Research On Quality Prediction Of Hole Making Process In Carbon Fiber Composite

Carbon fiber composites are widely used in aerospace,military equipment and other fields because of their very high strength-to-weight ratio,stiffness-to-weight ratio,and fatigue resistance.However,due to the high hardness and difficult processing characteristics of carbon fiber composites,the traditionally processed parts have defects such as rough surface and delamination,which affect the assembly quality and service life.Therefore,accurately predicting processing quality and realizing online monitoring of processing quality are essential for improving product quality and reducing the rate of non-conformance.The purpose of this paper is to establish a carbon fiber borehole quality prediction model,accurately predict borehole quality,and improve processing quality.On the basis of optimizing process parameters through experimental design,a processing quality prediction model based on PSO-BP neural network is established.Firstly,the fishbone diagram is used to qualitatively analyze the factors affecting the quality of carbon fiber board drilling,distinguishing controllable factors and uncontrollable factors.Secondly,using cutting speed,feed rate,and tool tip angle as factors,roughness and layering factors,roundness,and cylindricity as the responses,through the design factor test and the response surface test,the significance level of the carbon fiber hole making parameters and the interaction between the parameters are divided,and the satisfaction function is combined to obtain the factor level combination under the optimal quality characteristics,and pass the verification test confirmed the optimal parameter combination.Finally,based on the test data,the particle swarm algorithm was combined with the BP neural network to establish a predictive model between process parameters and quality characteristics.Through particle swarm optimization,the computational efficiency of the BP neural network is improved,and the local optimum is avoided.Comparing the models before and after optimization,the results show that the mean square error and average absolute percentage error of PSO-BP neural network are significantly reduced compared with BP neural network.The improved model showed good prediction accuracy and stability,which confirmed the superiority of the model in predicting the quality of carbon fiber holes.This paper proposes a carbon fiber composites hole quality prediction model based on previous studies.While improving the quality of the processing process and thus the product quality,it also provides a basis for the company to establish an effective quality improvement system.Provide effective reference strategies for quality management and improvement of similar enterprises.