Research On Robust Design Model Based On Quality Structure

Robust design is an important method to control product quality from the source. It candetermine the product performance and economic benefit to a large extent. Although robust design hasachieved a lot of effective research and application results in electronics, machinery and many otherfields since it was focused, it still has some problems to be solved. In this study, we focus onmodeling based on complex quality structure from quality and cost perspective. The main innovativepoints are as follows.(1)The quality structure is considered as the form of the complex quality structural relation in theproduct in this research, which can describe the relationship between quality factors conveniently andclearly. This paper presents the definition of quality structure based on the related quality concepts,and gives its classification and relationships between them. The quality structure can clearly performthe complex quality structural relations between parts and components, quality characteristics andother quality factor.(2)The first model established in this study is joint robust design of parameter and tolerancemodel based on dominant quality structure. This model is proposed for the known functionrelationship between quality characteristic and its variables. First we construct quality loss modelbased on quality structure, which considers dominant quality structural relations into qualitycharacteristic function and quality loss function. Then we give joint robust design model of parameterand tolerance, which can get the optimal values of parameters and its tolerances on the basis of highquality and low manufacturing cost.(3)The second model established in this study is DEA robust design based on invisible qualitystructure. This model is proposed for the unknown function relationship between quality characteristicand its variables. We build the model by orthogonal test and DEA game cross efficiency method. Themodel can give each relative efficiency value of quality to cost of each test programme, and theprogramme which has the highest efficiency value is the optimal combination of parameters and theirtolerances. This model can solve the robust design problem better when the relationship betweenquality characteristic and its variables is unknown.(4)The third model established in this study is DEA parameter design model based on parametercost perspective. The model is proposed based on the fact that different parameter levels usually havedifferent costs. However, this fact was often ignored by people. So in order to consider the cost impactin the parameter design phase, we build DEA parameter design model taking parameter cost intoaccount. Because the model considers the impact of different parameter level cost on parameter design, it could provide some references for a decision-maker.