Concrete 3D printer parameter rapid generation method based on large language model

Abstract

The invention discloses a concrete 3D printer parameter rapid generation method based on a large language model, and the method comprises the steps: deploying a large language model programming environment, and configuring a three-dimensional model of a nozzle of a concrete 3D printer; generating a process code of a mechanical simulation solving step of concrete 3D printing, and importing the process code of the mechanical simulation solving step into a large language model programming environment; the method comprises the following steps: inputting physical parameters of a concrete mixture and forming parameters of a printing strip in a large language model programming environment, and generating a custom solution step process code; and executing the custom solution step process code to obtain the movement information and the layer height information of the nozzle of the concrete 3D printer. The parameters of the concrete 3D printer can be scientifically and rapidly formulated, and the number of feasible experiments before construction is reduced.

Description

Technical Field

[0001] The present invention relates to the field of 3D printing concrete, and in particular to a method for quickly generating parameters of a concrete 3D printer based on a large language model. Background Art

[0002] 3D printed concrete is a concrete structure formed by additive manufacturing. It is characterized by not relying on the shape restrictions of the template and can quickly produce customized concrete components. The molding effect of traditional concrete components depends on controllable factors such as templates, materials, and curing conditions. The molding effect of 3D printed concrete is more complicated, among which the operating parameters of the 3D printer and the performance of the concrete mixture are the two key factors that determine the molding effect.

[0003] Before the construction of 3D printed concrete, a large number of feasibility tests are required until the relationship between the operating parameters of the concrete 3D printer and the concrete mixture ratio that meets the current construction conditions is determined. At present, for feasibility experiments with large engineering volumes, the main method is to rely on regular enumeration to obtain the parameters of the concrete 3D printer through repetitive experiments. Since the parameters of the concrete 3D printer need to be repeatedly adjusted for concrete discharge testing, this results in a large waste of materials and a long test cycle.

[0004] Therefore, in order to solve the problem of quickly determining the parameters of concrete 3D printers, a method for quickly generating parameters of concrete 3D printers based on a large language model is needed, which can scientifically and quickly formulate the parameters of concrete 3D printers and reduce the number of feasibility experiments before construction. Summary of the invention

[0005] In view of this, the purpose of the present invention is to overcome the defects in the prior art and provide a method for quickly generating parameters of a concrete 3D printer based on a large language model, which can scientifically and quickly formulate parameters of a concrete 3D printer and reduce the number of feasibility experiments before construction.

[0006] The method for quickly generating parameters of a concrete 3D printer based on a large language model of the present invention comprises:

[0007] Deploy a large language model programming environment to configure the three-dimensional model of the concrete 3D printer nozzle;

[0008] Generate process codes of mechanical simulation solution steps for concrete 3D printing, and import the process codes of mechanical simulation solution steps into a large language model programming environment;

[0009] Input the physical parameters of concrete mixture and the molding parameters of printing strip in the large language model programming environment to generate the process code of custom solution steps;

[0010] Execute the custom solution step process code to obtain the motion information and layer height information of the concrete 3D printer nozzle.

[0011] Furthermore, deploy a large language model programming environment, including:

[0012] In Visual Studio Code, open the code editor, enable the Cline plug-in, select DeepSeek as the program provider, enter the DeepSeek API Key code, and set the type to DeepSeek Coder. Configure the Fortran plug-in parameters and the Fortran compiler to enable Visual Studio Code to write code in Fortran and obtain the configured large language model programming environment.

[0013] Furthermore, the three-dimensional model of the concrete 3D printer nozzle is configured, including:

[0014] Import the 3D model of the concrete 3D printer nozzle in the geometry panel of the Flow 3D software. After the import is completed, configure the necessary parameters:

[0015] Specify the type of physical calculations: elastic-plastic calculations, gravity and non-inertial reference frames, moving and simply deforming objects, viscosity and turbulence;

[0016] Specify the start time of the printing process and the time change dimension;

[0017] Specifies the output of calculation results as liquid fraction, fluid velocity, and pressure indicators.

