Intelligent printing method for thermoregulating wall and thermoregulating wall printed by using this method

Abstract

This application relates to a method for intelligent printing of a thermally adjustable wall and a thermally adjustable wall printed by using this method, which solves the problem that previous technologies mostly rely on fixed preset parameters, resulting in a lack of real-time dynamic adjustment ability during the printing process and making it difficult to meet the complex and changeable printing environment and wall performance requirements. The method includes: during the aggregate switching process, adjusting the opening degree of the valve and the pressure of the conveying pipeline through a preset control method to enable the new aggregate to enter the printing nozzle. And during the aggregate switching process, an adaptive fuzzy PID algorithm is used to synchronously adjust the rotation speed of the stirring paddle and the aggregate conveying amount, so that the particle size distribution of the aggregate after mixing the new aggregate with the radiation cooling enhanced cement-based material is within the preset standard range of the aggregate particle size distribution. This application has the following effects: improving the printing quality and performance of the thermally adjustable wall.

Description

Technical Field

[0001] The present invention relates to the technical field of building construction, and more particularly to an intelligent printing method for a heat-regulating wall and a heat-regulating wall printed by using the method. Background Art

[0002] In the field of architecture, the wall, as an important part of the building structure, its performance plays a key role in the overall quality of the building, energy consumption and living comfort. With the continuous progress of technology, 3D printing technology has gradually been applied to the construction industry, bringing new ideas and methods for the construction of walls. As a wall with special functions, the heat-regulating wall can automatically adjust the thermal performance of the wall according to the change of the environmental temperature, thereby effectively reducing the energy consumption of the building and improving the indoor comfort. The research and innovation of its printing technology have attracted much attention.

[0003] At present, in the wall printing technology, there have been some related technical means. Some traditional wall printing methods mainly print ordinary cement-based materials layer by layer along a preset path through a single nozzle to construct the basic structure of the wall. There are also some technical attempts to add some ordinary aggregates to the printing materials to enhance the performance of the wall such as strength.

[0004] In view of the above related technologies, the inventors found that there are the following defects: Using a single nozzle can only transport single or simply mixed materials, which makes it impossible to achieve the efficient and precise mixing of the radiation cooling enhanced cement-based material and the aggregate when printing structures such as heat-regulating walls that have strict requirements on material diversity and precise proportioning, seriously affecting the heat-regulating performance of the wall. Moreover, in the data processing and printing parameter regulation link, the previous technologies mostly rely on fixed preset parameters, which results in the lack of real-time dynamic adjustment ability in the printing process and is difficult to meet the complex and changeable printing environment and wall performance requirements. Summary of the Invention

[0005] In order to improve the printing quality and performance of the heat-regulating wall, the present application provides an intelligent printing method for a heat-regulating wall and a heat-regulating wall printed by using the method.

[0006] In the first aspect, the present application provides an intelligent printing method for a heat-regulating wall, adopting the following technical solution:

[0007] An intelligent printing method for a heat-regulating wall, comprising:

[0008] The 3D printing device is started, and according to the preset printing path, the mixed material of the radiation cooling enhanced cement-based material and the aggregate is extruded and printed layer by layer through a special nozzle according to the preliminarily set printing parameters of each layer, wherein the special nozzle is a printing nozzle with two independent inlets, one of which is connected to the radiation cooling enhanced cement-based material, and the other is connected to the aggregate, and a stirring paddle is arranged inside the special nozzle;

[0009] During the printing process, a sensor matrix is ​​used to collect multi-source data, including environmental data, printing progress data, and wall thermal performance related data;

[0010] Collect historical climate data of the printing area, and perform data preprocessing on the historical climate data and the collected multi-source data;

[0011] Input the multi-source data and historical climate data that have completed data preprocessing into the pre-built multi-objective optimization model, and output the optimal parameter combination for printing each layer of the wall, and replace the initially set printing parameters of each layer according to the optimal parameter combination for printing each layer of the wall;

[0012] During the printing process, based on the multi-source data collected in real time and pre-processed, it is determined whether the pre-set trigger conditions for aggregate switching are met;

[0013] If yes, an aggregate switching instruction is issued to control the aggregate channel of the special nozzle to switch. During the aggregate switching process, the valve opening and closing degree and the conveying pipeline pressure are regulated by a preset control method to allow new aggregate to enter the printing nozzle. During the aggregate switching process, an adaptive fuzzy PID algorithm is used to synchronously adjust the stirring blade speed and the aggregate conveying amount, so that the aggregate particle size distribution after the new aggregate and the radiation cooling enhanced cement-based material are mixed is within the preset aggregate particle size distribution standard range;

[0014] If not, keep the original setting.

[0015] By adopting the above technical solution, the printing method uses multi-source data collection and processing, combined with historical climate data, and accurately adjusts the printing parameters through multi-objective optimization to improve the comprehensive performance of the wall. The aggregate switching conditions are judged in real time and finely regulated to ensure the adaptation of the new aggregate, optimize the mixing effect, ensure the reliability of the wall thermal temperature control function, and improve the printing quality.

[0016] Optionally, the method further includes a step after preprocessing the historical climate data and the collected multi-source data, which is as follows:

[0017] The multi-source data and historical climate data that have completed data preprocessing are unified in data format and standardized to form various data vectors;

[0018] Input various data vectors into a preset fuzzy membership function, calculate the membership degrees of various data vectors in different fuzzy sets, and obtain a comprehensive membership degree by weighted summation of multiple fuzzy membership functions;

[0019] Analyze whether the calculated comprehensive membership degree exceeds a trigger threshold. The calculation formula of the trigger threshold is as follows: Analyze whether the calculated comprehensive membership degree exceeds a trigger threshold. The calculation formula of the trigger threshold is as follows: , where is the trigger threshold, is an adjustment coefficient used to adjust the weight of the comprehensive membership degree mean when calculating the trigger threshold, is the mean of the comprehensive membership degree, is an adjustment coefficient used to adjust the weight of the comprehensive membership degree standard deviation when calculating the trigger threshold, is the standard deviation of the comprehensive membership degree;

[0020] If the answer is no, continue with the subsequent steps;

[0021] If the answer is yes, based on the multi-source data and historical climate data after completing data preprocessing, extract key state features, and use principal component analysis for dimensionality reduction to obtain the principal component feature vector representing the environment and equipment state. The key state features include the environmental change rate, printing progress deviation, and wall thermal performance deviation;

[0022] Use a fuzzy logic system to calculate the adjustment coefficients of the weights of each performance objective according to the principal component feature vector and preset rules, and update the weights of the multi-objective optimization model. The performance objectives include thermal regulation performance, structural stability, and material cost;

[0023] Use the trained deep Q network to input different principal component feature vectors to obtain the corresponding printing parameter combinations, and continue this operation until different printing parameter combinations in a preset number of groups are generated;

[0024] According to the updated weights of the multi-objective optimization model, perform performance evaluation on each generated printing parameter combination to obtain the scores of each performance objective, and perform weighted summation of the scores of each performance objective according to the adjusted weights to obtain the comprehensive score of each group of parameter combinations;

[0025] Sort all groups of parameter combinations according to the comprehensive scores, select the parameter combination with the highest comprehensive score as the optimal parameter combination for printing each layer of the wall, and replace the initially set printing parameters for each layer according to the optimal parameter combination for printing each layer of the wall.

[0026] Optionally, it also includes the steps after obtaining the comprehensive scores of each group of parameter combinations, which are specifically as follows:

[0027] Analyze whether the relative deviation between the comprehensive membership degree and the trigger threshold within a preset number of consecutive monitoring periods is always within the preset deviation range;

[0028] If yes, continue with the subsequent steps;

[0029] If no, screen out the parameter combinations with the top preset rankings in the comprehensive score from all the generated parameter combinations as candidate parameter combinations;

[0030] For the preliminary candidate parameter combinations printed for each layer of the wall, determine the basic parameters related to wall printing, run a random algorithm to generate random parameters within the floating range of the basic parameters related to wall printing of different categories, and integrate the random parameters and input them into the multi-physical-field coupled digital twin model;

[0031] For each candidate parameter combination of each layer of wall printing, run the digital twin model a preset number of times under each set scenario. Each time it runs, newly generated random parameters are called, and the data related to the wall performance output by each simulation is recorded;

[0032] According to the data related to the wall performance output by the simulation, use the preset performance index evaluation scoring method to obtain the performance evaluation scores of different categories;

[0033] Based on the obtained performance evaluation scores of different categories, calculate the average value of the comprehensive scores of different candidate parameter combinations under all simulation conditions according to the weight allocation scheme matched under different simulation conditions;

[0034] Calculate the coefficient of variation of the comprehensive scores of each candidate parameter combination under different simulation conditions;

[0035] Screen out the parameter combinations whose comprehensive scores rank before the preset position and whose coefficient of variation is less than the preset value under all simulation conditions as the optimal parameter combinations for each layer of wall printing, and replace the initially set printing parameters for each layer according to the optimal parameter combinations for each layer of wall printing.

