Jet flow electrochemical deposition rate self-adaptive control method and system

Abstract

The invention relates to a jet flow electrochemical deposition rate self-adaptive control method and system, and belongs to the technical field, and the method comprises the steps: carrying out the physical field model construction of collected data through a digital twinning technology, and constructing a jet flow electrochemical deposition rate self-adaptive control model based on deep learning and a physical model; therefore, optimized working parameters of the chemical additive manufacturing equipment are generated through the jet flow electrochemical deposition rate self-adaptive control model, a visualization system is constructed based on the collected data, and finally interaction control is conducted on the chemical additive manufacturing equipment through the visualization system. According to the method, a high-fidelity virtual model which synchronously runs with physical equipment is created, and before a complex spiral structure is printed, full-process simulation can be performed in a digital twinning body, which corners possibly have defects due to rate mismatch can be found in advance, and optimized manufacturing production parameters are generated; therefore, dynamic fine adjustment is carried out according to real-time current and image data, and closed-loop control is realized.

Description

Technical Field

[0001] This invention relates to the field of jet electrochemical deposition rate control technology, and in particular to an adaptive control method and system for jet electrochemical deposition rate. Background Technology

[0002] Jet electrochemical deposition (JetECD) is a technique that enhances ion convection in solution and suppresses concentration polarization by using a high-speed jet, thereby achieving a higher deposition rate. In traditional electrochemical additive manufacturing techniques, deposition rates are generally low, and the electrolytes are mostly static, limiting ion transport and deposition efficiency. JetECD, by introducing a high-speed jet, significantly improves the deposition rate. However, existing jetECD techniques are constrained by nonlinear time-varying rates, making it difficult to achieve voxel-level precise control of complex three-dimensional structures. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides an adaptive control method and system for jet electrochemical deposition rate.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of this invention provides an adaptive control method for jet electrochemical deposition rate, comprising: By configuring a data acquisition system in the chemical additive manufacturing equipment, data can be acquired and synchronized through the data acquisition system. Digital twin technology is introduced, and a physical field model is constructed using the digital twin technology on the collected data. An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning and the physical model. The optimized operating parameters of the chemical additive manufacturing equipment are generated using the jet electrochemical deposition rate adaptive control model, and a visualization system is constructed based on the collected data. The visualization system is used to interactively control the chemical additive manufacturing equipment.

[0005] Furthermore, in the adaptive control method for jet electrochemical deposition rate, a data acquisition system is configured in the chemical additive manufacturing equipment, specifically: By configuring a sensor array, computing platform and actuator in a chemical additive manufacturing equipment, the sensor array is equipped with a global shutter industrial camera and a pulsed laser, and the sensor array is used for structured light three-dimensional topography scanning. The working electrode, counter electrode, and reference electrode are integrated inside the deposition nozzle, and an ultrasonic sensor is attached to the nozzle tip. A sampling frequency is set, and data is acquired based on the sampling frequency to receive backscattered signals, thus forming a data acquisition system.

[0006] Furthermore, in the adaptive control method for jet electrochemical deposition rate, data acquisition and synchronization are performed through a data acquisition system, specifically as follows: A local coordinate system is established with the deposition point as the center. At the beginning of deposition, the industrial camera triggers a three-dimensional scan with a preset thickness each time to obtain the topographic information containing point cloud data. The current density time series data is collected through the working electrode, counter electrode and reference electrode set in the deposition nozzle. An ultrasonic sensor collects high-frequency acoustic emission signals, and statistically analyzes topographic information, current density time-series data, and high-frequency acoustic emission signals containing point cloud data to obtain multi-source heterogeneous data information. By using a precision hardware clock to timestamp and synchronize the multi-source heterogeneous data information, time-aligned multimodal data packets are formed.

[0007] Furthermore, in the adaptive control method for jet electrochemical deposition rate, digital twin technology is introduced. This digital twin technology is used to construct a physical field model from the collected data, specifically: Digital twin technology was introduced, and a quasi-static electric field model was established using the current density time series data in the collected data through COMSOL Multiphysics. Coupled with computational fluid dynamics, the Navier-Stokes equations are solved to simulate the three-dimensional diffusion flow of electrolyte jet after impacting the substrate, forming a flow field model; Convection-diffusion equations are established based on topographic information including point cloud data to describe the transport process of metal ions under the combined action of electric and flow fields. Ion consumption at the deposition interface is described by the Butler-Volmer kinetic equations to form a concentration field model. Running a twin on a physical server, the physical field model is driven to perform synchronous calculations using the real-time coordinates of the actuator, the actual output of the pulse power supply, and the time-series data of the current density as boundary and initial conditions. The result is the output of the three-dimensional distribution of electric field intensity, flow velocity, and ion concentration at any location in the twin space.

[0008] Furthermore, in the adaptive control method for jet electrochemical deposition rate, an adaptive control model for jet electrochemical deposition rate is constructed based on deep learning, specifically as follows: An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning and physical models. The convection-diffusion equation and Butler-Volmer kinetic equation are solved using digital twin technology based on the current three-dimensional distribution data of electric field intensity, flow velocity, and ion concentration. The deposition rate and deposition morphology are then calculated and predicted. Construct a state space, an action space, and a reward function, where the physical field parameters of electric field strength, flow velocity, and ion concentration are used as the state space, and the parameters of nozzle movement speed and pulse duty cycle are used as the action space. A reward function is constructed using deposition rate, target deposition rate, and predicted deposition morphology, while penalizing rate deviation and surface roughness. The jet electrochemical deposition rate adaptive control model is trained based on the state space, action space, and reward function. The process involves continuous exploration in the digital space to generate a set of optimal control parameter sequences that maximize future rewards. When the prediction accuracy of the jet electrochemical deposition rate adaptive control model for predicting the optimal control parameter sequence is higher than a preset prediction accuracy threshold, the training of the jet electrochemical deposition rate adaptive control model is complete.

[0009] Furthermore, in the jet electrochemical deposition rate adaptive control method, the optimized operating parameters of the chemical additive manufacturing equipment are generated using the aforementioned jet electrochemical deposition rate adaptive control model, specifically as follows: Real-time data information from the chemical additive manufacturing equipment is acquired, and the real-time data information from the chemical additive manufacturing equipment is input into the jet electrochemical deposition rate adaptive control model for judgment and prediction. By judging and predicting, the optimized operating parameters of the chemical additive manufacturing equipment are obtained, and the optimized parameters are sent to the physical equipment for execution. The actual deposition data collected by the physical equipment is compared with the predicted deposition rate and predicted deposition morphology to generate residuals. When the residual is greater than a preset residual threshold, the model parameters in the physical model are dynamically adjusted to correct the physical model.

[0010] Furthermore, in the adaptive control method for jet electrochemical deposition rate, a visualization system is constructed based on the collected data, specifically as follows: By using the topographic information containing point cloud data collected by industrial cameras, and combining it with the equipment encoder information, a three-dimensional point cloud map of the deposition environment is constructed. The three-dimensional ion concentration distribution is superimposed on the real deposition area in the form of a translucent, color-gradient "ion cloud" and visualized as a dynamic, flowing luminescent line, thus visualizing the electric field distribution calculated by the twin. A floating panel on the side of the field of view displays real-time deposition rate, current efficiency, and cumulative deposition amount, and AR glasses are used as the display and human-computer interaction terminal to form a visualization system.

[0011] Furthermore, in the adaptive control method for jet electrochemical deposition rate, the visualization system is used to interactively control the chemical additive manufacturing equipment, specifically as follows: The AR glasses use a built-in depth camera to recognize the operator's gesture commands and then use the recognition results to interactively control the chemical additive manufacturing equipment. When an operator detects a potential anomaly through the AR interface, they can mark it with a gesture, package it according to the marked timestamp and multimodal sensor data, and provide an anomaly notification.

[0012] Furthermore, the adaptive control method for jet electrochemical deposition rate also includes: All chemical additive manufacturing equipment is networked and data packets are uploaded to a cloud server according to a unified data standard. The data packets include chemical additive manufacturing equipment, process parameters, and quality index data. Automated feature extraction is performed on massive process parameter time series data in the cloud to obtain statistical features, frequency domain features, and time domain features, and feature vectors are constructed based on the statistical features, frequency domain features, and time domain features. Based on the feature vectors, single-point, isolated process experiences are formed and transformed into a structured, associative, and reasonable networked process knowledge graph.

[0013] A second aspect of the present invention provides a jet electrochemical deposition rate adaptive control system, including a memory and a processor. The memory includes a jet electrochemical deposition rate adaptive control method program. When the jet electrochemical deposition rate adaptive control method program is executed by the processor, it implements the steps of the jet electrochemical deposition rate adaptive control method described in any one of the present invention.

[0014] This invention addresses the shortcomings of the prior art and has the following beneficial effects: This invention configures a data acquisition system in a chemical additive manufacturing equipment. This system collects and synchronizes data, then introduces digital twin technology. The acquired data is used to construct a physical field model, and an adaptive control model for the jet electrochemical deposition rate is built based on deep learning and the physical model. This model generates optimized operating parameters for the chemical additive manufacturing equipment, and a visualization system is constructed based on the acquired data. Finally, the visualization system is used for interactive control of the chemical additive manufacturing equipment. This invention creates a high-fidelity virtual model that runs synchronously with the physical equipment. Before printing a complex spiral structure, a full-process simulation can be performed in the digital twin, identifying potential defects at corners due to rate mismatch and generating optimized manufacturing parameters. These parameters are then dynamically fine-tuned based on real-time current and image data, achieving a closed-loop control of "prediction-optimization-correction." Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0016] Figure 1 The overall flowchart of the adaptive control method for jet electrochemical deposition rate is shown; Figure 2 A system block diagram of the adaptive control system for jet electrochemical deposition rate is shown. Detailed Implementation

[0017] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0019] like Figure 1 As shown, the first aspect of the present invention provides an adaptive control method for jet electrochemical deposition rate, comprising: By configuring a data acquisition system in the chemical additive manufacturing equipment, data can be acquired and synchronized through the data acquisition system. Digital twin technology is introduced, and a physical field model is constructed using digital twin technology on the collected data. An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning and the physical model. An adaptive control model for jet electrochemical deposition rate is used to generate optimized operating parameters for chemical additive manufacturing equipment, and a visualization system is built based on the collected data. A visualization system is used to interactively control chemical additive manufacturing equipment.

[0020] It should be noted that this invention creates a high-fidelity virtual model that runs synchronously with the physical device. Before printing a complex spiral structure, the entire process can be simulated in the digital twin. This allows for the early detection of which corners may have defects due to rate mismatch, and the generation of optimized manufacturing parameters. These parameters are then dynamically fine-tuned based on real-time current and image data to achieve closed-loop control of "prediction-optimization-correction".

[0021] Furthermore, in the adaptive control method for jet electrochemical deposition rate, a data acquisition system is configured in the chemical additive manufacturing equipment, specifically: By configuring a sensor array, computing platform and actuator in a chemical additive manufacturing equipment, the sensor array is equipped with a global shutter industrial camera and a pulsed laser, and the sensor array is used for structured light three-dimensional topography scanning. The working electrode, counter electrode, and reference electrode are integrated inside the deposition nozzle, and an ultrasonic sensor is attached to the nozzle tip. A sampling frequency is set, data is acquired based on the sampling frequency, and backscattered signals are received to form a data acquisition system.

[0022] It should be noted that the sensor sequence consists of a 20-megapixel global shutter industrial camera equipped with a 5-megapixel pulsed laser for structured light 3D topography scanning. A custom three-electrode microcurrent sensor, comprising a working electrode, a counter electrode, and a reference electrode, is integrated inside the deposition nozzle, with a sampling frequency of 1MHz. An ultrasonic high-frequency sensor, with a center sampling frequency of 10MHz, is positioned close to the nozzle tip to receive backscattered signals.

[0023] Furthermore, in the adaptive control method for jet electrochemical deposition rate, data acquisition and synchronization are performed through a data acquisition system, specifically as follows: A local coordinate system is established with the deposition point as the center. At the beginning of deposition, the industrial camera triggers a three-dimensional scan with a preset thickness each time to obtain the topographic information containing point cloud data. The current density time series data is collected through the working electrode, counter electrode and reference electrode set in the deposition nozzle. An ultrasonic sensor collects high-frequency acoustic emission signals, and statistically analyzes topographic information, current density time-series data, and high-frequency acoustic emission signals containing point cloud data to obtain multi-source heterogeneous data information. By using a precision hardware clock to timestamp and synchronize heterogeneous data from multiple sources, time-aligned multimodal data packets are formed.

[0024] It should be noted that, during the chemical manufacturing process, jet electrochemical deposition occurs in the chemical manufacturing equipment. At this time, a local coordinate system is established with the deposition point as the center. The industrial camera triggers a three-dimensional scan at the beginning of deposition and after every 5μm thickness, with an estimated time of 80ms, to obtain topographic information including point cloud data. Meanwhile, the microcurrent sensor collects current density time-series data in real time, and the ultrasonic sensor collects high-frequency acoustic emission signals. All data are synchronized with nanosecond-level timestamps through a precision hardware clock (PXIe-6674T) to form a time-aligned multimodal data packet.

[0025] Furthermore, in the adaptive control method for jet electrochemical deposition rate, digital twin technology is introduced to construct a physical field model based on the collected data. Specifically: Digital twin technology was introduced, and a quasi-static electric field model was established using the current density time series data in the collected data through COMSOL Multiphysics. Coupled with computational fluid dynamics, the Navier-Stokes equations are solved to simulate the three-dimensional diffusion flow of electrolyte jet after impacting the substrate, forming a flow field model; Convection-diffusion equations are established based on topographic information including point cloud data to describe the transport process of metal ions under the combined action of electric and flow fields. Ion consumption at the deposition interface is described by the Butler-Volmer kinetic equations to form a concentration field model. Running a twin on a physical server, the physical field model is driven to perform synchronous calculations using the real-time coordinates of the actuator, the actual output of the pulse power supply, and the time-series data of the current density as boundary and initial conditions. The result is the output of the three-dimensional distribution of electric field intensity, flow velocity, and ion concentration at any location in the twin space.

[0026] It should be noted that the calculation method for solving the Navier-Stokes equations to simulate the three-dimensional diffusion flow of the electrolyte jet after impacting the substrate is as follows: in, and For mixing density and dynamic viscosity, For mixed density, The dynamic viscosity is determined by the volume fraction 'a'. For the i-th mixing density, Let y be the mixing density. The l-th is the dynamic viscosity. Let y be the dynamic viscosity.

[0027] Based on the above calculation scheme, a three-dimensional diffusion flow of electrolyte jet after impacting the substrate is simulated numerically to form a flow field model. In the flow field model, the mixing density and dynamic viscosity can be dynamically obtained, and digital twins are used to dynamicize the flow field of the electrolyte.

[0028] The convection-diffusion equation is calculated as follows: in, The value of D represents the rate of change of electrolyte concentration c at a certain point in space with time t. D is the diffusion coefficient, a physical property parameter characterizing the rate of diffusion of a substance in a medium. A larger D value indicates faster diffusion. It depends on the properties of the diffusing substance and the medium (such as temperature and viscosity). For gases, D ~ 10⁻ 5 m² / s; for liquids, D ~ 10⁻ 9 m² / s. The velocity field vector represents the macroscopic velocity of the medium (fluid) at various points in space. It may be a constant or a function of space and time. Let be the divergence of the gradient, a scalar differential operator. Applying this operator to c yields ∇²c, which characterizes the difference between the concentration at that point and the surrounding average concentration. The Nabra operator, acting on a scalar field c, yields its spatial gradient ∇c, pointing in the direction of the fastest increase in c. This method expresses the following equilibrium relationship: the instantaneous change in concentration at a point = net input / output due to diffusion + net input / output due to convection. For example, electrolytes slowly disperse in all directions, even in still water. Similarly, the flow of electrolytes can cause the electrolyte droplets to move as a whole.

[0029] The ion consumption at the deposition interface is described by the Butler-Volmer kinetic equations, forming a concentration field model. This model utilizes digital twin technology to simulate the concentration model of the concentration chemical manufacturing equipment during the manufacturing process. The Butler-Volmer kinetic equations are all existing technologies and can be found in existing formulas, books, or textbooks; they will not be elaborated upon in this embodiment. The relationships are as follows: Net current density; For exchange current density; Anode transfer coefficient; Where n is the cathode transfer coefficient; n is the electron transfer number. R is the Faraday constant, 96485 C / mol; R is the gas constant, 8.314 J / (mol·K); and T is the constant absolute temperature.

[0030] In the adaptive control method for jet electrochemical deposition rate, an adaptive control model for jet electrochemical deposition rate is constructed based on deep learning, specifically as follows: An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning and physical models. The convection-diffusion equation and Butler-Volmer kinetic equation are solved using digital twin technology based on the current three-dimensional distribution data of electric field intensity, flow velocity, and ion concentration. The deposition rate and deposition morphology are then calculated and predicted. Construct a state space, an action space, and a reward function, where the physical field parameters of electric field strength, flow velocity, and ion concentration are used as the state space, and the parameters of nozzle movement speed and pulse duty cycle are used as the action space. A reward function is constructed using deposition rate, target deposition rate, and predicted deposition morphology, while penalizing rate deviation and surface roughness. The adaptive control model for jet electrochemical deposition rate is trained based on the state space, action space, and reward function. The process involves continuous exploration in the digital space to generate a set of optimal control parameter sequences that maximize future rewards. When the prediction accuracy of the jet electrochemical deposition rate adaptive control model for predicting the optimal control parameter sequence is higher than the preset prediction accuracy threshold, the training of the jet electrochemical deposition rate adaptive control model is complete.

[0031] It should be noted that the deep learning network can be a convolutional neural network, including convolutional layers, activation functions (ReLU function), pooling layers, and fully connected layers. The core parameters of the convolutional layer are the number of filters, filter size, stride, etc. The number and size of filters vary between shallow and deep layers. Shallow layers have fewer filters (32-64) and capture basic features (edges, color); deep layers have more filters (128-512) and capture higher-level features. A convolutional neural network is used to train an adaptive control model for the jet electrochemical deposition rate based on the state space, action space, and reward function. It continuously explores in the digital space, generating a set of optimal control parameter sequences that maximize future rewards. These optimal control parameter sequences include nozzle movement speed and pulse duty cycle. This method enables automated optimal control of parameters, allowing for early detection and compensation of process defects, achieving true process feedforward optimization.

[0032] It should be noted that, based on the current state in the state space, the model outputs the actions in the action space.

[0033] Furthermore, in the adaptive control method of the jet with chemical deposition rate, the optimized operating parameters of the chemical additive manufacturing equipment are generated using the adaptive control model of the jet electrochemical deposition rate, specifically: Real-time data information from the chemical additive manufacturing equipment is acquired and input into the adaptive control model of jet electrochemical deposition rate for judgment and prediction. By judging and predicting, the optimized working parameters of the chemical additive manufacturing equipment are obtained, and the optimized parameters are sent to the physical equipment for execution. The actual deposition data collected by the physical equipment is compared with the predicted deposition rate and predicted deposition morphology to generate residuals. When the residual exceeds the preset residual threshold, the model parameters in the physical model are dynamically adjusted to correct the physical model.

[0034] It should be noted that this set of optimized parameters is sent to the physical equipment for execution. Simultaneously, the actual depositional data collected by the physical equipment, such as the actual morphology, is compared with the predicted data of the twin to generate residuals. These residuals are used to trigger an online parameter identification algorithm to dynamically adjust key parameters in the twin model, such as the diffusion coefficient and reaction rate constant, achieving model self-correction and reducing "model drift".

[0035] Furthermore, in the adaptive control method for jet electrochemical deposition rate, a visualization system is constructed based on the collected data, specifically as follows: By using the topographic information containing point cloud data collected by industrial cameras, and combining it with the equipment encoder information, a three-dimensional point cloud map of the deposition environment is constructed. The three-dimensional ion concentration distribution is superimposed on the real deposition area in the form of a translucent, color-gradient "ion cloud" and visualized as a dynamic, flowing luminescent line, thus visualizing the electric field distribution calculated by the twin. A floating panel on the side of the field of view displays real-time deposition rate, current efficiency, and cumulative deposition amount, and AR glasses are used as the display and human-computer interaction terminal to form a visualization system.

[0036] It should be noted that this method enables visualization of the production process.

[0037] Furthermore, in the adaptive control method for jet electrochemical deposition rate, a visualization system is used to interactively control the chemical additive manufacturing equipment, specifically: The AR glasses use a built-in depth camera to recognize the operator's gesture commands and then use the recognition results to interactively control the chemical additive manufacturing equipment. When an operator detects a potential anomaly through the AR interface, they can mark it with a gesture, package it according to the marked timestamp and multimodal sensor data, and provide an anomaly notification.

[0038] It's worth noting that the AR glasses utilize their built-in depth camera to recognize operator gesture commands. For example, a "tap in the air" gesture on the virtual hovering button allows for actions like "pausing deposition" and "switching view modes." A "pinch-drag" gesture allows the operator to "grab" a virtual ion concentration cloud and move it in space; the system interprets this as a desired adjustment of the deposition area and automatically generates a motion platform path plan. A "circle" gesture allows the operator to draw the desired deposition outline on the virtual substrate; the system converts this into G-code and executes it. If the operator detects a potential anomaly (such as an area of ​​excessively dark color) through the AR interface, they can mark it using voice or gestures. The timestamp and multimodal sensor data of this marked point are packaged and used for alerts.

[0039] Furthermore, the adaptive control method for jet electrochemical deposition rate also includes: All chemical additive manufacturing equipment is networked and data packets are uploaded to the cloud server according to a unified data standard. The data packets include chemical additive manufacturing equipment, process parameters, and quality index data. Automated feature extraction is performed on massive process parameter time series data in the cloud to obtain statistical features, frequency domain features, and time domain features, and feature vectors are constructed based on these features. Based on the feature vectors, single-point, isolated process experiences are formed and transformed into a structured, associative, and reasonable networked process knowledge graph.

[0040] It should be noted that all networked jet electrochemical deposition equipment uploads data packets to the cloud data lake according to a unified data standard (referencing the MTConnect standard). Each data packet contains operating parameters, process parameters, and quality indicators. Operating parameters include material type, electrolyte formulation, temperature, and initial pH value; process parameters include the motion trajectory sequence, pulse power supply parameter sequence, and flow rate parameter sequence throughout the deposition process; quality indicators include the deposition layer thickness distribution, surface roughness Ra value, microhardness, and resistivity measured by an offline coordinate measuring machine or scanning electron microscope after deposition. The cloud automatically extracts features from the massive time series of process parameters, including: statistical mean, variance, skewness, frequency domain features extracted by FFT, and time domain features (zero-crossing rate), ultimately representing each processing step as a feature vector of over 500 dimensions.

[0041] The knowledge graph is constructed around the core relationship between "process" and "performance". Nodes include: process characteristics (e.g., "high duty cycle pulse"), material combinations (e.g., "copper sulfate electrolyte / stainless steel substrate"), quality attributes (e.g., "low roughness"), and application scenarios (e.g., "high-frequency circuits"). Edges represent strong correlations between them (e.g., "process A" leads to "quality B"). When a user sets a goal, such as "depositing a dense copper layer on titanium alloy with a thickness of 50 μm and a roughness of <1 μm", the system first searches the knowledge graph for historical cases that perfectly match. If none are found, a user-based collaborative filtering algorithm is activated: it finds a group of "neighboring users" with similar historical preferences to the current user, and ranks and recommends the process solutions that this group has successfully used and are closest to the current goal. For the recommended basic process solutions, the system calls a Bayesian optimization algorithm deployed in the cloud. This algorithm takes process parameters (e.g., voltage, scan speed) as input and the predicted quality index as output, performing virtual iterative optimization in the cloud. Typically, after 50-100 iterations, the Pareto optimal parameter frontier for the current specific objective can be found, and the optimized parameter set can be distributed to the equipment. This method can transform single-point, isolated process experience into structured, associative, and reasonable networked process knowledge. Through data mining and collaborative intelligence, it provides rapid "from 1 to N" process solution generation and optimization for new tasks, greatly reducing the process development threshold and cycle time.

[0042] In addition, this method also includes: Store all process data and run large-scale training tasks of deep learning models, physical models generated by digital twins, and Bayesian optimization algorithms; manage, version control, and distribute a unified adaptive control model for jet electrochemical deposition rate. The system is configured with an edge layer and device terminals. The edge layer receives and runs a lightweight inference model from the cloud on a local server in the production workshop. It performs real-time fusion and rate prediction of data from multiple devices in the workshop, manages task scheduling between devices, and ensures production continuity based on a local cache model in the event of network outages or high latency. Each deposition device in the equipment is equipped with an embedded controller. The embedded controller is responsible for the lowest-level real-time control loop, and collects raw sensor data, receives control commands from the edge layer, and drives the motor and pump. A new version of the prediction model is trained using the stored process data. Differential packages are generated through incremental update technology and pushed to each edge server periodically / on demand. The device sends sensor data to the edge server, the edge server runs the model for inference, and sends the generated control commands to the device for execution. The process and quality data from the device are desensitized and initially compressed at the edge layer before being asynchronously uploaded to the cloud data lake for the next round of model training.

[0043] It should be noted that the three-tiered "cloud-edge-device" architecture decouples computational model training from real-time control. This achieves both centralized optimization and sharing of global process knowledge and ensures the real-time performance and reliability of individual device control, providing a core architecture for building a flexible and intelligent distributed manufacturing network.

[0044] In summary, this invention achieves precise feedforward and adaptive closed-loop control of the deposition rate by using a digital twin to simulate the multiphysics coupling process in real time and combining it with a deep learning model for prediction and optimization. This significantly improves the forming accuracy and surface quality of complex structures. Furthermore, by utilizing a precision clock to synchronize multimodal data such as morphology, current, and acoustic emission, it provides comprehensive and synchronized input for model construction, enhancing the accuracy and reliability of state perception.

[0045] like Figure 2 As shown, a second aspect of the present invention provides a jet electrochemical deposition rate adaptive control system, including a memory and a processor. The memory includes a jet electrochemical deposition rate adaptive control method program. When the jet electrochemical deposition rate adaptive control method program is executed by the processor, it implements any of the steps of the jet electrochemical deposition rate adaptive control method.

[0046] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0047] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0048] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0049] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0050] Alternatively, if the integrated units of the present invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0051] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An adaptive control method for jet electrochemical deposition rate, characterized in that, include: By configuring a data acquisition system in the chemical additive manufacturing equipment, data can be acquired and synchronized through the data acquisition system. Digital twin technology is introduced, and a physical field model is constructed using the digital twin technology on the collected data. An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning and the physical model. The optimized operating parameters of the chemical additive manufacturing equipment are generated using the jet electrochemical deposition rate adaptive control model, and a visualization system is constructed based on the collected data. The visualization system is used to interactively control the chemical additive manufacturing equipment.

2. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, By configuring a data acquisition system in the chemical additive manufacturing equipment, specifically: By configuring a sensor array, computing platform and actuator in a chemical additive manufacturing equipment, the sensor array is equipped with a global shutter industrial camera and a pulsed laser, and the sensor array is used for structured light three-dimensional topography scanning. The working electrode, counter electrode, and reference electrode are integrated inside the deposition nozzle, and an ultrasonic sensor is attached to the nozzle tip. A sampling frequency is set, and data is acquired based on the sampling frequency to receive backscattered signals, thus forming a data acquisition system.

3. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, Data acquisition and synchronization are performed through a data acquisition system, specifically as follows: A local coordinate system is established with the deposition point as the center. At the beginning of deposition, the industrial camera triggers a three-dimensional scan with a preset thickness each time to obtain the topographic information containing point cloud data. The current density time series data is collected through the working electrode, counter electrode and reference electrode set in the deposition nozzle. An ultrasonic sensor collects high-frequency acoustic emission signals, and statistically analyzes topographic information, current density time-series data, and high-frequency acoustic emission signals containing point cloud data to obtain multi-source heterogeneous data information. By using a precision hardware clock to timestamp and synchronize the multi-source heterogeneous data information, time-aligned multimodal data packets are formed.

4. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, Digital twin technology is introduced, and a physical field model is constructed using the collected data, specifically as follows: Digital twin technology was introduced, and a quasi-static electric field model was established using the current density time series data in the collected data through COMSOL Multiphysics. Coupled with computational fluid dynamics, the Navier-Stokes equations are solved to simulate the three-dimensional diffusion flow of electrolyte jet after impacting the substrate, forming a flow field model; Convection-diffusion equations are established based on topographic information including point cloud data to describe the transport process of metal ions under the combined action of electric and flow fields. Ion consumption at the deposition interface is described by the Butler-Volmer kinetic equations to form a concentration field model. Running a twin on a physical server, the physical field model is driven to perform synchronous calculations using the real-time coordinates of the actuator, the actual output of the pulse power supply, and the time-series data of the current density as boundary and initial conditions. The result is the output of the three-dimensional distribution of electric field intensity, flow velocity, and ion concentration at any location in the twin space.

5. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning, specifically as follows: An adaptive control model for jet electrochemical deposition rate is constructed based on deep learning and physical models. By using digital twin technology to solve the convection-diffusion equation and the Butler-Volmer kinetic equation based on the current three-dimensional distribution data of electric field intensity, flow velocity and ion concentration, the deposition rate and deposition morphology are calculated and predicted. Construct a state space, an action space, and a reward function, where the physical field parameters of electric field strength, flow velocity, and ion concentration are used as the state space, and the parameters of nozzle movement speed and pulse duty cycle are used as the action space. A reward function is constructed using deposition rate, target deposition rate, and predicted deposition morphology, while penalizing rate deviation and surface roughness. The jet electrochemical deposition rate adaptive control model is trained based on the state space, action space, and reward function. The process involves continuous exploration in the digital space to generate a set of optimal control parameter sequences that maximize future rewards. When the prediction accuracy of the jet electrochemical deposition rate adaptive control model for predicting the optimal control parameter sequence is higher than a preset prediction accuracy threshold, the training of the jet electrochemical deposition rate adaptive control model is complete.

6. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, The optimized operating parameters of the chemical additive manufacturing equipment are generated using the aforementioned jet electrochemical deposition rate adaptive control model, specifically: Real-time data information from the chemical additive manufacturing equipment is acquired, and the real-time data information from the chemical additive manufacturing equipment is input into the jet electrochemical deposition rate adaptive control model for judgment and prediction. By judging and predicting, the optimized operating parameters of the chemical additive manufacturing equipment are obtained, and the optimized parameters are sent to the physical equipment for execution. The actual deposition data collected by the physical equipment is compared with the predicted deposition rate and predicted deposition morphology to generate residuals. When the residual is greater than a preset residual threshold, the model parameters in the physical model are dynamically adjusted to correct the physical model.

7. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, A visualization system is built based on the collected data, specifically as follows: By using the topographic information containing point cloud data collected by industrial cameras, and combining it with the equipment encoder information, a three-dimensional point cloud map of the deposition environment is constructed. The three-dimensional ion concentration distribution is superimposed on the real deposition area in the form of a translucent, color-gradient "ion cloud" and visualized as a dynamic, flowing luminous line, showing the electric field distribution calculated by the twin. A floating panel on the side of the field of view displays real-time deposition rate, current efficiency, and cumulative deposition amount, and AR glasses are used as the display and human-computer interaction terminal to form a visualization system.

8. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, The visualization system is used to interactively control the chemical additive manufacturing equipment, specifically as follows: The AR glasses use a built-in depth camera to recognize the operator's gesture commands and then use the recognition results to interactively control the chemical additive manufacturing equipment. When an operator detects a potential anomaly through the AR interface, they can mark it with a gesture, package it according to the marked timestamp and multimodal sensor data, and provide an anomaly notification.

9. The adaptive control method for jet electrochemical deposition rate according to claim 1, characterized in that, Also includes: All chemical additive manufacturing equipment is networked and data packets are uploaded to a cloud server according to a unified data standard. The data packets include chemical additive manufacturing equipment, process parameters, and quality index data. Automated feature extraction is performed on massive process parameter time series data in the cloud to obtain statistical features, frequency domain features, and time domain features, and feature vectors are constructed based on the statistical features, frequency domain features, and time domain features. Based on the feature vectors, single-point, isolated process experiences are formed and transformed into a structured, associative, and reasonable networked process knowledge graph.

10. An adaptive control system for jet electrochemical deposition rate, characterized in that, The device includes a memory and a processor. The memory includes a jet electrochemical deposition rate adaptive control method program. When the jet electrochemical deposition rate adaptive control method program is executed by the processor, it implements the steps of the jet electrochemical deposition rate adaptive control method as described in any one of claims 1-9.

Images