Multi-field coupling closed-loop control method for 3D printing co-extrusion equipment

Abstract

The invention provides a field coupling closed-loop control method for 3D printing co-extrusion equipment. The field coupling closed-loop control method comprises the steps that target printing piece parameters are input; 3D printing co-extrusion equipment is started for preheating, temperature, pressure, material state field and flow velocity multi-field coupling state data are collected, and a coupling state value C is generated through fusion in combination with a multi-field coupling physical model; in the printing process, an inter-field coupling correlation factor K is calculated based on a partial correlation analysis algorithm by collecting multi-field data in real time, and a coupling state value corresponding to a preset printing layer number is predicted through a multi-field coupling physical model; combining the predicted coupling state value with the coupling correlation factor and an optimization algorithm to solve an optimal control parameter, and driving an actuator to respond according to the priority to realize closed-loop regulation and control; and after the preset number of printing layers is printed, the prediction error is determined by collecting the actual multi-field state value and the prediction coupling state value in real time, the optimization algorithm is corrected, precise control over the 3D printing co-extrusion technological process is achieved, and the printing effect is good.

Description

Technical Field

[0001] This invention relates to the field of 3D printing technology, and more specifically to a multi-field coupled closed-loop control method for 3D printing co-extrusion equipment. Background Technology

[0002] 3D printing co-extrusion technology, through the synergistic feeding and melt mixing of filaments and granules, can produce high-performance, multifunctional composite parts, with broad application prospects in aerospace, automotive manufacturing, and other fields. However, existing co-extrusion control technologies have the following drawbacks: Traditional control methods often adjust a single physical field (such as temperature or pressure) independently, ignoring the coupling relationship between temperature, pressure, material state, and flow field, which can easily lead to problems such as uneven material melting, inaccurate mixing ratios, and fluctuations in extruded filament width. Furthermore, the use of open-loop or semi-closed-loop control logic lacks real-time data feedback and algorithm self-optimization mechanisms, making it difficult to adapt to dynamic changes in different material combinations and printing conditions, resulting in lower part forming accuracy and consistency.

[0003] Current 3D printers only detect single temperature and pressure fields, without monitoring material state fields or flow fields. Temperature fluctuations affect pressure changes, which in turn affect material flow field changes. Furthermore, material state fields such as melt density and mixing uniformity are only indirectly inferred from temperature, lacking direct measurement methods. The relationship between temperature and melt density is not a simple linear one, but is significantly affected by material residence time and thermal conductivity. Relying solely on temperature cannot accurately determine the actual melt quality of the material. Mixing uniformity also lacks effective monitoring, and can only be indirectly assessed through destructive testing of the printed parts, without real-time feedback and adjustment. Therefore, it is necessary to develop a dynamic collaborative closed-loop control system based on multi-field coupled sensing to achieve precise control, intelligent optimization, and risk warning of the co-extrusion process.

[0004] For example, patent document with patent application number 202411827503.6 and publication date of April 18, 2025 discloses an optimized 3D printing process based on granular material melting. This optimized 3D printing process based on granular material melting uses an intelligent feeding device and a closed-loop control system to achieve precise control of the material supply, ensuring uniform material supply, thereby improving the interlayer bonding strength of the printed parts and reducing defects caused by uneven material supply. It adopts a multi-stage heating device with zoned temperature control and a dynamic temperature adjustment module, which can automatically adjust the temperature distribution according to the material characteristics and the thickness of the printed layers, ensuring uniform melting of the granular material and reducing the problems of local overheating or insufficient cooling, ultimately improving the mechanical properties of the printed parts. This optimized 3D printing process based on granular material melting utilizes a computer numerical control system and path optimization algorithm, combined with real-time adjustment of the nozzle movement speed and material extrusion rate, to ensure the geometric accuracy and surface quality of the printed parts, while shortening the printing time and improving production efficiency.

[0005] While the aforementioned literature achieves refined parameter control in stages such as granular material pretreatment, quantitative supply, and melt molding, it does not address the co-extrusion synergistic control scenario of filament and granular materials. It only adjusts the temperature distribution of a single granular material. However, temperature changes will lead to changes in pressure and the material state field, thus affecting the printing effect. Therefore, it lacks a multi-material mixing ratio control, multi-physical field coupling perception and dynamic collaborative closed-loop mechanism such as temperature field, pressure field, material state field, and flow field, making it difficult to adapt to the high-precision requirements of composite material co-extrusion molding. At the same time, its risk prediction relies only on single parameter feedback and does not establish a multi-feature fusion risk warning model. It lacks the ability to predict and handle co-extrusion-specific risks such as material blockage and uneven melting, and it does not set up an algorithm self-optimization iteration mechanism, making it difficult to dynamically adjust the control strategy according to different material combinations and printing conditions. Summary of the Invention

[0006] This invention provides a multi-field coupled closed-loop control method for 3D printing co-extrusion equipment. Based on a dynamic collaborative closed-loop control system with multi-field coupled sensing, it achieves precise control of the co-extrusion process and produces good printing results.

[0007] To achieve the above objectives, the technical solution of the present invention is: a multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment, used in a 3D printing co-extrusion equipment for printing two or more materials together, the 3D printing co-extrusion equipment being equipped with a temperature detection device, a pressure detection device, a flow rate detection device, and a material state field detection device, comprising the following steps: S1 Input the target print parameters; S2 starts the 3D printing co-extrusion equipment for preheating, and collects multi-field coupled state data such as temperature, pressure, material state field, and flow rate. It then combines the multi-field coupled physical model to generate the coupled state value C. During the S3 printing process, based on real-time acquisition of multi-field data, the inter-field coupling correlation factor K is calculated through a partial correlation analysis algorithm. The coupling state value corresponding to the preset number of printing layers is predicted through a multi-field coupling physical model. The predicted coupling state value, coupling correlation factor, and optimization algorithm are used to solve for the optimal control parameters and drive the actuator response according to priority to achieve closed-loop control. After the S4 printer presets the number of printing layers, it collects the actual multi-field state values ​​and the predicted coupled state values ​​to determine the prediction error and corrects and optimizes the algorithm.