[0018] Furthermore, the process code of the mechanical simulation solution step of concrete 3D printing is generated, and the process code of the mechanical simulation solution step is imported into the large language model programming environment, specifically including:

[0019] In the test panel of Flow 3D software, select the nozzle of concrete 3D printer as the calculation subject, enter the editing calculation mode, call the Fortran language compiler in the editing panel, and obtain the process code of the mechanical simulation solution step; and copy the process code of the mechanical simulation solution step to the dialog box of Cline plug-in.

[0020] Furthermore, the physical parameters of the concrete mixture and the molding parameters of the printed strip are input into the large language model programming environment to generate the custom solution step process code, including:

[0021] In the dialog box of the Cline plug-in, input the physical parameters of the concrete mixture to be printed measured by the experimental instrument; the physical parameters include density, viscosity and liquid fraction;

[0022] Enter the molding parameters of the printing strip in the dialog box of the Cline plug-in; the molding parameters of the printing strip include the printing strip width, the single layer stacking height and the printing strip stacking diffusion rate;

[0023] DeepSeek makes analysis and calculation inferences based on the logic and key parameters of the process code of the mechanical simulation solution step, and generates a custom solution step process code based on the physical parameters of the concrete mixture and the molding parameters of the printed strip.

[0024] Further, execute the custom solution step process code, including:

[0025] Open the file location of Flow 3D software, create a new subroutine by copying in the hydr3d folder, open the subroutine operation script, replace the original script content with the custom solution step process code, obtain the custom subroutine, and complete the associated operation script;

[0026] In the calculation preference options of Flow 3D software, adjust the calculation version to user double precision, open the Fortran compiler to execute the custom subroutine.

[0027] Furthermore, the motion information and layer height information of the concrete 3D printer nozzle are obtained, including:

[0028] After executing the custom subroutine, the horizontal moving speed of the printer nozzle is obtained according to the distance of the printing fluid accumulated in the Y-axis direction of the carrier plate per unit time, and the discharge speed and extrusion ratio of the printer nozzle are obtained according to the volume of the fluid passing through the cross section of the printing nozzle per unit time; the movement information includes the horizontal moving speed of the printer nozzle, the discharge speed of the printer nozzle and the extrusion ratio of the printer nozzle;

[0029] Select n calculation frames of the 15-24s part, take the print strip section in the X-axis direction in the calculation frame, obtain n print strip cross-sectional views, divide the maximum value of the Z-axis in the n cross-sectional views by n to obtain an average value, and use the average value as the layer height information; wherein, a coordinate system is established with the length direction of the carrier plate as the Y-axis, the width direction as the X-axis, the perpendicular to the plane of the carrier plate as the Z-axis, the intersection of the length and width of the upper left corner of the carrier plate as the origin, the horizontal movement direction of the nozzle is the positive direction of the Y-axis, and the print strip stacking direction is the positive direction of the Z-axis.

[0030] The beneficial effects of the present invention are as follows: the present invention discloses a method for quickly generating parameters of a concrete 3D printer based on a large language model. By using DeepSeek as a large language model for programming, errors in the code can be automatically repaired, and the consumption is much lower than other existing large language models. It also supports multi-modal input, adapts to code generation tasks, and can quickly generate high-quality code. The process code of the Flow 3D software solver is imported into the large language model programming environment, the semantic context is reasonably constructed, the search scope of the DeepSeek knowledge base is narrowed, the generated code is clearly directed, the DeepSeek logic divergence is reduced, and the code generation efficiency and accuracy are improved.

[0031] The concrete 3D printing task was executed by customizing the process code of the mechanical simulation solution steps, and it was not dependent on the actual experimental process, which could avoid material waste. The large language model was used to generate code, and the software executed the code to simulate the experimental process, avoiding the repetition of the experimental process through the regular enumeration method, reducing the number of experiments, and greatly shortening the test cycle, thereby quickly generating the parameters of the concrete 3D printer, providing technical support for scientifically predicting the molding effect of concrete 3D printed objects. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0033] Figure 1 A schematic diagram of a process flow of a method for quickly generating printer parameters according to the present invention;

[0034] Figure 2 A schematic diagram of a three-dimensional model of a concrete 3D printer nozzle introduced in the present invention;

