Multivariate Exploratory Techniques Helping to Increase the Reliability of Facing Clutch Material Burst Resistance

The necessity to improve the reliability of clutch facing in burst resistance characteristic was checked on a specific part number due to a high customer specification for burst resistance. It was decided to use multivariable analyses tools such as Tree Clustering diagram analysis, Principal Component's analysis to explore some of the process variables and to find correlation to burst resistance. Eight variables of the process were investigated and according to the analysis two variables showed a correlation to the burst resistance. A very interesting detail of these variables was hidden in the process and very difficult to vision in the correlation with the burst characteristics however one variable was positive and the other negative in correlation to the burst test results. This information was very important to make decisions and after some changes in the process based on this information the burst resistance increased as well as the reliability, according to the customer requirement. By means of the Multivariable analysis it can in fact help to explore the data directly from the process without any special treatment and therefore increase the reliability of the product.