Deviation propagation and reduction analysis for multi-operational machining processes

Product dimensional quality has been one of the most important challenges for machining processes. However, current practice in multi-operational machining is a monitoring or inspection oriented measurement strategy which could not provide solutions for deviation reduction while deviation propagation and compensation are taken into account. While the deviation of some quality characteristic is not satisfactory and could not be controlled in the last operation, a partial least squares regression method is proposed to set up the relations between upstream operations and the last one. The regression result helps engineers to identify key processes with respect to their impacts on the last one, and to figure out possible solutions to reduce the deviation and thus makes the quality characteristic achieve the acceptable level. An industrial case is presented to illustrate our proposed methodology.