Font Size: a A A

Research On Multi-response Robust Design Theory And Its Application In Spot Welding Process

Posted on:2009-09-23Degree:MasterType:Thesis
Country:ChinaCandidate:L BaoFull Text:PDF
GTID:2121360272984752Subject:Measuring and Testing Technology and Instruments
Abstract/Summary:PDF Full Text Request
The famous Taguchi method initiated in the 70s of the 20th century integrates signal-to-noise ratios(S/N) and design of experiment (DOE), and serves as an important part in robust design theory. With the increment of product complexity as well as more and more concern is paid to several quality characteristics'ability of being simultaneously insensitive to noise, the study on multi-response robust design is attracting more and more attention, which will eventually improve the traditional Taguchi method whose main target is solving univariate quality characteristic problems.After a review of robust design theory finished in chapter 1, this dissertation systematically lists the basic definition, principle, main feature and classification of multi-response robust design, and points out the main obstacle and problems still exiting in this field in chapter 2. Then 9 kinds of present multi-response robust design methods'feature and their advantage as well as disadvantage are comparatively studied. Based on the comparison study about the relationship between multi-response robust design and multiple attribute decision making problems, two new multi-response robust design methods are respectively elicited. One is the multi-response robust design method based on response surface methodology (RSM) and technique for order preference by similarity to ideal solution (TOPSIS), the other is the multi-response robust design method based on multiple genetic algorithm (MGA). Application steps of the first method which integrates RSM and TOPSIS as well as weighting method are put forward in chapter 3. In chapter 4 some basic knowledge about GA is introduced, and the concrete application procedures of how to use the popular non-dominated sorting algorithm and crowding function algorithm in multi-response robust design problems are described under the consideration of the special characteristics of multi-response robust design problems. A case study about spot welding process optimization is carried out in chapter 5, which specifically demonstrates how to put the two abovementioned new multi-response robust design methods into practice. Some experiments are designed and carried out as well. Two response surface models are fitted and the two proposed methods which finally come out with satisfying robust design results are implemented based on these models.The multi-response robust design method is a kind of relatively new design technique. With more and more attention is paid to product quality problems all over the world and the development of some related subjects, the importance of multi-response robust design in engineering will become more and more noticeable.
Keywords/Search Tags:multi-response robust design, response surface methodology, TOPSIS, genetic algorithm, spot welding
PDF Full Text Request
Related items