Interval estimation of capability index Cpmk for manufacturing processes with asymmetric tolerances

2009 ◽  
Vol 56 (1) ◽  
pp. 312-322 ◽  
Author(s):  
Y.C. Chang
1990 ◽  
Vol 19 (12) ◽  
pp. 4455-4470 ◽  
Author(s):  
N.F. Zhang ◽  
G.A. Stenback ◽  
D.M. Wardrop

1996 ◽  
Vol 118 (2) ◽  
pp. 252-258 ◽  
Author(s):  
X. D. Fang ◽  
Y. Zhang

The major advantage of using selective assembly in manufacturing is that it allows the use of low precision parts to achieve high precision assembly. However, two problems arise: (a) the surplus parts due to unbalanced mating parts in some selective groups (b) there is no quantitative method to predict the matchable degree before assembly, and correspondingly no quantitative criterion on how to adjust manufacturing processes so that the desired match between mating parts could be guaranteed. By adopting the concepts of intersection and union from set theory and incorporating with the probability method, this paper develops a predictive model for quantitatively estimating the matchable degree between mating parts for selective assembly. Based on such a quantitative reference, together with the criterion for guaranteeing process capability index Cpk, an optimal algorithm for adjusting biases of dimensional distributions can be achieved to assure the matchable degree, thus improving the effectiveness of selective assembly.


2014 ◽  
Vol 1039 ◽  
pp. 622-626
Author(s):  
Zai Fang Zhang ◽  
Xiao Song Wu ◽  
Rui Miao ◽  
Bei Xin Xia

Process capability index (PCI) has been widely applied in manufacturing industry as an effective management tool for quality evaluation and improvement, whose calculation in most existing research work is premised on the assumption that there exists no bias. In this paper, the bias of gauge which exerts an effect on the calculation of PCI is indicated inevitable. The influence on PCI caused by the bias is analyzed by constructing a comparative ratio R between the empirical process capability index and the PCI. A confidence interval estimation method is proposed to solve the underestimation problem of PCI.


Sign in / Sign up

Export Citation Format

Share Document