Research Of Gaussian Mutation Hybrid Genetic Algorithms And Its Application In Injection Molding

Injection molding is a forming method to use the injection machine and injection mould to transform the plastic raw materials into plastic products.Along with the injection products widely used in the fields of precision instrument and etc.,people's requirements on the quality of the product also increase gradually,and the quality of injection products is directly related with the setting of process parameters in shaping.But now there is not a good way to help people to set and adjust the process parameters.For a long time,operators are relying on their own experience or test mould repeatedly to determine process parameters,which expends a lot of time and money,and is not consistent with the development trends of today's manufacturing industry.The application based on molding CAE technology can improve the quality of the products to a certain extent,but there are many defects in the process parameters optimization by that method.In order to solve above problems,this thesis studies a new optimization algorithm to determine the process parameter groups of forming process,simultaneously optimize two quality target,the warp and linear shrinkage of products.The first investigative problem is the relevance between molding process parameters and molding defects,and the feasibility of using genetic algorithm.The first half article introduces the basic principle of injection molding and several common defects of injection products and their ways of improvement.Then introduces the multi-objectives optimization theory,expounds the basic process of genetic algorithm and the basic operation,and briefly introduces four typical multi-objectives genetic algorithms.According to good robustness and other advantages when solving global optimal problems used genetic algorithm,this thesis decide to use genetic algorithm to optimize process parameters.The second theme is improved research of the genetic algorithm,considering the shortage of genetic algorithm in solving complex problems,such as the lack of local search ability.So this thesis introduces hill-climbing and construct a new multi-objective hybrid genetic algorithm by gauss mutation to make up for the defect.And then use three test functions to test the performance of the new algorithm and verify the validity,which lay a theoretical foundation for the subsequent practical application.Finally realize the injection molding process parameters optimization system development and application cases,which verify the validity of this method.With an example of mobile hard shell,using enhanced translation propagation Latin hypercube design(ETPLHD)method to select the sample points,set up surrogate models using Gaussian process to receive the best process parameter combinations more accurately.The optimization results show that this method can really find optimal processing parameters so as to reach the target of optimizing the quality goal.