Document Type

Thesis - University Access Only

Award Date

2012

Degree Name

Master of Science (MS)

Department / School

Mechanical Engineering

First Advisor

Shanzhong Duan

Abstract

This thesis presents a new multibody model and efficient algorithm for computer modeling and simulation of dynamic behaviors associated with flexible structures of horizontal axis wind turbines (HA WT) using multibody dynamics. The model is based on multibody system formed by four rigid bodies linked by joints and springs. A four-rigidbody element has been introduced to represent flexible structures of HA WT. The element consists of four rigid bodies linked by three of joints. Two universal joints are placed at two ends and one cylindrical joint at the middle. So the four-rigid-body element has six degrees of freedom total which includes four degrees of freedom for bending, one degree of freedom for torsion, and one degree of freedom for axial stretch. For each degree of freedom, a spring is used to describe the stiffness of the structure such as wind turbine tower. Potential energy equivalence and Timoshenko beam theory have been used to determine the stiffness of the springs. With these considerations, flexible components of the HA WT may then be represented by connecting several such elements together. Based on the four-rigid-body element model, the tower of a HA WT has been constructed. Equations of motion of the modeled tower are then derived via Kane's dynamical method. To validate the model and algorithm, Matlab simulation codes for description of dynamical behaviors of the tower structure have been produced in commercial computational multibody dynamic analysis software Autolev under free vibration mode. Matlab simulation results were compared with bench mark simulation results produced from commercial software MD Adams. Comparison shows that Matlab simulation results basically agree with the bench mark data. To improve efficiency of simulation cases, the parallel computing code has been developed and implemented on SDSU cluster server. The simulation time differences between serial computing code and parallel computing code are compared. The new model and algorithm may overcome some shortcomings in current wind turbine design and analysis techniques and provide another approach for future development of wind turbine analysis tools.

Library of Congress Subject Headings

Wind turbines
Vibration
Dynamics -- Computer simulation

Publisher

South Dakota State University

Download

Share

COinS
 

Rights Statement

In Copyright