Rotor blade operational data analysis methods and applications for condition monitoring of vertical and horizontal axis wind turbines

Joshua Kusnick, Purdue University

Abstract

Wind turbines are among the largest rotating machines in the world and are subjected to highly variable loading and environmental conditions. In actively-controlled horizontal axis wind turbines, generator power output, wind speed, and wind direction are monitored and fed into a control system to pivot (yaw) the rotor into the oncoming wind and to rotate (pitch) the blades to maintain the most efficient rotational speed for power-generation and limit excessive loading in high winds. However, these systems do not take into account that wind loading can vary significantly along the span of the blades, and it is the blade responses to this loading that dictate both power output and reliability of the turbine. More detailed knowledge of the blade structural dynamic response has the potential to improve wind turbine control so that high power output is maintained while damaging loading conditions are mitigated, and the data analysis methods applied in this thesis seek to accomplish those goals using blade-mounted inertial and strain sensors. Both scheduled and unscheduled maintenance on wind turbines are costly, and downtime can significantly reduce revenue. Expanding current condition monitoring systems to integrate blade load monitoring can improve maintenance scheduling and allow for controlling or positioning the turbine so as to reduce asymmetric loading and rotor imbalances which can damage structural and drive line components. As a step toward achieving these goals, a set of data analysis tools were developed for processing generator power, rotor operational acceleration and strain measurements, and nacelle and drivetrain loads in both horizontal and vertical axis wind turbines (HAWTs and VAWTs, respectively). Modal filtering of blade responses was demonstrated to be effective in classifying the loading cycles of potentially damaging vibration in an experimental VAWT subjected to a variety of wind shear conditions and showed that proper placement of a VAWT in a wind field can reduce excitation of asymmetric modes by as much as 40% while increasing power output by 5%. Blade vibration analysis was also used to classify individual blade pitch error on an operating 900 watt HAWT. Computer models were created to verify this technique in utility scale HAWTs, as well as to demonstrate the effectiveness of a combination of condition monitoring methods for evaluating rotor mass imbalance and aerodynamic asymmetry in the form of blade pitch error. It was shown that even a single-blade pitch error of 1° in a utility scale turbine can increase the nacelle yaw moment by 250%, a 2° error increased the low speed shaft bending moment by 100%, and a 5° error reduced the power output by over 7%. A 0.5% blade mass imbalance increased the low speed shaft shear force by 60%, the tower top side-to-side shear force by 50%, and the tower-base side-to-side shear force by 250%. Therefore, it is critical that these imbalances be detected and accounted for to reduce or prevent potentially damaging loading conditions and maintain peak performance. Imbalance detection algorithms using only non-blade measurements, which are the current state of the art, could not distinguish between simultaneously applied mass imbalances and pitch errors, and could not locate the blade responsible for the pitch or mass imbalance in these simulations. A combination of blade and non-blade measurements were formulated into an algorithm for the detection and classification of pitch and mass imbalances in a wind turbine rotor - the algorithm quantifies the magnitude of the imbalance and locates the problematic blade or blades.

Degree

M.S.M.E.

Advisors

Adams, Purdue University.

Subject Area

Alternative Energy|Mechanical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS