An intelligent control method and system for distribution uniformity of electric atomization deposition

Abstract

The present application belongs to the field of printed display technology, and particularly relates to an intelligent control method and system for the uniformity of electric atomization deposition, comprising: obtaining a target deposition distribution uniformity and process parameters with their values that can be determined by the current printing process, inputting the process parameters and their values into a PID gain coefficient prediction model to obtain PID related gain coefficients and determine a PID control model; inputting the process parameters with their values and to-be-optimized process parameters with their initial values into a deposition distribution uniformity prediction model to obtain the deposition distribution uniformity under the current process parameter value combination; based on the target and the current predicted deposition distribution uniformity, updating the values of the to-be-optimized process parameters by using the PID control model; based on the to-be-optimized process parameters and their updated values, repeating the above process until the difference between the predicted deposition distribution uniformity and the target deposition distribution uniformity is within a preset range, and then being used for printing, the present application can adaptively adjust the printing parameter values to realize precise control of the uniformity.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of printed display technology, and more particularly, relates to an intelligent control method and system for deposition distribution uniformity of electro-spraying. BACKGROUND

[0002] The inkjet printing process is an additive manufacturing process that can directly spray micron-sized ink droplets onto a flexible or rigid substrate. Compared with traditional thin film preparation techniques such as evaporation and atomic layer deposition, inkjet printing technology has the advantages of high material utilization, non-contact manufacturing, high manufacturing precision, and flexible manufacturing, and is one of the technologies widely discussed and researched in the field of thin film preparation.

[0003] In inkjet printing technology, electrohydrodynamic atomization technology is a technology with good effect and good application prospect. In the electrohydrodynamic atomization process, the printing parameters have a great influence on the printing quality, and the mathematical rules are weak. Currently, the selection of parameters is usually based on the experience of engineers, and the printing parameters are adjusted manually by the quality of the printed film. Therefore, in the electrohydrodynamic atomization printing process, how to quickly select appropriate printing parameters to make the deposition distribution uniformity of printing meet the expectations has become a problem of concern.

[0004] Some feedback adjustment schemes for electrohydrodynamic atomization technology have been proposed in the prior art. For example, patent CN111106267A proposes a feedback adjustment scheme for electro-spraying process parameters by combining multiple modules to realize the control of the thickness and uniformity of the atomized liquid film. However, the feedback system proposed in the patent is based on an empirical formula, and its accuracy cannot be guaranteed, especially when a new solution is used, the process parameters need to be changed. Comprehensive research shows that the existing patents and papers mainly have the following shortcomings: the current electrohydrodynamic atomization control system often directly uses a self-defined empirical formula for control, and the setting of the empirical parameters is entirely based on the experience of engineers, which cannot adapt to changes in process environment, etc.

[0005] Therefore, how to design a new intelligent control method that can automatically adapt to changes in data to accurately control the deposition distribution uniformity of electrohydrodynamic atomization has become a key technical problem that needs to be solved in the field. SUMMARY

[0006] In view of the defects and improvement needs of the prior art, the present application provides an intelligent control method and system for deposition distribution uniformity of electro-spraying, which aims to solve the problem that the existing method cannot self-adaptively adjust the value of the printing parameter in electro-spraying printing, thereby limiting the control precision of the deposition distribution uniformity.

[0007] To achieve the above object, according to one aspect of the present application, an intelligent control method for deposition distribution uniformity of electric atomization is provided, comprising:

[0008] S1, obtaining a target deposition distribution uniformity and current printing process parameter whose value is determinable and the value thereof, and inputting a PID gain coefficient prediction model to predict a proportional gain coefficient, an integral gain coefficient and a differential gain coefficient for PID control, so as to determine a PID control model;

[0009] S2, inputting the process parameter whose value is determinable and the value thereof and a process parameter to be optimized and an initial value thereof into a deposition distribution uniformity prediction model to predict a deposition distribution uniformity under a current process parameter value combination;

[0010] S3, based on the target deposition distribution uniformity and the deposition distribution uniformity under the current process parameter value combination, updating the value of the process parameter to be optimized by using the PID control model;

[0011] S4, based on the process parameter whose value is determinable and the value thereof and the process parameter to be optimized and the updated value thereof, repeating S2 until a difference between the deposition distribution uniformity predicted by the deposition distribution uniformity prediction model and the target deposition distribution uniformity is within a preset range;

[0012] S5, performing electric atomization printing based on the process parameter whose value is determinable and the value thereof and the final value of the process parameter to be optimized to realize intelligent control of deposition distribution uniformity of electric atomization.