[0036] Optionally, determining whether the trigger conditions for aggregate switching are met includes:

[0037] Extract the key features of the environment, printing progress, and wall thermal performance, and use the key features of the environment, printing progress, and wall thermal performance as matrix elements to construct a correlation matrix , where there are m key features of environmental data, n key features of printing progress, and p key features of wall thermal performance;

[0038] At each monitoring moment, form a state vector according to the key feature states , where , is the status value of the i-th key feature, which takes the value of 1 when the feature meets the trigger condition and 0 when it does not;

[0039] The multi-condition correlation score formula is as follows:

[0040] ;

[0041] Among them, is the multi-condition correlation score, is the transpose of the status vector , which converts the row vector into a column vector for matrix multiplication operations, is the correlation matrix The element in the i-th row and j-th column of represents the correlation strength between the i-th feature and the j-th feature;

[0042] According to the comparison result of the multi-condition correlation score and the correlation score threshold, it is determined whether the trigger condition for aggregate switching preset is met.

[0043] Optionally, determining whether the trigger condition for aggregate switching preset is met also includes steps before constructing the correlation matrix as follows:

[0044] Analyze whether the relative deviation between the comprehensive membership degree and the trigger threshold is always within the preset deviation range within a preset number of consecutive monitoring periods;

[0045] If so, continue to execute the subsequent original steps;

[0046] If not, according to the multi-source data that has completed data preprocessing, use a preset clustering analysis method to determine the parameter combination category. Among them, the parameter combination category includes heat preservation priority type walls and structure priority type walls;

[0047] According to the mapping relationship between the parameter combination category and the key features, determine the key features that match the determined parameter combination category, and input the determined parameter combination category and the matching key features into the trigger condition judgment model constructed based on the decision tree algorithm, and output the trigger condition as the trigger condition for aggregate switching preset;

[0048] Based on the real-time collected and preprocessed multi-source data, determine whether the trigger condition for aggregate switching preset is met.

[0049] Optionally, it also includes steps before determining whether the trigger condition for aggregate switching preset is met, as follows:

[0050] According to the different functions of the building, determine the thermal performance requirements of the wall, and based on the preset energy-saving standards, determine the quantitative indicators of the wall thermal performance requirements;

[0051] Perform data preprocessing on the real-time environmental data and the determined quantitative value of the wall thermal performance requirements;

[0052] Match the collected real-time environmental data, the preset quantitative value of the wall thermal performance requirements, with the set aggregate type selection criteria, and screen out the eligible aggregate categories;

[0053] Determine whether to execute the subsequent steps according to the comparison result between the screened aggregate category and the currently used aggregate category.

[0054] Optionally, the analysis process of the trigger condition for aggregate switching is as follows:

[0055] According to the screened aggregate category, determine the characteristics of the corresponding aggregate category;

[0056] According to the mapping relationship between the characteristics of the aggregate category and the trigger condition setting information for aggregate switching, analyze and determine the trigger condition setting information for aggregate switching.

[0057] Optionally, regulating the valve opening degree and the conveying pipeline pressure through a preset regulation method includes:

[0058] Control the valve that is currently conveying the aggregate to close;

[0059] Collect and obtain new aggregate characteristic data;

[0060] According to the new aggregate characteristic data, match and obtain the valve opening degree adjustment plan from the preset valve opening degree adjustment plan analysis model, and execute the valve opening degree regulation;

[0061] According to the new aggregate characteristic data, match and analyze the predicted conveying resistance from the aggregate conveying resistance analysis model;

[0062] According to the predicted conveying resistance, adjust in real time, determine the additional power required by the conveyor pump located in the conveying pipeline, and make corresponding power adjustment to the conveyor pump located in the conveying pipeline.

[0063] In a second aspect, the present application provides a thermally regulated wall, adopting the following technical solution:

[0064] A thermally regulated wall, the thermally regulated wall is a layered structure, from the outside to the inside are the outer layer, the middle layer and the inner layer in sequence; the outer layer is composed of a radiation cooling enhanced cement-based material and polystyrene particle aggregate, the middle layer is composed of a radiation cooling enhanced cement-based material and 35°C phase change aggregate, and the inner layer is composed of a radiation cooling enhanced cement-based material and 25°C phase change aggregate. Description of the Drawings

[0065] Figure 1 It is a schematic flow chart of a method for intelligent printing of a thermally adjustable wall in an embodiment of the present application.

[0066] Figure 2 It is a schematic diagram related to the manufacture of a thermally adjustable wall in an embodiment of the present application. Detailed implementation manners

[0067] The present application will be further described in detail below with reference to the accompanying drawings.

[0068] Referring to Figure 1 , a method for intelligent printing of a thermally adjustable wall disclosed in the present application includes the following steps:

[0069] Step S100: Start the 3D printing device. According to the preset printing path, layer by layer extrude and print a mixture of a radiation cooling enhanced cementitious material and aggregate through a special nozzle according to the initially set printing parameters for each layer.

[0070] Preset printing path: The printing trajectory planned in 3D modeling software according to the designed shape and size of the thermally adjustable wall, which determines the movement route of the nozzle during printing and ensures the accurate formation of each part of the wall.

[0071] The special nozzle is a nozzle specially designed for printing thermally adjustable wall materials, having two inlets respectively connected to the radiation cooling enhanced cementitious material and aggregate, and is provided with a stirring paddle inside to uniformly mix the two during printing. For the specific setting of the special nozzle, reference can be made to Figure 2 Figure a, Figure b and Figure c in Figure 2 Figure d in

[0072] Initially set printing parameters for each layer: Parameters initially determined according to the wall design requirements before printing, including printing speed, nozzle temperature, thickness of each layer, etc. These parameters will be adjusted according to the results of the subsequent multi-objective optimization model.

[0073] General process description: Taking the printing of a residential wall in a tropical region as an example, after starting the 3D printing device, the device reads the printing path data pre-planned in the modeling software. The special nozzle sucks materials from the inlets connected to the radiation cooling enhanced cementitious material and aggregate according to the initially set parameters. For example, when printing the outer layer, the radiation cooling enhanced cementitious material and polystyrene particle aggregate are inhaled into the nozzle in a certain proportion. After being evenly stirred by the stirring paddle inside the nozzle, the mixed material is extruded and printed layer by layer along the preset path to form the outer layer structure of the wall.

[0074] Step S200: During the printing process, collect multi-source data by using a sensor matrix.

[0075] Among them, multi-source data includes environmental data, printing progress data, and wall thermal performance related data.

[0076] Sensor Matrix: An array of multiple different types of sensors used to collect various types of data during the printing process. These sensors work together to obtain information from different dimensions and provide a basis for subsequent analysis and decision-making.

[0077] Environmental data: refers to various physical parameters in the printing environment, including temperature, humidity, light intensity, wind speed, etc. These data will affect the performance of wall materials and printing quality. For example, temperature and humidity may affect the solidification speed of cement-based materials. The environmental data can be obtained as follows: Use professional environmental monitoring sensors, such as temperature and humidity sensors such as DHT11 and SHT11, and light intensity sensors such as BH1750. Install these sensors around the printing area, and connect them to the data acquisition device through the sensor's built-in interface to obtain environmental data.

[0078] Print progress data: Information reflecting the progress of the printing process, such as the number of layers printed, the completion ratio of the current printing layer, printing time, etc. By monitoring the printing progress data, the printing strategy can be adjusted in time to ensure the smooth progress of the printing process. The printing progress data is directly obtained from the control system of the 3D printing device. The 3D printing device usually records the relevant information during the printing process and reads the printing progress data through the communication interface of the device (such as USB, serial port, etc.) using the corresponding software tools.

