Mold-free production method for gypsum 3D printing by adding fibers and metal materials

Abstract

The invention discloses a gypsum 3D printing fiber and metal material adding mold-free production method, and relates to the technical field of material 3D printing, the method comprises the specific steps of raw material pretreatment, AI model construction, 3D model treatment, gradient ratio printing and post-treatment and optimization; according to the method, through fusion of AI self-adaptive matching and 3D printing, the problems that in traditional gypsum component production, mold dependence is caused, the shape is limited, and the material matching is rigid are solved, mold-free production is achieved through application of the 3D printing technology, and meanwhile, the production flexibility and integrity are improved in combination with the synchronous implanting technology of the metal hook; the AI self-adaptive matching model dynamically adjusts a matching scheme of fibers and metal materials through a three-layer network structure in combination with regional feature weights and material compatibility constraints according to differential performance requirements of a bearing area, a decoration area and a transition area of a component, so that the structural strength of the bearing area is guaranteed, and the flexibility and the surface texture of the decoration area are also considered.

Description

Technical Field

[0001] This invention relates to the field of 3D printing technology, specifically a moldless production method for adding fiber and metal materials to gypsum 3D printing. Background Technology

[0002] As the architectural decoration industry develops towards a fusion of personalization, artistry, and functionality, high-quality gypsum decorative components, especially glass fiber reinforced gypsum (GRG) components, have become one of the core decorative materials for public buildings such as theaters, museums, convention centers, star-rated hotels, and high-end commercial complexes, as well as private spaces such as high-end villas and clubs, due to their combination of decorative and functional properties. In recent years, architectural design has significantly increased the complexity and customization requirements for component shapes, while the concept of green building has driven the industry to continuously improve the environmental friendliness and performance adaptability of materials.

[0003] However, traditional gypsum component production largely relies on mold casting as its core process. Molds need to be designed and manufactured individually for different component shapes, and the process from design and drawing to molding often takes a long time. Furthermore, mold materials are mostly disposable or have low reusability, increasing both time and material costs in the early stages of production. Mold wear also generates significant waste. Additionally, the rigid structure of the molds inherently limits component shapes; for complex shapes such as curved surfaces, irregular shapes, and large spans, molds struggle to accurately reproduce design details, leading to molding errors and failing to meet the precision requirements of high-end architectural decoration. In terms of material application, traditional processes typically use a uniform proportioning model, failing to adjust material combinations according to the functional needs of different areas of the component. This results in some areas having excessive material properties while others are insufficient, causing material waste and affecting the overall performance of the component. Moreover, traditional production processes lack a standardized control system; raw material drying, component casting, and curing rely heavily on manual experience, leading to product quality fluctuations due to operational differences. Furthermore, the lack of a linkage feedback mechanism between testing results and production parameters prevents optimization of proportions or processes based on finished product performance defects, exacerbating the problem of poor product consistency. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a moldless production method for adding fiber and metal materials to gypsum 3D printing. This invention solves the pain points of mold dependence, shape limitation, and rigid material ratio in traditional gypsum component production by integrating AI adaptive proportioning with 3D printing. The application of 3D printing technology enables moldless production, eliminating the cumbersome steps of mold design, manufacturing, and demolding. At the same time, the simultaneous implantation process of metal hooks improves production flexibility and integrity. The AI ​​adaptive proportioning model, through a three-layer network structure, combined with regional feature weights and material compatibility constraints, dynamically adjusts the proportion of fiber and metal materials according to the differentiated performance requirements of the load-bearing area, decorative area, and transition area of ​​the component. This ensures the structural strength of the load-bearing area, takes into account the flexibility and surface texture of the decorative area, and achieves a smooth performance transition in the transition area to avoid the risk of cracking.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a moldless production method for adding fiber and metal materials to gypsum 3D printing, the specific steps of which are as follows: Raw material pretreatment: High-density Alpha gypsum powder is selected as the base material, combined with alkali-free glass fiber and metal components as raw materials, and the raw materials are pretreated by drying and removing impurities. AI Model Construction: Collect mechanical and acoustic performance parameters of different areas of building decorative components, establish a mechanical demand-material ratio correlation database, and construct an AI adaptive ratio model. The input parameters are the performance parameters of the component area, and the output parameters are the raw material ratio parameters. The AI ​​adaptive ratio model is trained and optimized. 3D model processing: A 3D model of the component is created using 3D modeling software, and slicing is completed using slicing software. The 3D model is divided into load-bearing area, decorative area and transition area by combining mechanical requirements-material ratio correlation database. At the same time, the implantation coordinates of metal hooks are set in the 3D model and slice files are generated. Gradient proportion printing: The pre-treated raw materials are placed in the multi-bin of the 3D printer, and the printing parameters are set based on the loaded AI adaptive proportion model and slice file. During the printing process, the AI ​​adaptive proportion model adjusts the output ratio of each bin in real time, and the metal hook is implanted by the robotic arm according to the implantation coordinate. Post-processing and optimization: The printed components are placed in a specific humidity environment to air dry naturally. After drying, the surface of the components is polished and repaired. The performance of the finished product is tested, and the test results are fed back to the AI ​​adaptive mixing model to optimize the mixing ratio.

[0006] Furthermore, in the raw material pretreatment, the whiteness of the high-density Alpha gypsum powder is ≥88%, and its air-dry specific gravity is 1.85~1.91 g / cm³. 3Alkali-free glass fiber includes short fibers and long fibers. The length of short fibers is 0.4~2.1mm, and the length of long fibers is 2.8~8.2mm. The diameter of the single filament is 4~21μm. The metal components include metal hooks and metal particles. The material of the metal components is stainless steel or aluminum alloy. The drying and impurity removal process is as follows: gypsum powder is placed in a constant temperature drying oven at 55~85℃ and dried for 3~7 hours, controlling the moisture content to ≤0.6%. Glass fiber is removed by a vibrating screening machine to remove broken impurities and dust. Metal components are cleaned by an ultrasonic cleaner to remove surface oil and oxide layer, and then air-dried naturally.

[0007] Furthermore, in the construction of the AI ​​model, the mechanical performance parameters of different regions include compressive strength ≥14MPa, tensile strength ≥1.8MPa, and impact strength ≥4.8KJ / m in the load-bearing area. 2 The bending deformation of the decorative area is ≥2.8mm; the acoustic performance parameters include a sound insulation of ≥26dB at 100Hz and a sound insulation of ≥22dB at 500Hz; the mechanical requirements-material ratio correlation database includes 900~1100 sets of sample data, each set of sample data includes performance indicators, fiber length, metal addition density and target ratio parameters of mixing ratio.

[0008] Furthermore, in the AI ​​model construction, the AI ​​adaptive matching model is a three-layer network structure, consisting of an input layer, a hidden layer, and an output layer. The input layer receives performance parameters of the component regions; the hidden layer performs feature extraction and nonlinear transformation on the input performance parameters; the output layer outputs the final raw material matching parameters; and the output value of each neuron in the hidden layer is calculated using a region feature weighted input-hidden layer mapping formula to strengthen the performance requirement weights of different regions. The region feature weighted input-hidden layer mapping formula is as follows: ,in, For the hidden layer The output value of each neuron For activation function, For the input layer The first neuron to the hidden layer The connection weights of each neuron For the input performance parameters, , For compressive strength, For tensile strength, For impact resistance, This represents the amount of bending deformation. For sound insulation, The total number of performance parameters. For region type coefficients, For load-bearing areas, For decoration area, As a transition zone, For the hidden layer Bias terms for each neuron, This refers to the index of the input layer neurons, with a value ranging from 1 to... , This refers to the sequence number of the hidden layer neurons.

[0009] Furthermore, in the construction of the AI ​​model, based on the output value of the hidden layer, the raw material ratio parameters output by the output layer are calculated using a material compatibility-constrained hidden-output layer mapping formula. The material compatibility-constrained hidden-output layer mapping formula is as follows: ,in, For the output layer The output value of each neuron, i.e., the raw material ratio parameter, is as follows: These correspond to fiber percentage, metal addition density, and gypsum-based material flow rate, respectively. For activation function, For the hidden layer The nth neuron to the output layer The connection weights of each neuron For the output layer Bias terms for each neuron, The material compatibility coefficient. This represents the index of the output layer neuron, ranging from 1 to 3. This represents the total number of neurons in the hidden layer.

[0010] Furthermore, in the AI ​​model construction, during the AI ​​adaptive proportioning model training phase, 75%~85% of the sample data is extracted from the mechanical requirements-material proportioning correlation database as the training set, and 15%~25% of the sample data is used as the validation set. Training parameters are set, and the number of iterations is 4500~5500. During the training process, the output raw material proportioning parameters are compared. Adjust the connection weights according to the target matching parameters. , and bias terms , The error of the output raw material ratio parameters is controlled within ±4%.

[0011] Furthermore, in the gradient ratio printing, the 3D printer's multi-material bins include 3 to 5 independent material control bins, which respectively store gypsum-based materials, short fibers, long fibers, and metal components. The printing parameters include the printing layer height being consistent with the slice thickness, the printing speed being 45 to 155 mm / s, and the 3D printer's nozzle temperature being 23 to 37°C.

[0012] Furthermore, in the gradient proportion printing process, the specific discharge ratio of each hopper is as follows: the density of the metal components added to the load-bearing area is 0.45~0.85 kg / m³. 3The long fibers account for 2.8%~5.2% of the gypsum substrate mass, and the gypsum substrate discharge flow rate is 1.1~1.6L / min; the addition density of metal components in the decorative area is 0.08~0.32kg / m³. 3 The short fibers account for 0.8%~2.2% of the gypsum substrate mass, and the gypsum substrate discharge flow rate is 0.7~1.1L / min; the addition density of metal components in the transition zone is 0.30~0.50kg / m³. 3 The fiber length is 1.5~3.0mm, the fiber accounts for 1.5%~3.5% of the mass of the gypsum substrate, the gypsum substrate discharge flow rate is 1.0~1.2L / min, and the fiber length and metal density change continuously and gradually from the load-bearing area to the decorative area.

[0013] Furthermore, in the post-processing and optimization, the ambient humidity for natural air drying is 38%~62%, the temperature is 18~27℃, and the drying time is 70~74 hours; sanding and repair are performed using 700~900 grit sandpaper for dry sanding, and a waterproofing agent is sprayed at a rate of 0.18~0.32 kg / m². 2 Finished product performance testing includes mechanical properties, fire resistance, acoustic properties, and environmental performance testing.

[0014] Furthermore, in the post-processing and optimization, the test results are organized and fed back to the AI ​​adaptive mixing model. By comparing the test results with the qualified performance indicators, the parameter combinations that do not meet the test results are marked and the mixing ratio is corrected.

[0015] Compared with existing technologies, this moldless production method for adding fiber and metal materials to gypsum 3D printing has the following advantages: I. This invention solves the pain points of traditional plaster component production, such as mold dependence, limited shape, and rigid material ratio, by integrating AI adaptive proportioning with 3D printing. The application of 3D printing technology enables moldless production, eliminating the cumbersome steps of mold design, manufacturing, and demolding. At the same time, the simultaneous implantation process of metal hooks improves production flexibility and integrity. The AI ​​adaptive proportioning model, through a three-layer network structure, combined with regional feature weights and material compatibility constraints, dynamically adjusts the proportioning scheme of fiber and metal materials according to the differentiated performance requirements of the load-bearing area, decorative area, and transition area of ​​the component. This ensures the structural strength of the load-bearing area while taking into account the flexibility and surface texture of the decorative area, and achieves a smooth performance transition in the transition area to avoid the risk of cracking.

[0016] Second, this invention establishes a standardized control logic for each stage, from drying and impurity removal in the raw material pretreatment stage, to training the AI ​​adaptive mixing model based on sample data, to adjusting the mixing ratio during the 3D printing process, and finally to natural air drying and polishing in the post-processing stage. Combined with the feedback of finished product testing results to the AI ​​adaptive mixing model for optimized design, this allows for continuous correction of mixing parameters, constantly improving the consistency and stability of product performance. In terms of materials and performance assurance, the combination of gypsum, alkali-free glass fiber, and metal components enhances the mechanical properties and durability of the components through the synergistic effect of fibers and metal, making them suitable for various public buildings and commercial spaces, thus improving the product's market competitiveness.

