YAW DETECTION IN WIND TURBINES WITH SCADA DATA-DRIVEN ANALYSIS, QUADRATIC REGRESSION AND ULTRASONIC SENSOR VALIDATION

Yaw control in wind turbines optimizes energy production and protects mechanical components from damage. The control accuracy depends on the quality of the yaw-angle measurement, which is compromised by the so-called sensor's zero-point position, causing errors and losses. This work proposes an approach based on SCADA data and quadratic regression for diagnosing the yaw error. The data filtering procedures in terms of i) wind speed range and ii) angle bins are used to identify the parabola-vertex as the error indicator. Validated with ultrasonic sensors (iSpin), the approach was applied in turbines at Rio Grande do Norte, detecting critical deviations (e.g., 9.48°) and exemplary cases (0.07°). The proposed approach eliminates the need for advanced sensor solutions, leading to cost reduction and offering an affordable tool for wind farms. Aligned with the demand for sustainable, renewable energy, the approach transforms operational data into precise corrections, increasing the efficiency and reliability of operations and maintenance routines.