Masters Theses

Joel Chesser

Date of Award

8-1995

Degree Type

Thesis

Degree Name

Master of Science

Major

Engineering Science

Major Professor

J. F. Wasserman

Committee Members

Reid Kress

Abstract

Current force sensing technology is dependent on a deformable element to measure force. The deformation of the element can be detected mechanically, electrically or optically. Examples of the current state of the art for force measurement include strain gage or piezoelectric load cells These are all based on measuring a finite deformation of the load cell. The goal of this thesis is to lay the groundwork for developing a load cell which includes an active feedback loop to compensate for the deformation of the load cell caused by the load. The resulting force sensor will appear to be infinitely stiff over some finite bandwidth to an applied load. The motivation for developing this instrument is to improve the accuracy of mechanical testing of small, stiff specimens. A current problem with testing of such specimens is that compliance of the load cell contributes to errors in determining the mechanical properties such as elastic modulus or stress-strain relations. The concept was developed, analyzed, and tested. The analysis indicated that significant reductions in compliance could be achieved. The controller designed, a gain stage in series with a pole zero cancellation stage, was calculated to reduce the load cell compliance to 6.2 μm/N. This is a reduction in compliance by a factor of three over the commercial load cells surveyed. The addition of a lag-lead compensator could further reduce the compliance to 0.01 μm/N, several orders of magnitude better than the commercial load cells. The analyzed load cell was thai built and tested. The electronics performed as designed. The mechanical plant, however, had significant hysteresis limiting its performance. Several attempts were made to reduce the mechanical hysteresis. The diaphragm flexures were annealed, and slotted with limited improvement. The load cell tested did, however, successfully demonstrate the concept. It exhibited reduced compliance due to the control system action. The compliance was reduced from 333 μm/N to 38 μm/N by the control system tested, a simple gain stage. The compliance of 38 μm/N is of the same order as that of commercial load cells. It was determined by simulation that the mechanical hysteresis limited the load cell accuracy achievable with a more sophisticated controller, gain followed by pole-zero cancellation. Therefore that controller was not built and tested. Because of the limitations of the mechanical plant, simulation studies were conducted to determine the potential compliance improvements of the load cell without any mechanical hysteresis. The simulations included the frequency response, noise, and non-linearities of the electronics. Simulation indicated that if the mechanical non-linearities are removed performance is limited by the electronic noise and non-linearities. A problem of the load cell designed is the introduction of systematic errors if load is assumed to be simply proportional to current. A method of compensating for systematic errors caused by displacement and actuator non-linearities was simulated. These simulations indicated that a measurement error of 0.005 N is achievable with the electronics designed in the course of this thesis if the mechanical non-linearities can be eliminated.

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

Share

COinS