PREDICTIVE ESTIMATION OF DYNAMIC DIMENSIONAL SPECIFICATIONS IN THE ASSEMBLING OF COMPONENTS COMING FROM LOW PRECISION MANUFACTURING PROCESSES

Carlos Hernández, Rainer Tutsch
Abstract:
Low precision manufacturing processes usually give rise to items that present a high (dimensional) variation. Therefore, when assembling these items special techniques have to be employed to minimize either the resulting stacked variation or the scrap level. However, full inspection (100%) is commonly required. The Statistical Feed-Forward Control Model (SFFCM) divides the system in two to apply iteratively the Statistical Dynamic Specifications Method (SDSM) on small groups of component items that are produced consecutively in a short-time interval. Ina parallel manufacturing configuration, where the components are produced simultaneously in different lines, the adjustment offset generates an undesired bias. To approach theproblem three different predictive models to estimate the adjustments were developed. Their properness was tried by means of designing a set of specific experiments and by simulating the production of lots of 1,000 assemblies made of two components having high dimensional variation. The effectiveness of the models was measured in terms of the reduction achieved in the mean shift and the standard deviation of the resulting assemblies’ length and in the improvement achieved in the capability indices of the processes. Simulation results showed that whereas the predictive models helped reduce the average mean shift between 71% and 83%, the average standard deviation varied increased between 4% and 28%. In conclusion, the proposed approach helped reduce significantly the mean shift but not the standard variation. The resulting process capability indices, cp and cpk, revealed that none of the predictive models performed well enough to get rid of the offset problem.
Keywords:
assembling technique, high variation manufacturing processes, statistical dynamic specifications method, statistical feed-forward control model, predictive specification estimation
Download:
IMEKO-TC14-2013-11.pdf
DOI:
-
Event details
IMEKO TC:
TC14
Event name:
TC14 ISMQC 2013
Title:

11th International Symposium on Measurement and Quality Control

Place:
Cracow and Kielce, POLAND
Time:
11 September 2013 - 13 September 2013
 Back