Operator Related Causes for Low Correlation Between CAT Simulations and Physical Results
The objective of this study was to explore correlation between CAT (Computer Aided Tolerancing) simulation and physically measured results in running production with focus on operator dependant factors. Therefore, the manual assembly of 25 different system solutions (locating scheme, tolerances, fasteners etc for a part) was analyzed. The study has been performed in the automotive industry and the system solutions are from 3 different cars in 2 different factories, all manual assembly in a paced line. The analysis shows several interesting results; in running production 33% of the measurements are not ok although 28% had their tolerance zone adjusted according to the measured results to make them ok. The conclusion is that the CAT simulations do not predict all the variation and therefore additional factors need to be included to enable accuracy improvement. Further relationships between additional factors such as operator influence and bad geometrical quality can be proven. A short term solution is suggested as well as a long term solution involving the need for development of additional functions in CAT tools, the overall goal being to decrease the difference between simulation results and actual physical results.