Simulation Study On Optimization Of Injection Molding For Automobile Quick Connector

The lightweight of automobile is one of the most important subjects in the development of automobile industry, which has promoted the wide application of plastics in automobiles. Plastic products are prone to various defects in the molding process, so how to improve the quality of products is an important problem faced by the automotive lightweight.CAD/CAE technology is rapidly developing and applied to the injection molding industry, which efficiently improve the quality of products as well as cut the cost of product development cycle.The automobile quick connector shell and lock spring are chosen as research objects in this paper. The injection molding process of shell is optimized by using orthogonal test method and gray correlation theory,and the realization of injection molding process for lock spring is by the combination method of neural network and genetic algorithm. The main contents of this paper are as follows:(1) Combined with product analysis and Moldflow software analysis, determine injection mould design of the two key parts for automobile quick plug connector: shell and the lock spring, including the gate design, cavity number, cavity arrangement, cooling system and molding material selection. Based on this scheme, through the simulation of injection molding process for the two parts under their own original recommended process parameters by using Moldflow software, analyzed the reason of the dimple phenomenon for shell and the break phenomenon for lock spring.(2) Based on Moldflow simulation analysis and orthogonal test,analyzed the influence of process parameters on the weight and volume shrinkage for automobile quick connector shell. But because the choice of linear weighted multi-objective optimization method is artificial distribution, which will affect objectivity of the results, so this paper uses grey relational analysis to verify the analysis results of range and variance analysis. Finally, the optimized process parameters were obtained on the basis of experiment, under which the the product weight was closer to the theoretical value, the weight repeatability was higher, and the volume shrinkage was smaller.(3 ) As for the phenomenon of being easy to break, the evaluation indexes of different process conditions were obtained by Moldflow software and orthogonal test. Then through training the neural network model, the model between the process parameters and evaluation indexes is established. However, the neural network can not get proper weight distribution because of its own defects, and the outputs of the network can not meet the requirements, so the genetic algorithm is used to optimize the connection weights of the neural network.(4) Based on the obtained neural network model, a genetic algorithm for multi-objective optimization of injection molding process is established to find the combination of process parameters to achieve the best comprehensive quality.