A hybrid modeling method for 3D printing concrete pumping and printing processes

Abstract

The present application provides a hybrid modeling method for the pumping and printing process of 3D-printed concrete. By introducing intelligent aggregates, the present invention realizes real-time monitoring and data collection of particle motion behavior during the 3D concrete pumping and printing process. The collected multi-dimensional dynamic data of particle velocity, acceleration, and collision frequency are compared and analyzed with discrete element-computational fluid dynamics (DEM-CFD) coupled simulation results, thereby realizing efficient inversion and precise calibration of the input parameters of the DEM-CFD coupled simulation model. Compared with the unobservability and uncertainty of the particle flow state during traditional concrete pumping, the present invention can achieve "transparency" of the particle motion process. The parameter inversion method proposed in the present invention is based on the fitting optimization of measured particle motion data and numerical simulation results, which significantly reduces manual intervention and experimental overhead.

Description

Technical Field

[0001] The present application relates to the field of civil engineering and building construction technology, and in particular to a hybrid modeling method for 3D printing concrete pumping and printing processes. Background Art

[0002] In recent years, with the development of additive manufacturing technology, 3D-printed concrete has been gradually applied to various fields, including construction, bridges, and prefabricated components, due to its advantages such as high geometric freedom in the construction of complex structures and a high degree of construction automation. However, in the actual application of 3D-printed concrete, the movement of aggregate particles during the pumping and printing process is highly nonlinear, uncertain, and random. During transportation through the pumping pipeline, 3D-printed concrete particles may accumulate, separate, or become blocked, resulting in pumping interruptions. During the printing stage, the uneven distribution of 3D-printed concrete particles may lead to a decrease in interlayer bonding, thereby affecting the molding quality and the overall mechanical properties of the structure.

[0003] Numerical simulation has become an important auxiliary tool to address the above issues. Compared with traditional methods such as the finite element method (FEM) and smoothed particle hydrodynamics (SPH), the discrete element method (DEM) has a natural advantage in simulating the contact behavior and multi-body interactions of granular materials such as aggregates; while computational fluid dynamics (CFD) can effectively capture the continuous medium flow behavior of mortar. The coupling of the two can more realistically reconstruct the particle-pasture flow mechanism. However, the practical problem is that the input parameters required for DEM-CFD coupled simulation are diverse and mutually coupled, and they are extremely dependent on high-quality experimental data for calibration and verification. For example, the inter-particle friction coefficient, fluid viscosity, boundary condition setting, etc., all require repeated experiments and simulations to determine, resulting in long modeling cycles and high computational overhead, which seriously limits its promotion and application in actual engineering.

[0004] Therefore, there is an urgent need for an efficient hybrid modeling method that can integrate real-time measured data and numerical simulation systems to improve parameter calibration efficiency and realize real-time prediction and feedback control of pumping process risks (such as pipe blockage) and printing quality. Summary of the Invention

[0005] The purpose of this application is to provide a hybrid modeling method for the pumping and printing process of 3D printed concrete. This method effectively solves the problem in the existing technology that it is impossible to obtain the movement state of particles and fluids in real time and accurately, and provides comprehensive data support for parameter calibration, optimization and reliability verification of numerical simulation models, which is used to achieve accurate simulation and parameter optimization of 3D printed concrete during the pumping and printing process.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] A hybrid modeling method for 3D printing concrete pumping and printing processes includes the following steps:

[0008] S1. Determine the initial input parameters required for DEM-CFD coupled simulation;

[0009] S2. Setting multiple groups of different parameter value combinations, wherein some parameters have different values ​​in different groups, while the remaining parameters remain unchanged or vary within a small range, thereby effectively covering the input parameter space; performing DEM-CFD coupled numerical simulations on each group to obtain corresponding particle motion simulation results;

[0010] S3. Smart aggregate replaces some aggregates in 3D printing concrete raw materials. During the pumping and printing process, the smart aggregate collects dynamic behavior data of 3D printing concrete raw material particles;

[0011] S4. Compare the simulation results with the measured data and inversely optimize the parameters to achieve high-precision calibration of the DEM-CFD coupled simulation model;

[0012] S5. Build a digital twin control system based on the calibrated DEM-CFD coupling simulation model.

[0013] Furthermore, in step S1, the initial input parameter types include: particle contact parameters, fluid parameters, particle-fluid coupling interaction parameters, and material parameters of 3D printing concrete raw materials;

[0014] The particle contact parameters include normal stiffness, tangential stiffness, friction coefficient, and bonding strength; the fluid parameters include slurry viscosity, density, and flow velocity boundary conditions; the coupling interaction parameters include particle-fluid resistance and drag coefficient; and the material parameters include aggregate particle size distribution, volume fraction, and water-cement ratio.

[0015] Furthermore, in step S2, multiple sets of parameter value combinations are constructed for the required input parameters based on literature and engineering experience. In each simulation, a few key parameters within these parameter value combinations are varied, while the remaining parameters remain fixed or randomly selected within a set range. Orthogonal design or Latin hypercube sampling is used to construct representative sets of parameter value combinations to cover key influencing factors and improve simulation efficiency. DEM-CFD coupled numerical simulations are then performed to obtain particle motion simulation results corresponding to velocity, trajectory, and collision state. The "key influencing factors" in this section include the following parameters: particle contact parameters (normal stiffness, tangential stiffness, friction coefficient), slurry fluid parameters (viscosity, density, boundary velocity), particle-fluid coupling parameters (resistance, drag coefficient), and material composition parameters (particle size distribution, aggregate volume fraction, water-cement ratio).

[0016] Furthermore, in step S3, the smart aggregate replaces part of the coarse aggregate with a particle size greater than 10 mm, with the replacement ratio being 10% to 30% of the volume of the aggregate in this particle size range. Smart aggregates are divided into two categories: each embedded with a type of sensor, the first category embedded with inertial sensors, and the second category embedded with piezoelectric sensors, with each category accounting for half of the total amount of smart aggregate. The 3D printing concrete raw material includes all types of sensors. The dynamic behavior data collected by the smart aggregate includes particle motion speed, trajectory, acceleration, rotational angular velocity, collision frequency, and contact force between particles.

[0017] Furthermore, in step S4, the numerical simulation results of each group in step S2 are compared and fitted with the dynamic behavior data actually obtained in step S3 on a unified time axis, and an error function is constructed. The error function is minimized through iterative optimization, and the DEM-CFD initial input parameters closest to the measured working conditions are inferred, thereby achieving high-precision calibration of the DEM-CFD coupled simulation model.

[0018] Furthermore, in step S4, during the iterative optimization process, a Bayesian optimization algorithm is used, and the mapping relationship between the input parameter value combination and the error function is regarded as a "black box model". A proxy model of the error function is constructed through Gaussian process regression, and the next set of simulation parameters is intelligently selected in combination with the expected improvement acquisition function, so as to efficiently search for the optimal parameter value combination within a limited number of simulations; a set of DEM-CFD numerical simulations is run in each iteration, the error function value under the current parameters is calculated, and the proxy model is updated with it; the error function uses the root mean square error as the main evaluation indicator, and the maximum relative error of the particle velocity, acceleration or collision frequency in the key area is combined as an auxiliary reference indicator; if the root mean square error is less than 5%, and the maximum relative error in the key area is less than 10%, it is considered that the parameter value combination of the current DEM-CFD coupling simulation model has achieved the required fitting accuracy.

[0019] Furthermore, in step S5, the digital twin control system runs synchronously with the measured construction process, wherein: while the system collects dynamic data of intelligent aggregate in real time, it runs a DEM-CFD simulation model consistent with the current construction parameter configuration to predict the particle flow behavior of concrete under current pumping and printing conditions; the real-time collected data does not directly drive the simulation calculation, but is dynamically compared with the simulation results as input, and is input into a risk identification and prediction model trained based on historical simulation data; the risk identification and prediction model is used to determine whether there is a pipe blocking trend or forming anomaly, and when an anomaly is detected, it is automatically identified as a risky working condition; based on the identification results, the system generates control instructions in real time, automatically adjusts the pumping rate, printing speed, path trajectory, or triggers a temporary shutdown, thereby realizing closed-loop intelligent construction control with simulation prediction as the core.

[0020] Furthermore, the construction parameters include pumping flow, pumping pressure, printing rate, nozzle movement path, nozzle height, nozzle outlet diameter, printing layer thickness, printing pitch, pumping frequency, switching valve timing, material ratio, printing beat synchronization control parameters and concrete outlet temperature.

[0021] Furthermore, the DEM-CFD coupling simulation model is a typical Euler-Lagrange multi-physics field coupling framework, in which the fluid phase is solved using the finite volume method based on the Navier-Stokes equations, and the particle phase is modeled using mass point dynamics based on Newton's second law. The coupling process realizes a two-way feedback effect between particles and fluid through the momentum source term, including buoyancy, resistance, drag force and stress transfer. A soft ball contact model is used inside the particle phase and the bonding failure behavior is considered to realize the dynamic prediction and control of particle distribution and clogging trend in 3D printed concrete.

[0022] Furthermore, the smart aggregate includes a shell, a PCB board fixedly installed in the shell, a control unit arranged on the PCB board, a sensor, a communication module, and a battery assembly power supply system for powering the components on the PCB board; the shell is made of thermoplastic resin material PLA, and the shell shape can be any shape. The outer dimensions of the smart aggregate match the coarse aggregate with a particle size of 10-20 mm and larger than 10 mm that it replaces, so as to meet the requirements for effective simulation of the movement characteristics of particles in concrete slurry; the PCB board and the components thereon are fixed in the shell by potting glue; the surface of the shell is sandblasted to a roughness of Ra2.5-6.3μm to simulate the surface texture of the coarse aggregate and improve the authenticity of its movement behavior and mechanical interaction effect in the slurry.

[0023] The technical solution of this application has the following beneficial effects:

[0024] 1. Visualization of particle flow: Compared with the unobservability and uncertainty of particle flow during traditional concrete pumping, this invention incorporates smart aggregates into the aggregates of 3D-printed concrete to achieve real-time and continuous collection of key parameters such as aggregate particle velocity, trajectory, acceleration, and force. This effectively improves the visualization and explainability of particle flow behavior during pumping, making the originally "black box" particle movement process "transparent."

[0025] 2. High efficiency and precision in numerical simulation parameter calibration: The parameter setting of traditional DEM-CFD coupled simulation usually relies on a large number of trial-and-error experiments or empirical estimates, which is complex and inefficient. The parameter inversion method proposed in this invention is based on the fitting optimization of measured particle motion data and numerical simulation results. It can automatically iteratively correct the model input parameters, significantly reducing manual intervention and experimental overhead, and improving the efficiency and physical rationality of simulation construction.

[0026] 3. After completing high-precision calibration of the DEM-CFD coupled simulation model, the hybrid modeling system constructed by this invention simultaneously runs a simulation model consistent with the current construction status based on digital twin technology. Combined with a risk identification and prediction model trained with historical simulation data, this system enables real-time prediction and dynamic identification of potential pipe blockage risks during pumping and concrete forming anomalies during printing. Based on simulation trends and sensor data analysis results, the system automatically generates control recommendations for key construction parameters such as pumping pressure, flow rate, printing speed, and path. This system then implements online dynamic optimization and feedback execution through a control interface, thereby enhancing the intelligence level of the entire construction process and the stability and controllability of structural forming quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings and descriptions that constitute part of this application are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. Among them:

[0028] Figure 1 Schematic diagram of the state of the smart aggregate during production according to an embodiment of the present invention.

[0029] Figure 2 Schematic diagram of circuit connection of smart aggregate according to an embodiment of the present invention.

[0030] Explanation of reference numerals: 1-housing, 2-potting compound, 3-PCB board. DETAILED DESCRIPTION

[0031] The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments. Each example is provided by way of explanation of the present application and does not limit the present application. In fact, it will be clear to those skilled in the art that modifications and variations can be made in the present application without departing from the scope or spirit of the present application. For example, a feature shown or described as part of one embodiment can be used in another embodiment to produce yet another embodiment. Therefore, it is expected that the present application includes such modifications and variations within the scope of the appended claims and their equivalents.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art in the art of the present disclosure. The terms used herein are only for the purpose of describing the embodiments of the present disclosure and are not intended to limit the present disclosure.

[0033] The terms "connected", "connected" and "set" used in this application should be understood in a broad sense. For example, it can be a fixed connection or a detachable connection; it can be a direct connection or an indirect connection through an intermediate component; it can be a wired electrical connection, a radio connection, or a wireless communication signal connection. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0034] A hybrid modeling method for 3D printing concrete pumping and printing processes includes the following steps:

[0035] S1. Determine the initial input parameters required for DEM-CFD coupled simulation.

[0036] S2. Set multiple groups of different parameter value combinations and perform DEM-CFD coupled numerical simulations respectively to obtain corresponding particle motion simulation results; the parameter value combinations include some parameters with different values ​​in different groups, while the remaining parameters remain unchanged or vary within a small range, thereby constituting effective coverage of the input parameter space; note that the parameter types in each group of parameter value combinations include all types of the initial input parameters in step S1.

[0037] S3. Smart aggregate replaces some aggregate in the raw materials of 3D-printed concrete. During the pumping and printing process, the smart aggregate is used to collect dynamic behavior data of the raw materials of 3D-printed concrete. The smart aggregate replaces some coarse aggregate with a particle size greater than 1 cm, and the replacement ratio is 10% to 30% of the volume of the aggregate in this particle size range. Smart aggregates are divided into two categories: each embedded with a type of sensor, the first category embedded with inertial sensors, and the second category embedded with piezoelectric force sensors, and the two categories each account for half of the total amount of smart aggregate. The inertial sensors include, for example, triaxial accelerometers and gyroscopes. The dynamic behavior data collected by the smart aggregate include particle motion speed, trajectory, acceleration, rotational angular velocity, collision frequency, and inter-particle contact force. The inertial sensor is used to collect particle motion state data, such as particle motion speed, trajectory, acceleration, rotational angular velocity, etc. The force sensor is used to collect the mechanical response during inter-particle contact and collision, including collision frequency and inter-particle contact force. The inertial sensor includes, for example, triaxial accelerometers and gyroscopes. The raw materials of 3D-printed concrete include all types of sensors. Through the coordinated arrangement of these two types of sensing particles, a multi-dimensional dynamic monitoring system is constructed, while ensuring that the geometric and physical properties of each embedded aggregate are close to those of real fine aggregate, avoiding disruption of the flow pattern of the pumping system.

[0038] S4. Compare the simulation results with the measured data and inversely optimize the parameters to achieve high-precision calibration of the DEM-CFD coupled simulation model.

[0039] S5. Build a digital twin control system based on the calibrated DEM-CFD coupling simulation model.

[0040] By introducing intelligent aggregates, this invention realizes real-time monitoring and data collection of particle motion behavior during 3D concrete pumping and printing. The collected multi-dimensional dynamic data such as particle velocity, acceleration, and collision frequency are compared and analyzed with the discrete element method-computational fluid dynamics (DEM-CFD) coupled simulation results, thereby realizing efficient inversion and precise calibration of the input parameters of the DEM-CFD coupled simulation model and the construction of a twin model.

[0041] In step S1, the initial input parameter types include: particle contact parameters, fluid parameters, particle-fluid coupling interaction parameters, and material parameters of 3D printing concrete raw materials;

[0042] The coupling interaction parameters include the resistance and drag coefficient between particles and fluid; the material parameters include aggregate particle size distribution, volume fraction, and water-binder ratio.

[0043] In step S2, the input parameter range required for the DEM-CFD coupled simulation is set based on relevant literature and engineering experience, and multiple sets of parameter value combinations are constructed. While maintaining the consistency of the input parameter types, the key parameters are combined and varied within a reasonable range. The input parameters include: particle contact parameters (such as normal stiffness, tangential stiffness, friction coefficient, and bond strength), fluid parameters (such as slurry viscosity, density, and flow boundary conditions), coupling interaction parameters (such as the particle-fluid resistance coefficient and drag coefficient), and material parameters of the 3D printed concrete raw materials (aggregate particle size distribution, water-cement ratio, volume fraction, etc.). Each set of parameter value combinations is used to drive the DEM-CFD coupled simulation to obtain the corresponding particle motion state output, including particle velocity, trajectory, acceleration, rotational angular velocity, collision frequency, and inter-particle contact behavior.

[0044] In step S3, the present application introduces smart aggregate particles equipped with sensors, which are used to collect the movement and force data of the particles in real time during the pumping and printing process of 3D printed concrete. After preliminary investigation, the particle size range of concrete aggregate particles is generally 5mm to 20mm. Among them, aggregates with a particle size of less than 10mm are light in weight and have low inertia, and usually do not cause pipe blockage or significantly disturb the slurry flow during the pumping process. Therefore, in order to avoid affecting its natural movement behavior, the present invention stipulates that aggregates with a particle size of less than 10mm are not embedded in any sensors and only exist as reference particles. For coarse aggregates with a particle size of more than 10mm, since they are more likely to become the cause of local accumulation or blockage during the pumping process, it is necessary to focus on monitoring their movement behavior and the force state between particles. Therefore, in step S3, the smart aggregate is used to replace part of the coarse aggregate with a particle size of more than 10mm. Considering that integrating multiple types of sensors within a single smart aggregate particle would significantly increase its volume and alter its motion characteristics, the present invention adopts a separate sensor embedding strategy. Preferably, only one sensor is placed within a single smart aggregate. If the sensor is relatively small and does not increase the volume of the smart aggregate, one or more sensors can be placed within a single smart aggregate particle. This approach effectively avoids the volume redundancy caused by integrating multiple sensors within a single smart aggregate, ensures that the structural dimensions of the smart aggregate remain consistent with the measured dimensions of coarse aggregate particles in 3D-printed concrete, and enhances the stability of the experimental system and the credibility of the test results.

[0045] In one embodiment, the smart aggregate comprises a housing 1, a PCB board 3 fixedly mounted within the housing 1, a control unit fixed to the PCB board 3, a sensor, a communication module, and a power supply system for supplying power to the components on the PCB board 3. Preferably, the PCB board 3 and the components thereon are fixed within the housing 1 by potting with a potting compound 2; the potting compound 2 can be epoxy resin potting compound 2, silicone potting compound 2, polyurethane potting compound 2, or silicone rubber potting compound 2. Specifically, given that the pumping pipeline for 3D-printed concrete on actual construction sites is relatively short (generally within 10-30 meters) and the duration of the entire pumping process is limited (usually no more than tens of minutes), a power supply system such as a lithium battery is used. Other existing wireless power supply methods can also be used to power the smart aggregate. The communication module can be a Bluetooth module, Wi-Fi module, ZigBee module, GPRS module, etc.

[0046] The manufacturing process of the smart aggregate described in this invention includes the following steps: First, based on the morphological characteristics of coarse aggregate with a particle size greater than 1 cm, an appropriate molding process (such as 3D printing, injection molding, or other precision machining methods) is selected to fabricate a housing 1 structure. The housing 1 is preferably made of PLA (polylactic acid), a thermoplastic resin. The specific manufacturing method can be flexibly determined based on performance requirements and production scale. Subsequently, a customized PCB assembly integrating sensors, a control unit, a communication module, and a power supply system is installed within the housing 1 and encapsulated with a low-viscosity, high-sealing potting compound 2 (preferably epoxy resin or silicone gel) to ensure the stability and durability of the electronic components. If necessary, a layer of cement mortar can be applied to the surface of the housing 1 to further enhance its consistency with natural aggregate in terms of surface roughness, density, and flow behavior. This structure ensures that the smart aggregate closely matches the dimensions, kinematic characteristics, and mechanical response of real coarse aggregate, while also possessing sufficient impact resistance and wireless communication capabilities, making it suitable for online monitoring during the pumping and printing of 3D-printed concrete.

[0047] To more realistically simulate the surface friction and irregular structure of coarse or fine aggregate in 3D-printed concrete, the present invention sandblasts the surface of the aggregate shell 1 after it is formed, creating a rough texture with microscopic irregularities. Glass beads with a particle size of 100-200 μm are used as the blasting medium, and the blasting pressure is controlled between 0.2-0.3 MPa (2-3 bar), which falls within the low-pressure, flexible treatment range. Unlike traditional metal sand or alumina sand, glass bead sandblasting offers gentle surface impact, low cutting forces, and no damage to the shell structure. Even when used on PLA shells, it does not cause structural damage or performance degradation. After sandblasting, the surface roughness of the smart aggregate reaches Ra2.5-6.3 μm, effectively improving its random rotation, stacking, and frictional response in the flow field, approaching the surface characteristics of actual aggregate, thereby enhancing the representativeness of the collected particle motion behavior data. Figure 1 This is a structural diagram of the smart aggregate. The shape of the smart aggregate can be any shape, such as a hexahedron, a frustum, a prism, etc. It is best to refer to the shape of the actual aggregate.

[0048] The sensors, power supplies, and communication modules within the smart aggregate used in the present invention are capable of miniaturized design. In one embodiment, a small, high-density microbattery is used to power data acquisition, Bluetooth communication, and short-term continuous operation. The microbattery can be a Murata G1 series ultra-thin solid-state battery, with dimensions of 3.0 mm in diameter and 0.9 mm in thickness, and a capacity of approximately 0.7-1.0 mA; or a suitable model from the LiPol Battery Company's Li-Pol 4.5 mm series, which has the advantages of small size, high stability, and adaptability to micro-packaging. The Bluetooth module uses a low-power Bluetooth communication module, such as the nRF52 series, and an intermittent sampling strategy, which can support the continuous operation of the sensing system throughout the pumping and printing cycle, and achieve real-time data transmission through the Bluetooth low-power BLE protocol. Figure 2 This is a schematic diagram of the internal circuit connections of the smart aggregate.

[0049] In step S4, the numerical simulation results of each group in step S2 are compared and fitted with the dynamic behavior data (including velocity, acceleration, trajectory, force, etc.) obtained by actual measurement in step S3 on a unified time axis. An error function containing multiple behavioral variables is systematically constructed. This error function is minimized through an iterative optimization algorithm (such as the least squares method, genetic algorithm, or particle swarm optimization algorithm). The initial DEM-CFD input parameters that are closest to the measured working conditions are inferred, and the initial DEM-CFD input parameters that best match the working conditions under real laboratory test conditions are determined, thereby achieving high-precision calibration of the DEM-CFD coupled simulation model. The present invention uses a numerical simulation platform based on the EDEM-OpenFOAM coupling framework. By setting different parameter value combinations (i.e., different value settings for key parameters under a fixed set of input parameter types), and combining historical literature and engineering experience, a systematic analysis is performed on the influence of each parameter on the coupling of particle motion and slurry flow. This simulation platform is based on the coupling analysis technology of discrete element method (DEM) and computational fluid dynamics (CFD). It can simultaneously simulate the contact-collision behavior of aggregate particles and the flow response of mortar fluid, thereby accurately evaluating the influence of different input parameter values ​​on the macro-rheological properties of concrete materials.

[0050] In step S4, an iterative optimization strategy based on the combination of error analysis and Bayesian optimization is generally adopted. In the iterative optimization process, the Bayesian optimization algorithm is used, and the mapping relationship between the input parameter value combination and the error function is regarded as a "black box model". The proxy model of the error function is constructed through Gaussian process regression, and the next set of simulation parameters is intelligently selected in combination with the expected improvement acquisition function, so as to efficiently search for the optimal parameter value combination within a limited number of simulations; a set of DEM-CFD numerical simulations is run in each iteration to calculate the error function value under the current parameters and use it to update the proxy model; the error function uses the root mean square error (RMSE) as the main evaluation indicator, and the maximum relative error of the particle velocity, acceleration or collision frequency in the key area is combined as an auxiliary reference indicator; if the root mean square error is less than 5%, and the maximum relative error (Max err ) is less than 10%, the parameter set of the current DEM-CFD coupled simulation model is considered to have achieved the required fitting accuracy. This optimization method can significantly reduce the extensive trial-and-error process in traditional parameter calibration, improve model calibration efficiency and simulation accuracy, and enhance the predictive capability and reliability of the DEM-CFD coupled simulation model in actual engineering applications.

[0051] In a preferred embodiment, in step S5, the digital twin control system runs synchronously with the actual construction process. While collecting dynamic data of intelligent aggregate in real time, the system runs a DEM-CFD simulation model consistent with the current construction parameter configuration to predict the particle flow behavior of concrete under current pumping and printing conditions; the real-time collected data does not directly drive the simulation calculation, but is dynamically compared with the simulation results as input, and is input into a risk identification and prediction model trained based on historical simulation data; the risk identification and prediction model is used to determine whether there is a pipe blocking trend or forming anomaly. When abnormal particle accumulation, sudden drop in speed, abnormal collision frequency and other characteristics are detected, it is automatically identified as a risky working condition; based on the identification results, the system generates control instructions in real time, automatically adjusts the pumping rate, printing speed, path trajectory or triggers a temporary shutdown, and realizes closed-loop intelligent construction control with simulation prediction as the core. The construction parameters include pumping flow, pumping pressure, printing rate, nozzle motion path, nozzle height, nozzle outlet diameter, print layer thickness, print pitch, pumping frequency, valve on / off timing, material mix ratio (such as water-cement ratio and admixture dosage), print beat synchronization control parameters, and concrete outlet temperature. Digital twin technology constructs a virtual-physical fusion system for the 3D-printed concrete pumping and printing process, enabling dynamic linkage and intelligent control between the physical equipment and the simulation model. Specifically, the intelligent aggregate collects real-time motion data of 3D-printed concrete particles, such as velocity, acceleration, and force, during the pumping process and transmits this data to a local receiving system via a low-power Bluetooth communication module. The receiving system maps the sensor data to a pre-built DEM-CFD coupled simulation model, enabling dynamic calibration and synchronous updating of the virtual model. The DEM-CFD coupled simulation model can predict potential anomalies such as pipe blockage and accumulation based on real-time input and generate optimization recommendations when risks are detected. The system feeds back construction parameter adjustment instructions to the pumping equipment through the control interface, achieving adaptive control of pump pressure, flow rate, and printing rate, thereby building a closed-loop control system that integrates physical and virtual elements and improves the intelligence and stability of the construction process. Through this digital twin control system, not only is the complex manual parameter calibration workload in numerical simulations significantly reduced, the response time between simulation and construction is shortened, but the model's predictive capabilities and the intelligence level of engineering construction are also effectively improved. At the same time, this method can be used to predict concrete printing quality and potential construction problems in advance, providing a precise adjustment basis for actual construction.

[0052] The DEM-CFD coupled simulation model utilizes an Euler-Lagrange coupling framework. The fluid phase is modeled using the Euler method, employing the finite volume method to solve momentum conservation based on the continuity equation and the Navier-Stokes equations. The particle phase is modeled using the Lagrangian method, constructing the governing equations for linear and angular momentum of the particles based on Newton's second law (i.e., modeling based on mass-point dynamics). A soft-sphere contact model is employed within the particle phase, and cohesive failure behavior is considered. The fluid and particles are bidirectionally coupled via a momentum source term. Forces acting on the particles include fluid drag, buoyancy, and wall reaction, while forces acting on the fluid include volume forces fed back by the particle population (reflecting bidirectional feedback and stress transfer between particles and the fluid). The coupled calculation achieves the coordinated evolution of the particle-fluid field through time-step synchronization and source term updates. This method can be used to dynamically predict and control particle distribution and clogging trends in 3D-printed concrete.

[0053] This method first uses smart aggregates with built-in sensors to monitor particle behavior (such as speed, trajectory, and collision frequency) in real time during the pumping and printing process, thereby obtaining high-precision dynamic data. Subsequently, this monitoring data is used to drive discrete element (DEM)-computational fluid dynamics (CFD) coupled numerical simulations and construct a digital twin control system. Through this digital twin control system, not only is the complex manual parameter calibration workload in numerical simulations significantly reduced, the response time between simulation and construction is shortened, but the predictive ability of the twin model and the level of intelligent engineering construction are also effectively improved. The constructed twin model can be used to predict the printing quality and potential construction problems of 3D printed concrete under the same working conditions in advance, providing a precise adjustment basis for actual construction.

[0054] The foregoing description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are readily apparent to those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A hybrid modeling method for 3D printing concrete pumping and printing process, characterized in that: The following steps are involved: S1. Determine the initial input parameters required for DEM-CFD coupled simulation; S2. Setting multiple groups of different parameter value combinations, wherein some parameters have different values ​​in different groups, while the remaining parameters remain unchanged or vary within a small range, thereby effectively covering the input parameter space; performing DEM-CFD coupled numerical simulations on each group to obtain corresponding particle motion simulation results; S3. Smart aggregate replaces some aggregates in 3D printing concrete raw materials. During the pumping and printing process, the smart aggregate collects dynamic behavior data of 3D printing concrete raw material particles; The smart aggregate replaces some coarse aggregate with a particle size greater than 10 mm, with the replacement ratio being 10% to 30% of the volume of the aggregate in this size range. The smart aggregates are divided into two categories: one embedded with a sensor type, the first category embedded with inertial sensors, and the second category embedded with piezoelectric sensors, with each category accounting for half of the total smart aggregate volume. The 3D printing concrete raw material includes all types of sensors. The dynamic behavior data collected by the smart aggregate includes particle motion speed, trajectory, acceleration, rotational angular velocity, collision frequency, and inter-particle contact force. S4. Compare and fit the numerical simulation results of each group in step S2 with the dynamic behavior data actually obtained in step S3 on a unified time axis, construct an error function, minimize the error function through iterative optimization, and inversely deduce the DEM-CFD initial input parameters that are closest to the measured working conditions, thereby achieving high-precision calibration of the DEM-CFD coupled simulation model; S5. Build a digital twin control system based on the calibrated DEM-CFD coupling simulation model.

2. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 1, characterized in that: In step S1, the initial input parameter types include: particle contact parameters, fluid parameters, particle-fluid coupling interaction parameters, and material parameters of 3D printing concrete raw materials; The particle contact parameters include normal stiffness, tangential stiffness, friction coefficient, and bonding strength; the fluid parameters include slurry viscosity, density, and flow velocity boundary conditions; the coupling interaction parameters include particle-fluid resistance and drag coefficient; and the material parameters include aggregate particle size distribution, volume fraction, and water-cement ratio.

3. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 1, characterized in that: In step S2, based on literature and engineering experience, multiple groups of parameter value combinations are constructed for the required input parameters. In each group of simulations, some key parameters in the parameter value combination are selected and changed, and the remaining parameters are kept fixed or randomly selected within a set range. A representative set of parameter value combinations is constructed using an orthogonal design or Latin hypercube sampling method; DEM-CFD coupled numerical simulations are performed separately to obtain particle motion simulation results corresponding to the speed, trajectory, and collision state.

4. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 1, characterized in that: In step S4, during the iterative optimization process, a Bayesian optimization algorithm is used to treat the mapping relationship between the input parameter value combination and the error function as a "black box model", and a proxy model of the error function is constructed through Gaussian process regression. The next set of simulation parameters is intelligently selected in combination with the expected improvement acquisition function, thereby efficiently searching for the optimal parameter value combination within a limited number of simulations. In each iteration, a set of DEM-CFD numerical simulations is run to calculate the error function value under the current parameters and use it to update the proxy model. The error function uses the root mean square error as the main evaluation indicator, and combines the maximum relative error of particle velocity, acceleration or collision frequency in the key area as an auxiliary reference indicator; if the root mean square error is less than 5% and the maximum relative error in the key area is less than 10%, it is considered that the parameter value combination of the current DEM-CFD coupled simulation model has achieved the required fitting accuracy.

5. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 1, characterized in that: In step S5, the digital twin control system runs synchronously with the measured construction process, wherein: while the system collects dynamic data of intelligent aggregate in real time, it runs a DEM-CFD simulation model consistent with the current construction parameter configuration to predict the particle flow behavior of concrete under current pumping and printing conditions; the real-time collected data does not directly drive the simulation calculation, but is dynamically compared with the simulation results as input, and is input into a risk identification and prediction model trained based on historical simulation data; the risk identification and prediction model is used to determine whether there is a pipe blocking trend or forming anomaly. When an anomaly is detected, it is automatically identified as a risky working condition; based on the identification results, the system generates control instructions in real time, automatically adjusts the pumping rate, printing speed, path trajectory, or triggers a temporary shutdown, thereby realizing closed-loop intelligent construction control with simulation prediction as the core.

6. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 5, characterized in that: The construction parameters include pumping flow, pumping pressure, printing rate, nozzle movement path, nozzle height, nozzle outlet diameter, printing layer thickness, printing pitch, pumping frequency, switching valve timing, material ratio, printing beat synchronization control parameters and concrete outlet temperature.

7. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 1, characterized in that: The DEM-CFD coupled simulation model is a typical Euler-Lagrange multi-physics field coupling framework, in which the fluid phase is solved using the finite volume method based on the Navier-Stokes equations, and the particle phase is modeled using mass point dynamics based on Newton's second law. The coupling process realizes a two-way feedback effect between particles and fluid through the momentum source term, including buoyancy, resistance, drag force and stress transfer. A soft sphere contact model is used within the particle phase, and bond failure behavior is considered to achieve dynamic prediction and control of particle distribution and clogging trends in 3D printed concrete.

8. The hybrid modeling method for 3D printing concrete pumping and printing process according to claim 1, characterized in that: The smart aggregate comprises a housing (1), a PCB board (3) fixedly mounted in the housing (1), a control unit, a sensor, a communication module, and a power supply system arranged on the PCB board (3) for supplying power to components on the PCB board (3); the housing (1) is made of thermoplastic resin material, and the housing can be of any shape. The dimensions of the smart aggregate match the coarse aggregate with a particle size greater than 10 mm that it replaces, so as to meet the requirements for effectively simulating the movement characteristics of particles in concrete slurry.

Images