Constable, Gordon K.; Cleary, Michael J.; Tickel, Craig; Zhang, Gongxu (1988, ASQC) Wright State University, Dayton, OH; PQ Systems, Dayton, OH; Beijing Institute of Post and Telecommunications, Beijing, China
The U.S. auto industry has experienced heavy competition from abroad during the last decade. Both cost and quality have been at the root of their inability to compete on an even footing. The attempt at improvement in quality has taken many forms, including the increased use of statistics in the production area. Both manufacturers and auto suppliers are now applying Shewhart Charts in an attempt to improve the quality of their processes. A component part is normally processed through several operations before it is assembled on the final product. At a particular stage of production (operation), there is an "overall quality" of the part due to the preceding operations, and a "specific quality" about to be imparted to the part at the current operation. If the previous operations affect the quality independently of the current operation, the quality imparted by the current operation is simply additive. However, if the quality of the prior operations affect the quality imparted by the current operation, the quality imparted by the current operation is dependent and is not simply additive. The standard Shewhart control charts can be used to control the overall quality and to control the specific quality of an independent operation, but it cannot discriminate between the quality imparted by prior operations and that imparted by a current operation where they are dependent. The Cause-Selecting chart proposed by Professor Gongxu Zhang is one method of dealing with this problem. Since so many automotive processes have stages that are not independent, the cause-selecting chart shows promise for increasing the ability to analyze and improve processes. This paper will include a complete mathematical presentation of cause-selecting charts, along with an application from the auto industry.