[0079] Wall thermal performance related data: data closely related to the thermal regulation function of the wall, such as temperature distribution, thermal conductivity, specific heat capacity, etc. of each layer of the wall. These data are used to evaluate the thermal performance of the wall and determine whether it meets the design requirements. Wall thermal performance related data are obtained with the help of thermal performance test sensors, such as heat flow sensors, infrared temperature sensors, etc. These sensors are installed inside the wall or at specific locations on the surface to monitor the thermal performance data of the wall in real time and transmit the data to the data acquisition equipment.

[0080] Step S300, collecting historical climate data of the printing area, and performing data preprocessing on the historical climate data and the collected multi-source data.

[0081] Historical climate data of the printing area: refers to the climate information of the area where the thermal temperature control wall is printed over a period of time, covering meteorological elements such as temperature, humidity, sunshine duration, wind speed and direction, etc. These data reflect the long-term climate characteristics of the area and are of great significance for analyzing the performance of the wall under different climate conditions.

[0082] Data preprocessing: Operations such as cleaning, transforming, and normalizing the collected raw data are performed to remove noise and outliers in the data, unify the data format and dimension, improve the data quality, and make it more suitable for subsequent analysis and model calculations.

[0083] Step S400: Input the multi-source data after completing data preprocessing and the historical climate data into the pre-constructed multi-objective optimization model, output the optimal parameter combinations for printing each layer of the wall, and replace the initially set printing parameters for each layer according to the optimal parameter combinations for printing each layer of the wall.

[0084] Multi-objective optimization model: A mathematical model that comprehensively considers multiple conflicting objectives (such as wall heat regulation performance, structural stability, material cost, etc.) and seeks the optimal solution or non-dominated solution set under certain constraint conditions. In this application, it is used to determine the optimal parameter combinations for printing each layer of the wall according to the input data. The multi-objective optimization model is constructed based on common multi-objective optimization algorithms, such as the non-dominated sorting genetic algorithm (NSGA-II), multi-objective particle swarm optimization algorithm (MOPSO), etc., using relevant scientific computing libraries in Python (such as DEAP, PyGMO, etc.). It can also be built and solved through professional optimization software (such as the optimization toolbox of MATLAB).

[0085] Optimal parameter combination: A set of parameters that makes multiple performance objectives (such as heat regulation performance, structural stability, material cost, etc.) reach an overall optimum for printing each layer of the heat-regulating wall, including printing speed, thickness of each layer, aggregate ratio, dosage of radiative cooling material, etc.

[0086] Step S500: During the printing process, based on the multi-source data collected in real time and preprocessed, determine whether it meets the pre-set trigger conditions for aggregate switching. If yes, execute Step S600; if no, execute Step S700.

[0087] Pre-set trigger conditions for aggregate switching: Criteria for determining whether to switch aggregates determined in advance based on wall design, material characteristics, and printing process. It is the basis for deciding when to replace aggregates and can ensure that the wall performance meets the standards.

[0088] The general process is described as follows:

[0089] Existing aggregate switching situation: Taking the printing of residential building walls in tropical regions as an example, polystyrene pellet aggregate is used when printing the outer layer. During the printing process, the sensor collects data in real time. If the ambient temperature continuously exceeds 30°C and the temperature of the middle layer of the wall approaches 35°C, and at the same time the printing progress reaches 80% of the designed thickness of the outer layer, these data are preprocessed and compared with the preset conditions. When the triggering condition of "high ambient temperature and the middle layer of the wall is about to reach the phase change temperature of 35°C and the outer layer printing is nearly completed" is met, the system determines that the aggregate needs to be switched, from polystyrene pellet aggregate to 35°C phase change aggregate, to meet the temperature regulation requirements of the middle layer of the wall.

[0090] Switching from no aggregate to having aggregate: Assume that no aggregate is used when printing the wall base layer (which may be a pure radiation cooling enhanced cementitious material layer). When 80% of the base layer printing is completed and the ambient temperature fluctuates greatly, and it is expected that the subsequent wall requires stronger temperature regulation ability, the sensor collects this data. After preprocessing, it is matched with the preset conditions (such as a certain proportion of the base layer completion and the ambient temperature fluctuation exceeding the threshold). The system determines that aggregate needs to be added and switches to adding polystyrene pellet aggregate to enhance the heat insulation performance of the wall and start the subsequent printing process with the participation of aggregate.

[0091] In step S600, an aggregate switching instruction is issued to control the switching of the aggregate channel of the special nozzle. During the aggregate switching process, the opening degree of the valve and the pressure of the conveying pipeline are regulated by the preset regulation method, so that the new aggregate enters the printing nozzle. And during the aggregate switching process, the adaptive fuzzy PID algorithm is used to synchronously adjust the rotation speed of the stirring paddle and the aggregate conveying volume, so that the particle size distribution of the aggregate after mixing the new aggregate and the radiation cooling enhanced cementitious material is within the preset standard range of the aggregate particle size distribution.

[0092] Aggregate switching instruction: A command signal issued by the control system to control the switching of the aggregate channel of the printing nozzle, ensuring the orderly supply of different types of aggregates according to the printing requirements.

[0093] Preset regulation method: A strategy set in advance for accurately controlling the opening degree of the valve and the pressure of the conveying pipeline, ensuring that the new aggregate enters the printing nozzle smoothly and accurately and is well mixed with the radiation cooling enhanced cementitious material.

[0094] Adaptive fuzzy PID algorithm: An algorithm that combines the advantages of fuzzy control and PID control, which can automatically adjust the control parameters according to the real-time error and the error change rate, realize the precise synchronous adjustment of the rotation speed of the stirring paddle and the aggregate conveying volume, and ensure that the particle size distribution of the mixed aggregate meets the standard.

[0095] General process description:

[0096] Existing aggregate switching situation: Assume that when printing the wall of a residential building in the temperate zone, polystyrene pellet aggregate is used for the outer layer. When step S500 determines that a switch to 23°C phase change aggregate is required, the control system issues an aggregate switching instruction. After receiving the instruction, the special nozzle controls the valve to close the polystyrene pellet aggregate channel and simultaneously opens the 23°C phase change aggregate channel. The preset regulation method is activated, and the valve opening degree is adjusted according to the characteristics of the new aggregate, enabling the 23°C phase change aggregate to stably enter the conveying pipeline, and the pipeline pressure is adjusted to ensure smooth conveying. The adaptive fuzzy PID algorithm works synchronously. According to the real-time monitored aggregate mixing situation, the rotation speed of the stirring paddle and the conveying volume of the 23°C phase change aggregate are adjusted to ensure that the particle size distribution of the mixed aggregate meets the standard and achieve efficient and uniform mixing.

[0097] Situation of switching from no aggregate to having aggregate: For example, at the starting layer of printing the wall, a pure radiation cooling enhanced cementitious material layer may be printed first. When the switching condition is met (such as reaching a certain thickness and significant ambient temperature change), the system issues an aggregate switching instruction. The special nozzle opens the aggregate channel (assuming polystyrene pellet aggregate is connected). The preset regulation method adjusts the valve opening degree and pipeline pressure to allow the polystyrene pellet aggregate to smoothly enter the nozzle. The adaptive fuzzy PID algorithm starts to work, adjusting the rotation speed of the stirring paddle and the conveying volume of the polystyrene pellet aggregate to uniformly mix the polystyrene pellets and the radiation cooling enhanced cementitious material, ensuring the quality of subsequent printing.

[0098] In step S700, maintain the original settings.

[0099] Among them, maintaining the original settings: During the printing process of the thermally regulated wall, when the multi-source data collected in real time and pre-processed do not meet the pre-set aggregate switching trigger conditions, keep the aggregate conveying state of the current printing nozzle, the rotation speed of the stirring paddle, the valve opening degree, the pipeline pressure, and the printing parameters of each layer unchanged, and continue the printing operation according to the established printing process.

[0100] An intelligent printing method for a thermally regulated wall further includes steps after pre-processing the historical climate data and the collected multi-source data, specifically as follows:

[0101] In step S301, unify the data formats of the multi-source data and historical climate data that have completed data pre-processing, and perform data standardization processing to form various data vectors.

[0102] Data format unification: Organize the multi-source data (environmental data, printing progress data, wall thermal performance data, etc.) and historical climate data that have completed pre-processing according to the same standards and specifications to make them have a consistent data format, facilitating subsequent processing and analysis. For example, unify temperature data in different units to degrees Celsius and unify time data in different formats to the standard date and time format.