[0013] Further, the deposition distribution uniformity prediction model is constructed by using the following method:

[0014] The value range of the process parameter whose value is determinable and the process parameter to be optimized is respectively divided into equal steps to construct multiple sets of process parameter values;

[0015] Actual electric atomization printing is performed under each set of process parameter value combination, and m deposition images are collected in the stable electric atomization printing area, wherein the image collection method is as follows: one or more shooting paths perpendicular to the movement direction of the electric atomization nozzle are determined in the stable electric atomization printing area, and multiple deposition images are continuously and uninterruptedly collected from one end of each shooting path perpendicular to the movement direction of the electric atomization nozzle to the other end;

[0016] Based on the m deposition images corresponding to each set of process parameter value combination, the deposition distribution uniformity under the set of process parameter value combination is determined, and each set of process parameter value combination and the corresponding deposition distribution uniformity is taken as a training sample to construct a training sample set;

[0017] Train the neural network by using the training sample set to obtain the deposition distribution uniformity prediction model.

[0018] Further, the deposition distribution uniformity is characterized by a coefficient of variation of the electric atomization deposition distribution rate, wherein the coefficient of variation is a ratio between a standard deviation and a mean value of m electric atomization deposition distribution rates calculated from m deposition images.

[0019] Further, the PID gain coefficient prediction model is obtained by using the following method:

[0020] A plurality of target deposition distribution uniformities and values of the process parameters with determinable values are constructed.

[0021] For each target deposition distribution uniformity and value of the process parameter with a determinable value, a plurality of values of proportional gain coefficients, integral gain coefficients and differential gain coefficients are set; based on the value of the process parameter with a determinable value and a preset value of the to-be-optimized parameter in the group, the deposition distribution uniformity prediction model is used to obtain a corresponding deposition distribution uniformity; based on the deposition distribution uniformity and the value of the target deposition uniformity in the group, a PID control model composed of the values of the proportional gain coefficients, the integral gain coefficients and the differential gain coefficients corresponding to the group is used to update the value of the to-be-optimized parameter; based on the value of the process parameter with a determinable value and each updated value of the to-be-optimized parameter in the group, the deposition distribution uniformity prediction model is used to predict the deposition distribution uniformity, and the gap between each predicted deposition distribution uniformity and the target deposition uniformity is compared; the values of the proportional gain coefficients, the integral gain coefficients and the differential gain coefficients corresponding to the smallest gap are taken as the label corresponding to the target deposition distribution uniformity and the value of the process parameter with a determinable value, and a training sample is constructed.

[0022] The PID gain coefficient prediction model is trained by using each training sample.

[0023] Further, the gap corresponding to the values of the proportional gain coefficients, the integral gain coefficients and the differential gain coefficients as the label is also less than 3%.

[0024] Further, the to-be-optimized process parameter is voltage, and the process parameters with determinable values include atomization gas pressure and height of the nozzle to the deposition substrate.

[0025] The application also provides an electric atomization deposition distribution uniformity intelligent control system for executing the electric atomization deposition distribution uniformity intelligent control method, which comprises a process parameter optimization module and a printing module.

[0026] The process parameter optimization module is configured to perform value optimization of the to-be-optimized parameter based on the target deposition distribution uniformity and the process parameters and their values that can be determined in the current printing process, and based on a PID gain coefficient prediction model and a deposition distribution uniformity prediction model.

[0027] The printing module is configured to perform electro-spraying printing based on the process parameters and their values that can be determined and the final value of the to-be-optimized process parameter.

[0028] Overall, the above technical solutions conceived by the present application can achieve the following beneficial effects:

[0029] (1) The present application proposes an intelligent control method for electro-spraying deposition distribution uniformity, which mainly focuses on the value optimization of key parameters that have an important influence on deposition distribution uniformity. After the optimal value of the key parameter is obtained, electro-spraying printing is performed in combination with the values of other parameters, and intelligent control of deposition uniformity is realized. The value optimization of the key parameter combines a deposition distribution uniformity prediction model based on artificial neural network, which takes process parameters and their values as input and deposition distribution uniformity as output, and a PID gain coefficient prediction model based on artificial neural network, which takes target deposition distribution uniformity and process parameters and their values excluding the key parameter as input and PID gain coefficient as output. Specifically, first, based on the target deposition distribution uniformity and the process parameters and their values that can be determined in the current printing process, the PID gain coefficient prediction model is used to predict appropriate proportional gain coefficient, integral gain coefficient and differential gain coefficient for PID control, thereby determining an adaptive PID control model. The adaptive PID control model is used to update the value of the key parameter based on the difference between the predicted deposition distribution uniformity and the target deposition distribution uniformity. Therefore, the deposition distribution uniformity prediction model and the adaptive PID control model are used to iteratively optimize the key parameter, that is, to give the appropriate value of the key parameter based on the target deposition distribution uniformity and the known process parameters, thereby for actual electro-spraying printing, and intelligent control of electro-spraying deposition distribution uniformity is realized, solving the problem of limited deposition distribution uniformity control precision caused by the inability of existing methods to adaptively adjust the value of the printing parameter in electro-spraying printing.

[0030] (2) The present application constructs a deposition distribution uniformity prediction model based on an artificial neural network, wherein the input is process parameters and their values, and the output is deposition distribution uniformity, by collecting deposition area images to obtain deposition distribution uniformity through electro-spraying printing under different printing parameters, so as to construct a deposition distribution uniformity prediction model based on an artificial neural network, wherein when collecting images, the present application proposes to determine one or more shooting paths perpendicular to the movement direction of the electro-spraying nozzle in the stable printing area of the electro-spraying, and continuously and uninterruptedly collect multiple deposition images from one end of each shooting path perpendicular to the movement direction of the electro-spraying nozzle to the other end, so as to ensure that the accuracy of the trained prediction model is above 95%.

[0031] (3) The present application defines a calculation method of electro-spraying deposition distribution uniformity, that is, the deposition distribution uniformity is represented by the coefficient of variation of the electro-spraying deposition distribution rate, wherein the value of the coefficient of variation is the ratio between the standard deviation and the mean value of the m electro-spraying deposition distribution rates calculated from m deposition images, so as to ensure the accuracy of the trained prediction model.

[0032] (4) When constructing the PID gain coefficient prediction model, the present application uses a large amount of data for fitting, sets multiple sets of proportional gain coefficient, integral gain coefficient and differential gain coefficient values for each set of target deposition distribution uniformity and the values of the process parameters with certain values, selects one set of proportional gain coefficient, integral gain coefficient and differential gain coefficient values as a label from the multiple sets of proportional gain coefficient, integral gain coefficient and differential gain coefficient values, and constructs a training sample, so as to ensure the convergence speed and accuracy of the iterative optimization process and reduce the time cost when manually selecting PID parameters. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 A flow chart of an intelligent control method of electro-spraying deposition distribution uniformity is provided for the embodiments of the present application.

[0034] Figure 2 A module schematic diagram of an intelligent control method of electro-spraying deposition distribution uniformity is provided for the embodiments of the present application.

[0035] Figure 3 An image processing schematic diagram when calculating deposition distribution uniformity is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0037] Embodiment one

[0038] An intelligent control method for deposition distribution uniformity of electric atomization, as shown by the formula: Figure 1

[0039] S1, obtain the target deposition distribution uniformity and the current printing process value of the process parameters and their values, and input the PID gain coefficient prediction model to predict the proportional gain coefficient, integral gain coefficient and differential gain coefficient for PID control, thereby determining the PID control model;

[0040] S2, input the above process parameters and their values and the process parameters to be optimized and their initial values into the deposition distribution uniformity prediction model to predict the deposition distribution uniformity under the current process parameter value combination;

[0041] S3, based on the above target deposition distribution uniformity and the deposition distribution uniformity under the current process parameter value combination, update the value of the process parameter to be optimized using the above PID control model;

[0042] S4, based on the above process parameters and their values and the process parameters to be optimized and their updated values, repeat S2 until the difference between the deposition distribution uniformity predicted by the deposition distribution uniformity prediction model and the target deposition distribution uniformity is within the preset range;

[0043] S5, based on the above process parameters and their values and the final value of the process parameter to be optimized, perform electric atomization printing to realize intelligent control of the deposition distribution uniformity of electric atomization.

[0044] ​This embodiment takes into account that real-time downward observation feedback cannot be achieved during actual printing. Therefore, in the offline modeling stage, electro-atomization printing is performed under different printing parameters to obtain the deposition distribution uniformity, and a prediction model is constructed with process parameters and their values ​​as input and deposition distribution uniformity as output. In the online control stage, the target deposition distribution uniformity is used as the input, the deposition distribution uniformity predicted by the prediction model is used as the feedback, and the difference between the input and feedback is used as the error. Intelligent adaptive optimization of key process parameters in electro-atomization printing based on PID is performed. The difference between the obtained deposition distribution uniformity and the target deposition distribution uniformity is within a certain error allowable range. The relevant gain parameters in the PID control based on the difference between the input and feedback are predicted in advance based on the target deposition distribution uniformity and some process parameters combined with the PID gain parameter prediction model constructed offline. Both the PID gain coefficient prediction model and the deposition distribution uniformity prediction model have high accuracy during construction. During the iterative optimization process, the optimal solution that meets the accuracy requirements can be quickly determined. Therefore, the method in this embodiment can obtain suitable electro-atomization printing parameters more quickly and accurately, and has a certain anti-disturbance capability. It is suitable for applications such as manufacturing uniform films using electro-atomization.

[0045] like Figure 2 As shown, after determining the PID control model using the PID gain coefficient prediction model, the subsequent key parameter optimization and updates are performed using the deposition distribution uniformity prediction model and the PID control model until the optimal solution is found before electro-atomization printing. Therefore, in this embodiment, the deposition distribution uniformity prediction model and the PID control model can be considered as a whole, as an intelligent control module. This intelligent control module aims at achieving the target deposition distribution uniformity. And parameters with definite values ​​and their values ​​are used as input quantities. And the sedimentation uniformity predicted by the sedimentation uniformity prediction model based on the current printing parameters The difference As error The proportional gain in the PID control model is The integral gain is The differential gain is Then the first In this PID iteration, the error is , No. Voltage in the next iteration The calculation method is shown in the following formula:

[0046] ;

[0047] After selecting the appropriate , , After parameterization, the intelligent control module can provide a suitable voltage value after multiple iterations. Under this voltage value, the uniformity of the deposition distribution in electro-atomization printing is similar to the target deposition distribution uniformity. The difference is within a certain allowable error range.

[0048] As a preferred embodiment, the above-mentioned sediment distribution uniformity prediction model is constructed in the following manner:

[0049] The range of values ​​for the process parameters with definite values ​​and the process parameters to be optimized are divided into equal step sizes to construct multiple sets of process parameter values;

[0050] Under each combination of process parameters, actual electro-atomization printing was performed. m deposition images were collected in the electro-atomization stable printing area. The image acquisition method was as follows: one or more shooting paths perpendicular to the movement direction of the electro-atomization nozzle were determined in the electro-atomization stable printing area. Multiple deposition images were continuously and without interval from one end of the path to the other end of each shooting path perpendicular to the movement direction of the electro-atomization nozzle.

[0051] Based on the m deposition images corresponding to each combination of process parameters, the deposition distribution uniformity under that combination of process parameters is determined; each combination of process parameters and its corresponding deposition distribution uniformity are used as a training sample to construct a training sample set.

[0052] The neural network was trained using the above training sample set to obtain a prediction model for sediment distribution uniformity.

[0053] As a further preferred embodiment, the coefficient of variation of the electro-atomized deposition distribution rate is used to characterize the uniformity of the deposition distribution, wherein the coefficient of variation is the ratio between the standard deviation and the mean of the m electro-atomized deposition distribution rates calculated from the m deposition images.

[0054] For example, after one electro-atomization printing process, taking m photos of the deposited substrate (perpendicular to the direction of movement of the electro-atomization nozzle) with an industrial camera can yield... Each image capture area, such as Figure 3 As shown, this The area of ​​each mapping region is [missing information]. , of which Within each mapping area, there are Droplet deposition region ,this The first droplet deposition region The area of ​​each droplet deposition region is .

[0055] It can be done through the first Within the mapping area The total area of ​​each droplet deposition region Occupying the entire mapping area To quantitatively characterize the proportion of the first Electro-atomization deposition distribution rate in each mapping area :

[0056] ;

[0057] It can be done Electro-atomization deposition distribution rate in each mapping area The coefficient of variation is used to quantitatively characterize the uniformity of electro-atomized deposition distribution. :

[0058] In a single electro-atomization printing process, the solution was obtained using the method described above. Electro-atomization deposition distribution rate in each mapping area Then it can be calculated according to the following formula. Electro-atomization deposition distribution rate in each mapping area mean and standard deviation :

[0059]

[0060]

[0061] The uniformity of electro-atomized deposition distribution in this electro-atomization printing... It can be calculated using the following formula:

[0062]

[0063] In the offline modeling stage, electro-atomization printing parameters include, for example, voltage. air pressure ,high The ranges of voltage, air pressure, and altitude are divided into multiple sets of parameter combinations with equal step sizes. Electro-atomization printing is performed under each set of printing parameters. An industrial camera is used to capture images, and multiple sets of electro-atomization deposition distribution uniformity are obtained using the above method, thereby constructing an electro-atomization deposition distribution uniformity dataset for modeling.

[0064] Prediction model for sediment distribution uniformity (i.e.) Figure 2 Model 1 in the example is constructed using an artificial neural network approach. Specifically, it involves constructing a network with three input layer neurons (voltage, voltage, and so on). air pressure ,high ), 1 output layer neuron (uniformity of electro-atomization deposition distribution) ), One hidden layer neuron network structure, and the activation function adopts a Sigmoid function;

[0065] voltage , air pressure , height and deposition distribution uniformity The prediction model Model1 is expressed as:

[0066] ;

[0067] In the formula, represents the network connection weight of the first input layer neuron and the hidden layer neuron, represents the network connection weight of the second input layer neuron and the hidden layer neuron, represents the network connection weight of the third input layer neuron and the hidden layer neuron, represents the network connection weight of the hidden layer neuron and the output layer neuron, represents the threshold value of the hidden layer neuron, represents the threshold value of the output layer neuron.

[0068] As a preferred embodiment, the above PID gain coefficient prediction model is constructed by using the following method:

[0069] Constructing multiple sets of target deposition distribution uniformity and the value of the above process parameters which can be determined;

[0070] For each group of target deposition distribution uniformity and the values ​​of the process parameters whose values ​​can be determined, multiple sets of proportional gain coefficients, integral gain coefficients, and derivative gain coefficients are set. Based on the values ​​of the process parameters whose values ​​can be determined in the group and the preset values ​​of the parameters to be optimized, a deposition distribution uniformity prediction model is used to obtain the corresponding deposition distribution uniformity. Based on the deposition distribution uniformity and the value of the target deposition uniformity in the group, a PID control model composed of the values ​​of the proportional gain coefficients, integral gain coefficients, and derivative gain coefficients corresponding to the group is used to update the values ​​of the parameters to be optimized. Based on the values ​​of the process parameters whose values ​​can be determined in the group and the updated values ​​of the parameters to be optimized, a deposition distribution uniformity prediction model is used to predict the deposition distribution uniformity. The difference between each predicted deposition distribution uniformity and the target deposition uniformity is compared. The set of proportional gain coefficients, integral gain coefficients, and derivative gain coefficients corresponding to the smallest difference is used as the label corresponding to the target deposition distribution uniformity and the values ​​of the process parameters whose values ​​can be determined in the group, forming a training sample.

[0071] A PID gain coefficient prediction model was obtained by training each training sample.

[0072] PID gain coefficient prediction model (i.e. Figure 2 Model 2 in the model is constructed using an artificial neural network approach. Specifically, it involves constructing a model with three input layer neurons (representing the target deposition distribution uniformity). air pressure ,high ), 3 output layer neurons ( , , ), One hidden layer neuron The network structure uses the Sigmoid function as the activation function, and Model 2 is represented as follows:

[0073]

[0074]

[0075]

[0076] In the formula, Represents the first input layer neuron and the The network connection weights of each hidden layer neuron This represents the second input layer neuron. and the The network connection weights of each hidden layer neuron Represents the third input layer neuron and the network connection weight of the first hidden layer neuron,

[0077] As a preferred embodiment, the above-mentioned gap corresponding to the values of the proportional gain coefficient, the integral gain coefficient and the differential gain coefficient of each group of the above-mentioned labels also satisfies that the gap is less than 3%.

