阅读说明：本技术 一种金属材料加工系统 (Metal material processing system ) 是由 王怀志 于 2021-10-08 设计创作，主要内容包括：一种金属材料加工系统,包括可见光摄像机、红外摄像机、发光二极管、沉积平台、计算机、压电驱动器、喷嘴、感应加热线圈、压电管、注射器、金属粉料箱和开关阀,压电管中的液位将反映施加到金属粉料上的压力,压电管中的金属粉料箱进入喷嘴后,在感应加热线圈的作用下形成液滴,滴向沉积平台上的界面；计算机通过红外摄像机采集液滴的温度并控制感应加热线圈对喷嘴进行加热使金属粉料变为金属液体,计算机控制注射器的推进压力；可见光摄像机捕获由喷嘴喷射的图像,计算机基于喷射的图像进行图像处理和分析；从一系列图像中相应地获得动态液滴行为。(A metal material processing system comprises a visible light camera, an infrared camera, a light emitting diode, a deposition platform, a computer, a piezoelectric driver, a nozzle, an induction heating coil, a piezoelectric tube, an injector, a metal powder box and a switch valve, wherein the liquid level in the piezoelectric tube reflects the pressure applied to metal powder; the computer collects the temperature of the liquid drop through the infrared camera and controls the induction heating coil to heat the nozzle so as to change the metal powder into metal liquid, and the computer controls the propelling pressure of the injector; the visible light camera captures the image sprayed by the nozzle, and the computer performs image processing and analysis based on the sprayed image; dynamic drop behavior is accordingly obtained from a series of images.)

1. A metal material processing system comprises a visible light camera, an infrared camera, a light emitting diode, a deposition platform, a computer, a piezoelectric driver, a nozzle, an induction heating coil, a piezoelectric tube, an injector, a metal powder box and a switch valve, wherein the metal powder box enters the piezoelectric tube from the metal powder box under the pushing of the injector, metal powder is pumped into the piezoelectric tube by adjusting the injector, the liquid level in the piezoelectric tube reflects the pressure applied to the metal powder, liquid drops are formed under the action of the induction heating coil after the metal powder box in the piezoelectric tube enters the nozzle and drop to an interface on the deposition platform, and the deposition platform performs three-dimensional motion under the control of the computer to form an expected part; the computer collects the temperature of the liquid drop through the infrared camera and controls the induction heating coil to heat the nozzle so as to change the metal powder into metal liquid, and the computer controls the propelling pressure of the injector; the computer generates two pulses, the first pulse is used as a driving voltage signal with a trapezoidal waveform to control the piezoelectric driver to apply pressure on the piezoelectric tube so as to enable the metal liquid to be ejected from the nozzle to form liquid drops, the second pulse is used for controlling the light emission of the gated light-emitting diode, the period of the second pulse is the same as that of the first pulse, the delay time between the first pulse and the second pulse is adjustable so as to capture images at required time, when the light-emitting diode is triggered to emit light, the visible light camera captures the images ejected by the nozzle, and the computer performs image processing and analysis based on the ejected images; when the delay time between the first pulse and the second pulse is extended, the image will show the drop falling later, by continuously scanning the delay time, the dynamic drop behavior is correspondingly obtained from a series of images, the nozzle and the visible camera are fixed in position, the deposition platform moves relatively, and the drop characteristics are not affected by the movement of the nozzle.

2. A metallic material processing system as set forth in claim 1, wherein: the jetting state of the liquid drop depends on the parameters including driving voltage, nozzle orifice, propelling pressure, liquid viscosity and surface tension, in order to control the manufacturing process to minimize defects, a closed-loop controller integrating image processing and a neural network is built in a computer, the closed-loop controller includes a feature mapping module, a neural network model, a PID unit and an image processing module, a target feature is formed by referring to a liquid drop feature input feature mapping module, the neural network model establishes a relationship between a voltage level as an output value and a liquid drop feature as an output value, jetting behavior is mapped to a controllable voltage value without manual intervention, an image sequence of the liquid drop captured by a visible light camera is fed back to the image processing module, the liquid drop feature is extracted, the liquid drop feature is compared with the target feature by using the neural network model, the PID unit is used to adjust the drive voltage, applied to the piezo actuator, to adjust the ejection of the droplets, and the above process is repeated until the desired ejection behaviour is reached.