[0103] Data standardization processing: Scale the data after unified formatting to make it have specific statistical characteristics, such as a mean of 0 and a standard deviation of 1. This can eliminate the dimensional differences between different data features and avoid some features having too large or too small an impact on subsequent analysis due to overly large or small numerical ranges.

[0104] Data vector: Combine the data that has undergone unified formatting and standardization processing into a vector form according to certain rules. Each vector represents a set of related data information, facilitating input into subsequent models or algorithms for calculation.

[0105] Step S302: Input various data vectors into a preset fuzzy membership function, calculate the membership degrees of various data vectors in different fuzzy sets, and obtain a comprehensive membership degree by weighted summation of multiple fuzzy membership functions.

[0106] Preset fuzzy membership function: A mathematical function used to describe the degree to which a data element belongs to a certain fuzzy set, and its value ranges from 0 to 1. In this step, the preset function is used to quantify the association degree between various data vectors and different fuzzy sets. For example, for an environmental temperature data vector, a fuzzy membership function can be set to determine the degree to which it belongs to fuzzy sets such as "high temperature", "medium temperature", and "low temperature".

[0107] Fuzzy set: A set with unclear boundaries. Different from traditional sets, the membership relationship of an element to a fuzzy set is not simply "belonging" or "not belonging", but is represented by a membership degree. For example, "comfortable temperature" is a fuzzy set, and different temperature values have different membership degrees to it.

[0108] Membership degree: A numerical value representing the degree to which a data element belongs to a certain fuzzy set, which is the calculation result of the fuzzy membership function. The closer the value is to 1, the higher the degree to which the data element belongs to the fuzzy set; the closer it is to 0, the lower the degree to which it belongs to the fuzzy set.

[0109] The general process is described as follows: Suppose during the printing process, there is a data vector containing environmental temperature, printing progress, and wall thermal performance data, such as [28, 0.6, 0.1] (representing environmental temperature 28°C, printing progress 60%, and wall thermal performance deviation 0.1 respectively). First, for the environmental temperature of 28°C, according to the preset fuzzy membership function for temperature (such as a triangular function), calculate its membership degree in the "high temperature" fuzzy set. Suppose the calculation result is 0.8; for the printing progress of 60%, according to the corresponding fuzzy membership function, the membership degree calculated in the "normal progress" fuzzy set is 0.7; for the wall thermal performance deviation of 0.1, the membership degree in the "low deviation" fuzzy set is 0.9.

[0110] If the weights of the three fuzzy membership functions of ambient temperature, printing progress, and wall thermal performance deviation are preset to be 0.4, 0.3, and 0.3 respectively.

[0111] Then the comprehensive membership degree = 0.8 × 0.4 + 0.7 × 0.3 + 0.9 × 0.3 = 0.8. Through such calculation, the comprehensive membership degree of the data vector is obtained for subsequent analysis and judgment.

[0112] Step S303: analyzing whether the calculated comprehensive membership exceeds a trigger threshold.

[0113] The calculation formula of the trigger threshold is as follows: ,in, is the trigger threshold, is the adjustment coefficient, which is used to adjust the weight of the comprehensive membership mean when calculating the trigger threshold. is the mean of the comprehensive membership, is the adjustment coefficient, which is used to adjust the weight of the comprehensive membership standard deviation when calculating the trigger threshold. is the comprehensive membership standard deviation. If it is no, then execute step S304; if it is yes, then execute step S305.

[0114] Step S304, continue with subsequent steps.

[0115] Continue to the next steps: When the comprehensive membership does not exceed the trigger threshold, no additional parameter adjustment or status analysis operation is performed, but the established printing process is continued. That is, the current print head working state, material delivery status, printing parameter settings, etc. are kept unchanged, and the subsequent printing control logic is executed in sequence, such as continuing data monitoring and judging whether the aggregate switching conditions are met.

[0116] Step S305 , based on the multi-source data that have completed data preprocessing and the historical climate data, key state features are extracted, and principal component analysis is used to reduce the dimension to obtain the principal component feature vector representing the environment and equipment status.

[0117] Among them, the key status characteristics include environmental change rate, printing progress deviation, and wall thermal performance deviation.

[0118] Environmental change rate: An indicator that measures how fast the printing environment (such as temperature, humidity, etc.) changes over time. It can be calculated by the ratio of the difference in environmental data at adjacent moments to the time interval. A large environmental change rate may affect material performance and printing quality.

[0119] Print progress deviation: The difference between the actual printing progress and the preset printing progress. For example, if a certain proportion of wall printing should be completed within a preset time period, the difference between the actual completion proportion and the preset printing progress is the printing progress deviation, which reflects whether the printing process is normal.

[0120] Deviation degree of wall thermal performance: The degree of deviation between the actual thermal performance of the wall (such as thermal conductivity, specific heat capacity, etc.) and the designed required thermal performance, which reflects whether the wall thermal performance meets the expectations.

[0121] Principal component analysis for dimensionality reduction: A data dimensionality reduction technique that converts multiple related original data features into a few uncorrelated comprehensive indicators (i.e., principal components), reducing the data dimension while retaining most of the data information and simplifying subsequent analysis.

[0122] Principal component eigenvector: The new eigenvector obtained by principal component analysis, and each vector represents a comprehensive feature that can effectively summarize the state information of the environment and equipment.

[0123] In step S306, use a fuzzy logic system to calculate the adjustment coefficients of the weights of each performance objective according to the principal component eigenvector and preset rules, and update the weights of the multi-objective optimization model.

[0124] Among them, the performance objectives include thermal regulation performance, structural stability, and material cost.

[0125] Fuzzy logic system: A system based on fuzzy set theory and fuzzy inference rules. It can process inaccurate and fuzzy information, and convert the input fuzzy data into output results through fuzzy inference. In this step, it is used to calculate the adjustment coefficients of the weights of each performance objective according to the principal component eigenvector and preset rules.

[0126] Preset rules: A series of conditional statements pre-established according to factors such as the design requirements of the wall, material properties, and printing process before printing the thermally adjustable wall, which are used to guide the fuzzy logic system to perform reasoning calculations. For example, if the principal component eigenvector shows a large environmental change rate and a slow printing progress, then appropriately increase the weight of the thermal regulation performance objective.

[0127] Performance objective weight adjustment coefficient: A value used to adjust the weights of performance objectives such as thermal regulation performance, structural stability, and material cost. By changing these coefficients, the importance of different performance objectives in the multi-objective optimization model can be dynamically adjusted.

[0128] Multi-objective optimization model weights: The weight values corresponding to different performance objectives in the multi-objective optimization model. These weights determine the relative importance of each objective in the model. By adjusting the weights, the model can be made to be more inclined to optimize certain objectives in different situations.

[0129] The general process is described as follows: Assume that during the printing of the thermally regulated wall, the principal component feature vector shows a relatively large environmental change rate, the printing progress deviation is negative (i.e., the printing progress is behind), and the deviation of the wall's thermal performance is small. These pieces of information are input into a fuzzy logic system constructed using the Scikit-fuzzy library. The system makes inferences based on preset rules. For example, the preset rule stipulates that when the environmental change rate is large and the printing progress is behind, the weight of the thermal regulation performance target should be appropriately increased, and the weight of the material cost target should be decreased. After fuzzy inference calculation, the adjustment coefficient of the weight of the thermal regulation performance target is 1.2, the adjustment coefficient of the weight of the structural stability target is 1.0, and the adjustment coefficient of the weight of the material cost target is 0.8. Then, these adjustment coefficients are used to update the weights of each performance target in the multi-objective optimization model, making the model pay more attention to the optimization of thermal regulation performance in subsequent calculations and relatively reducing the consideration of material costs to adapt to the current printing state.

[0130] Step S307: Using the trained deep Q-network, input different principal component feature vectors to obtain the corresponding combinations of printing parameters. Continue this operation until different combinations of printing parameters in a preset number of groups are generated for different states.

[0131] Deep Q-Network (DQN): A neural network model based on reinforcement learning that makes decisions by learning the environmental state and the corresponding optimal behavior strategy. In the printing of thermally regulated walls, it is used to output the corresponding combinations of printing parameters according to different principal component feature vectors (representing environmental and equipment states).

[0132] Preset number of groups: The number of combinations of printing parameters preset according to printing requirements and experimental experience, used to ensure that a sufficient number of parameter combinations in different states are obtained to meet the needs of multi-objective optimization.

