Analytical And Experimental Study On Prediction Model For Cone Jet Forming Of Electrohydrodynamic Printing

Electrohydrodynamic printing technology is defined as the process of utilizing printing ink driven by electric field force to generate micro/nano-scale jets to print flexible electronic components,precision sensors,and other micro/nano-scale devices,with the advantages of high resolution,low cost and strong controllability.The stability of the electrohydrodynamic jet forming was affected by ink properties,working voltage,working distance,nozzle size and many other factors,and could be featured by the bending moon surface height and the half angle of Taylor cone.However,an accurate analytical model has not been established to describe the stability of the cone jet process.Therefore,the influence law of process parameters on the bending moon surface height and half angle of Taylor cone in the cone jet process was worth studying,and an accurate analytical model to predict the bending moon surface height and half angle of Taylor cone was established in order to evaluate the stability of the cone jet.In this paper,a stable cone jet prediction model based on backpropagation neural network and genetic algorithm was proposed to judge the stability of cone jet molding by the bending moon surface height and the half cone angle of Taylor cone for the first time.The cone jet mechanism analysis,numerical simulation,experimental research and model building of electrohydrodynamic printing were researched.The main research contents and conclusions were as follows:(1)The forming mechanism of cone jet in electrohydrodynamic jet printing was analyzed,and the method of building prediction model by neural network was presented.The forming process and force of Taylor cone were analyzed,and the process of appearing cone jet at the top of Taylor cone is described in detail.The necessary conditions and influence laws of cone jet forming were explained from the perspectives of critical voltage and critical equivalent flow rate.The method of training the model by artificial neural network was introduced,and genetic algorithm was applied to optimize the artificial neural network for global optimization to improve the model prediction accuracy.(2)The key characterization parameters of electrohydrodynamic printing were simulated by the model of "nozzle-substrate".The "nozzle-substrate" model was optimized design to replace the traditional complex model of electrohydrodynamic printing,and effectively saved computing resources and shortens the simulation time.Taylor cone produced internal reflux,and the edge velocity was higher than the internal velocity which would cause a cone jet.Numerical simulations were carried out to investigate the effects of working voltage,working distance,nozzle inner diameter and print ink conductivity on jet forming.(3)The electrohydrodynamic printing experimental platform was designed to carry out validation experiments.The electrohydrodynamic printing experiment platform was built to carry out the electrohydrodynamic printing experiment by using different conductivity ink.The parameters of Taylor cone edge were extracted by image processing method.The bending moon surface height and half cone Angle of Taylor cone were proposed as evaluation indexes for generating stable cone jet.Under the conditions of conductivity of 4 μS/cm,working voltage of 5 k V,working distance of 2.0 mm and inner diameter of 0.25 mm,stable cone jet is easily formed.(4)A cone-jet forming prediction model with high accuracy and low loss was constructed based on neural network.The BP neural network model with the optimal parameters had an accuracy of 91.52%.The genetic algorithm was used to optimize the BP neural network model,which has a 94.35% accuracy.The accuracy of BPNN model and GANN model was tested with the test set.The accuracy of GANN model in the test set was 94.3%,significantly higher than89.1% of BPNN model.Through mechanism analysis,simulation research and experimental research,the influences of ink conductivity,working distance,working voltage and nozzle size on cone jet forming were investigated.Furthermore,the prediction model of cone jet forming based on genetic algorithm and neural network algorithm was innovatively constructed to accurately predict the cone jet forming under different parameters.The bending moon surface height and the half cone angle of the Taylor cone are accurately predicted by the GANN model.The research results indicated that the GANN model can effectively predict the cone jet forming under different parameter conditions,avoiding a large amount of work such as numerical simulation and repetitive experimental research,and is expected to be promoted in the field of micro/nano precision manufacturing.