3. A metallic material processing system as set forth in claim 1, wherein: the computer obtains each temperature through the infrared camera to control the heating coil to carry out temperature control to the liquid drop, make the production of avoiding the hole under the condition that calculates the non-dimensional temperature parameter and can form flat surface, make the work piece surface of production comparatively smooth.

4. A metal-cutting machining system according to claim 1, characterized in that: the method comprises the steps that a visible light camera obtains an image, the image is processed by an image processing module to obtain a binary image based on liquid drops, a color image is converted into a gray image according to the brightness of the image, the gray image is converted into the binary image by setting a proper threshold value, the value of a pixel is 0 or 1, the value higher than the threshold value is mapped to 1, the value lower than the threshold value is mapped to 0, the liquid drops are extracted and analyzed from a gray background image by using the binary image, the background is 1 or 0 when the pixel of the liquid drops is 0 or 1, the number, the size and the position of the liquid drops are analyzed, and the image area is a 2D array of pixel data.

5. A metallic material processing system as set forth in claim 2, wherein: the visible light camera captures the ejected liquid drop images to form a liquid drop image sequence, the image processing module analyzes each liquid drop image respectively, liquid drop characteristics are extracted from each liquid drop image, and the characteristics extracted from the liquid drop images in different time step lengths are gathered into attributes through a group of rules; for a given liquid drop image, defining an interest area below a nozzle of the liquid drop image, selecting a threshold value through a histogram, distinguishing the foreground and the background of the liquid drop image, acquiring the image characteristics of liquid drops in the foreground, and forming an image sequence by a plurality of liquid drop images; the image processing module extracts droplet characteristics, the droplet characteristics comprise scattering points, toughness, speed and volume, the volume and the toughness express the geometric information of the droplets, the scattering points express the morphological state of the droplets, the speed expresses the motion behavior of the droplets, a single characteristic is insufficient to determine the process state, a plurality of characteristics are utilized to represent random defects, and the sensitivity and the accuracy of the system are improved.

6. A metallic material processing system as set forth in claim 5, wherein: the droplet characteristics are expressed as:

the scatter is equal to the maximum value of the number of connected components in all static images at a certain time;

the magnitude of the toughness is defined as the maximum value of the aspect ratio of each connected component in all images;

the total volume of the drop is the maximum of the total area of the component in all images;

when the scatter and spray velocity suddenly decrease over time, the droplets leave the region of interest, and when the scatter suddenly decreases, the toughness suddenly increases, the volume suddenly increases, or the velocity suddenly increases, an uncontrollable random disturbance occurs.

7. A metallic material processing system as set forth in claim 2, wherein: the neural network model adjusts the drive voltage to minimize defects caused by random variables, establishes a relationship between droplet features and the drive voltage, and during training of the neural network model, all input values are features extracted from droplet images in a training data set, and target data are drive voltage levels under which each droplet image is captured. A trained neural network model is used to characterize offset voltage values based on the acquired drop images to adjust the drive voltage applied to the piezoelectric actuator.

8. A metallic material processing system as set forth in claim 3, wherein the detection and control of the temperature of the liquid droplets is embodied as:

firstly, establishing a liquid drop model, and establishing a function representing the volume fraction of fluid of each grid unit, an incompressible fluid continuity equation of the liquid drop, a momentum conservation equation of a fluid velocity component, a surface tension model and an energy-saving equation;

and secondly, establishing a parameter relation of metallurgical bonding between adjacent droplets and analyzing the interaction between the adjacent droplets. The dynamic behavior of the liquid drop impacting the interface surface is characterized by a series of dimensionless parameters including Weber number, Auger number and Reynolds number, and three different conditions of the liquid drop impacting the interface are characterized by Sophia parameters: rebound, sedimentation and splashing; in order to obtain a smooth surface and avoid the generation of hole defects, good metallurgical bonding between adjacent droplets should be achieved, interfaces between adjacent droplets should be remelted and fused together in a continuous deposition process, and the remelting behavior is estimated by the interface temperature;

thirdly, analyzing the formation of holes, establishing dimensionless temperature parameters and dimensionless time, quantitatively describing the diffusion of the liquid drops in the continuous deposition process, defining the dimensionless parameters as diffusion factors, and analyzing the melting states of the liquid drops with different dimensionless temperature parameters;

and fourthly, controlling the temperature of the heating coil to meet the dimensionless temperature parameter without defects.

9. A metal material processing system according to claim 4, wherein the algorithm of image processing is: determining a suitable threshold, the position of the droplet being identified according to pixel positions in the droplet image, the region of interest defining local pixel positions such that the upper left corner position of the region of interest becomes a local origin position, the droplet position being locally identified in the user-defined region of interest, the upper left corner pixel position of the region of interest being added to the locally identified position such that the identified droplet position remains unchanged; identifying the position and the diameter of the liquid drop by overlapping in the original image to check whether the identified result is matched with the liquid drop image; to calculate the velocity of the droplet, it is necessary to convert the analysis result into a distance in μm, one pixel corresponding to about 1.03 μm, calculate the droplet velocity and estimate the droplet position.

Technical Field

The invention belongs to the field of processing, and particularly relates to a metal material processing system.

Background

The use of molten metal for deposition processes can create complex objects with a variety of materials and functions, opening up a great deal of opportunity for a wide range of applications including aerospace, automotive, defense and biomedical industries. There are a number of problems with the deposition process of most current metallic materials: (1) during the layer-by-layer deposition process, if the process deviation is not corrected in time, the defects are propagated to the subsequent layers, so that the functional integrity (fatigue property, strength property and geometric integrity) of the part is influenced, the problems of low reliability and low quality are caused, and the large-scale commercialization is not facilitated; (2) parts deposited by metal droplets are often rough in surface due to their solidification morphology and the step effect, and rough surfaces prevent this technique from being used to manufacture parts with high quality surface finish requirements, particularly certain microwave devices, such as waveguides and feedhorns, which require smooth interior surfaces to ensure the propagation of electromagnetic waves; (3) to be able to control the generation of droplets requires very accurate measurement and control of ejection properties such as droplet velocity and droplet volume, there is currently a lack of accurate droplet capture.

Disclosure of Invention

In order to solve the above problems, the present invention provides a metal material processing system which overcomes the deviation in the processing process by precisely capturing the form of droplets and controlling the generation of droplets by a control system, and thus ensures the integrity of parts and prevents the occurrence of holes in the parts.

In order to achieve the purpose, the technical scheme of the invention is as follows: a metal material processing system comprises a visible light camera, an infrared camera, a light emitting diode, a deposition platform, a computer, a piezoelectric driver, a nozzle, an induction heating coil, a piezoelectric tube, an injector, a metal powder box and a switch valve, wherein the metal powder box enters the piezoelectric tube from the metal powder box under the pushing of the injector, metal powder is pumped into the piezoelectric tube by adjusting the injector, the liquid level in the piezoelectric tube reflects the pressure applied to the metal powder, liquid drops are formed under the action of the induction heating coil after the metal powder box in the piezoelectric tube enters the nozzle and drop to an interface on the deposition platform, and the deposition platform performs three-dimensional motion under the control of the computer to form an expected part; the computer collects the temperature of the liquid drop through the infrared camera and controls the induction heating coil to heat the nozzle so as to change the metal powder into metal liquid, and the computer controls the propelling pressure of the injector; the computer generates two pulses, the first pulse is used as a driving voltage signal with a trapezoidal waveform to control the piezoelectric driver to apply pressure on the piezoelectric tube so as to enable the metal liquid to be ejected from the nozzle to form liquid drops, the second pulse is used for controlling the light emission of the gated light-emitting diode, the period of the second pulse is the same as that of the first pulse, the delay time between the first pulse and the second pulse is adjustable so as to capture images at required time, when the light-emitting diode is triggered to emit light, the visible light camera captures the images ejected by the nozzle, and the computer performs image processing and analysis based on the ejected images; when the delay time between the first pulse and the second pulse is extended, the image will show the drop falling later, by continuously scanning the delay time, the dynamic drop behavior is correspondingly obtained from a series of images, the nozzle and the visible camera are fixed in position, the deposition platform moves relatively, and the drop characteristics are not affected by the movement of the nozzle.

Wherein the jetting state of the liquid droplet depends on the parameters including driving voltage, nozzle orifice, propelling pressure, liquid viscosity and surface tension, in order to control the manufacturing process to minimize defects, a closed-loop controller integrating image processing and a neural network is built in a computer, the closed-loop controller includes a feature mapping module, a neural network model, a PID unit and an image processing module, a reference liquid droplet feature input feature mapping module forms target features, the neural network model establishes a relationship between a voltage level as an output value and liquid droplet features as output values, jetting behavior is mapped to controllable voltage values, without manual intervention, an image sequence of the liquid droplet captured by a visible light camera is fed back to the image processing module, the liquid droplet features are extracted, the liquid droplet features are compared with the target features by using the neural network model, the PID unit is used to adjust the drive voltage, applied to the piezo actuator, to adjust the ejection of the droplets, and the above process is repeated until the desired ejection behaviour is reached.

In the machining process, the computer acquires each temperature through the infrared camera and controls the heating coil to control the temperature of the liquid drops, so that the calculation of dimensionless temperature parameters can avoid the generation of holes under the condition of forming a flat surface, and the surface of a generated workpiece is smooth.

The method comprises the steps of obtaining a binary image based on liquid drops after an image obtained by a visible light camera is processed by an image processing module, converting a color image into a gray-scale image according to the brightness of the image, converting the gray-scale image into the binary image by setting a proper threshold value, wherein the value of a pixel is 0 or 1, the value higher than the threshold value is mapped into 1, the value lower than the threshold value is mapped into 0, extracting and analyzing the liquid drops from a gray background image by using the binary image, the background is 1 or 0 if the pixel of the liquid drops is 0 or 1, and the number, the size and the position of the liquid drops are analyzed, and an image area is a 2D array of pixel data.

The invention has the beneficial effects that:

the processing system of the invention is provided with a monitoring system consisting of a camera, a light emitting diode, a computer, a piezoelectric driver, a nozzle and an induction heating coil, has the functions of droplet detection based on vision, neural network technology and PID technology, realizes the monitoring of droplet ejection, including the volume, the speed and the position in ejection of the droplets, and ensures the geometric and functional integrity of processed parts. The system can capture high fidelity data of droplets and extract critical information for downstream decision making for correction, ultimately improving process reliability, and enablingRepeatability ofAndquality of processing. The system continuously monitors and adjusts the pattern of droplets during processingApplied drive voltageLevel to compensate for ringDifferences between the observed pattern and the desired pattern due to environmental changes and accidents effectively solve the problems encountered during processing. Opens up a new way for additive manufacturing. According to the invention, the temperature information is collected through the infrared camera, the fluid equation is established, the dimensionless temperature parameter and the dimensionless time temperature are established, so that the temperature control is realized, and the heating coil is controlled to control the temperature of the liquid drop, so that the flat surface of the workpiece is ensured, and the generation of holes is avoided; the invention measures the speed of the liquid drop through the high-speed camera, detects the sudden change of the pixel intensity value along the line of the region of interest by utilizing a detection algorithm, identifies the position and the size of the liquid drop, calculates the speed of the liquid drop through the images of two light-emitting diodes with two different trigger delay times, and the two different trigger delay times are synchronous with the sprayed signal.

Drawings

FIG. 1 is a system engineering block diagram of the present invention;

FIG. 2 is a schematic block diagram of a closed loop controller of the present invention;

Detailed Description

The invention is further described with reference to the following figures and examples.

Embodiments of the present invention are illustrated with reference to fig. 1-2.

A metal material processing system comprises a visible light camera, an infrared camera, a light emitting diode, a deposition platform, a computer, a piezoelectric driver, a nozzle, an induction heating coil, a piezoelectric tube, an injector, a metal powder box and a switch valve, wherein the metal powder box enters the piezoelectric tube from the metal powder box under the pushing of the injector, metal powder is pumped into the piezoelectric tube by adjusting the injector, the liquid level in the piezoelectric tube reflects the pressure applied to the metal powder, liquid drops are formed under the action of the induction heating coil after the metal powder box in the piezoelectric tube enters the nozzle and drop to an interface on the deposition platform, and the deposition platform performs three-dimensional motion under the control of the computer to form an expected part; the computer collects the temperature of the liquid drop through the infrared camera and controls the induction heating coil to heat the nozzle so as to change the metal powder into metal liquid, and the computer controls the propelling pressure of the injector; the computer generates two pulses, the first pulse is used as a driving voltage signal with a trapezoidal waveform to control the piezoelectric driver to apply pressure on the piezoelectric tube so as to enable the metal liquid to be ejected from the nozzle to form liquid drops, the second pulse is used for controlling the light emission of the gated light-emitting diode, the period of the second pulse is the same as that of the first pulse, the delay time between the first pulse and the second pulse is adjustable so as to capture images at required time, when the light-emitting diode is triggered to emit light, the visible light camera captures the images ejected by the nozzle, and the computer performs image processing and analysis based on the ejected images; when the delay time between the first pulse and the second pulse is extended, the image will show the drop falling later, by continuously scanning the delay time, the dynamic drop behavior is correspondingly obtained from a series of images, the nozzle and the visible camera are fixed in position, the deposition platform moves relatively, and the drop characteristics are not affected by the movement of the nozzle.

Wherein the jetting state of the liquid droplet depends on the parameters including driving voltage, nozzle orifice, propelling pressure, liquid viscosity and surface tension, in order to control the manufacturing process to minimize defects, a closed-loop controller integrating image processing and a neural network is built in a computer, the closed-loop controller includes a feature mapping module, a neural network model, a PID unit and an image processing module, a reference liquid droplet feature input feature mapping module forms target features, the neural network model establishes a relationship between a voltage level as an output value and liquid droplet features as output values, jetting behavior is mapped to controllable voltage values, without manual intervention, an image sequence of the liquid droplet captured by a visible light camera is fed back to the image processing module, the liquid droplet features are extracted, the liquid droplet features are compared with the target features by using the neural network model, the PID unit is used to adjust the drive voltage, applied to the piezo actuator, to adjust the ejection of the droplets, and the above process is repeated until the desired ejection behaviour is reached.

The ideal jetting behavior is that each pulse of the input signal generates only one single drop of sufficient volume, and no scatter follows.

The visible light camera captures the jetted liquid drop images to form a liquid drop image sequence, the image processing module analyzes each liquid drop image respectively, liquid drop characteristics are extracted from each liquid drop image, and the characteristics extracted from the liquid drop images in different time step lengths are gathered into attributes through a group of rules; for a given liquid drop image, defining an interest area below a nozzle of the liquid drop image, selecting a threshold value through a histogram, distinguishing the foreground and the background of the liquid drop image, acquiring the image characteristics of liquid drops in the foreground, and forming an image sequence by a plurality of liquid drop images; the image processing module extracts droplet characteristics, the droplet characteristics comprise scattering points, toughness, speed and volume, the volume and the toughness express the geometric information of the droplets, the scattering points express the morphological state of the droplets, the speed expresses the motion behavior of the droplets, a single characteristic is insufficient to determine the process state, a plurality of characteristics are utilized to represent random defects, and the sensitivity and the accuracy of the system are improved.

Wherein, when the amplitude of the driving voltage is increased, the volume and the speed of the liquid drop are increased,wherein, V_pIs the volume of the piezoelectric tube, Δ V_pIs the volume change of the piezoelectric tube, d_εIs the piezoelectric strain constant, U is the drive voltage, t_nThe thickness of the piezoelectric tube is equal to that of the piezoelectric tube, the piezoelectric driver is arranged on the piezoelectric tube, the volume change of the piezoelectric tube is increased due to higher voltage, and the metal liquid is sprayed out due to the extrusion of the piezoelectric tube.

In practice droplet characteristics are much more complex and the same droplet characteristics may be caused by different dispensable factors and random disturbances, such as: the abnormal velocity of the drop caused by the instability and uncontrollable random deflection of the back pressure is; the volume of the liquid drop is not enough to form a sharp drop to increase the number of scattered points and cannot form a liquid drop due to the lower driving voltage and the lower deposition rate; the droplets are subjected to fluid dynamics and the same droplet will have different volume characteristics at different times.