[0133] The general process is described as follows: Assume that when printing the thermally regulated wall in a cold region, multiple principal component feature vectors are obtained from step S305. A deep Q-network model has been built and trained using TensorFlow. The first principal component feature vector is input into the deep Q-network model, and the model outputs a set of printing parameter combinations after calculation. For example, the printing speed is 50 mm / s, the nozzle temperature is 200 °C, the thickness of each layer is 5 mm, and the ratio of phase change aggregate to radiation cooling enhanced cement-based material is 3:7. Then, the next principal component feature vector is input into the model, and another set of different printing parameter combinations is obtained. This process continues until different combinations of printing parameters in a preset number of groups (assumed to be 20 groups) for different states are generated. In this way, multiple parameter combination schemes suitable for different printing states are obtained, laying a foundation for selecting the optimal parameter combination subsequently.

[0134] Step S308: Based on the updated multi-objective optimization model weights, perform performance evaluation on each generated set of printing parameter combinations to obtain the scores of each performance objective, and sum up the scores of each performance objective weighted by the adjusted weights to obtain the comprehensive score of each set of parameter combinations.

[0135] Updated multi-objective optimization model weights: In step S306, the weight values obtained by modifying the original multi-objective optimization model weights according to the performance objective weight adjustment coefficients calculated by the fuzzy logic system. These weights are used to measure the relative importance of performance objectives such as thermal regulation performance, structural stability, and material cost in the comprehensive evaluation.

[0136] Performance evaluation: According to certain criteria and methods, quantitatively evaluate the performance of each set of printing parameter combinations in terms of thermal regulation performance, structural stability, material cost, etc., to obtain the scores of each performance objective, and thus judge the influence degree of the parameter combination on the wall performance.

[0137] Performance objective scores: The quantitative evaluation results of each set of printing parameter combinations on performance objectives such as thermal regulation performance, structural stability, and material cost. The higher the score, the better the performance on that performance objective.

[0138] Comprehensive score: The value obtained by summing up the scores of each performance objective of each set of printing parameter combinations weighted by the updated multi-objective optimization model weights, which comprehensively reflects the overall quality of each set of printing parameter combinations and is used for subsequent ranking and selection of different parameter combinations.

[0139] Step S309: Sort all sets of parameter combinations according to the comprehensive scores, select the parameter combination with the highest comprehensive score as the optimal parameter combination for printing each layer of the wall, and replace the initially set printing parameters for each layer according to the optimal parameter combination for printing each layer of the wall.

[0140] Comprehensive score: In step S308, the value obtained by summing up the scores of each performance objective of each set of printing parameter combinations weighted by the updated multi-objective optimization model weights, which is used to comprehensively measure the quality of the printing parameter combinations.

[0141] Parameter combination ranking: Arrange all generated printing parameter combinations in order according to the level of the comprehensive scores. The combinations with high scores are ranked in the front, and the combinations with low scores are ranked in the back, which is convenient for screening out the optimal combination.

[0142] Optimal parameter combination: Among all printing parameter combinations, the set of parameters with the highest comprehensive score, which achieves the best balance among multiple performance objectives such as thermal regulation performance, structural stability, and material cost, and is most suitable for printing each layer of the wall.

[0143] A method for intelligent printing of a thermal temperature-controlled wall further includes steps after obtaining a comprehensive score for each parameter combination, specifically as follows:

[0144] Step S30A, analyzing whether the relative deviation between the comprehensive membership and the trigger threshold value is always within the preset deviation range within the preset number of continuous monitoring cycles. If yes, execute step S30B; if no, execute step S30C.

[0145] Preset number of continuous monitoring cycles: The number of pre-set time periods for continuous monitoring of the comprehensive membership. In each monitoring cycle, the comprehensive membership is calculated and compared with the trigger threshold to determine the stability of the printing process.

[0146] Preset deviation range: The preset allowable range of relative deviation between the comprehensive membership and the trigger threshold according to the printing process requirements and wall quality standards. If the relative deviation is within this range, it means that the printing process is relatively stable, otherwise the printing parameters may need to be adjusted.

[0147] Step S30B, continue with subsequent steps.

[0148] Continue to the next steps: When the relative deviation between the comprehensive membership and the trigger threshold is always within the preset deviation range within the preset number of continuous monitoring cycles, no additional parameter adjustment or screening operation is performed, but the established printing process is continued. This means that the current print head working state, material delivery situation, printing parameter settings, etc. are kept unchanged, and the subsequent printing control logic is continued, such as continuing data monitoring and judging whether other conditions are met.

[0149] Step S30C: Filter out, from all generated parameter combinations, parameter combinations with comprehensive scores ranked in a preset position as candidate parameter combinations.

[0150] All generated parameter combinations: In step S307, multiple groups of printing parameter combinations are obtained by inputting different principal component eigenvectors using the trained deep Q network, covering different value combinations of parameters such as printing speed, nozzle temperature, thickness of each layer, and aggregate ratio.

[0151] Comprehensive score ranking: the order of each parameter combination after all generated parameter combinations are sorted from high to low according to the comprehensive scores (calculated in step S308).

[0152] Preset ranking: A ranking limit value is preset according to printing requirements and actual conditions. Parameter combinations ranked before this ranking are considered to have relatively good comprehensive performance and have the potential to become the object of further screening.

[0153] Candidate parameter combinations: From all the generated parameter combinations, screen out those parameter combinations whose comprehensive scores rank among the top preset positions. These combinations have been preliminarily screened and are relatively outstanding in terms of thermal regulation performance, structural stability, material cost, etc., and are used for subsequent more refined evaluation and selection.

[0154] Step S30D: For the preliminary candidate parameter combinations printed for each layer of the wall, determine the basic parameters related to wall printing, and run a random algorithm to generate random parameters within the floating range of the basic parameters related to wall printing of different categories, and integrate the random parameters and input them into the multi-physical-field coupled digital twin model.

[0155] Basic parameters related to wall printing: The basic parameters directly related to the wall printing process, such as printing speed, nozzle temperature, printing thickness per layer, proportion of different aggregates, etc. These parameters directly affect the printing quality and performance of the wall.

[0156] Random algorithm: An algorithm that can generate random numbers. In this step, it is used to generate random parameters that fluctuate within a specific range to simulate various uncertain factors that may occur during the actual printing process.

[0157] Floating range: The preset fluctuating interval for each basic parameter related to wall printing. The random parameters generated by the random algorithm will take values within this range to consider the uncertainties during the printing process.

[0158] Multi-physical-field coupled digital twin model: A model based on digital twin technology that couples and simulates multiple physical fields (such as thermal field, mechanical field, etc.) during the wall printing process. By inputting different parameter combinations, the performance of the wall during the printing process and subsequent use can be simulated.

[0159] Step S30E: For each candidate parameter combination of each layer of wall printing, run the digital twin model the preset number of times under each set scenario. Each time it runs, newly generated random parameters are called, and the data related to the wall performance output by each simulation is recorded.

[0160] Preset number of times: According to the requirements of model accuracy and the comprehensiveness of simulation, the number of times the digital twin model is run for each candidate parameter combination under each scenario is preset in advance. Running multiple times can consider the uncertain influence brought by random parameters and improve the reliability of the simulation results.

[0161] Random parameters: In step S30D, the parameter values within the floating range of the basic parameters related to wall printing generated by the random algorithm are used to simulate the variable factors that may occur during the actual printing process, such as slight differences in material properties and small fluctuations in environmental factors.

[0162] Wall performance-related data: Various data reflecting the wall performance output by the digital twin model during operation, such as the temperature distribution, structural stress, thermal conductivity, compressive strength, etc. of the wall. These data are used to evaluate the impact of candidate parameter combinations on the wall performance.

[0163] Step S30F, according to the wall performance-related data output by the simulation, adopt a preset performance index evaluation scoring method to obtain performance evaluation scores of different categories.

[0164] Wall performance-related data output by the simulation: In step S30E, the data output by the digital twin model after running for each candidate parameter combination in different scenarios, including multi-faceted information such as the thermal performance, structural performance, and durability of the wall, such as data on temperature distribution, stress and strain, material aging degree, etc.

[0165] Preset performance index evaluation scoring method: A set of rules and algorithms formulated in advance according to the design objectives and quality requirements of the wall for quantitatively evaluating the wall performance. This method converts the various wall performance-related data output by the simulation into specific scores to measure the performance level of the wall in different performance aspects.

