METHODOLOGY TO COMPARE 3-D AND 2-D PARAMETERS FOR THE OPTIMIZATION OF HARD TURNED SURFACES

ABSTRACT

Hard Turning provides an alternative to grinding in some finishing applications. Average Surface Roughness R_a has been widely used in industry to establish surface texture needed for a given application. It is known that the single parameter R_a is inadequate to define the functionality of a surface texture. The quality of a surface can be determined by the nature of its interaction with another surface. Thus a surface with significant peaks will not make as good a bearing surface as a surface with deep valleys and low peaks. Two different surfaces with similar values of R_a can behave differently under fatigue loading conditions. 3-D visualization of expected surface texture will facilitate optimization of machining parameters to produce function-specific surfaces. The advantages and shortcomings of some current surface texture prediction models are discussed. A new method based on neuro-fuzzy techniques is proposed. Optimization using some 3-D surface parameters was carried out and compared with the results of those obtained using 2-D parameters.

Keywords:

• Surface Roughness
• Surface Texture
• Modeling of Surface Roughness
• Surface Prediction
• 2-D and 3-D Parameters
• Neuro-fuzzy Systems

ACKNOWLEDGMENT

This project was funded by Dana Spicer Driveshaft Inc. and is part of a comprehensive project on providing an alternative process to finish grinding operations. Mr. Sudheer Kodem, The University of Toledo and Mr. Ranganath Kothamasu, The University of Cincinnati played a vital role in combining the manufacturing and computational aspects.