Wherein m are differentImage of a drop at a time in time sequence t_mAnd ordered in each still image from top to bottomA connected component representing a single drop appearing in the image, the drop characteristic then being represented as:

the scatter San is equal to the number # seg of connected components in all still images at time t_tMax { # seg, namely San ═ max { # seg_t}；

The size of toughness Ren is defined as the maximum value of the aspect ratio of each connected component in all images, i.e., Ren max h_i,t/w_i,t},Wherein h is_i,tDenotes the height, w, of the ith connecting component at time t_i,tRepresents the width of the ith connecting component at the moment t;

the total volume Vol of the drop is the maximum of the total area of the components in all images, i.e., Vol max {. Sigma_iA_i,1,∑_iA_i,2,...,∑_iA_i,t}. Wherein A is_i,tShowing the volume of the ith connecting component at the time t;

the evaluation values from the above steps are used for screening purposes, and when the scatter and jetting speed suddenly decrease with time t, the droplet leaves the region of interest, and when the scatter suddenly decreases, the toughness suddenly increases, the volume suddenly increases, or the speed suddenly increases, it indicates that an uncontrollable random disturbance has occurred.

Among them, in order to analyze the droplet characteristics, the visible light camera captures images of droplets ejected continuously at a frequency of 300Hz, the continuous droplets have the same behavior in a short time, more images are captured more robustly, and the minimum number of captured images is 3.

The neural network model adjusts the driving voltage, defects caused by random variables are reduced to the maximum extent, the neural network model establishes a relation between droplet characteristics and the driving voltage, in the training process of the neural network model, all input values are characteristics extracted from droplet images in a training data set, and target data are driving voltage levels under which each droplet image is captured. A trained neural network model is used to characterize offset voltage values based on the acquired drop images to adjust the drive voltage applied to the piezoelectric actuator.

In order to train a neural network model, collecting droplet images under different driving voltages and labeling one by one, firstly, as an initialization step, setting to obtain a standard droplet mode, wherein the standard droplet mode is San-1 and Ren-1, the driving voltage is 45V, and the spraying frequency is 300 Hz; the visible light camera captures the droplet images and adjusts the driving voltage, the voltage is adjusted step by an offset value, for example, 1V, 2V, 3, …, 10V, and marks the corresponding droplet images to form an output; finally, four features are extracted from the drop image, including scatter, toughness, volume, and velocity, which are used as inputs to the network; this process is repeated several times to minimize the impact of incorrect initial settings; the above procedure was repeated and 800 sets of data were collected, 70% for training, 15% for validation and another 15% for testing.

Wherein the output of the neural network module is an offset voltage value used to characterize the injection behavior, the input to the PID unit regulates the drive voltage applied to the piezo actuator, and the PID control equation is:

where err (t) is the feedback value output from the neural network, op is the output voltage, K_P、K_IAnd K_DProportional coefficient, integral coefficient and differential coefficient, after PID parameter is adjusted, the PID unit controls the driving voltage to realize stable liquid drop ejection.

Wherein, in the course of working, the computer acquires each temperature through infrared camera to control heating coil carries out temperature control to the liquid drop, makes to calculate dimensionless temperature parameter and can avoid the production of hole under the condition that forms flat surface, makes the workpiece surface of production comparatively smooth, detects and controls the temperature of liquid drop specifically to be:

firstly, establishing a liquid drop model, wherein the speed of the liquid drop before impacting an interface is constant, the temperature of the liquid drop before impacting the interface is uniform and constant, the liquid drop is incompressible liquid, the convection and radiation of the liquid drop are ignored, and the shrinkage and deformation of the liquid drop in the solidification process after deposition are ignored; representing the fluid volume fraction of each grid cell using the function F, where F1 represents the grid cell filled with liquid; and F ═ 0 indicates that the grid cell is empty, the function F satisfies:

wherein, V_FIs the volume fraction of the flow, and u, v and w are the velocity components in the x, y and z directions, respectively, of the cartesian coordinate system. A. the_x，A_yAnd A_zFractional regions of the fluid in the x, y and z directions, respectively; the incompressible fluid continuity equation for a droplet is:

where ρ is_dIs the density of the droplet.