[0166] Taking the thermal regulation performance as an example, temperature fluctuation: within 24 hours, if the indoor surface temperature fluctuation ≤ 2°C, it gets 10 points; if it is 2 - 4°C, it gets 8 points; if it is 4 - 6°C, it gets 6 points; if it > 6°C, it gets 4 points. For example, if the fluctuation is 3°C, this item gets 8 points.

[0167] Temperature difference regulation: If the average indoor-outdoor temperature difference ≥ 8°C, it gets 10 points; if it is 6 - 8°C, it gets 8 points; if it is 4 - 6°C, it gets 6 points; if it < 4°C, it gets 4 points. If the temperature difference is 7°C, it corresponds to 8 points.

[0168] Taking the structural stability as an example, compressive strength: If the actual compressive strength reaches 100% and above of the design value, it gets 10 points; if it is 90% - 100%, it gets 8 points; if it is 80% - 90%, it gets 6 points; if it < 80%, it gets 4 points. The design strength is 20 MPa, and the actual is 18 MPa (the compliance rate is 90%), it gets 8 points. Deformation degree: If the maximum deformation amount < 0.1% of the wall height, it gets 10 points; if it is 0.1% - 0.3%, it gets 8 points; if it is 0.3% - 0.5%, it gets 6 points; if it > 0.5%, it gets 4 points. If the deformation amount is 0.2% of the wall height, it gets 8 points.

[0169] Performance evaluation scores of different categories: Specific scores calculated according to the preset performance index evaluation scoring method for different performance categories of the wall (such as thermal regulation performance, structural stability, waterproof performance, etc.), which are used to analyze in detail the impact of candidate parameter combinations on various aspects of the wall performance.

[0170] Step S30G: Based on the obtained performance evaluation scores of different categories, calculate the average value of the comprehensive scores of different candidate parameter combinations under all simulation conditions according to the weight allocation scheme matched under different simulation conditions.

[0171] Different simulation conditions: Different scenarios set in step S30E, such as different climate conditions (high temperature, low temperature, humidity, etc.) and different building usage scenarios (residential, commercial, industrial, etc.).

[0172] Weight allocation scheme: The weights preset for the performance evaluation scores of different categories for different simulation conditions, which reflects the importance of each performance under different conditions.

[0173] Comprehensive score: The score obtained by weighted summing the performance evaluation scores of different categories according to the corresponding weights, which is used to comprehensively measure the advantages and disadvantages of candidate parameter combinations under specific simulation conditions.

[0174] Average value of comprehensive scores: The arithmetic average of the comprehensive scores of each candidate parameter combination under all simulation conditions, which can more comprehensively reflect the overall performance of the parameter combination.

[0175] Suppose there are 3 candidate parameter combinations when printing a thermally regulated wall, and two simulation conditions, namely "high-temperature dry residential scenario" and "low-temperature humid commercial scenario", are set. There are three types of performance evaluation scores, namely thermal regulation performance, structural stability, and waterproof performance, under each condition.

[0176] For the "high-temperature dry residential scenario", the weight allocation scheme is that the weight of thermal regulation performance is 0.6, the weight of structural stability is 0.3, and the weight of waterproof performance is 0.1; for the "low-temperature humid commercial scenario", the weight allocation scheme is that the weight of thermal regulation performance is 0.4, the weight of structural stability is 0.3, and the weight of waterproof performance is 0.3.

[0177] Taking the first candidate parameter combination as an example, in the "high-temperature dry residential scenario", the thermal regulation performance score is 8 points, the structural stability score is 7 points, and the waterproof performance score is 6 points. Then the comprehensive score under this condition = 8×0.6 + 7×0.3 + 6×0.1 = 7.5 points. In the "low-temperature humid commercial scenario", the thermal regulation performance score is 7 points, the structural stability score is 8 points, and the waterproof performance score is 7 points. Then the comprehensive score under this condition = 7×0.4 + 8×0.3 + 7×0.3 = 7.3 points. The average value of the comprehensive scores of this candidate parameter combination under all simulation conditions = (7.5 + 7.3)÷2 = 7.4 points.

[0178] Perform the same calculation process for the other two candidate parameter combinations, so as to obtain the average value of the comprehensive scores of each candidate parameter combination under all simulation conditions, providing a basis for subsequent screening of the optimal parameter combination.

[0179] Step S30H, calculate the coefficient of variation of the comprehensive scores of each candidate parameter combination under different simulation conditions.

[0180] Coefficient of variation: A statistic used to measure the degree of data dispersion. In this step, by calculating the coefficient of variation of the comprehensive scores of the candidate parameter combinations under different simulation conditions, the stability of the comprehensive performance of this parameter combination under different usage environments is reflected. The smaller the coefficient of variation, the more stable the comprehensive score, that is, the less this parameter combination is affected by different working conditions. The calculation formula for the coefficient of variation is: Coefficient of variation = (Standard deviation of comprehensive score ÷ Mean of comprehensive score) × 100%.

[0181] Step S30I, screen out the parameter combinations whose comprehensive scores rank before the preset position and whose coefficient of variation is less than the preset value under all simulation conditions, as the optimal parameter combinations for printing each layer of the wall, and replace the initially set printing parameters for each layer according to the optimal parameter combinations for printing each layer of the wall.

[0182] Preset value: Determined by technicians according to the actual requirements and experience of wall printing, stored in the configuration file of the 3D printing equipment control system, and called when executing Step S30I.

[0183] Judging whether the trigger conditions for aggregate switching are met includes:

[0184] Step S510, extract the key features of the environment, printing progress, and wall thermal performance, and use the key features of the environment, printing progress, and wall thermal performance as matrix elements to construct a correlation matrix , where there are m key features of environmental data, n key features of printing progress, and p key features of wall thermal performance.

[0185] Step S520, at each monitoring moment, form a state vector according to the key feature states , where, , is the state value of the i-th key feature, taking the value of 1 when this feature meets the trigger condition and 0 when it does not.

[0186] The multi-condition correlation score formula is as follows:

[0187] ;

[0188] where, is the multi-condition correlation score, is the transpose of the state vector , which converts the row vector into a column vector for matrix multiplication operations, is the correlation matrix The element in the i-th row and j-th column of, representing the correlation strength between the i-th feature and the j-th feature.

[0189] Monitoring time: During the wall printing process, it is the time point set at a certain time interval for checking and analyzing the status of various key features, so as to obtain data in real time and determine whether the trigger conditions are met.

[0190] Status of key features: It refers to the status of key features such as environment, printing progress, and wall thermal performance at the monitoring time, which is manifested as whether the pre-set trigger conditions are met.

[0191] State vector: A vector composed of the status values of each key feature. The number of elements in the vector is equal to the total number of key features. The value of each element is 0 or 1, indicating that the corresponding feature does not meet or meets the trigger condition respectively.

[0192] Status value of the i-th key feature: In the state vector, it represents the status value of the i-th key feature (one of the key features of environment, printing progress, or wall thermal performance) at the monitoring time. 1 indicates that the feature meets the trigger condition, and 0 indicates that it does not.

[0193] Step S530, according to the multi-condition correlation score Compare the result with the correlation score threshold to determine whether the pre-set trigger conditions for aggregate switching are met.

[0194] Multi-condition correlation score: A numerical value obtained by performing matrix operations on the state vector in step S520 and the correlation matrix in step S510. This score comprehensively reflects the correlation degree among multiple key features such as environment, printing progress, and wall thermal performance and their satisfaction of the trigger conditions.

[0195] Correlation score threshold: A pre-set standard value used to compare with the multi-condition correlation score. If the multi-condition correlation score reaches or exceeds this threshold, it is considered that the trigger conditions for aggregate switching are met; otherwise, they are not. This threshold is determined based on the actual requirements, experience, and a large amount of experimental data of wall printing. It balances various factors in the printing process to ensure aggregate switching at the appropriate time.

[0196] Judging whether the pre-set trigger conditions for aggregate switching are met also includes steps before constructing the correlation matrix Specifically as follows:

[0197] Step S5A0, analyze whether the relative deviation between the comprehensive membership degree and the trigger threshold is always within the pre-set deviation range within a preset number of consecutive monitoring periods. If so, execute step S5B0; if not, execute step S5C0.

