In the manufacturing process of high-precision electronic devices through inkjet printing,precise control of droplet status is crucial for device performance.However,acquiring a large amount of droplet status data can be time-consuming and labor-intensive.Therefore,a prediction method for droplet status based on small sample data was proposesd in this study.Firstly,a single-pole trapezoidal wave was used to drive the nozzle to produce droplets in different states as sample data by changing the amplitude and duration of the driving voltage.Then,five methods including Support Vector Machine(SVM),Back Propagation Neural Network(BPNN),Decision Tree(DT),Random Forest(RF)and Extreme Gradient Boosting(XGBoost)were used to predict the droplet status after preprocessing.Experimental results showed that the XGBoost method has a significant advantage in small sample situations,with an average absolute error,root mean square error,and average absolute percentage error of 0.088,0.123,and 1.85%,respectively,all being the best among the five methods.The XGBoost method can be used as a prediction method for droplet status of inkjet printing based on small sample data.
Small samplePrinted electronicsDroplet statusExtreme Gradient Boosting
NETL
NSTL
万方数据
