Open Collections

An optical CMM for flexible on-machine inspection

University of British Columbia

This thesis presents the design and testing results of an optical coordinate measurement machine for on-machine inspection of machined parts. Inspecting parts on the machine which produces them avoids lost productivity by identifying malfunctioning production machines quickly, thereby reducing the number of out of spec parts produced. The optical coordinate measurement machine designed for this project uses two cameras to track a handheld probe which is used to probe the inspected part. The cameras are designed specifically for the project. They are able to capture target measurements at a very high frequency, which allows averaging of many measurements for increased accuracy. Inspecting parts on-machine requires the system configuration to be flexible, as different machines will need to have the cameras in different locations relative to one another. The cameras are designed to communicate using the flexible EtherCAT fieldbus protocol, and are attached to a modular frame, so that the system can be reconfigured. The probe has four LED targets which are tracked by the cameras and a stylus tip with a probing ball. The reconstruction of the probe tip position requires two stages. First the LED target positions are calculated from the camera measurements using a camera model, and then a geometric fitting is done to match the known probe geometry to the measurements and find the probe tip location. The reconstruction is done by software running in the TwinCAT real-time operating system on a host PC. The calibration of the camera model parameters is done using a CMM with <5um accuracy. The accuracy of the camera calibration is 23.6um at a distance of 1m from the cameras. The probe is calibrated using a procedure developed by the author. The probe ball center calibration accuracy is estimated to be less than 10um. The system is tested by performing a series of measurement tasks probing flat and spherical surfaces. The system shows an accuracy of 30um for these tasks.

Text

2018-04-23

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/65580

Mechanical Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/