Six sigma methodologies in ignition coil manufacturing using experimental design

You are here: Homepage > Search

Six sigma methodologies in ignition coil manufacturing using experimental design - improve step

Lynch, D.P. / Fleischmann, J.E. / Duff, W.S.

Quality issues in magnet wire stripping and soldering have led to continuous improvement efforts in ignition coil manufacturing using Six Sigma methodologies resulting in the investigation of an alternative product and a micro-joining process design. The confirmation of the results is accomplished by response surface methodologies, mathematical modelling, and optimisation of the process. Non-linear design of experiments have been used to confirm a breakthrough in the micro-joining process as an alternative to soldering. The statistical methods used to develop the process are based on the documented research efforts. These results significantly extend the level of knowledge of designed experiment application in the welding field and summarise a typical analysis performed in the improvement phase of a Six Sigma project. It was shown that interaction effects as high as 5-way and squared terms could be statistically significant in micro-joining of magnet wire. The apparent reason for this significance of higher order interactions is the nature of the physics of the process. All of the controllable predictor values used in this study are physically linked via empirical equations and are expected to have a high degree of cross-correlation, with sets of higher interaction effects being statistically significant. Quantifying the amount of cross-correlation and using these values to re-specify the model is another opportunity for future investigation. Three separate optimisation methods were applied to the various models developed in this research. The similar results from the different optimisation methods further confirmed the results of the analysis.