[0035] Figure 3 It is a schematic diagram demonstrating the printing strip effect of the present invention. DETAILED DESCRIPTION

[0036] The present invention is further described below in conjunction with the accompanying drawings, as shown in the drawings:

[0037] This embodiment discloses a method for quickly generating parameters of a concrete 3D printer based on a large language model, comprising the following steps:

[0038] Deploy a large language model programming environment to configure the three-dimensional model of the concrete 3D printer nozzle;

[0039] Generate process codes of mechanical simulation solution steps for concrete 3D printing, and import the process codes of mechanical simulation solution steps into a large language model programming environment;

[0040] Input the physical parameters of concrete mixture and the molding parameters of printing strip in the large language model programming environment to generate the process code of custom solution steps;

[0041] Execute the custom solution step process code to obtain the motion information and layer height information of the concrete 3D printer nozzle.

[0042] In this embodiment, a programming environment capable of running a large language model is first established to support Fortran programming. The deployment of the large language model programming environment specifically includes:

[0043] In Visual Studio Code, open the code editor, enable the Cline plug-in, select DeepSeek as the program provider, enter the DeepSeek API Key code, and set the type to DeepSeek Coder. Configure the Fortran plug-in parameters and the Fortran compiler to enable Visual Studio Code to write code in Fortran and obtain the configured large language model programming environment.

[0044] In this embodiment, the three-dimensional model of the nozzle of the concrete 3D printer is configured, specifically including:

[0045] Import the 3D model of the concrete 3D printer nozzle into the geometry panel of the Flow 3D software. The 3D model has the same size and main components as the nozzle used in the feasibility experiment. The main components include: hopper, nozzle, and carrier plate. Figure 2 After the import is complete, configure the necessary parameters:

[0046] Specify the type of physical calculation: elastic-plastic calculation, gravity and non-inertial reference system, moving and simple deforming objects, viscosity and turbulence; for the specified physical calculation types, such as elastic-plastic calculation, gravity and non-inertial reference system, moving and simple deforming objects, Flow 3D software can perform the calculation items of elastic-plastic mechanics and theoretical mechanics included in the concrete 3D printing process; for viscosity and turbulence, Flow 3D software can perform calculation items related to the fluid mechanics properties of concrete.

[0047] Specify the start time and time change dimension of the printing process; the start time of the printing process is 0s, and the time change dimension is 0.001s;

[0048] Specify the output method of calculation results as liquid fraction, fluid velocity and pressure index. By setting the output parameters as liquid fraction, fluid velocity and pressure index, the software can be restricted to output only mechanical indicators related to the concrete 3D printing process, preventing other unnecessary physical quantities from interfering with the results.

[0049] In this embodiment, the process code of the mechanical simulation solution step of concrete 3D printing is generated, and the process code of the mechanical simulation solution step is imported into the large language model programming environment, specifically including:

[0050] In the test panel of Flow 3D software, select the nozzle of the concrete 3D printer as the calculation subject, enter the editing calculation mode, and call the Fortran language compiler in the editing panel. The compiler can display the code of all calculation steps to be executed in the simulation solution process of the current project, so as to obtain the process code of the mechanical simulation solution step; wherein, the mechanical simulation solution of the process code of the mechanical simulation solution step refers to the solution of the process of concrete 3D printing, and the printing process refers to the entire printing process of concrete being extruded from the nozzle and laid on the carrier plate; then, by copying the process code of the mechanical simulation solution step into the dialog box of the Cline plug-in, the process code of the mechanical simulation solution step is imported into the large language model programming environment.

[0051] In this embodiment, the physical parameters of the concrete mixture and the molding parameters of the printing strip are input in the large language model programming environment to generate a custom solution step process code, which specifically includes:

[0052] In the Cline plug-in dialog box, enter the physical parameters of the concrete mixture to be printed measured by the experimental instrument; the physical parameters include density, viscosity and liquid fraction; the density can be 2300kg / m 3 , viscosity is 60Pa·s, liquid fraction is 1;

[0053] Enter the forming parameters of the print strip in the dialog box of the Cline plug-in; the forming parameters of the print strip include the print strip width, the single layer stacking height and the print strip stacking diffusion rate; wherein the print strip width can be 10 cm, the single layer stacking height is 12 cm, and the print strip stacking diffusion rate is 5-10%;

[0054] DeepSeek makes analysis and calculation inferences based on the logic and key parameters of the process code of the mechanical simulation solution step, and generates a custom solution step process code in combination with the physical parameters of the concrete mixture and the molding parameters of the printed strip. Among them, the key parameters are the parameters that play a decisive role or have a significant impact in the analysis or calculation, which can be determined according to the actual working conditions. For example, the key parameters include the dimensional information, material characteristic parameters, time parameters, physical conditions and rheological model of the entire operation process; the dimensional information and material characteristic parameters determine the performance of concrete in the 3D printing process; the time parameters and physical conditions establish the operation environment of the process; the rheological model determines the theoretical basis of the mechanical simulation.

[0055] In this embodiment, the process code of executing the custom solution step specifically includes:

[0056] Open the file location of Flow 3D software, create a new subroutine by copying in the hydr3d folder, open the subroutine operation script, replace the original script content with the custom solution step process code, obtain the custom subroutine, and complete the associated operation script;

[0057] In the calculation preference options of Flow 3D software, adjust the calculation version to user double precision, open the Fortran compiler to execute the custom subroutine.

[0058] After executing the custom solution step process code, the printed strips manufactured by the concrete 3D printing process can meet the requirements of "printed strip width of 10cm, single layer stacking height of 12cm, and printed strip stacking diffusion rate of 5-10%". For example, after calculation, the simulated 3D printed concrete forming strips are obtained, such as Figure 3 shown.

[0059] Then, the motion information and layer height information of the concrete 3D printer nozzle are obtained, including:

[0060] After executing the custom subroutine, the horizontal movement speed V1 of the printer nozzle is obtained as 35 mm / s according to the distance accumulated by the printing fluid in the Y-axis direction of the carrier plate per unit time, and the discharge speed V2 of the printer nozzle is obtained as 70 mm / s and the extrusion ratio is 250% according to the volume of fluid passing through the cross section of the printing nozzle per unit time; the motion information includes the horizontal movement speed of the printer nozzle, the discharge speed of the printer nozzle and the extrusion ratio of the printer nozzle; wherein, the product of the discharge speed V2 and the discharge coefficient is the extrusion ratio, and the discharge coefficient is the ratio of the maximum discharge speed to the maximum horizontal movement speed under the condition that the printer hopper is fully loaded; however, the factory settings of different concrete 3D printers will adjust the discharge coefficient, and the discharge coefficient used in this embodiment is a parameter provided by the printer.

[0061] 10 calculation frames of the 15-24s calculation part are selected, and the printed strip section in the X-axis direction of the calculation frame is taken to obtain 10 printed strip cross-sectional views. The maximum value of the Z-axis in the 10 cross-sectional views is divided by 10 to obtain an average value of 13.1 cm, and the average value is used as the layer height information; wherein, a coordinate system is established with the length direction of the carrier plate as the Y-axis, the width direction as the X-axis, the Z-axis perpendicular to the plane of the carrier plate, and the intersection of the length and width of the upper left corner of the carrier plate as the origin, the horizontal movement direction of the nozzle is the positive direction of the Y-axis, and the stacking direction of the printed strip is the positive direction of the Z-axis.

[0062] By inputting the movement information and layer height information of the printer nozzle into the control end of the concrete 3D printer, printing parameters can be formulated.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for quickly generating parameters of a concrete 3D printer based on a large language model, characterized in that: include: Deploy a large language model programming environment to configure the three-dimensional model of the concrete 3D printer nozzle; Generate process codes of mechanical simulation solution steps for concrete 3D printing, and import the process codes of mechanical simulation solution steps into a large language model programming environment; Input the physical parameters of concrete mixture and the molding parameters of printing strip in the large language model programming environment to generate the process code of custom solution steps; Execute the custom solution step process code to obtain the motion information and layer height information of the concrete 3D printer nozzle.

2. The method for rapid generation of parameters of a concrete 3D printer based on a large language model according to claim 1, characterized in that: Deploy a large language model programming environment, including: In Visual Studio Code, open the code editor, enable the Cline plug-in, select DeepSeek as the program provider, enter the APIKey code of DeepSeek, and set the type to DeepSeek Coder; configure the Fortran plug-in parameters and the Fortran compiler to enable Visual Studio Code to write code in Fortran and obtain the configured large language model programming environment.

3. The method for rapid generation of parameters of a concrete 3D printer based on a large language model according to claim 1 is characterized in that: Configure the 3D model of the concrete 3D printer nozzle, including: Import the 3D model of the concrete 3D printer nozzle in the geometry panel of the Flow 3D software. After the import is completed, configure the necessary parameters: Specify the type of physical calculations: elastic-plastic calculations, gravity and non-inertial reference frames, moving and simply deforming objects, viscosity and turbulence; Specify the start time of the printing process and the time change dimension; Specify the output method of calculation results as liquid fraction, fluid velocity and pressure index.

4. The method for rapid generation of parameters of a concrete 3D printer based on a large language model according to claim 1, characterized in that: Generate the process code of the mechanical simulation solution steps of concrete 3D printing, and import the process code of the mechanical simulation solution steps into the large language model programming environment, specifically including: In the test panel of Flow 3D software, select the nozzle of concrete 3D printer as the calculation subject, enter the editing calculation mode, call the Fortran language compiler in the editing panel, and obtain the process code of the mechanical simulation solution step; and copy the process code of the mechanical simulation solution step to the dialog box of Cline plug-in.

5. The method for rapid generation of parameters of a concrete 3D printer based on a large language model according to claim 1, characterized in that: Input the physical parameters of the concrete mixture and the molding parameters of the printed strip in the large language model programming environment to generate the custom solution step process code, including: In the dialog box of the Cline plug-in, input the physical parameters of the concrete mixture to be printed measured by the experimental instrument; the physical parameters include density, viscosity and liquid fraction; Enter the molding parameters of the printing strip in the dialog box of the Cline plug-in; the molding parameters of the printing strip include the printing strip width, the single layer stacking height and the printing strip stacking diffusion rate; DeepSeek makes analysis and calculation inferences based on the logic and key parameters of the process code of the mechanical simulation solution step, and generates a custom solution step process code based on the physical parameters of the concrete mixture and the molding parameters of the printed strip.

6. The method for rapid generation of parameters of a concrete 3D printer based on a large language model according to claim 1, characterized in that: Execute the custom solution step process code, including: Open the file location of Flow 3D software, create a new subroutine by copying in the hydr3d folder, open the subroutine operation script, replace the original script content with the custom solution step process code, obtain the custom subroutine, and complete the associated operation script; In the calculation preference options of Flow 3D software, adjust the calculation version to user double precision, open the Fortran compiler to execute the custom subroutine.

7. The method for rapid generation of parameters of a concrete 3D printer based on a large language model according to claim 6, characterized in that: Get the motion information and layer height information of the concrete 3D printer nozzle, including: After executing the custom subroutine, the horizontal moving speed of the printer nozzle is obtained according to the distance of the printing fluid accumulated in the Y-axis direction of the carrier plate per unit time, and the discharge speed and extrusion ratio of the printer nozzle are obtained according to the volume of the fluid passing through the cross section of the printing nozzle per unit time; the movement information includes the horizontal moving speed of the printer nozzle, the discharge speed of the printer nozzle and the extrusion ratio of the printer nozzle; Select n calculation frames of the 15-24s part, take the print strip section in the X-axis direction in the calculation frame, obtain n print strip cross-sectional views, divide the maximum value of the Z-axis in the n cross-sectional views by n to obtain an average value, and use the average value as the layer height information; wherein, a coordinate system is established with the length direction of the carrier plate as the Y-axis, the width direction as the X-axis, the perpendicular to the plane of the carrier plate as the Z-axis, the intersection of the length and width of the upper left corner of the carrier plate as the origin, the horizontal movement direction of the nozzle is the positive direction of the Y-axis, and the print strip stacking direction is the positive direction of the Z-axis.

Images