Process Parameters Intelligent Setting And Optimization For Plastic Injection Molding

As the most important processing method for plastic products,injection molding is widely used in national pillar industries such as consumer electronics,transportation,instrumentation,and national strategic fields of defense military and aerospace.Process parameters are the most critical factors affecting the product quality and forming efficiency of injection molding.The traditional process parameter setting heavily relies on the experience of the molding personal and requires repeated “trial and error”,which is long cycle,high cost and difficult to guarantee product quality.This dissertation focused on theprocess parameters of injection molding and systematically research around the intelligent setting of initial process parameters,automatic correction of molding defects,stable production,and transplantation of process parameters were carried out.The proposed mehtods break through the technical bottleneck,achieve a series of original results,and realize the engineering application of the process parameters intelligent setting and optimization system of injection molding.The main research work and results are listed as follow:Aiming at the difficulty in characterizing the mold forming characteristics in the initial setting of process parameters,an initial setting method of process parameters using the pressure curve at the injection position to characterize the forming characteristics of the mold was proposed.Take the pressure curve at the injection position as a description of the problem of the injection case,obtainthe process parameters of the qualified product as the case solution,and adopt case-based reasoning technique to obtain the initial process parameters.Typical cases show that compared with the traditional representation method,artificial extraction features such as maximum flow length and average wall thickness,the proposed method could characterize the small differences between mold forming characteristics better and has higher retrieval accuracy.Aiming at the problem of expert knowledge representation and evolution in the process of defect correction,a fuzzy reasoning method based on rule engine was presented.Through the fuzzification of defects and process parameters,and the establishment of a fuzzy rule inference engine based on knowledge base,reasoning process is separated from knowledge logic,and multiple defects are divided and solved.Also,the priority and reasoning history of reasoning knowledge were taken into account.In the end,the proposed methods effectively solve the problem that the forming defects and process parameters are strongly coupled,nonlinear,and the difficulty of product forming defects.Aiming at the problem of poor product quality stability during mass production,a new process parameters optimization method based on small sample was proposed.By modeling the sample data obtained from the test mold,the maximum forming process window was calculated to achieve stable process parameter optimization.The experimental results of typical parts showed that compared with the traditional fuzzy inference method,the optimized forming capacity index of the proposed method was increased from 0.52 to 3.78,which ensures the stability of the molding process.Aiming at the coupling problem between process parameters and injection machine equipment during process parameter transplantation,a deep reinforcement learning method for process parameter transplantation was proposed.The cavity pressure curve was selected as the standardized measurement and learning goal of product quality.By establishing the Markov decision process model of deep reinforcement learning,the action network,evaluation network,experience playback mechanism and expert model were constructed to realize the automatical setting of process parameters for target cavity pressure.Compared with the traditional manual transplantation method,the process parameters of the proposed method are more standardized,and the product quality is more stable.Based on the above methods,software and hardware related to process parameters intelligent setting and optimization system for plastic injection molding were developed.Using RS232/RS485,CAN,TCP/IP,OPCUA and other communication methods,the process intelligent system communicates with various injection machine control systems such as KEBA,Easton and HNC,and related integration and demonstration applications have successfully implemented in Borch,Zhenxiong,Hai Tai,Enruide,Haixing,Delitian and other multi-brand,multi-model injection machine.