ABSTRACT

In 2021, the electric power generated from wind turbines is estimated to have increased to 380 billion kWh, which is about 9.2% of the total U.S. utility-scale electricity generation. It is currently the largest installed renewable energy technology, and about 17% of the planned utility-scale generating capacity additions for 2022 are wind-based. The primary reason for this growth is the need to utilize free energy resources that are also environmentally clean. Wind-generated power, however, is uncertain and varies from time to time and season to season. Dealing with such uncertainty requires the wind-based electrical generating system to adapt to changes in wind speed and direction, and to adjust its pitching angle in such a way that maximizes output power and efficiency. This study explores and experiments with the relationship between wind speed, pitching angle, and generated electrical power. The data collected during experimentation is used to train and evaluate supervised machine learning (ML) models to capture and approximate those relationships. The study utilizes a WINDLAB™ electrical generation system, a self-contained wind tunnel with various wind profiles. Many experiments using different pitching angles were carried out with the hope of training a predictive model that, given a wind speed, outputs a pitching angle that maximizes system efficiency and generated output power.

INTRODUCTION

Wind energy has become one of the fastest growing sources of energy in the world. In 2021, wind power grew by 1.8% compared to the previous year, adding 93.6 GW to the global cumulative wind power capacity (Frangoul, 2022). The current wind power capacity (both onshore and offshore) is at 837 GW (Frangoul, 2022). Concerns regarding climate change have fueled great interest in the renewable energy sector. Countries are currently implementing initiatives to fight climate change and lessen its impacts by installing renewable energy facilities.

This content is only available via PDF.
You can access this article if you purchase or spend a download.