[0078] As a preferred embodiment, the above-mentioned process parameter to be optimized is voltage, and the above-mentioned process parameter with a determinable value includes atomization gas pressure and the height of the nozzle to the deposition substrate.

[0079] It should be noted that in the printing process parameters, the normal range of the working voltage is generally 2600V-4000V, the range of the gas pressure (generally referred to as the atomization gas pressure) is generally 10kPa-50kPa, the range of the height of the nozzle to the deposition substrate is generally 20mm-50mm, and the range of the target deposition distribution uniformity is generally 0.6-1. In addition, the proportional gain coefficient in the present embodiment generally ranges from 1 to 50, the integral gain coefficient generally ranges from 0 to 3, and the differential gain coefficient ​​​​​​​​​​​​​​​​​​Generally 0-3. When constructing the training sample of the PID gain coefficient prediction model, when constructing the training sample of the deposition distribution uniformity prediction model, and after the model training and actually performing the intelligent regulation and control of the deposition distribution uniformity of the electro-spraying, the value setting of each parameter can be within the respective range, and when combining the value of each parameter, a complete combination can be used, such as voltage 2600, 2800, 3000, 3200, 3400, 3600, 3800, 4000, air pressure 10, 20, 30, 40, 50, height 20, 30, 40, 50, then design 8x5x4, a total of 160 groups of data, which can be increased or decreased according to the situation.

[0080] Overall, from the entire process, the method proposed in the embodiment can be performed through the following steps:

[0081] (1) electro-spraying is performed using the nozzle under different electro-spraying printing parameters, and the deposition distribution uniformity is obtained to construct a deposition distribution uniformity data set;

[0082] (2) a prediction model Model1 is constructed according to the data set, with the input being, for example, voltage, air pressure, and height, and the output being deposition distribution uniformity;

[0083] (3) an intelligent regulation and control module based on PID is constructed with the target deposition distribution uniformity as the input and the deposition distribution uniformity predicted by the prediction model as the feedback;

[0084] (4) according to the experience of selecting the optimal parameters multiple times 、 、 , a prediction model Model2 is constructed from the target deposition distribution uniformity, air pressure, and height to 、 、 ;

[0085] (5) for any target deposition distribution uniformity and air pressure and height, the appropriate 、 、 can be obtained according to the prediction model Model2 of S4, and the intelligent regulation and control system constructed according to S3 can calculate the optimal voltage, and the difference between the deposition distribution uniformity of the electro-spraying printing under the voltage and the given air pressure and height and the target deposition distribution uniformity is within a certain error allowable range.

[0086] Embodiment Two

[0087] The application discloses an intelligent control system for uniformity of electric atomization deposition distribution, which is used for executing an intelligent control method for uniformity of electric atomization deposition distribution, and comprises a process parameter optimization module and a printing module. The process parameter optimization module is used for optimizing the value of a to-be-optimized parameter based on a PID gain coefficient prediction model and a deposition distribution uniformity prediction model according to a target deposition distribution uniformity, current printing process parameters and values of the process parameters. The printing module is used for performing electric atomization printing based on the process parameters and the values of the process parameters and the final value of the to-be-optimized process parameter.

[0088] The related technical scheme is the same as that in the first embodiment, and details are not repeated here.

[0089] Those skilled in the art can easily understand that the above description is only the preferred embodiment of the present application, and is not used to limit the present application, and any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for intelligently controlling the uniformity of electro-atomized deposition distribution, characterized in that, include: S1. Obtain the target deposition distribution uniformity and the process parameters and their values ​​that can be determined in the current printing process, and input them into the PID gain coefficient prediction model to predict the proportional gain coefficient, integral gain coefficient and derivative gain coefficient used for PID control, thereby determining the PID control model. S2. Input the process parameters with definite values ​​and their values, as well as the process parameters to be optimized and their initial values, into the deposition distribution uniformity prediction model to predict the deposition distribution uniformity under the current combination of process parameter values. S3. Based on the target deposition distribution uniformity and the deposition distribution uniformity under the current combination of process parameter values, the PID control model is used to update the values ​​of the process parameters to be optimized. S4. Based on the determined process parameters and their values, as well as the process parameters to be optimized and their updated values, repeat S2 until the difference between the deposition distribution uniformity predicted by the deposition distribution uniformity prediction model and the target deposition distribution uniformity is within a preset range. S5. Based on the process parameters with definite values ​​and their values, as well as the final values ​​of the process parameters to be optimized, electro-atomization printing is performed to achieve intelligent control of the uniformity of electro-atomization deposition distribution. The sediment distribution uniformity prediction model was constructed using the following method: The value ranges of the process parameters with definite values ​​and the process parameters to be optimized are divided into equal step sizes to construct multiple sets of process parameter values. Under each combination of process parameters, actual electro-atomization printing was performed. m deposition images were collected in the electro-atomization stable printing area. The image acquisition method was as follows: one or more shooting paths perpendicular to the movement direction of the electro-atomization nozzle were determined in the electro-atomization stable printing area. Multiple deposition images were continuously and without interval from one end of the path to the other end of each shooting path perpendicular to the movement direction of the electro-atomization nozzle. Based on the m deposition images corresponding to each combination of process parameters, the deposition distribution uniformity under that combination of process parameters is determined; each combination of process parameters and its corresponding deposition distribution uniformity are used as a training sample to construct a training sample set. The neural network is trained using the training sample set to obtain the deposition distribution uniformity prediction model; The PID gain coefficient prediction model was constructed in the following manner: Construct multiple sets of target deposition distribution uniformity and the values ​​of the process parameters whose values ​​can be determined; For each group of target deposition distribution uniformity and the values ​​of the definite process parameters, multiple sets of proportional gain coefficients, integral gain coefficients, and derivative gain coefficients are set. Based on the values ​​of the definite process parameters in this group and the preset values ​​of the parameters to be optimized, the deposition distribution uniformity prediction model is used to obtain the corresponding deposition distribution uniformity. Based on the deposition distribution uniformity and the value of the target deposition uniformity in this group, a PID control model is used, composed of the values ​​of the proportional gain coefficients, integral gain coefficients, and derivative gain coefficients corresponding to each group. The model updates the values ​​of the parameters to be optimized. Based on the values ​​of the process parameters with definite values ​​in the group and the updated values ​​of the parameters to be optimized, the deposition distribution uniformity prediction model is used to predict the deposition distribution uniformity. The difference between each predicted deposition distribution uniformity and the target deposition uniformity is compared. The values ​​of the proportional gain coefficient, integral gain coefficient and differential gain coefficient corresponding to the smallest difference are used as the labels corresponding to the target deposition distribution uniformity and the values ​​of the process parameters with definite values, forming a training sample. The PID gain coefficient prediction model is obtained by training each training sample.

2. The intelligent control method for the uniformity of electro-atomized deposition distribution according to claim 1, characterized in that, The uniformity of the deposition distribution is characterized by the coefficient of variation of the electro-atomized deposition distribution rate, wherein the coefficient of variation is the ratio between the standard deviation and the mean of the m electro-atomized deposition distribution rates calculated from m deposition images.

3. The intelligent control method for the uniformity of electro-atomized deposition distribution according to claim 1, characterized in that, The difference between the values ​​of the proportional gain coefficient, integral gain coefficient, and differential gain coefficient of each group used as the label is also satisfied that it is below 3%.

4. The intelligent control method for the uniformity of electro-atomized deposition distribution according to claim 1, characterized in that, The process parameter to be optimized is voltage, and the process parameters whose values ​​can be determined include atomizing gas pressure and the height of the nozzle from the deposition substrate.

5. A smart control system for the uniformity of electro-atomized deposition distribution, characterized in that, A method for intelligent control of the uniformity of electro-atomization deposition distribution as described in any one of claims 1 to 4, comprising: a process parameter optimization module and a printing module; The process parameter optimization module is used to optimize the values ​​of the parameters to be optimized based on the target deposition distribution uniformity and the process parameters and their values ​​that can be determined by the current printing process, and based on the PID gain coefficient prediction model and the deposition distribution uniformity prediction model. The printing module is used to perform electro-atomization printing based on the definite process parameters and their values, as well as the final value of the process parameters to be optimized.

Images