Economic design of X̄ control charts for non-normal data using variable sampling policy

Liao

2013

sajie

Generalised cumulative count of conforming (GCCC) charts have been proposed for monitoring a high-yield process that allows the items to be inspected sample by sample and not according to the production order. Recent study has shown that the GCCC chart with a variable sampling interval (VSI) is superior to the traditional one with a fixed sampling interval (FSI) because of the additional flexibility of sampling interval it offers. However, the VSI chart is still costly when used for the prevention of defective products. This paper presents an economic model for the design problem of the VSI GCCC chart, taking into account the correlation of the production outputs within the same sample. In the economic design, a cost function is developed that includes the cost of sampling and inspection, the cost of false alarms, the cost of detecting and removing the assignable cause, and the cost when the process is out-of-control. An evolutionary search method using the cost function is presented for finding the optimal design parameters of the VSI GCCC chart. Comparisons between VSI and FSI charts for expected cost per unit time are also made for various process and cost parameters. OPSOMMINGVeralgemeende Kumulatiewe Telling van Konformasie-grafieke (VKTK) word voorgehou as metode om ŉ hoë opbrengs proses, wat toelaat dat items monster-vir-monster eerder as in produksievolgorde geïnspekteer word, te monitor. Onlangse navorsing toon dat die VKTKgrafiek met 'n wisselende monsterneminginterval (WMI) beter is as dié met 'n vasgestelde monsterneminginterval (VMI) vanweë die addisionele aanpasbaarheid wat dit bied ten opsigte van die monsterneminginterval. Die VKTK-grafiek met WMI is egter duur wanneer dit gebruik word vir die voorkoming van defektiewe produkte. Hierdie artikel bied 'n ekonomiese model vir die ontwerpprobleem van die WMI VKTK-grafiek. Die model neem die korrelasie van produksieuitsette met die huidige monster in ag. In die ekonomiese ontwerp word 'n kostefunksie ontwikkel wat die koste van monsterneming en inspeksie, die koste van vals alarms, die koste van identifisering en verwydering van toerekenbare oorsake en die koste wanneer die proses buite beheer is, in ag neem. 'n Evolusionêre soekmetode wat van die kostefunksie gebruik maak om die optimale ontwerp parameters van die WMI VKTKgrafiek te vind, word voorgehou. Vergelykings tussen die WMI-en VMI-grafieke vir verwagte koste per tydseenheid word ook voorgehou vir verskillende proses-en kosteparameters. a a a a a AA 1

Section: Optimisation Problem and Solution Proceduresmentioning

confidence: 97%

Section: Optimisation Problem and Solution Proceduresmentioning

confidence: 99%

Economic Design of Vsi GCCC Charts for Correlated Samples

Liao

2013

sajie

“…That is, if the mean, standard deviation, skewness coefficient and kurtosis coefficient of the process characteristic can be estimated with reasonably accuracy, then Burr's density function may be applied to fit this data set. Burr's density function has been successfully used to stand for various normal and non-normal probability distributions in SPC research, including economic design of control charts and product tolerance design, e.g., Tsai [19], Chou et al [20,21], Chen [22], and Chen and Yeh [23].…”

Section: Introductionmentioning

confidence: 99%

Economic Design of Process Mean, Standard Deviation and Screening Limits Based on Burr Distribution

2017

AUSMT

Abstract:The traditional quality screening problem considers the correlated and normal quality characteristics between product and screening characteristics. However, non-normal characteristics may exist in many practical production and inspection processes. In this paper, the author applies the bivariate Burr's density function in the quality level setting and screening problem to obtain the optimal process mean, standard deviation, and screening limits based on the minimization of the expected total product cost including quality loss for conforming products and rework costs for non-conforming products. A numerical example is given for illustration and a sensitivity analysis is conducted to assess the effects of model parameters on the optimal problem solution.

“…But, sometimes the quality characteristic may not have normal distribution. Considering this situation, Rahim (1985), Chou et al (2001), Chen (2004), Chen and Yeh (2006) have developed economic design for x control chart for non-normal data, under different situations.…”

Section: Introductionmentioning

confidence: 99%

Economic design of moving average control chart for non-normally distributed data: a comparative study

Patil¹,

Shirke

2017

IJOR

An economic cost function for moving average control chart under non-normal quality characteristic is obtained using continuous and ceased production modes. This cost function can be used to obtain optimal design parameters which minimise the loss cost per hour of the production. Burr distribution has been used to transform the non-normal data to normal. Sensitivity analysis of the economic design is also discussed. The ceased process is observed to be dominant with significant saving in the cost over the continuous process.