A Mathematical Framework for the Key Characteristic Process

Abstract

To maximize product quality, a product design team selects concepts and dimensions to minimize a product’s sensitivity to variation. However, even for the most robust products, it is rarely possible to transition a product into production without encountering any variation-related problems. In a complex product, it is not economically or logistically feasible to control and/or monitor the thousands of tolerances specified in a product’s drawing set. To address this problem, many organizations are using Key Characteristic (KCs) methods to identify where excess variation will most significantly affect product quality, and what product features and tolerances require special attention from manufacturing. As simple as this principle seems, most companies struggle to effectively implement KC methods because no quantitative methods to prioritize KCs exist. This paper develops a mathematical definition of a KC based on a variation propagation model. In addition, it develops a quantitative effectiveness measure used to prioritize where verification, variation reduction, and on-going monitoring should be applied. The effectiveness measure incorporates the cost of control, the benefit of control, and the expected change in process capability. The methods are illustrated using an automotive door assembly.