The conservation of momentum equation for the fluid velocity components (u, v, w) is expressed as:

wherein G is_x、G_yAnd G_zIs the volumetric force acceleration. S_x、S_yAnd S_zIs the surface tension, K is the solidification resistance,

wherein F_sIs the solids content of the fluid:

wherein T is_{Liquid for treating urinary tract infection}，T_{Fixing device}And T is the liquidus temperature, solidus temperature and instantaneous temperature, respectively;

surface tension S (S)_x,S_y,S_z) Expressed as: s ═ σ κ δ n;

where σ and κ are the surface tension coefficient and the curvature of the free surface, respectively, the distribution function δ represents the surface tension centered on the free surface of the drop, and n is the unit normal vector to the free surface of the drop, then:

the energy-saving equation is as follows:

where I is the macroscopic internal energy expressed as, then: i ═ C_lgT+(1-F_s)gCLHT

Wherein, C_lAnd CLHT is the specific heat capacity and latent heat of fusion, respectively;

T_DIFis the heat conduction spread, expressed as:where k is the thermal conductivity of the fluid.

The molten droplets form pores on the interface surface due to air entrapment and local solidification of the droplets, and air entrapment is a contributing factor to the internal surface morphology in wick-supported droplet deposition.

And secondly, establishing a parameter relation of metallurgical bonding between adjacent droplets and analyzing the interaction between the adjacent droplets. The dynamic behavior of a droplet impacting an interface surface is characterized by a series of dimensionless parameters, including the weibo number (We), the oerg number (Oh), and the reynolds number (Re), defined as:

V_dσ, μ and D_dRespectively representing the velocity, surface tension coefficient, dynamic viscosity and droplet diameter of the droplets, and the impact interface of the droplets is divided into three different cases: rebound, deposition and spattering. These three droplet behaviors are characterized by the Sofitie parameter K, which is defined as:when the value of K is equal to K ≈ 7, the droplets deposit without bounce or splash.

Wherein, in order to obtain a smooth surface without hole defects, a good metallurgical bond between adjacent droplets should be achieved, and in a continuous deposition process the interfaces between adjacent droplets should be remelted and fused together, passing through the interface temperature T_{Boundary of China}To estimate the remelting behaviour:

wherein, T_dIs the initial drop temperature; r_c；k_sIn order to be a thermal conductivity coefficient,

T_{watch (A)}And alpha_sSurface temperature and thermal diffusivity, respectively, T heat transfer time, by setting T_{Boundary of China}＝T_{Fixing device}And T_{Boundary of China}＝T_{Liquid for treating urinary tract infection}The formed parameter map is used for predicting the remelting temperature condition between adjacent metal drops; when T is_{Boundary of China}≥T_{Fixing device}At the time, remelting and joining between adjacent metal droplets starts; when T is_{Boundary of China}≥T_{Liquid for treating urinary tract infection}In time, adjacent droplets in the interface region remain in an intact liquid state during the fusion process, resulting in excessive remelting conditions.

Thirdly, analyzing the formation of the holes, establishing a dimensionless temperature parameter beta and a dimensionless time t^*Is defined as follows

Wherein T is_mDenotes the fusion point temperature, T, of the metal_sDenotes the interface temperature, T_dRepresents the droplet temperature;

dimensionless time t^*The calculation is as follows:

t^*0 denotes the transient contact between the first drop and the second drop, V_dMeans the average statistical velocity of the drops, D_dIndicating the droplet diameter. Parameter set (T) according to the above parameter mapping analysis_d＝1423K,T_s323K) are located in the re-melt zone, indicating that the interfacial regions of adjacent droplets will re-melt and fuse together in a continuous deposition process.

When β is 1.3, the following is the deposition process of four adjacent droplets, at t^*At 0, the second droplet starts to spread along the first droplet and at the same time melting takes place in the contact area on the respective surface of the first droplet. From t^*Starting at 0.2, the second drop started to contact and spread out on the substrate. The heat conduction at the interface rapidly cools the bottom region of the second droplet, t^*At 0.46, the solid fraction is almost equal to 100% everywhere on the bottom surface of the second drop, the deposition interface is completely static, incomplete filling of the liquid metal leads to the formation of a cold bend between the two adjacent drops, when We and Oh are 1.42 and 1.02 x 10, respectively^-3When the liquid drop is dispersed, the liquid drop is driven by the conversion of the kinetic energy and the surface energy of the liquid drop; t is t^*The solidification front of the first drop moves upward in an L-shaped manner, and the residual liquid phase in the second drop remains in a tilted oscillation until its kinetic energy is completely dissipated by solidification. The deposition behavior of the third droplet is similar to the deposition behavior of the second droplet; the third droplet is mostly impinged and spread on the bottom surface of the first droplet; t is t^*1000.2, the third drop comes into contact with the interface and most of the liquid drips onto the interface, a cold trap can form on the bottom surface of the third and first drops due to incomplete filling of the liquid metal, and the fourth drop simultaneously with the first dropThe second and third droplets partially overlap and the portion of the fourth droplet that flows directly onto the interface is the smallest, the surface area of the fourth droplet is the smallest, and the center of the bottom surface surrounded by four adjacent droplets is the most difficult point to fill. As a result, the hole located at the center is the most, indicating remelting and poor fusion between droplets.

