On-line Quality Monitoring And Optimization Of The Injection Molding Products Based On Cavity Pressure Feedback

As the increasing demand of injection molding products with high quality andproductivity, the control aimed at injection quality and molding process has become aresearch concerns. However, because lack of sensor, with which we can measure the qualityof injection products online, and a general, accurate mathematical model between the processvariables and parameters of the real products, the on-line quality control system in anacceptable range cost without manual intervention has not been achieved. The CAE andartificial neural network technology have been extensively applied in quality control ininjection product, and it is realized in a certain extent via them, but there still exist limitations.Aiming at solving the limitation of nowadays quality control methods, a quality controlsystem of injection molding products based on cavity pressure feedback is built, combinedwith the artificial neural network technology. Through the data obtained from factorialexperiment, we analyze the changes of cavity pressure in each process stage; establish theprocess parameters of the cavity pressure online and optimization of mechanical model.This paper first introduces the main parameters injection process involved, parametersrole in the quality of products, and the latest development of quality control.Introduces the principle and equipment of data acquisition next, including the factorialexperiment method, acquiring the data acquisition, and the corresponding mechanicalproperties of work piece cavity pressure data. Via the data changed between differentinjections periods of cavity pressure, analyzed the effects of the process parameters role in thecavity pressure value. Find out the pressure (maximum system pressure) is the key of thecavity pressure, by fitting the numerical cavity pressure on the response model. At the end ofthis part, the relationship between the mechanical properties and the cavity pressure isanalyzed.Then, an BP neural network prediction model is established based on the cavity pressuredata collected from the factorial experiment, the process parameters and part of cavitypressure value as input, the cavity pressure value remaining in the other periods as output.After building the communication between MATLAB platform and data acquisition system,MATLAB platform and PLC, use the on-line molding cycle pressure data to predict other period’s cavity values on the same cycle. Then compared with the established numericalconstituted by10groups standard cavity pressure that have optimal performance. Accordingto the result of comparison, adjust pressure value and conversion, then output to PLC. UsingPLC closed-loop feedback control system of injection molding machine to adjust thedeviation value according to the PID regulation between computer output pressure and theoriginal output from injection molding machine control system, to obtain the desired output,and can execute on the injection molding machine.Finally, through the experimental verification, we find, the cavity pressure as a record ofthe melt’s change from viscous flow into a solid, is the comprehensive reflection of processparameters in the mold cavity, the cavity pressure prediction value as the on-line qualitycontrol system feedback, to adjust the corresponding process parameters, and it prove: thequality online control and optimization is feasible.