[0017] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0019] Figure 1 A flowchart of a moldless production method for adding fiber and metal materials to plaster 3D printing. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] Example 1: Raw material pretreatment: In the production of gypsum board decorative components for the ceiling of the city's grand theater auditorium, materials with a whiteness ≥88% and an air-dry specific gravity of 1.87 g / cm³ were selected. 3High-density Alpha gypsum powder is used as the base material, combined with alkali-free glass fiber and metal components as raw materials. The alkali-free glass fiber includes short fibers and long fibers. The short fibers are 1mm long and the long fibers are 4mm long, with a single filament diameter of 6μm. The metal components include metal hooks and metal particles, and are made of 304 stainless steel to meet the rust prevention requirements of theaters in long-term air-conditioned environments. The raw materials undergo pre-treatment to dry and remove impurities. The gypsum powder is placed in a constant temperature drying oven at 55~85℃ for 4 hours to control the moisture content to ≤0.6%. The glass fiber is screened by an 80-mesh vibrating screener to remove broken impurities and dust to avoid affecting the flatness of the ceiling surface. The metal components are cleaned by a 40kHz ultrasonic cleaner to remove surface oil and oxide layer, and then air-dried.

[0022] AI Model Construction: Collect mechanical and acoustic performance parameters of different areas of the ceiling components in the auditorium of the grand theater. This ceiling needs to cover 120m. 2 The arc-shaped area, with its load-bearing zone corresponding to the installation nodes connecting to the roof steel frame, has mechanical performance parameters including a compressive strength ≥14MPa, a tensile strength ≥1.8MPa, and an impact strength ≥4.8KJ / m². 2 The decorative area corresponds to the curved surface of the ceiling and the surrounding sound reflection groove area, with a bending deformation ≥2.8mm. Acoustic performance parameters include a sound isolation level ≥26dB at 100Hz and ≥22dB at 500Hz. A mechanical requirement-material ratio correlation database is established, containing 900-1100 sets of sample data, each set including performance indicators, fiber length, metal addition density, and target ratio parameters for the mixing proportion. An AI adaptive ratio model is then constructed, consisting of a three-layer network structure: an input layer, a hidden layer, and an output layer. The input parameters are the performance parameters of the component area, and the output parameters are the raw material ratio parameters. The output value of each neuron in the hidden layer is calculated using a region feature weighted input-hidden layer mapping formula to strengthen the performance requirement weights of different areas. The region feature weighted input-hidden layer mapping formula is as follows: ,in, For the hidden layer The output value of each neuron For activation function, For the input layer The first neuron to the hidden layer The connection weights of each neuron For the input performance parameters, , For compressive strength, For tensile strength, For impact resistance, This represents the amount of bending deformation. For sound insulation, The total number of performance parameters. For region type coefficients, For load-bearing areas, For decoration area, As a transition zone, For the hidden layer Bias terms for each neuron, This refers to the index of the input layer neurons, with a value ranging from 1 to... , Let be the sequence number of the hidden layer neurons; then, calculate the raw material ratio parameters of the output layer using the material compatibility-constrained hidden-output layer mapping formula, which is: ,in, For the output layer The output value of each neuron, i.e., the raw material ratio parameter, is as follows: These correspond to fiber percentage, metal addition density, and gypsum-based material flow rate, respectively. For activation function, For the hidden layer The nth neuron to the output layer The connection weights of each neuron For the output layer Bias terms for each neuron, The material compatibility coefficient. This represents the index of the output layer neuron, ranging from 1 to 3. The total number of neurons in the hidden layer is given. Then, 75% to 85% of the sample data is extracted from the mechanical requirements-material ratio association database as the training set and 15% to 25% of the sample data is extracted as the validation set. The training parameters are set to have 4500 to 5500 iterations. During the training process, the connection weights and bias terms are adjusted by comparing the output raw material ratio parameters with the target ratio parameters. The error of the output raw material ratio parameters is controlled within ±4% to complete the training and optimization of the AI ​​adaptive ratio model.

[0023] 3D Model Processing: A 3D model of the theater auditorium ceiling components was created using Rhino software. The modeling process focused on reproducing the details of the ceiling's curved surface and sound reflection channels, ensuring precise alignment with the installation holes of the roof steel frame. AutoCAD software was used for slicing. Based on a mechanical requirements-material ratio database, the 3D model was divided into load-bearing, decorative, and transition zones. The load-bearing zone corresponds to the ceiling installation nodes and stress support areas; the decorative zone corresponds to the ceiling surface shape and exposed areas; and the transition zone is located between the load-bearing and decorative zones. Simultaneously, the insertion coordinates of the metal hooks were set in the 3D model, ensuring that the coordinates matched the bolt hole positions of the ceiling installation steel frame. Slice files were then generated, such as... Figure 1 As shown.

[0024] Gradient proportioning printing: Pre-treated gypsum-based material, short fibers, long fibers, and metal components are placed in four independent material control bins of the 3D printer. Based on the AI ​​adaptive proportioning model and slicing file, printing parameters are set, with the layer height consistent with the slice thickness, a printing speed of 45~155mm / s, and a 3D printer nozzle temperature of 23~37℃. During printing, the material output ratio of each bin is adjusted through the AI ​​adaptive proportioning model, and the density of the metal components in the load-bearing area is 0.45~0.85kg / m³. 3 The long fibers account for 2.8%~5.2% of the gypsum substrate mass, the gypsum substrate discharge flow rate is 1.1~1.6L / min, and the addition density of metal components in the decorative area is 0.08~0.32kg / m³. 3 The short fibers account for 0.8%~2.2% of the gypsum substrate mass, the gypsum substrate discharge flow rate is 0.7~1.1L / min, and the addition density of metal components in the transition zone is 0.30~0.50kg / m³. 3 The fiber length is 1.5~3.0mm, the fiber accounts for 1.5%~3.5% of the mass of the gypsum substrate, the gypsum substrate discharge flow rate is 1.0~1.2L / min, and the fiber length and metal density change continuously and gradually from the load-bearing area to the decorative area. The metal hook is implanted by a six-axis robotic arm according to the implantation coordinates.

[0025] Post-processing and optimization: The printed theater ceiling components are placed in a constant temperature and humidity curing room with a humidity of 38%~62% and a temperature of 18~27℃ to air dry naturally for 70~74 hours to avoid deformation caused by subsequent humidity fluctuations on-site in the theater. After air drying, the surface of the components is dry-sanded with 800-grit sandpaper to polish and repair any imperfections, paying special attention to cleaning the printing marks on the edges of the sound reflection grooves. An environmentally friendly water-based waterproofing agent is then sprayed on at a rate of 0.2~0.3 kg / m². 2， The product is adapted to the condensate water environment of the theater air conditioning system. Then, the mechanical properties, fire resistance, acoustic properties and environmental performance are tested. The test results are compiled and fed back to the AI ​​adaptive mixing model. By comparing the test results with the qualified performance indicators, the parameter combinations that do not meet the test results are marked and the mixing ratio is corrected.

[0026] In summary, for the production scenario of gypsum decorative components for the ceiling of the auditorium of the city's grand theater, high-density Alpha gypsum powder was used as the base material, combined with alkali-free glass fiber and stainless steel metal components. The raw materials were pre-treated by constant temperature drying, vibration screening, and ultrasonic cleaning. By collecting mechanical and acoustic parameters of the load-bearing and decorative areas of the ceiling, a related database was established and a three-layer AI adaptive proportioning model was constructed. The model was trained and optimized, and Rhino modeling and AutoCAD slicing and area division were used. 3D printers with four independent material control bins were used to print according to gradient proportions, and a six-axis robotic arm was used to implant metal hooks. Finally, constant temperature and humidity drying, dry grinding repair, and performance testing were carried out. The test results were fed back to the AI ​​adaptive proportioning model for optimization. The entire process was adapted to the theater's curved ceiling, acoustic requirements, and rust prevention environment, ensuring the quality of the components and their suitability for the scene.

[0027] Example 2: Raw material pretreatment: In the production of relief plaster components for the main wall of the hotel lobby, materials with a whiteness ≥88% and an air-dry specific gravity of 1.85~1.91g / cm³ are selected. 3 High-density Alpha gypsum powder is used as the base material, combined with alkali-free glass fiber and metal components as raw materials. The alkali-free glass fiber includes short fibers and long fibers. The short fibers are 0.4 mm long and the long fibers are 2.8 mm long, with a single filament diameter of 5 μm. The metal components include metal hooks and metal particles, and are made of 6061 aluminum alloy to reduce the weight of the relief and adapt to the load-bearing capacity of the wall. The raw materials are pre-treated by drying and removing impurities. The gypsum powder is placed in a constant temperature drying oven at 55~85℃ for 7 hours to control the moisture content to ≤0.4%. The glass fiber is screened by a 100-mesh vibrating screener to remove broken impurities and dust, ensuring the forming accuracy of delicate parts such as relief petals. The metal components are cleaned by a 35kHz ultrasonic cleaner to remove surface oil and oxide layer, and then air-dried.

[0028] AI Model Construction: Mechanical and acoustic performance parameters of different areas of the relief components on the main wall of the five-star hotel lobby were collected. The load-bearing area corresponds to the connection area between the relief and the light steel keel of the wall. Mechanical performance parameters include compressive strength ≥14MPa, tensile strength ≥1.8MPa, and impact strength ≥4.8KJ / m². 2The decorative area corresponds to the three-dimensional shapes of flower petals and leaves on the relief surface, with a bending deformation ≥2.8mm. Acoustic performance parameters include a sound isolation of ≥26dB at 100Hz and ≥22dB at 500Hz to prevent lobby noise from penetrating into adjacent areas. A mechanical requirement-material ratio correlation database is established, containing 900-1100 sets of sample data, each including performance indicators, fiber length, metal addition density, and target ratio parameters for the mixing proportion. An AI adaptive ratio model is then constructed, consisting of a three-layer network structure: an input layer, a hidden layer, and an output layer. The input parameters are the performance parameters of the component area, and the output parameters are the raw material ratio parameters. The output value of each neuron in the hidden layer is calculated using a region feature weighted input-hidden layer mapping formula to strengthen the performance requirement weights of different areas. The region feature weighted input-hidden layer mapping formula is as follows: Then, the raw material ratio parameters of the output layer are calculated using the material compatibility-constrained implicit-output layer mapping formula. The material compatibility-constrained implicit-output layer mapping formula is as follows: Then, 75%~85% of the sample data were extracted from the mechanical requirements-material ratio association database as the training set and 15%~25% of the sample data as the validation set. The training parameters were set to have 4500~5500 iterations. During the training process, the connection weights and bias terms were adjusted by comparing the output raw material ratio parameters with the target ratio parameters. The error of the output raw material ratio parameters was controlled within ±4% to complete the training and optimization of the AI ​​adaptive ratio model.

[0029] 3D Model Processing: A 3D model of the relief components on the hotel lobby wall was created using Rhino software. During modeling, the natural curvature of the flower petals and the texture of the leaf veins were reproduced to ensure consistency with the installation spacing of the light steel keel on the wall. Then, AutoCAD software was used to complete the slicing process. Combining the mechanical requirements-material ratio correlation database, the 3D model was divided into load-bearing area, decorative area, and transition area. The load-bearing area corresponds to the connection area of ​​the relief fixing the wall, the decorative area corresponds to the surface pattern and three-dimensional shape area of ​​the relief, and the transition area is located between the load-bearing area and the decorative area. At the same time, the implantation coordinates of the metal hooks were set in the 3D model to ensure that the coordinates match the bolt hole positions of the light steel keel on the wall, and a slice file was generated.

[0030] Gradient proportioning printing: Pre-treated gypsum-based material, short fibers, long fibers, and metal components are placed in four independent material control bins of the 3D printer. Based on the AI ​​adaptive proportioning model and slicing files, printing parameters are set, with the layer height consistent with the slice thickness. The printing speed is 45~155mm / s, with 45~80mm / s for relief patterns to ensure modeling accuracy and 120~155mm / s for planar areas to improve efficiency. The 3D printer nozzle temperature is 23~37℃. During printing, the material output ratio of each bin is adjusted through the AI ​​adaptive proportioning model. The density of the metal components in the load-bearing area is 0.45~0.85kg / m³. 3 The long fibers account for 2.8%~5.2% of the gypsum substrate mass, the gypsum substrate discharge flow rate is 1.1~1.6L / min, and the addition density of metal components in the decorative area is 0.08~0.32kg / m³. 3 The short fibers account for 0.8%~2.2% of the gypsum substrate mass, the gypsum substrate discharge flow rate is 0.7~1.1L / min, and the addition density of metal components in the transition zone is 0.30~0.50kg / m³. 3 The fiber length is 1.5~3.0mm, the fiber accounts for 1.5%~3.5% of the mass of the gypsum substrate, the gypsum substrate discharge flow rate is 1.0~1.2L / min, and the fiber length and metal density change continuously and gradually from the load-bearing area to the decorative area. The metal hook is implanted by a six-axis robotic arm according to the implantation coordinates.

[0031] Post-processing and optimization: The printed hotel wall relief components are placed in a curing room with a humidity of 38%~62% and a temperature of 18~27℃ to air dry naturally for 70~74 hours to avoid localized dehydration caused by lobby ventilation later. After air drying, the surface of the components is dry-sanded with 900-grit sandpaper to polish and repair, focusing on repairing printing layer marks on the edges of the petals. An environmentally friendly water-based waterproofing agent is then sprayed at a rate of 0.18~0.32 kg / m². 2 To adapt to the humid environment of the lobby, the finished product is then tested for mechanical, fire-resistant, acoustic, and environmental performance. The test results are then compiled and fed back to the AI ​​adaptive mixing model. By comparing the test results with the qualified performance indicators, the parameter combinations that do not meet the test results are marked and the mixing ratio is corrected.

[0032] In summary, for the production scenario of relief plaster components for the main wall of a hotel lobby, high-density Alpha plaster powder is used as the base material, combined with alkali-free glass fiber and aluminum alloy metal components. Pre-treatment ensures the stability of raw material performance. Performance parameters of the load-bearing and decorative areas of the relief are collected, a database is established, and an AI adaptive mixing model is trained. Rhino modeling and AutoCAD slicing are used to divide the area. A 3D printer is used to adjust the printing speed by area, and gradient mixing printing is achieved through four material control bins. A six-axis robotic arm is embedded with metal hooks. After drying in a curing workshop, sanding with fine sandpaper, and spraying with an environmentally friendly waterproofing agent, the AI ​​adaptive mixing model is optimized based on performance testing feedback to meet the requirements of detail precision, wall load-bearing capacity, and adaptability to humid environments in the hotel lobby relief.

[0033] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A moldless production method for adding fiber and metal materials to plaster 3D printing, characterized in that, The specific steps of this method are as follows: Raw material pretreatment: High-density Alpha gypsum powder is selected as the base material, combined with alkali-free glass fiber and metal components as raw materials, and the raw materials are pretreated by drying and removing impurities. AI Model Construction: Collect mechanical and acoustic performance parameters of different areas of building decorative components, establish a mechanical demand-material ratio correlation database, and construct an AI adaptive ratio model. The input parameters are the performance parameters of the component area, and the output parameters are the raw material ratio parameters. The AI ​​adaptive ratio model is trained and optimized. 3D model processing: A 3D model of the component is created using 3D modeling software, and slicing is completed using slicing software. The 3D model is divided into load-bearing area, decorative area and transition area by combining mechanical requirements-material ratio correlation database. At the same time, the implantation coordinates of metal hooks are set in the 3D model and slice files are generated. Gradient proportion printing: The pre-treated raw materials are placed in the multi-bin of the 3D printer, and the printing parameters are set based on the loaded AI adaptive proportion model and slice file. During the printing process, the AI ​​adaptive proportion model adjusts the output ratio of each bin in real time, and the metal hook is implanted by the robotic arm according to the implantation coordinate. Post-processing and optimization: The printed components are placed in a specific humidity environment to air dry naturally. After drying, the surface of the components is polished and repaired. The performance of the finished product is tested, and the test results are fed back to the AI ​​adaptive mixing model to optimize the mixing ratio.

2. The moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the raw material pretreatment, the whiteness of the high-density Alpha gypsum powder is ≥88%, and the air-dry specific gravity is 1.85~1.91 g / cm³. 3 Alkali-free glass fiber includes short fibers and long fibers. The length of short fibers is 0.4~2.1mm, and the length of long fibers is 2.8~8.2mm. The diameter of the single filament is 4~21μm. The metal components include metal hooks and metal particles. The material of the metal components is stainless steel or aluminum alloy. The drying and impurity removal process is as follows: gypsum powder is placed in a constant temperature drying oven at 55~85℃ and dried for 3~7 hours, controlling the moisture content to ≤0.6%. Glass fiber is removed by a vibrating screening machine to remove broken impurities and dust. Metal components are cleaned by an ultrasonic cleaner to remove surface oil and oxide layer, and then air-dried naturally.

3. The moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the construction of the AI ​​model, the mechanical performance parameters of different regions include compressive strength ≥14MPa, tensile strength ≥1.8MPa, and impact strength ≥4.8KJ / m in the load-bearing area. 2 The bending deformation of the decorative area is ≥2.8mm; the acoustic performance parameters include a sound insulation of ≥26dB at 100Hz and a sound insulation of ≥22dB at 500Hz; the mechanical requirements-material ratio correlation database includes 900~1100 sets of sample data, each set of sample data includes performance indicators, fiber length, metal addition density and target ratio parameters of mixing ratio.

4. The moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the AI ​​model construction, the AI ​​adaptive matching model is a three-layer network structure, consisting of an input layer, a hidden layer, and an output layer. The input layer receives performance parameters of the component regions; the hidden layer performs feature extraction and nonlinear transformation on the input performance parameters; the output layer outputs the final raw material matching parameters; and the output value of each neuron in the hidden layer is calculated using a region feature weighted input-hidden layer mapping formula to strengthen the performance requirement weights of different regions. The region feature weighted input-hidden layer mapping formula is as follows: ,in, For the hidden layer The output value of each neuron For activation function, For the input layer The first neuron to the hidden layer The connection weights of each neuron For the input performance parameters, , For compressive strength, For tensile strength, For impact resistance, This represents the amount of bending deformation. For sound insulation, The total number of performance parameters. For region type coefficients, For load-bearing areas, For decoration area, As a transition zone, For the hidden layer Bias terms for each neuron, This refers to the index of the input layer neurons, with a value ranging from 1 to... , This refers to the sequence number of the hidden layer neurons.

5. A moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 4, characterized in that, In the construction of the AI ​​model, based on the output value of the hidden layer, the raw material ratio parameters output by the output layer are calculated using the material compatibility-constrained hidden-output layer mapping formula. The material compatibility-constrained hidden-output layer mapping formula is as follows: ,in, For the output layer The output value of each neuron, i.e., the raw material ratio parameter, is as follows: These correspond to fiber percentage, metal addition density, and gypsum-based material flow rate, respectively. For activation function, For the hidden layer The nth neuron to the output layer The connection weights of each neuron For the output layer Bias terms for each neuron, The material compatibility coefficient. This represents the index of the output layer neuron, ranging from 1 to 3. This represents the total number of neurons in the hidden layer.

6. A moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 4, characterized in that, In the AI ​​model construction, during the AI ​​adaptive proportioning model training phase, 75%–85% of the sample data is extracted from the mechanical requirements-material proportioning correlation database as the training set, and 15%–25% of the sample data is used as the validation set. Training parameters are set, and the number of iterations is 4500–5500. During the training process, the output raw material proportioning parameters are compared. Adjust the connection weights according to the target matching parameters. , and bias terms , The error of the output raw material ratio parameters is controlled within ±4%.

7. A moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the gradient ratio printing, the 3D printer's multi-material bin contains 3 to 5 independent material control bins, which respectively store gypsum-based materials, short fibers, long fibers, and metal components. The printing parameters include the printing layer height being consistent with the slice thickness, the printing speed being 45 to 155 mm / s, and the 3D printer's nozzle temperature being 23 to 37°C.

8. A moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the gradient proportion printing process, the specific material output ratio of each hopper is as follows: the density of metal components added to the load-bearing area is 0.45~0.85 kg / m³. 3 The long fibers account for 2.8%~5.2% of the gypsum substrate mass, and the gypsum substrate discharge flow rate is 1.1~1.6L / min; the addition density of metal components in the decorative area is 0.08~0.32kg / m³. 3 The short fibers account for 0.8%~2.2% of the gypsum substrate mass, and the gypsum substrate discharge flow rate is 0.7~1.1L / min; the addition density of metal components in the transition zone is 0.30~0.50kg / m³. 3 The fiber length is 1.5~3.0mm, the fiber accounts for 1.5%~3.5% of the mass of the gypsum substrate, the gypsum substrate discharge flow rate is 1.0~1.2L / min, and the fiber length and metal density change continuously and gradually from the load-bearing area to the decorative area.

9. A moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the post-treatment and optimization, the ambient humidity for natural air drying is 38%~62%, the temperature is 18~27℃, and the drying time is 70~74 hours; sanding and repair are carried out using 700~900 grit sandpaper for dry sanding, and a waterproofing agent is sprayed at a rate of 0.18~0.32 kg / m². 2 Finished product performance testing includes mechanical properties, fire resistance, acoustic properties, and environmental performance testing.

10. A moldless production method for adding fiber and metal materials to plaster 3D printing according to claim 1, characterized in that, In the post-processing and optimization, the test results are organized and fed back to the AI ​​adaptive ratio model. By comparing the test results with the qualified performance indicators, the parameter combinations that do not meet the test results are marked and the ratio is corrected.

Images