To prevent the formation of holes on the bottom surface of the continuously deposited droplets, a more precise temperature selection considering heat dissipation is required.

To quantitatively describe the spreading of droplets during continuous deposition, a dimensionless parameter ξ is defined as the spreading factor, defined as:wherein S is_BottomAnd S_{Top roof}Respectively the area of the bottom surface of the deposited droplet and the area of the top surface of the droplet. When β is 5.6, the diffusion factors of four consecutive deposited droplets are increased by 13.7%, 6.1%, 6% and 7%, respectively, compared to the case of β being 1.3, and the solidification times of the bottom layers of the second, third and fourth droplets are increased by 0.58, 0.58 and 0.46, respectively, indicating that the remaining liquid metal of the new droplet has more time to fill the gap surrounded by four adjacent droplets, in which case the holes on the bottom surface are effectively eliminated due to the guaranteed complete filling of the remaining liquid metal; when beta is 8.2, the holes on the bottom surface can be effectively eliminated;

when β is 10.9, the degree of remelting between adjacent droplets is too high to effectively form a flat surface. Multiple droplets are randomly merged together into larger droplets, which results in the formation of a discontinuous surface.

Fourthly, controlling the temperature of the heating coil to enable the temperature to meet a dimensionless temperature parameter beta without defects, wherein beta is more than or equal to 1.3 and less than 10.9.

Preferably 1.3. ltoreq. beta. ltoreq.8.2, further: β ═ 1.3 or β ═ 8.2.

The method comprises the steps of obtaining a binary image based on liquid drops after an image obtained by a visible light camera is processed by an image processing module, converting a color image into a gray-scale image according to the brightness of the image, converting the gray-scale image into the binary image by setting a proper threshold value, wherein the value of a pixel is 0 or 1, the value higher than the threshold value is mapped into 1, the value lower than the threshold value is mapped into 0, extracting and analyzing the liquid drops from a gray background image by using the binary image, the background is 1 or 0 if the pixel of the liquid drops is 0 or 1, and the number, the size and the position of the liquid drops are analyzed, and an image area is a 2D array of pixel data.

Processing time is reduced by analyzing the region of interest rather than the entire image area.

The algorithm of image processing is as follows: determining a suitable threshold k that satisfies:

wherein mu₁Is the average of all pixel values between 0 and k, μ₂Is the average of all pixel values lying between k +1 and 255.

The position of the droplet is identified from pixel positions in the droplet image, the region of interest defines local pixel positions such that an upper left corner position of the region of interest becomes a local origin position, the droplet position is locally identified in the user-defined region of interest, and the upper left corner pixel position of the region of interest is added to the locally identified position such that the identified droplet position remains unchanged.

The droplet position and diameter are identified by superimposition in the original image to check whether the identified result matches the droplet image.

To calculate the velocity of the droplet, the analysis result needs to be converted into a distance in μm. One pixel corresponds to about 1.03 μm, the drop velocity v is calculated:

wherein (Px)₁，Py₁) And (Px)₂、Py₂) And t is the center of the drop in the x and y directions, respectively₁And t₂Image of timeThe pixel position, the ejection direction theta is evaluated from the identified drop position,

the alignment of the center of the region of interest with respect to the droplet ejection direction, when the center of the region of interest is 3 ° off the ejection direction, the estimated velocity has an error rate of about 3%.

The above-described embodiment merely represents one embodiment of the present invention, but is not to be construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention.