Automatic crack detection in stamped auto-parts

The objective of the project is to develop a non-destructive test protocol to accurately and reliably detect cracked, pre-welded automotive parts within the constraints of an industrial assembly cell. This will involve reviewing, developing and testing one or more test procedures based on vibration excitation of the part and the measurement and analysis of the response. In particular the work will focus on exploring existing potential test protocols, defining the capabilities of different sensors that could be used in the given environment (non-contact vibration transducers), developing appropriate vibration signal analysis algorithms for crack detection and decision making. The final result of the work will be a test protocol suitable for implementation in an actual assembly cell. The goal will be to achieve crack detection performance that is 100% accurate and reliable. Kirchhoff Van-Rob will benefit through participation and guidance of an investigation into the possibility of developing a new completely automated, accurate and reliable crack detection system. A successful system could be used in a wide range of applications where inspection is done manually. Reliable detection of cracked parts early in the manufacturing process will result in improved efficiency and improved customer satisfaction.

Faculty Supervisor:

Dr. Chris Mechefske

Student:

Fei Geng

Partner:

Van-Rob Inc.

Discipline:

Engineering - mechanical

Sector:

Automotive and transportation

University:

Queen's University