Developing Capability Indices for Measuring the Performance of a Multidimensional Machining Process under Non-normal Distributions

碩士 === 國立成功大學 === 統計學系 === 107 === Process capability index (PCI) is commonly used to measure the capability of a manufacturing process and evaluate whether it can meet the engineering specifications in industries. Generally speaking, the multidimensional machining process has a specific specificati...

Full description

Bibliographic Details
Main Authors: Ling-YuLiao, 廖聆禹
Other Authors: Jeh-Nan Pan
Format: Others
Language:zh-TW
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/akcn7s
Description
Summary:碩士 === 國立成功大學 === 統計學系 === 107 === Process capability index (PCI) is commonly used to measure the capability of a manufacturing process and evaluate whether it can meet the engineering specifications in industries. Generally speaking, the multidimensional machining process has a specific specification called the positional tolerance. In the past decades, many scholars including Krishamoorthi (1990), Davis et al. (1992), Pan and Li (2014), etc. have devoted themselves to developing capability indices for multidimensional machining processes based on the assumption of normality. However, in practical applications, many manufacturing processes, such as nano-cutting process may not follow normal distribution. Thus, by relieving the normality assumption, this paper aims to propose new non-normal capability indices that can correctly reflect the true nonconforming rate for multidimensional machining processes. We first use Scaled Weighted Variance (SWV) method to revise NPC_p and NPC_pk indices proposed by Pan and Li (2014). Then, the two new indices RNPC_p and RNPC_pk can be derived accordingly. In the numerical calculation studies, we compare the performance among various multidimensional machining process capability indices under different parameter combinations for non-normal distributions. The numerical calculation results show that our proposed indices can properly reflect the actual performance for non-normal multidimensional machining processes. Finally, a nano-cutting example is used to demonstrate that the proposed indices are suitable to assess the risk of non-normal multidimensional machining processes.