GRINDING WHEEL WEAR PREDICTION USING A DIGITAL CAMERA TO INVESTIGATE THE FEASIBILITY OF A MACHINE VISION SYSTEM
Open Access
Author:
Cutter, Eric T
Area of Honors:
Mechanical Engineering
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Eric Russell Marsh, Thesis Supervisor Eric Russell Marsh, Thesis Supervisor Professor Matthew M Mench, Thesis Honors Advisor
Keywords:
grinding wheel machine vision camera texture image processing
Abstract:
Many people agree that the grinding process is very complex and requires the expertise of a skilled operator to be performed satisfactorily. Attaining a high quality surface finish on a ground part requires that the grinding wheel remains free of metallic chips and other debris for the entirety of the process. One way of ensuring the wheel is clean is to dress it frequently, but this is a time consuming solution that is economically inefficient. The goal of this research is to investigate the use of machine vision as part of a possible solution to maintaining the grinding wheel to its peak performance. For this research MATLAB image processing techniques are used to predict when the grinding wheel will begin to pick up metallic chips. Eventually the machine controls with the software built in will interpret the images from the camera and determine when the wheel has reached the end of its life, so that the operator will only need to dress the wheel, saving time and money.
Using a high speed camera with a telecentric lens, high quality images of the grinding wheel surface were obtained and analyzed in MATLAB. The analysis showed that five of the twelve calculated statistical parameters exhibit strong trends. Since these trends were also observed to be repeatable, it shows that statistical image analysis can be a useful tool for predicting grinding wheel wear and identifying a wheel’s end-of-life point.