[0008] The above settings first determine the parameters of the target printed part, then collect parameters from multiple fields such as temperature, pressure, material state field, and flow rate, and form coupled state values ​​from these multiple field parameters. Based on these coupled state values, inter-field coupling correlation factors are determined, establishing the coupling correlation between temperature and pressure, pressure and flow rate, and flow rate and the object state field. A prediction model then predicts the coupled state values, and the adjustment parameters are determined by comparing the predicted coupled state values ​​with the coupling correlation factors. This ensures that the predicted coupled state values ​​correspond to the coupling correlation factors, determining the final optimal control parameters. These optimal control parameters are then adjusted in a prioritized order, ensuring that significantly influential control parameters are adjusted promptly. This parameter adjustment achieves closed-loop control of the 3D printing co-extrusion equipment, ensuring coordinated control of temperature, pressure, flow rate, and melt density. Furthermore, after printing a preset number of layers, real-time collection of multi-field data values ​​and comparison with predicted values ​​determines the error, further refining the control model and making the prediction model more accurate, facilitating more precise subsequent control.

[0009] Furthermore, the target printing parameters mentioned in step S1 include the target coupling state value C. tar =0-100 points, target temperature T tar Target pressure P tar Target mixing ratio γ tar and target melt density M tar ≥95%, manufacturing dimensional accuracy value.

[0010] The above settings, in step S1, clearly configure the relevant modules, target parameters and process templates for four-field coupling control. The target parameters cover core dimensions such as coupling state, temperature and pressure, mixing ratio and melt degree, and clearly define the quantitative standards of melt degree ≥95% and coupling state value 0-100 points, avoiding initial setting deviations caused by parameter ambiguity, and ensuring that the basic conditions for co-extrusion of filament and granules are accurate and controllable.

[0011] Furthermore, the multi-field coupled physical model includes a temperature-material state coupling model, a material state field and viscosity coupling model, a pressure field and flow field coupling model, a flow field and mixing uniformity coupling model, and a multi-field overall coupling model. The temperature-material state coupling model, the material state field and viscosity coupling model, the pressure field and flow field coupling model, and the flow field and mixing uniformity coupling model determine the correlation functions between temperature, material state field, viscosity, pressure, and flow velocity. The multi-field overall coupling model represents the relationship between the correlation functions.

[0012] The above settings, through the correlation coefficients between temperature and material state, material state field and viscosity, pressure and flow rate, and flow rate and mixing uniformity, comprehensively obtain a multi-field total coupling model. This first determines the preliminary correlations, and finally performs a comprehensive determination, so that temperature, material state field, viscosity, pressure, flow rate, and mixing uniformity can be comprehensively correlated.

[0013] Furthermore, the temperature-material state coupling model is as follows: M is the melt density of the material. max T represents the maximum melting point of the material, and T represents the zone temperature. mo t is the critical melting temperature of the material, t is the residence time of the material in this region, and k1 is the material specificity coefficient determined experimentally. The coupling model of material state field and viscosity is as follows , For material viscosity, The reference viscosity is the viscosity at temperature T0 when M=100%. The reference temperature is k2, and k3 are material specificity coefficients. The coupled pressure field and flow field model is ΔP=(8 η L A v) / (π R 4 ), Let L be the pressure difference between the two ends of the flow channel, L be the length of the flow channel, A be the cross-sectional area of ​​the flow channel, v be the average flow velocity, and R be the radius of the flow channel. The coupled model of flow field and mixing uniformity is as follows , For a mixing uniformity of 0-100%, To achieve minimum mixing uniformity, To achieve maximum mixing uniformity, For the velocity gradient, k is the mixing time of the material in the mixing chamber, and k4 is the mixing coefficient; Multi-field total coupling model f1, f2 and f3 are the coupling correlation functions of each field.

[0014] The above settings, through the temperature and material state coupling model, introduce a material-specific coefficient k1, which clarifies the influence of temperature and residence time on melt density M, and also distinguishes the melting characteristics of different materials through experimental calibration, ensuring that the melt density calculation closely matches the actual material properties. The material state field and viscosity coupling model quantifies the correlation between melt density M, temperature T, viscosity η, and mixing uniformity U, and provides coefficient examples for specific materials such as PEEK, providing accurate viscosity data support for the subsequent coordinated control of pressure field and flow field. The pressure field and flow field coupling model directly correlates pressure difference, viscosity, channel parameters, and volumetric flow rate, while the flow field and mixing uniformity coupling model links velocity gradient, mixing time, and uniformity, realizing the quantitative correlation of multiple physical field parameters. The multi-field overall coupling model integrates various sub-models through correlation functions to form a complete multi-field coupling logic, providing reliable model support for the precise control of the co-extrusion process of filament and granular materials. This model shows that the higher the material melt density M and the higher the temperature, the lower the material viscosity, providing a basis for the correlation between pressure field and flow field.

[0015] The multi-field total coupling model F is a composite function of T, P, and M, through... The two intermediate variables, v and v, are linked together to ultimately output the "state vector" of the flow field (including flow velocity magnitude, direction, distribution uniformity, and shear rate, etc.).

[0016] The multi-field total coupling model F is a composite function of T, P, and M, through... The two intermediate variables, v and v, are linked together to ultimately output the "state vector" of the flow field (including flow velocity magnitude, direction, distribution uniformity, and shear rate, etc.).

[0017] First, through Calculate physical viscosity (By calling the material state field and viscosity coupling model) we get: = = ; Then through Calculate the flow velocity v (using the pressure field and flow field coupling model) to obtain: v = = ; Then through The flow field F is calculated by calling the coupled flow field and mixing homogeneity model, resulting in: F = = ,in For shear rate, The velocity gradient together constitutes the flow field state vector; Final coupling: The multi-field total coupling model F consists of v, , They are jointly defined, and all three are derived from T, P, and M through f1 and f2, achieving deep integration of the four fields.

[0018] Furthermore, step S2 also includes: determining whether the initial coupling state value C0 is within a preset range. If it is not, it is determined as an outlier and removed. The preset range is C0 ≥ 90 points, and the temperature deviation ≤ ±0.3℃ and the material melting degree M ≥ 95%.

[0019] The above settings, including the preset standards, data processing methods, and acquisition frequency for the initial coupling state value C0 in step S2, are tailored to the actual needs of co-extrusion printing and offer significant advantages: clearly defining the core threshold of C0 ≥ 90 points, and simultaneously requiring temperature deviation ≤ ±0.3℃ and material melt density M ≥ 95%. By strictly controlling the initial state before printing starts through multi-dimensional quantitative indicators, it ensures that both filament and granular materials reach the ideal conditions of stable melting and temperature-pressure synergy, thus avoiding defects such as poor interlayer bonding and fluctuating extruded filament width caused by substandard initial states from the source.

[0020] Furthermore, the inter-field coupling correlation factor K includes six core factors, namely temperature and material factors K. tm Temperature and pressure factor K tp Temperature and flow field factor K t f. Material and pressure factor K mp Material and flow field factor K m f. Pressure and flow field factor K p f, where each factor takes values ​​ranging from 0 to 1, is calculated using a partial correlation analysis algorithm. It is normally updated every 100ms, but when the coupling relationship changes abruptly, the update frequency is shortened to once every 50ms.

[0021] The above settings comprehensively cover all coupling relationships between temperature, pressure, material state, and flow field with six core factors, leaving no dimension out. They can clearly quantify the intensity of the mutual influence of each physical field, providing a clear basis for subsequent priority control. The value range of each factor from 0 to 1 is simple and intuitive. The calculation through partial correlation analysis algorithm ensures the scientific nature of the results and can accurately identify the inter-field correlations that have a key impact on the co-extrusion state. The regular 100ms update frequency can meet the real-time control requirements of a stable printing process, and the update frequency is shortened to 50ms when the coupling relationship changes abruptly, which can quickly capture drastic changes in inter-field correlations.

[0022] Furthermore, step S3, "predicting the coupling state value corresponding to the preset number of printing layers through a multi-field coupled physical model, and solving for the optimal control parameters by combining the predicted coupling state value with the coupling correlation factor and optimization algorithm," includes: S31 constructing a multi-field coupled physical model to predict the coupling state value corresponding to the preset number of layers; S32 determining the cost function as... w1, w2, w3, and w4 are all weighting coefficients, which are strongly correlated with the coupling correlation factor K. The sum of squares of the deviations between the coupling state values ​​and the target value over the next 5 cycles; The square of the mixing uniformity deviation; To produce dimensions with squared accuracy; S33 defines the sum of squares of the adjustment amounts of each actuator; S33 determines the constraint conditions, which are the temperature change ΔT∈[-10,10]℃ and the speed change Δv∈[-20%,20%]; S34 uses the gradient descent algorithm to solve for the optimal control parameters, which include temperature, feeding speed, and stirring intensity.

[0023] The above settings, K tm The higher the temperature and material factor, the greater the weight of w1, prioritizing the coordination between temperature and material state, making the control strategy more targeted. The combination of the prediction dimension of the future preset period and the gradient descent algorithm not only achieves the foresight of the control and avoids the risk of parameter fluctuation in advance, but also quickly solves the optimal adjustment of core parameters such as temperature, feeding speed and stirring intensity, ensuring that the control command is accurate and efficient.

[0024] Furthermore, the priority-driven actuator response in step S3 includes: coupling the temperature and material factor K among the correlation factors. tm When the pressure is ≥0.7, prioritize adjusting the temperature control and feeding coordination parameters; when the pressure and flow field factor K p When f > 0.8, priority is given to adjusting the pressure and flow rate parameters together. When the material and pressure factor K mp When the value is >0.6, the material state field and pressure parameters are optimized simultaneously. The actuator includes a dynamic temperature control unit, a precise feeding and mixing unit, and an adaptive flow channel unit, with a response delay ≤0.2s.

[0025] The above settings, including the priority allocation rules and high-performance actuator configuration clearly defined in step S3, provide a strong guarantee for the precise and efficient control of the co-extrusion process. The priority adjustment level is determined according to the values ​​of temperature and material factors, pressure and flow field factors, and material and pressure factors, ensuring that the parameter adjustment of each link is orderly and highly targeted. The configuration of the dynamic temperature control unit, the precise feeding and mixing unit, and the adaptive flow channel unit accurately matches the control requirements of different priorities, while the fast response delay of ≤0.2s ensures that the optimal control parameters are implemented in a timely manner, avoiding the imbalance of the inter-field coupling state due to response lag.

[0026] Furthermore, step S4 also includes: after printing a preset number of layers, the system optimizes the calculation model of the weight coefficients w1, w2, w3, w4 and the coupling correlation factor K of the cost function based on the part quality score; every 10 layers of printing data are accumulated, the system automatically generates and saves the optimized process parameter template for the current material combination; at the same time, the parameters k1, k2, k3, k4 of the multi-field coupling physical model are calibrated online using a melt flow indexer.