[0198] Step S5B0, continue to execute the subsequent original steps.

[0199] Continue to execute the subsequent original steps: When the relative deviation between the comprehensive membership degree and the trigger threshold remains within the preset deviation range in a preset number of consecutive monitoring periods, the printing process is determined to be relatively stable, and no additional special processing or adjustment is required. At this time, the 3D printing device will continue to execute the subsequent operation steps according to the established printing program and process, including but not limited to continuously monitoring the printing environment, advancing the printing progress, and real-time detecting the thermal performance of the wall, etc., to ensure the smooth progress of the printing work until the wall printing task is completed.

[0200] Step S5C0, according to the multi-source data that has completed data preprocessing, use a preset clustering analysis method to determine the parameter combination categories, where the parameter combination categories include heat preservation priority type walls and structure priority type walls.

[0201] Preset clustering analysis method: A data mining technique that groups data objects into clusters of similar objects, so that data objects within the same cluster have a high degree of similarity, while data objects between different clusters have large differences. In this step, it is used to analyze the multi-source data that has completed data preprocessing and divide similar parameter combinations into the same category. Common clustering analysis methods include K-Means clustering, hierarchical clustering, etc.

[0202] Parameter combination category: Different categories obtained by processing multi-source data through the clustering analysis method, which reflect the internal structure and characteristics of the data. In the wall printing scenario, it mainly includes heat preservation priority type walls and structure priority type walls. The parameter combination of the heat preservation priority type wall focuses on improving the heat preservation performance of the wall, such as using more heat preservation materials, adjusting the printing process to optimize the heat insulation structure of the wall, etc.; the parameter combination of the structure priority type wall pays more attention to the structural strength and stability of the wall, such as adjusting the aggregate ratio, increasing the support structure, etc.

[0203] Step S5D0, according to the mapping relationship between the parameter combination category and the key features, determine the key features that match the determined parameter combination category, and input the determined parameter combination category and the matching key features into the trigger condition judgment model constructed based on the decision tree algorithm, and output the trigger condition as the trigger condition for aggregate switching set in advance.

[0204] Key feature: An important feature that is closely related to the wall printing process and performance, covering multiple dimensions such as the environment (such as temperature, humidity), printing progress (such as the number of printed layers, printing speed), and wall thermal performance (such as wall surface temperature, thermal conductivity).

[0205] Mapping relationship: A correspondence between parameter combination categories and key features, which indicates which key features play important roles under different parameter combination categories and how these key features affect the performance of the wall.

[0206] Trigger condition judgment model: Built by training with a large amount of historical data based on the decision tree algorithm using tools such as the Scikit-learn library in Python. The trained model is stored in the control system of the 3D printing device in the form of a file and is loaded and used when needed.

[0207] The general process is described as follows:

[0208] Suppose in step S5C0, two parameter combination categories, namely heat preservation priority type walls and structure priority type walls, are obtained through cluster analysis. Through analyzing historical data, technicians determine that the key features corresponding to the heat preservation priority type walls are ambient temperature, wall surface temperature, and heat preservation material ratio; the key features corresponding to the structure priority type walls are the number of printed layers, printing speed, and aggregate strength.

[0209] In a certain printing project, according to the current printing situation, it is determined that the current parameter combination category is the heat preservation priority type wall. Key feature data such as the current ambient temperature, wall surface temperature, and heat preservation material ratio are collected from sensors. The parameter combination category of the heat preservation priority type wall and the corresponding key feature data are input into the trigger condition judgment model built based on the decision tree algorithm.

[0210] The model makes judgments according to the rules learned during training. For example, if the ambient temperature is lower than a certain threshold, and the wall surface temperature is higher than another threshold, and at the same time the heat preservation material ratio is lower than a certain value, the model outputs the trigger condition for aggregate switching. The control system stores these trigger conditions in the configuration file for subsequent judgment of whether to perform the aggregate switching operation.

[0211] Step S5E0, based on the multi-source data collected in real time and pre-processed, judge whether the trigger conditions for aggregate switching set in advance are met.

[0212] The general process is described as follows:

[0213] Suppose when printing a thermally adjusted wall, the trigger conditions for aggregate switching set in advance are: when the ambient temperature is lower than 10 °C, and the number of printed layers reaches 30 layers, and at the same time the wall surface temperature is lower than 25 °C, aggregate switching is required. During the printing process, the ambient temperature data is collected in real time by the ambient sensor, the number of printed layers data is recorded by the monitoring system of the printing device, and the wall surface temperature data is obtained by the wall thermal performance detection device. These data are transmitted to the data processing module in real time, and after pre-processing operations such as cleaning, denoising, and normalization, they are stored in the data storage area of the control system.

[0214] At a certain moment, the data collected by the sensor and preprocessed shows that the ambient temperature is 8°C, the number of printed layers is 32, and the wall surface temperature is 23°C. The control system reads the preset trigger conditions from the configuration file and compares the real-time data with them. It is found that the current ambient temperature of 8°C is lower than 10°C, the number of printed layers of 32 reaches 30, and the wall surface temperature of 23°C is lower than 25°C, meeting all the preset trigger conditions. At this time, the control system determines that an aggregate switching operation is required, and then issues a corresponding instruction to start the aggregate switching process. If at a certain moment, for example, the ambient temperature is 12°C, even if the number of printed layers and the wall surface temperature meet the conditions, but since the ambient temperature does not meet the condition of being lower than 10°C, the control system determines that the trigger conditions are not met, and the printing process continues according to the current aggregate usage situation.

[0215] A smart printing method for a thermally regulated wall further includes steps before determining whether the preset trigger conditions for aggregate switching are met, specifically as follows:

[0216] Step SA00, according to the different functions of the building, determine the thermal performance requirements of the wall, and based on the preset energy-saving standards, determine the quantitative indicators of the wall thermal performance requirements.

[0217] Different functions of the building: various functions that the building has due to different usage purposes. For example, a residential building is mainly used for living and requires good thermal insulation performance to ensure indoor comfort; a commercial building such as a shopping mall has a large flow of people and has high requirements for the ability to ventilate and quickly adjust the indoor temperature; an industrial building may have specific requirements for the fire prevention, moisture-proof, thermal insulation and other performances of the wall due to different production processes.

[0218] Thermal performance requirements of the wall: the performance requirements that the wall should achieve in terms of heat transfer, heat insulation, etc. determined according to the building function. For buildings with different functions, the thermal performance requirements of their walls are different. For example, the walls of a residential building need to effectively prevent the exchange of heat between indoors and outdoors and keep the indoor temperature stable; the walls of a cold storage building need extremely strong heat insulation performance to prevent cold loss.

[0219] Preset energy-saving standards: a series of specifications and indicators for building energy conservation formulated by the state or the industry. These standards stipulate the minimum requirements that different types of buildings should meet in terms of energy consumption, thermal performance, etc., aiming to promote the building industry to develop in the direction of energy conservation and emission reduction.

[0220] Quantification Index of Wall Thermal Performance Requirements: Represent the wall thermal performance requirements with specific numerical values to more precisely guide the parameter selection and material configuration during the wall printing process. Common quantification indices include heat transfer coefficient, thermal resistance, heat storage coefficient, etc. The lower the heat transfer coefficient, the better the heat insulation performance of the wall; the greater the thermal resistance, the stronger the ability of the wall to prevent heat transfer.

[0221] The general process is described as follows: Suppose we want to print the thermally conditioned wall of a residential building. First, we learn from the architectural design drawings that the building is an ordinary residence with the function of providing a comfortable living space. Then, we query the local energy-saving standards for residential buildings. For example, the local requirement is that the heat transfer coefficient of the exterior wall of a residential building should not exceed 1.5 W / (m²・K). Next, we use the Hongye load calculation software and input the relevant parameters of the residence, including information such as the building orientation being due south and the exterior wall being an ordinary brick wall. The software calculates according to the energy-saving standards and the input parameters, and obtains the quantification index of the wall thermal performance requirements of this residence, such as the heat transfer coefficient needs to be controlled within 1.2 W / (m²・K), and the thermal resistance should reach above 0.83 m²・K / W. Through such a process, the quantification index of the wall thermal performance requirements of this residential building is clarified, providing a basis for material selection and process control during the subsequent wall printing process.

[0222] Step SB00, perform data preprocessing on the real-time environmental data and the determined quantification value of the wall thermal performance requirements.

