| Open mill is the earliest rubber mixing equipment, due to reasons of low automation, poor working environment and workers labor intensity, its application in rubber factory is more and more marginalized, and is gradually replaced by internal mixer which has higher degree of automation and more environmental protection. Open mill has the advantage of low mixing temperature, less oxidative destruction to rubber molecular chain, and the mixing quality is better than high-temperature mixing method such as internal mixer and so on. Especially the development of low-temperature one time rubber mixing process in recent years, which makes the full use of open mill advantages, and puts open mill’s application into a very high level, thus higher process optimization is put forward to open mill. But because of traditional open mill’s simple struction and backforward control method, can’t contral the technological parameter effectively, make the traditional open mill can’t satisfy the request of low-temperature one time rubber mixing process. The XK-160E open mill intelligent rubber mixing experiment platform developed by Qingdao University of Science and Technology can control the technological parameter effectively, it become possible to increase the quality of mixing rubber by changing the technological parameter.The process of open mill mixing is complex, the technological parameter like roll nip, roller speed, speed ratio, temperature and mixing time all have influence. Now, some document use Orthogonal Experiment to analyze one target in multi-factor, find a technological parameter combination toward one target. But because it have not only one targets to mesure the mixing rubber, this method can’t make all the target reach the best at the same time. Aimed at this situation, this papper brings the weighted summation method in the study of open mill’s technological parameter optimize, and use BP neural network method and GA to optimize the open mill’s technological parameter. In this paper, the main contents are as follows: 1. This paper consulted a lot of materials, analyzed the open mill mixing mechanism, choice the mixing rubber’s qualiry index and the technological parameters which need to optimize. By analysis the research status of open mill’s process optimization and commonly used method of multi-objective optimization, determined the process of open mill’s multi-objective optimization process optimization.2. Conduct an orthogonal experiment on XK-160E open mill intelligent rubber mixing experiment platform, get the data sample, to be range analysis and fuzzy comprehensive qualty weighted score analysis for every qualiry index, study every technological parameter’s influence to qualiry index. And use weighted sum method to proceed a weighted composite score for every experiment data, transforme the multi-objective optimization problem to a single objective optimization problem, get the open mill’s multi-objective optimization best technology parameter under orthogonal experiment:roll nip:0.6mm, roller speed:32r/min,, speed ratio:1:1.25, temperature:50℃, mixing time:18.5min.3. Bring the neural network in open mill’s process optimization, across the contrast experiment about BP neural network’s number of hidden layer neurons and learning algorithm, build the BP neural network apply to this papper’s problem, supply a mathematical model for the optimization of process parameters.4. Use the idea of natural evolution of genetic algorithm, with mixed rubber quality fuzzy comprehensive scoring function as fitness function, and five parameters as design variables, do global optimization for the mathematical model established by neural network. The results show that:the bese process parameters combination for the XK-160E open mill intelligent rubber mixing experiment platform is:roll nip:0.685mm, roller speed:33.314r/min,, speed ratio:1:1.3, temperature:53.175℃, mixing time:19.15min. The corresponding Comprehensive score value is71.4076. The experiment proves that this method is feasible. |
PDF Full Text Request