[0027] The above settings optimize the calculation model based on the part quality score, using the weight coefficients w1, w2, w3, and w4 of the algorithm and the coupling correlation factor K. This promptly corrects control deviations caused by fluctuations in material properties and changes in operating conditions during the printing process, ensuring the algorithm always closely matches the actual printing status. Every 10 layers of data are accumulated, an optimized process template for the current material combination is automatically generated and stored. Subsequent similar printing tasks can directly call this template without repeated debugging, significantly improving work efficiency and parameter adaptability. The melt flow indexer is used to calibrate the parameters k1, k2, k3, and k4 of the four-field coupled physical model online, ensuring that the model calculations highly match the actual melting, viscosity, and mixing characteristics of the material, providing reliable model support for closed-loop control.

[0028] Furthermore, step S4 also includes: S5 determining the risk level based on multiple field parameters such as temperature, pressure, material state field, and flow rate, as well as LSTM neural network prediction, and processing it according to a graded handling strategy. The graded handling strategy is as follows: Early warning level: risk probability 30%-50%, automatically fine-tuning the parameters of the corresponding control actuator and continuously monitoring; Alarm level: risk probability 50%-80%, pausing printing and starting the correction program; Emergency level: risk probability >80%, immediately cutting off the heating power supply, starting the dynamic pressure relief valve to perform emergency handling, and using a quick-release anti-blocking component to assist in fault diagnosis; After the fault handling is completed, the system records the handling process data and updates the risk prediction model.

[0029] The above settings include: Step S4, the early warning level automatically fine-tunes parameters and continuously monitors to mitigate minor risks without interrupting printing; the alarm level pauses printing and initiates a dedicated correction program to precisely address moderate risks and prevent defects from escalating; the emergency level immediately cuts off the heating power, activates the dynamic pressure relief valve, and uses quick-release anti-blocking components to minimize damage to equipment and parts caused by serious risks such as material blockage and melt leakage, balancing production continuity and safety. Furthermore, the system's design of recording data and updating the risk prediction model after fault handling allows for the continuous accumulation of risk response experience.

[0030] Furthermore, the temperature detection device is installed in the material conveying channel, the mixing chamber, and the nozzle flow channel; the material state field measurement device includes a miniature dielectric sensor; the pressure detection device consists of eight flexible thin-film pressure sensors arranged in a gradient along the inner wall of the nozzle flow channel in the mixing chamber; and the flow field detection device includes two laser Doppler velocimetry sensors.

[0031] The above setup includes a temperature detection device with 12 temperature measurement points along the entire process of the filament material channel, granular material channel, mixing chamber, and nozzle flow channel. Each measurement point uses a dual-mode sensor combining a K-type thermocouple and infrared temperature measurement, which can accurately collect temperature field data T1 from different dimensions in key areas, avoiding errors from a single temperature measurement method and ensuring comprehensive temperature monitoring of filament and granular materials throughout the entire process of conveying, mixing, and extrusion. Four micro dielectric sensing points specifically collect material state field data M1, enabling real-time monitoring of material melt changes in different channels and the mixing chamber, providing a direct basis for subsequent viscosity calculations and parameter adjustments. Eight flexible thin-film pressure sensing points are gradient-distributed along the inner wall of the mixing chamber and the nozzle flow channel, which can accurately capture the gradient distribution of pressure field data P1 within the flow channel and promptly detect abnormal pressure fluctuations. Two laser Doppler velocity sensing points, combined with pressure data, infer the overall flow velocity distribution to obtain flow field data v1, achieving comprehensive perception of the flow field state without the need for a full-area sensor deployment, reducing hardware costs while ensuring monitoring coverage. Attached Figure Description

[0032] Figure 1 This is a flowchart of the present invention.

[0033] Figure 2 This is a block diagram of the key channel structure of the 3D printing equipment of the present invention. Detailed Implementation

[0034] like Figure 1-2 As shown, a multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment is applied to a 3D printing co-extrusion equipment that involves the mixing and printing of two or more materials. The 3D printing co-extrusion equipment is equipped with a temperature detection device, a pressure detection device, a flow rate detection device, and a material state field detection device. The temperature detection device is located in the material conveying channel, the mixing chamber, and the nozzle flow channel. The material state field measurement device includes a micro dielectric sensor. The pressure detection device consists of eight flexible thin-film pressure sensors arranged in a gradient along the inner wall of the nozzle flow channel in the mixing chamber. The flow field detection device includes two laser Doppler velocimetry sensors.

[0035] In this embodiment, the two materials are filaments and granules, which are fed by a filament feeding device and a granule feeding device, respectively. The filament feeding device includes a filament feeding channel, and the granule feeding device includes a granule feeding channel. A temperature detection device is set with 12 temperature measuring points arranged sequentially along the filament material channel, the granule material channel, the mixing chamber, and the nozzle flow channel. Each temperature measuring point is equipped with a K-type thermocouple and an infrared temperature measurement dual-mode sensor to collect temperature field data T1. The material state field includes 4 miniature dielectric sensors to collect material state field data M1. The pressure detection device includes 8 flexible thin film pressure sensors arranged in a gradient along the inner wall of the nozzle flow channel of the mixing chamber to collect pressure field data P1. The flow field detection device includes 2 laser Doppler velocity sensors, which, combined with the pressure data, infer the flow velocity distribution over the entire flow channel to obtain flow field data v1.

[0036] A multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment includes the following steps: S1 inputs the target print parameters, which include the target coupling state value C. tar =0-100 points, target temperature T tar Target pressure P tar Target mixing ratio γ tar and target melt density M tar ≥95%, manufacturing dimensional accuracy value; S2 starts the 3D printing co-extrusion equipment for preheating, and collects multi-field coupled state data such as temperature, pressure, material state field, and flow rate. It then combines the multi-field coupled physical model to generate the coupled state value C. During the S3 printing process, based on real-time acquisition of multi-field data, the inter-field coupling correlation factor K is calculated through a partial correlation analysis algorithm. The coupling state value corresponding to the preset number of printing layers is predicted through a multi-field coupling physical model. The predicted coupling state value, coupling correlation factor, and optimization algorithm are used to solve for the optimal control parameters and drive the actuator response according to priority to achieve closed-loop control. After the S4 printer presets the number of printing layers, it collects the actual multi-field state values ​​and the predicted coupled state values ​​to determine the prediction error and corrects and optimizes the algorithm. The weight coefficients in the specific correction algorithm can be implemented using the PID algorithm.

[0037] S5 determines the risk level based on multiple field parameters including temperature, pressure, material state field, and flow rate, as well as LSTM neural network prediction, and processes the risk according to a graded handling strategy. The graded handling strategy is as follows: Early warning level: risk probability 30%-50%, automatically fine-tunes the parameters of the corresponding control actuator and continuously monitors; Alarm level: risk probability 50%-80%, suspends printing and starts the correction program; Emergency level: risk probability >80%, immediately cuts off the heating power supply, starts the dynamic pressure relief valve to perform emergency handling, and uses a quick-release anti-blocking component to assist in fault diagnosis; after the fault handling is completed, the system records the handling process data and updates the risk prediction model.

[0038] The multi-field coupled physical model includes a temperature-material state coupling model, a material state field and viscosity coupling model, a pressure field and flow field coupling model, a flow field and mixing homogeneity coupling model, and a multi-field overall coupling model. The temperature-material state coupling model, the material state field and viscosity coupling model, the pressure field and flow field coupling model, and the flow field and mixing homogeneity coupling model determine the correlation functions between temperature, material state field, viscosity, pressure, and flow velocity. The multi-field overall coupling model represents the relationship between the correlation functions.

[0039] The temperature-material state coupling model is as follows: Where M is the melt density of the material (0-100%), M max T represents the maximum melt density of the material (100%), and T represents the zone temperature (°C). mo t is the critical melting temperature of the material (°C), t is the residence time of the material in this region (s), and k1 is the material specificity coefficient determined experimentally. The meltability M of the material increases with increasing temperature and residence time, and the melting characteristics of different materials are distinguished by the k1 coefficient. For example, when PEEK material is held at 380°C for 2 seconds, the melt density M = When held for 1.5 seconds, the degree of melt flow is M = 90.1%.

[0040] The coupling model of material state field and viscosity is as follows , Material viscosity (Pa) s), Reference viscosity (Pa) s, viscosity at temperature T0 when M=100%) Here, k2 and k3 are the reference temperature (°C), and k2 and k3 are the material specificity coefficients (e.g., for PEEK material, k2 = 0.8, k3 = 0.01°C). - ¹). This model shows that the higher the melt density M and the higher the temperature, the lower the material viscosity, providing a basis for the correlation between the pressure field and the flow field; for example, for PEEK material, k2=0.8, k3=0.01℃. - ¹.

[0041] Pressure field and flow field coupled model , Let L be the pressure difference between the two ends of the flow channel (Pa), L be the length of the flow channel (m), Q be the volumetric flow rate (m³ / s), Q = A * V, and R be the radius of the flow channel (m). For example, the flow path parameters are L=0.1m, R=0.005m, and A=7.85×10⁻⁶. -5 m², material viscosity =500Pa Given an average flow velocity v = 0.05 m / s, the pressure difference can be calculated. =203718Pa≈0.2Mpa.

[0042] The coupled model of flow field and mixing uniformity is as follows , For a mixing uniformity of 0-100%, To achieve minimum mixing uniformity, To achieve maximum mixing uniformity, The velocity gradient (s) - ¹), k is the mixing time (s) of the material in the mixing chamber, and k4 is the mixing coefficient; Multi-field total coupling model f1, f2, and f3 are the coupling correlation functions of each field, which are obtained through the correlation values ​​of the functions.

[0043] The multi-field total coupling model F is a composite function of T, P, and M, through... The two intermediate variables, v and v, are linked together to ultimately output the "state vector" of the flow field (including flow velocity magnitude, direction, distribution uniformity, and shear rate, etc.).

[0044] First, through Calculate physical viscosity (By calling the material state field and viscosity coupling model) we get: = = ; Then through Calculate the flow velocity v (using the pressure field and flow field coupling model) to obtain: v = = ; Then through The flow field F is calculated by calling the coupled flow field and mixing homogeneity model, resulting in: F = = ,in For shear rate, The velocity gradient together constitutes the flow field state vector; Final coupling: The multi-field total coupling model F consists of v, , They are jointly defined, and all three are derived from T, P, and M through f1 and f2, achieving deep integration of the four fields.

[0045] In step S2, it is determined whether the initial coupling state value C0 is within the preset range. If it is not, it is determined as an outlier and is removed. The preset range is C0 ≥ 90 points, and the temperature deviation ≤ ±0.3℃ and the material melting degree M ≥ 95% are simultaneously satisfied. The initial data acquisition frequency is 200Hz.

