Measurement Scheme Synthesis in Multi-Station Machining Systems

Abstract

Different sets of measurements carry different amounts of information about the root causes of quality problems in machining. The selection of measurements in multi-station machining systems is currently a slow and error-prone process based on expert human knowledge. In this paper, we propose systematic procedures for synthesizing measurement schemes that carry the most information about the root causes of dimensional machining errors. The amount of root cause information conveyed by a given set of measurements was assessed using the recently introduced formal methods for quantitative characterization of measurement schemes in multi-station machining systems. The newly proposed measurement scheme synthesis procedures were applied to devising measurement schemes in an automotive cylinder head machining process. It was observed that the measurement scheme synthesis procedure based on a genetic algorithm robustly outperformed the synthesis procedures based on the heuristics of successive measurement removal.