A Study Of Parameters Optimization For Dyeing Production Process With Multiple Quality Characteristics

Since 2015, because the global economy is still sluggish, Chinese economic is more and more week. Asian emerging industrial countries with its cheap wage and abundant labor caused a strong impact on Chinese garment export industry. Due to various reasons, as well as the problems of insufficient labor and rising wages in Chinese textile industry, the competitiveness of the industry in China continued to decline. Printing and dyeing, a kind of textile technology, also suffer the challenge. The key process of printing and dyeing is setting parameters. The selection of parameters directly determines the quality of printing and dyeing products. But the printing and dyeing factory still rely on the experience of some skillful workers. This often dissatisfied customer and make a lot of rework and refund. How to make a standard setting of parameters for enterprise has become an urgent problem.The research process has four stages. At the first stage, discuss with the engineers at related enterprise, and then determine the impact factors and quality characteristics; At the second stage, find the relationship between the construction factors and quality characteristics by response surface method, use the orthogonal table to confirm the configuration of regression function is suitable for one order regression model or not, if not, use the central composite design as experiment to constructed two order regression model, and then use the total expectation function, improved the total expected function and quality loss function to integrate two quality indicators into a single indicator, and these three functions would been used as the fitness function of the genetic algorithm; At the third stage, use the orthogonal test to find the best parameters for genetic algorithm; At the last stage, use genetic algorithm to find the best parameter combination.This paper put printing and dyeing enterprises as the research object, using the genetic algorithm toolbox in MATLAB. Optimize the model which set by response surface method in order to solve the question caused by using the error method to set parameters. What more, set a standard process for enterprise so that solve the problem of setting parameters in the production process. All the work makes a certain contribution to reducing product rework loss and refund.