Evaluation model for multi-process capabilities of stranded wire

Abstract

Stranded wire is the most important component of familiar mechanical equipment such as elevators, cable cars, and cranes. The quality of these products that are used on a daily basis are mainly affected by the tensile strength of stranded wire. In order to attain the purpose of economical design and a long life span of stranded wire, a less relaxation property of strand type is suitable for manufactured tools. Thus, the manufacturing industries of stranded wire need to reach the goals of high tensile strength and low relaxation. To ensure the required quality of stranded wire, the strand pull test and the long period relaxation test are two important quality assurance tests. There are three specific items of the tensile strength test that belong to the larger-the-better quality type. The quality type of the smaller-the-better is for the long period relaxation test. However, many existing methods are able to measure process capability for the product with a single quality characteristic although it cannot be applied to most products with multiple properties. Thus, the indices of C _pu and C _pl , for the larger-the-better and the smaller-the-better quality type respectively proposed by Kane [5], are quoted and combined to propose a new index to evaluate the quality of multiple characteristics of stranded wire in this article. The principle of statistics is then used to derive the one-to-one mathematical relationship of this new index and ratio of satisfactory production process. Finally, the procedure and criteria to evaluate the quality of stranded wire is proposed. This integrated multi-quality property capability analysis model can be used to evaluate the multi-process capabilities and provide continuous improvements on the manufacturing process of stranded wire.

Appendix

/***************************************************************/;

/* SAS PROGRAMING--p-value */;

/***************************************************************/;

/* H0:Cpoi>=1.436 (P-VALUE) */;

OPTIONS REPLACE PAGESIZE=58 LINESIZE=78 NODATE;

DATA NCENT;

INPUT w @@;

n=16;v=1.436;bn=0.949;

D=3*SQRT(n)*v;

x=3*SQRT(n)*w/bn;

PV=PROBT(X,15,D);

FORMAT PV 6.4;

CARDS;

0.1359 0.2468 0.0484 −0.0145 1.7553 1.8348 1.4810 0.6879 1.5900 1.6867 1.7774 1.6805

;

PROC PRINT DATA=NCENT;

VAR W PV;

RUN;