been cited in this paper.