Stream of Variations Analysis

Abstract

Stream of Variation (SoV) theory is a unified, model-based method for modeling, analyzing, and controlling variation in multistage manufacturing systems. A SoV model represents variation and its propagation in a multistage system using the recursive structure of state space models; such models can be derived from physical knowledge and/or estimated empirically using system operational data. Immediately, the SoV model enables integrated design and optimization for product and process tolerancing, allocation of distributed sensors in production lines, and evaluation of multistage system designs. With the help of these functions, the SoV method fulfills the objectives of system monitoring, diagnosis, and control and, ultimately, reduces a system's variation during its operation. The SoV method can be further extended to model the interactions among product quality and tooling reliability, known as the Quality and Reliability chain effects, which is the crucial element in carrying out quality-ensured maintenance, as well as system reliability evaluation and optimization. The SoV theory has been successfully implemented in assembly, machining, and semiconductor manufacturing processes. More research and development are needed to extend the SoV theory to manufacturing systems with complex configurations.