Masters Theses

Date of Award

5-1991

Degree Type

Thesis

Degree Name

Master of Science

Major

Mechanical Engineering

Major Professor

Remi C. Engels

Committee Members

John Caruthers, Louis Deken

Abstract

Whenever large quantities of a particular part are manufactured, production tolerances can give rise to physical variations between parts. Physical part-to-part variations can in turn lead to differences in both static and dynamic stresses, for a given loading. The variations in stress measured on a sample of parts within the population can be used to statistically describe the behavior of the larger population of parts. Strain gages are often used to determine stresses and are applied at locations defined by analysis or at locations where a known abnormality, such as a fracture, may occur. Because the application of strain gages is subject to various application tolerances, measured values of strain can be somewhat different even within a constant strain field. Therefore, the potential for differences between indicated part-to-part stress variation and actual part-to-part stress variation exist. Because it is the variations in stress that may be used to describe the behavior of a larger population of parts, these differences can introduce error. It is the purpose of this research to identify how much part-to-part stress variation can result from application of strain gages within the application tolerance ranges. Knowing this, it would then be possible to quantify on a relative basis, how much this contributes to overall part-to-part stress variations. An analytical approach employing finite element techniques was used so that variables producing real part-topart stress variations and variables inherent in the test process were eliminated. The aim of the approach was to define the stress field within the small dimensions of a predetermined tolerance range. Having done this, the maximum range of part-to-part stress variation resulting from application variations would be the range of stress within the tolerances. Because the control variable of the study was the application of the strain gage, it was necessary to identify an appropriate strain gage size and to establish realistic bounds for strain gage application tolerances. For the part considered in this study, a compressor blade, an appropriate strain gage size would measure 0.020" x 0.020". The placement tolerance range was found to be ± 0.010 inches (50% of the gage size). The orientation tolerance used was ± 22.5 degrees. Three load cases were run which provided different stress gradients within the specified tolerance ranges. Results of the study indicated that significant part-to-part stress variation can result from small variations in strain gage placement and orientation. The amount of stress variation ranged between 12 and 96% for the cases considered. To reduce this effect, the technique used in this research can be used pretest to define allowable tolerances so that variation is minimized. If reduced tolerances are not practical, then the technique can be used to find a location which produces acceptable stress variation within the established tolerances. A technique for removing this effect posttest from given measured data was also demonstrated. In this case, actual application tolerances and stress variations are much more difficult to determine. Under certain circumstances however, it is still possible to establish reasonable error bounds.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS