Modeling and analysis of stream-of-variation in multistage manufacturing processes.

摘要

Multistage manufacturing processes (MMP) are complicated processes involving more than one workstation or operation to produce products. The control of MMP requires dealing with issues of increasing complexities, coupled with shorter product life cycle. This has created an immense need for analysis and validation of process design and system performance to avoid process failure during production, and also the need for utilization of in-line sensing information to achieve rapid product/process failure detection and isolation. The limitations of current process/product development techniques are: a large number of iterative engineering re-designs due to a poor understanding of the MMP's response to uncertainty, and the current focus of Statistical Process Control on monitoring, as opposed to root cause identification and fault prevention.; Stream-of-Variation (SOV) modeling and analysis is the methodology to model, analyze, diagnose, and control complex MMP for quality and productivity improvement. Critical techniques that must be targeted are design and in-line root cause identification, both based on the system-level model of MMP. Fixture design and diagnostics in autobody assembly system are used to demonstrate the proposed methodology. The general framework includes (1) a station-indexed state space modeling of MMP, which integrates the product/process design information, (2) design evaluation using the concept of multi-layer sensitivity in MIMO (Multi-Input-Multi-Output) control, (3) process-oriented tolerance synthesis facilitated by the state space model to integrate product and process characteristics at minimum cost, (4) statistical methods driven by engineering design information for diagnosing root causes, and (5) a further diagnosability study for evaluation of sensor distribution strategy and determination of optimal sensor distribution. Both methodology and implementation are presented.; The SOV methodology can be potentially applied to other multistage manufacturing industries. Development of the SOV methodology will provide substantial benefits to the field of manufacturing. It is our goal that this tool will enable the domestic manufacturer to continue the improvement in quality and productivity and reduce the overall cost so that they can outperform their international competitors.