Research On Intelligent Decision Method Of Liquid Bridge Force And Liquid Bridge Rupture For Flexible Manipulation Of Micro-Objects

The demand for manipulating micro-objects increases with the miniaturization trend in the fields of microelectromechanical systems,optics,biology and medicine.Micromanipulation is a key technology in the microscopic fields.Micro-objects have a size ranging from millimeters to sub-millimeters or even sub-microns,such as microchips in integrated electronic circuits,micro pumps and micro valves in fluid control fields.Droplet-based micro-object manipulation methods are promising with advantages of operation flexibility,self-centering and self-calibration.To monitor the liquid bridge force and liquid bridge rupture during the manipulation of micro-objects in real time,a method based on artificial neural network to predict liquid bridge force and liquid bridge rupture was proposed.The artificial neural network model was optimized by various optimization algorithms.The validity of the established model was verified by conducting experiments of micro-objects manipulation.For the nonlinear solution of the liquid bridge force between two plates,an artificial neural network model optimized by genetic algorithm was established to predict the liquid bridge force and the contact diameter of the liquid bridge.The theoretical solution of the Young-Laplace equation and the simulation solution based on the minimum energy method are compared.The mean square error MSE and the correlation coefficient R~2 are used to evaluate the prediction accuracy of the established model.The influence of input parameters including the liquid volume and separation distance was analyzed.Based on the weights of the established artificial neural network model,a sensitivity analysis was performed to investigate the influence of input parameters on the liquid bridge force and contact diameter.To investigate the rupture distance and liquid transfer ratio of the quasi-static liquid bridge after rupture,a liquid bridge at critical rupture distance between sphere and spherical concave was established based on the Young-Laplace equation and the minimum energy method to obtain the liquid bridge profile and calculate the transfer ratio.The effects of radius ratio,liquid volume and contact angle on the rupture distance and transfer ratio were analyzed experimentally.The simulation results were compared with the experimental data.An artificial neural network model of predicting rupture distance and transfer ratio of quasi-static liquid bridge was established.Input parameters included the radius ratio and volume.Output parameters included the rupture distance and transfer ratio.MSE calculation and regression analysis were performed to verify the effectiveness of the artificial neural network model in predicting the quasi-static liquid bridge rupture.For the prediction of dynamic stretching of the liquid bridge rupture between two plates,an artificial neural network model was constructed with input parameters including contact angles on the top and bottom plates,liquid volume,stretching speed and liquid viscosity,and output parameters including rupture distance and transfer ratio.The initial weights of the model were optimized using the grey wolf algorithm.The iterative comparison and the prediction comparison with the standard BP algorithm were performed.By calculating MSE and R~2,the effectiveness of the proposed method that the weights of the model optimized by the grey wolf algorithm was demonstrated.The feasibility of the artificial neural network model to predict dynamic stretching of the liquid bridge rupture was verified.A micro-object flexible manipulation platform was built,consisting of a hardware control system and a liquid bridge monitoring system.Experiments on the pick-up operation of micro-object were conducted.By adjusting the separation distance and liquid volume to regulate the liquid bridge force,the pick-up operation of the micro-object was completed.The results shown that the liquid bridge force predicted by the artificial neural network model was in good agreement with the measured liquid bridge force in experiments.The artificial neural network model can be used for the prediction of micro-object pickup,which exhibited an important role for the application of artificial neural network in the liquid drop-based micro-object manipulation.