[0046] The specific calculation process for determining the fraction of the coupling state value is as follows: Data from each field is standardized to the [0,1] interval. For example, the standardization formula for temperature data is... T min T max The minimum and maximum temperatures for this material combination are given; the standardized formula for pressure data is... P is the measured pressure. min To achieve the minimum stable co-extrusion pressure, P max Maximum pressure that the flow channel can withstand; pressure gradient normalization ΔP min To minimize the pressure gradient during stable extrusion, ΔP max To block the precursor critical pressure gradient; to standardize the melt density of the object. M represents the measured melt density (0-100%), directly mapped to [0,1]; the mixing uniformity is standardized. U represents the measured mixing uniformity (0-100%), directly mapped to [0,1]; flow velocity data are standardized. v is the measured flow velocity, v min For the minimum steady flow velocity (5 mm / s), v max Maximum safe flow rate (200 mm / s); flow rate gradient normalization , v is the measured velocity gradient. v min The minimum velocity gradient is required for uniform mixing. v max For the mixed limiting velocity gradient (50s) - ¹); then construct a coupled state value calculation model. C represents the multi-field coupling state value (0-100 points), with a higher score indicating a better overall synergistic state of the co-extrusion system; T'_avg represents the standardized mean of the temperature field, calculated by weighted averaging the temperature mean of multiple core areas, including the filament material channel, granular material channel, blending cavity, and nozzle flow channel. ΔP'_avg is the pressure gradient normalized value, M'_avg is the mean melt density of the four micro-dielectric sensing points (normalized mean), U' is the mixing uniformity normalized value, and v'_avg is the flow rate normalized mean of 10 flow rate data points. v'_avg is the standardized value of the velocity gradient; a, b, c, d, e, f, and g are weighting coefficients, covering more than 20 types including PEEK, PA, and PLA, based on over 5000 sets of experimental data. For example, the coefficients for the PEEK+CF-PEEK combination are a=0.25, b=0.2, c=0.25, d=0.15, e=0.1, f=0.03, and g=2.5. The inter-field coupling correlation factor K includes six core factors, namely temperature and material factors K. tm Temperature and pressure factor K tp Temperature and flow field factor K t f. Material and pressure factor K mp Material and flow field factor K m f. Pressure and flow field factor K p f, where each factor takes values ​​ranging from 0 to 1, is calculated using a partial correlation analysis algorithm. Normally, the algorithm updates every 100ms, but when the coupling relationship changes abruptly, the update frequency is shortened to 50ms. In this embodiment, the partial correlation analysis algorithm can be calculated by inputting two sets of related data into SPSS statistical software, while keeping other sets of data unchanged.

[0047] Partial correlation analysis can eliminate interference from other fields and accurately reflect the degree of direct correlation between two fields. The calculation formula is as follows: , where r x When z is the control variable, the partial correlation coefficient between variables x and y; r x r x z、r z represents the Pearson correlation coefficients between x and y, x and z, and y and z, respectively. For example, calculating K... tm When determining the correlation factor between the temperature field and the material state field, the controlled variables are the pressure field and flow field data. The specific steps are: calculate the Pearson correlation coefficient r between the standardized temperature field data (T') and the standardized material state field data (M'). tm ; Calculate the correlation coefficient r between temperature field data and pressure field data (P'). tp The correlation coefficient r between temperature field data and flow field data (F') t f; Calculate the correlation coefficient r between the material state field data and the pressure field data. mp The correlation coefficient r between material state field data and flow field datam f; Using the extended formula of multivariate partial correlation analysis, calculate the partial correlation coefficient between T' and M' when controlling for P' and F', which is K. tm .

[0048] The optimization algorithm described in step S3 is an improved model predictive control (MPC) algorithm, specifically including: S31 constructing a multi-field coupled physical model, specifically including a temperature-material state coupling model, a material state field and viscosity coupling model, a pressure field and flow field coupling model, a flow field and mixing homogeneity coupling model, and a total multi-field coupling model. The correlation functions between temperature, material state field, viscosity, pressure, and flow velocity are determined through the temperature-material state coupling model, the material state field and viscosity coupling model, the pressure field and flow field coupling model, and the flow field and mixing homogeneity coupling model. The total multi-field coupling model represents the relationship between these correlation functions, predicting the coupling state values ​​corresponding to a preset number of layers; S32 determining the cost function as... w1, w2, w3, and w4 are all weighting coefficients, which are strongly correlated with the coupling correlation factor K. The sum of squares of the deviations between the coupling state values ​​and the target value over the next 5 cycles; The square of the mixing uniformity deviation; To produce dimensions with squared accuracy; S33 defines the sum of squares of the adjustment amounts of each actuator; S33 determines the constraint conditions, which are the temperature change ΔT∈[-10,10]℃ and the speed change Δv∈[-20%,20%]; S34 uses the gradient descent algorithm to solve for the optimal control parameters, which include temperature, feeding speed, and stirring intensity. The gradient descent algorithm in this embodiment is a prior art technique.

[0049] In the S35 coupling correlation factor, temperature and material factor K tm When the pressure is ≥0.7, prioritize adjusting the temperature control and feeding coordination parameters; when the pressure and flow field factor K p When f > 0.8, priority is given to adjusting the pressure and flow rate parameters together. When the material and pressure factor K mp When the value is >0.6, the material state field and pressure parameters are optimized simultaneously. The actuator includes a dynamic temperature control unit, a precise feeding and mixing unit, and an adaptive flow channel unit, with a response delay ≤0.2s.

[0050] In step S4, after printing 5 layers of parts, the system optimizes the calculation model of the weight coefficients w1, w2, w3, and w4 of the MPC algorithm and the coupling correlation factor K based on the part quality score. Specifically, the optimization is performed by the correspondence between the weight coefficients of the MPV algorithm corresponding to the quality score and the coupling correlation factor. After accumulating 10 layers of printing data, the system automatically generates and stores the optimized process parameter template for the current material combination. At the same time, the parameters k1, k2, k3, and k4 of the multi-field coupled physical model are calibrated online using a melt flow indexer.

[0051] Specific examples are as follows: 3D printing equipment: MAGIC-HT-PRO (fused filament extrusion process, dual nozzle system, nozzle diameter 0.6mm, printing platform temperature 200℃, positioning accuracy X / Y: 12.5μm, Z: 1.25μm); Filament material: PEEK (diameter 1.75mm, melting temperature 380℃, thermal decomposition temperature 500℃, density 1.32g / cm³). Granular material: CF-PEEK (particle size 120μm, carbon fiber filler ratio 35%, melt temperature 410℃, density 1.45g / cm³). Component requirements: Lightweight aerospace structural component, dimensions 100mm×50mm×20mm, dimensional accuracy ±0.005mm, surface roughness Ra≤0.3μm, tensile strength≥140MPa, compressive strength≥180MPa, internal porosity≤0.5%. The system calls the "PEEK+CF-PEEK+Aerospace" parameter template, and the final process parameters generated after dynamic correction are as follows: Dynamic temperature control parameters: Filament material channel: inlet 375℃, middle section 385℃, outlet 388℃ (power distribution ratio 30%:40%:30%). Granular material channel: inlet 400℃, middle section 415℃, outlet 420℃ (power distribution ratio 25%:45%:30%). Mixing cavity gradient: inlet 420℃, middle 428℃, outlet 425℃ (dynamic gradient mode); Nozzle flow path: inlet 428℃, middle section 430℃, outlet 435℃; Nozzle temperature compensation: +5℃; Pressure threshold range: Blending chamber inlet: 2.5-3.0MPa; Blending chamber middle: 3.0-3.5MPa; Blending chamber outlet: 2.8-3.3MPa; Nozzle outlet: 2.0-2.8MPa; Precise feeding parameters: Filament feed rate: 55mm / s; Particle feed rate: 23.6mm / s; Mixing ratio (mass ratio): 7:3; Feeding accuracy: ±0.1mm / s; Mixing intensity parameters: Stirring speed: 200-250 r / min (dynamically adjustable); Magnetic field strength: 1.0-1.3T (dynamically adjustable); Adaptive flow channel parameters: Flow channel shrinkage ratio: first section 1.1, second section 1.2, third section 1.1; Nozzle outlet diameter: 0.6 mm; Vibration frequency: 50 Hz (anti-adhesion); Printing auxiliary parameters: Printing speed: 80 mm / s; Layer thickness: 0.1 mm; Fill rate: 80%; Printing path: Spiral path. The specific method and process are as follows: 1. Initialization phase: The system starts preheating according to the above parameters. After 20 minutes, the temperature of each area reaches the set value, the material melt degree is ≥95%, the multi-field coupling state value is 92 points, and it enters the standby state. 2. Printing Operation Phase: The multi-field coupling sensing module collects multi-field data in real time, and the multi-field dynamic collaborative control unit continuously optimizes control parameters. When printing to the 30th layer, the pressure in the middle of the mixing chamber rises to 3.6 MPa. The system automatically adjusts the flow channel shrinkage ratio and heating power, and the pressure recovers to 3.4 MPa after 1 second. When printing to the 60th layer, the mixing uniformity drops to 87%. The system increases the stirring intensity, and the mixing uniformity recovers to 93% after 2 seconds. During the printing process, a warning-level risk (slight material adhesion) occurs once. The system automatically activates flow channel vibration and temperature compensation, and the risk is quickly resolved without affecting the printing process. 3. Part Inspection Results: Dimensional Accuracy: Measured using a coordinate measuring machine (accuracy ±0.001mm), the actual dimensions of the part are 100.002mm × 50.001mm × 19.998mm, with a dimensional accuracy of ±0.002mm, meeting the ±0.005mm requirement; Surface Quality: Measured using a roughness tester (accuracy ±0.01μm), the surface roughness Ra=0.25μm, meeting the Ra≤0.3μm requirement; Internal Structure: Measured using a micro-focus CT (resolution 0.01mm), the internal porosity is 0.2%, with no obvious delamination or defects; Mechanical properties: Tested using a universal testing machine (maximum load 100kN), the tensile strength was 152MPa and the compressive strength was 185MPa, both far exceeding the target requirements; 4. Production efficiency and material utilization rate: Total printing time: 8 hours (including 20 minutes of warm-up); Material consumption: 180g of filament material, 77g of granules, total consumption 257g; Material waste amount: 4g, material waste rate: 1.5%; Printing success rate: 100%. The printing test results of this invention and traditional co-extrusion equipment are shown in Table 1.

[0052] Table 1 The working principle of this invention is as follows: First, the parameters of the target printed part are determined. Then, parameters of multiple fields such as temperature, pressure, material state field, and flow rate are collected, and coupling state values ​​are formed through the parameters of multiple fields. Then, the coupling correlation factors between fields are determined based on the coupling state values, and the coupling correlation relationships between temperature and pressure, pressure and flow rate, and flow rate and object state field are determined. Then, the coupling state values ​​are predicted by a prediction model. The adjustment parameter values ​​are determined by the predicted coupling state values ​​and coupling correlation factors, so that the predicted coupling state values ​​correspond to the coupling correlation factors, and the final optimal control parameters are determined. The optimal control parameters are adjusted in priority according to the set optimal control parameter order, so that the control parameters with greater influence can be adjusted in a timely manner. Thus, closed-loop control of the 3D printing co-extrusion equipment is achieved through parameter adjustment, ensuring coordinated control of multiple aspects such as temperature, pressure, flow rate, and melt density. After printing a preset number of layers, the multi-field data values ​​are collected in real time and compared with the predicted values ​​to determine the error, thereby further correcting the control model, making the prediction model more accurate, and facilitating more precise control in the future.

Claims

1. A multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment, used in a 3D printing co-extrusion equipment for printing two or more materials mixed together, characterized in that: The 3D printing co-extrusion equipment is equipped with a temperature detection device, a pressure detection device, a flow rate detection device, and a material state field detection device, and includes the following steps: S1 Input the target print parameters; S2 starts the 3D printing co-extrusion equipment for preheating, and collects multi-field coupled state data such as temperature, pressure, material state field, and flow rate. It then combines the multi-field coupled physical model to generate the coupled state value C. During the S3 printing process, based on real-time acquisition of multi-field data, the inter-field coupling correlation factor K is calculated through a partial correlation analysis algorithm. The coupling state value corresponding to the preset number of printing layers is predicted through a multi-field coupling physical model. The predicted coupling state value, coupling correlation factor, and optimization algorithm are used to solve for the optimal control parameters and drive the actuator response according to priority to achieve closed-loop control. After the S4 prints the preset number of printing layers, it collects the actual multi-field state values ​​and the predicted coupled state values ​​to determine the prediction error and corrects and optimizes the algorithm.

2. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: The target printing parameters mentioned in step S1 include the target coupling state value C. tar =0-100 points, target temperature T tar Target pressure P tar Target mixing ratio γ tar and target melt density M tar ≥95%, manufacturing dimensional accuracy value.

3. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: The multi-field coupled physical model includes a temperature-material state coupling model, a material state field and viscosity coupling model, a pressure field and flow field coupling model, a flow field and mixing homogeneity coupling model, and a multi-field overall coupling model. The temperature-material state coupling model, the material state field and viscosity coupling model, the pressure field and flow field coupling model, and the flow field and mixing homogeneity coupling model determine the correlation functions between temperature, material state field, viscosity, pressure, and flow velocity. The multi-field overall coupling model represents the relationship between the correlation functions.

4. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: The temperature-material state coupling model is M is the melt density of the material. max T represents the maximum melting point of the material, and T represents the zone temperature. mo t is the critical melting temperature of the material, t is the residence time of the material in this region, and k1 is the material specificity coefficient determined experimentally. The coupling model of material state field and viscosity is as follows , For material viscosity, The reference viscosity is the viscosity at temperature T0 when M=100%. The reference temperature is k2, and k3 are material specificity coefficients. The coupled pressure field and flow field model is ΔP=(8 η L A v) / (π R 4 ), Let L be the pressure difference between the two ends of the flow channel, L be the length of the flow channel, A be the cross-sectional area of ​​the flow channel, v be the average flow velocity, and R be the radius of the flow channel. The coupled model of flow field and mixing uniformity is as follows , For a mixing uniformity of 0-100%, To achieve minimum mixing uniformity, To achieve maximum mixing uniformity, For the velocity gradient, k is the mixing time of the material in the mixing chamber, and k4 is the mixing coefficient; Multi-field total coupling f1, f2 and f3 are the coupling correlation functions of each field.

5. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: Step S2 further includes: determining whether the initial coupling state value C0 is within a preset range. If not, it is determined as an outlier and removed. The preset range is C0 ≥ 90 points, and the temperature deviation ≤ ±0.3℃ and the material melting degree M ≥ 95%.

6. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: The inter-field coupling correlation factor K includes six core factors, namely temperature and material factors K. tm Temperature and pressure factor K tp Temperature and flow field factor K t f. Material and pressure factor K mp Material and flow field factor K m f. Pressure and flow field factor K p f, where each factor takes values ​​ranging from 0 to 1, is calculated using a partial correlation analysis algorithm. It is normally updated every 100ms, but when the coupling relationship changes abruptly, the update frequency is shortened to once every 50ms.

7. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: Step S3, "predicting the coupling state values ​​corresponding to the preset number of printing layers using a multi-field coupled physical model, and solving for the optimal control parameters by combining the predicted coupling state values ​​with the coupling correlation factor and optimization algorithm," includes: S31 constructing a multi-field coupled physical model to predict the coupling state values ​​corresponding to the preset number of layers; S32 determining the cost function as... w1, w2, w3, and w4 are all weighting coefficients, which are strongly correlated with the coupling correlation factor K. The sum of squares of the deviations between the coupling state values ​​and the target value over the next 5 cycles; The square of the mixing uniformity deviation; To produce dimensions with squared accuracy; S33 defines the sum of squares of the adjustment amounts of each actuator; S33 determines the constraint conditions, which are the temperature change ΔT∈[-10,10]℃ and the speed change Δv∈[-20%,20%]; S34 uses the gradient descent algorithm to solve for the optimal control parameters, which include temperature, feeding speed, and stirring intensity.

8. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: The priority-driven actuator response in step S3 includes: the temperature and material factor K in the coupling correlation factors. tm When the pressure is ≥0.7, prioritize adjusting the temperature control and feeding coordination parameters; when the pressure and flow field factor K p When f > 0.8, priority is given to adjusting the pressure and flow rate parameters together. When the material and pressure factor K mp When the value is >0.6, the material state field and pressure parameters are optimized simultaneously. The actuator includes a dynamic temperature control unit, a precise feeding and mixing unit, and an adaptive flow channel unit, with a response delay ≤0.2s.

9. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: Step S4 also includes: after printing a preset number of layers, the system optimizes the calculation model of the weight coefficients w1, w2, w3, w4 and the coupling correlation factor K of the cost function based on the part quality score; every 10 layers of printing data are accumulated, the system automatically generates and saves the optimized process parameter template for the current material combination; at the same time, the parameters k1, k2, k3, k4 of the multi-field coupling physical model are calibrated online using a melt flow indexer.

10. The multi-field coupled closed-loop control method for a 3D printing co-extrusion equipment according to claim 1, characterized in that: Step S4 also includes: S5 determining the risk level based on multiple field parameters such as temperature, pressure, material state field, and flow rate, as well as LSTM neural network prediction, and processing it according to a graded handling strategy. The graded handling strategy is as follows: Early warning level: risk probability 30%-50%, automatically fine-tuning the parameters of the corresponding control actuator and continuously monitoring; Alarm level: risk probability 50%-80%, pausing printing and starting the correction program; Emergency level: risk probability >80%, immediately cutting off the heating power supply, starting the dynamic pressure relief valve to perform emergency handling, and using a quick-release anti-blocking component to assist in fault diagnosis; After the fault handling is completed, the system records the handling process data and updates the risk prediction model. Temperature detection devices are installed in the material conveying channel, mixing chamber, and nozzle flow channel; material state field measurement devices include miniature dielectric sensors; pressure detection devices consist of eight flexible thin-film pressure sensors arranged in a gradient along the inner wall of the nozzle flow channel in the mixing chamber; flow field detection devices include two laser Doppler velocimetry sensors.

Images