A Study of Structural Characteristics of a Stewart Platform Based Machine Tool
 
 

Vivek Saxena | MS | 1997

ABSTRACT:

Traditional machine tool designs, such as the base and tower design architecture, suffer from certain drawbacks which adversely affect their performance and restrict their usage. One of the major limitations of these conventional machine tool designs is their inability to adequately machine with varying spindle orientations- and not being able to machine with maximum of six possible degrees of freedom. Such an ability is quintessential for properly and efficiently machining parts with highly contoured or complex free from surfaces. Some of the other limitations of the conventional machine tools are low rigidity, low accuracy and low natural frequencies. The research literature suggests that the Stewart Platform, a parallely actuated robotic manipulator with six degrees of freedom, has properties which overcome some of the above mentioned drawbacks. This thesis investigates the use of Stewart Platform for machine tool applications. A study of some key structural characteristics, which critically influence the behavior of such a Stewart platform based milling machine tool and also determine its feasibility for the purpose, from the focus of this thesis.

The properties of the Stewart Platform which have been studied are workspace, stiffness, dexterity, singularities, forces, and natural frequencies. Since many of these properties are theoretically derived from the kinematics of the mechanism, the inverse kinematics and Jacobian solution of the Stewart Platform are derived. The workspace, stiffness, dexterity, and singularities have been calculated theoretically for various angular configurations of the platform and their values plotted as three dimensional plots. Probable explanations for the exhibited behavior of these properties have been provided. The codes for the calculation of the property values and generation of plots were written in MATLAB. Forces in the legs of the manipulator are estimated by modeling it in mechanical dynamic mechanical software DADS, which solves the dynamically modeled problem by iterative methods. The natural frequencies have been estimated using Finite Element Analysis (FEA) (SDRC I-DEAS Masterseries). The natural frequencies are obtained for various configurations of the Stewart Platform. The study of all these properties also favorably indicates the feasibility of improved next generation machine tools based on six degrees of freedom parallel robotic manipulator.

If you have any comments or suggestions please e-mail jwsuther@mtu.edu.