[0223] Data Preprocessing: Perform a series of processing operations on the collected raw data to improve data quality and make it more suitable for subsequent analysis and use. Common data preprocessing operations include data cleaning (removing noise data, outliers, duplicate data), data normalization (converting data to a specific numerical range for easy comparison and operation between different data), data encoding (converting non-numerical data to numerical data), etc.

[0224] Step SC00, match the collected real-time environmental data, the preset quantification value of the wall thermal performance requirements, with the set aggregate type selection criteria, and screen out the eligible aggregate categories.

[0225] Set Aggregate Type Selection Criteria: A series of pre-established rules for selecting suitable aggregate categories based on real-time environmental data and the quantification value of the wall thermal performance requirements. These criteria consider the influence of different aggregates on the wall thermal performance under different environmental conditions.

[0226] Eligible Aggregate Categories: The aggregate types that are screened out after being matched with the set aggregate type selection criteria and can meet the requirements of the current real-time environmental data and the quantification value of the wall thermal performance requirements.

[0227] Step SD00: Determine whether to execute subsequent steps based on the comparison result between the selected aggregate category and the currently used aggregate category.

[0228] Selected aggregate category: In step SC00, after matching the real-time environmental data, the preset quantitative value of the wall thermal performance requirement, and the set aggregate type selection criteria, the qualified aggregate category is determined, such as lightweight thermal insulation aggregate, composite phase change aggregate, ordinary aggregate, etc.

[0229] Currently used aggregate category: The type of aggregate currently being used for printing during the wall printing process. As the printing process progresses, the aggregate may need to be changed due to various factors, and this category will change accordingly.

[0230] Comparison result: By comparing the selected aggregate category and the currently used aggregate category, it is judged whether the two are the same. If they are the same, it means that the currently used aggregate meets the current environmental and wall thermal performance requirements; if they are different, it indicates that the aggregate may need to be changed to better meet the requirements.

[0231] Determine whether to execute subsequent steps: Decide whether to continue with the subsequent steps related to aggregate switching based on the comparison result. If the comparison result is the same, usually there is no need to execute the subsequent complex aggregate switching judgment and operation steps, and the printing process can continue in the current state; if the comparison result is different, it may be necessary to further evaluate whether the trigger conditions for aggregate switching are met and other subsequent steps.

[0232] The analysis process for the trigger conditions for aggregate switching is as follows:

[0233] Step S5a0: Determine the characteristics of the corresponding aggregate category according to the selected aggregate category.

[0234] Selected aggregate category: In the previous step (such as SC00), after matching the real-time environmental data, the preset quantitative value of the wall thermal performance requirement, and the set aggregate type selection criteria, the qualified aggregate type is determined, such as lightweight thermal insulation aggregate, composite phase change aggregate, ordinary aggregate, etc.

[0235] Characteristics of the corresponding aggregate category: The unique physical and chemical properties of each aggregate category, which will affect the performance of the aggregate during the printing process and the final performance of the wall. For example, lightweight thermal insulation aggregate usually has the characteristics of low density and good thermal insulation performance; composite phase change aggregate can absorb or release heat according to the environmental temperature change and has good thermal regulation ability; ordinary aggregate has its own characteristics in terms of strength, cost, etc. Characteristics include but are not limited to the density, particle size distribution, thermal conductivity, specific heat capacity, water absorption, etc. of the aggregate.

[0236] The general process is described as follows: Assume that when printing a thermally conditioned wall, in step SC00, the qualified aggregate category is screened out as lightweight thermal insulation aggregate. When step S5a0 is executed, the 3D printing equipment control system reads from the data storage area that the screening result is lightweight thermal insulation aggregate. Then, the control system queries the database for records on the characteristics of the lightweight thermal insulation aggregate, which may be from previous laboratory test data and information provided by the aggregate manufacturer. It is found that the density of this lightweight thermal insulation aggregate is 300 kg / m³, the particle size is mainly distributed between 1 - 3 mm, the thermal conductivity is 0.05 W / (m・K), the specific heat capacity is 1000 J / (kg・K), and the water absorption is low. By determining these characteristics, it provides the basic data for setting information on the trigger conditions for aggregate switching in subsequent analysis (step S5b0), because different aggregate characteristics correspond to different trigger conditions for when to switch aggregates.

[0237] In step S5b0, based on the mapping relationship between the characteristics of the aggregate category and the setting information of the trigger conditions for aggregate switching, analyze and determine the setting information of the trigger conditions for aggregate switching.

[0238] Setting information of the trigger conditions for aggregate switching: A series of pre-set rules and conditions used to determine when an aggregate switching operation is required during the wall printing process. This information is closely related to the characteristics of the aggregate category. Different aggregates with different characteristics have different trigger switching conditions under different printing environments and wall performance requirements. For example, when the thermal performance requirements of the wall change and the current aggregate's thermal conductivity and other characteristics cannot meet the new requirements, it may trigger aggregate switching.

[0239] Mapping relationship: The corresponding connection established between the characteristics of the aggregate category and the setting information of the trigger conditions for aggregate switching. Through this mapping relationship, based on the characteristics of a specific aggregate, the reasonable trigger conditions for aggregate switching can be quickly determined. For example, for aggregates with a higher thermal conductivity, when the wall's insulation performance requirement is improved to a certain extent, it triggers switching; while for aggregates with strong water absorption, when the environmental humidity reaches a certain threshold and affects the wall quality, it triggers switching.

[0240] Regulating the valve opening degree and the pressure of the conveying pipeline through a pre-set regulation method includes:

[0241] In step S601, control the valve that is currently conveying the aggregate to close.

[0242] The valve that is currently conveying the aggregate: During the wall printing process, the valve responsible for controlling the current aggregate being conveyed from the storage container to the print head of the printing equipment. This valve plays a role in regulating the aggregate flow rate and on / off to ensure that the aggregate is conveyed to the printing position according to the set requirements.

[0243] Control valve closing: Send a closing instruction from the control system of the 3D printing device to the driving device of the valve, causing the valve flap or valve core of the valve to move to the closed position, thereby blocking the conveying channel of the aggregate and stopping the current aggregate conveyance.

[0244] Step S602: Collect and obtain new aggregate characteristic data.

[0245] The general process is described as follows:

[0246] Suppose during the wall printing process, it is necessary to switch from ordinary aggregate to lightweight thermal insulation aggregate. After the valve for the currently conveyed ordinary aggregate is closed in step S601, step S602 is started.

[0247] First, the operator or the equipment control system determines that the new aggregate is lightweight thermal insulation aggregate and locates the corresponding storage container. Then, some basic characteristic data of the lightweight thermal insulation aggregate, such as a density of 300 kg / m³ and a particle size mainly distributed between 1 - 3 mm, are obtained from the product manual attached to the storage container.

[0248] Next, to ensure the accuracy and timeliness of the data, some laboratory tests may be carried out on the new aggregate. For example, a moisture meter is used to measure its moisture content, and it is found that the moisture content is 2%. The data obtained through different methods are sorted and summarized to obtain the complete new aggregate characteristic data. These data will be input into the subsequent analysis model for determining the valve opening / closing degree adjustment plan and predicting the conveying resistance.

[0249] Step S603: Match and obtain the valve opening / closing degree adjustment plan from the preset valve opening / closing degree adjustment plan analysis model according to the new aggregate characteristic data, and perform the valve opening / closing degree regulation.

[0250] Preset valve opening / closing degree adjustment plan analysis model: A mathematical model constructed based on a large amount of experimental data, theoretical analysis, and actual engineering experience. This model inputs the characteristic data of the new aggregate and outputs the matching valve opening / closing degree adjustment plan. Its purpose is to ensure that the new aggregate can be conveyed to the printing nozzle at an appropriate flow rate and velocity to meet the printing process requirements.

[0251] Valve opening / closing degree adjustment plan: Specifically stipulates the opening / closing degree to which the conveying valve needs to be adjusted to adapt to the characteristics of the new aggregate to ensure the smooth conveyance of the aggregate and a stable flow rate. For example, it may be stipulated that the valve needs to be opened to 70% of the opening degree, or the current opening degree needs to be reduced / increased by a certain proportion, etc.

[0252] Perform valve opening / closing degree regulation: According to the obtained valve opening / closing degree adjustment plan, send an instruction from the control system of the 3D printing device to the valve driving device to actually adjust the opening / closing degree of the valve so that the valve reaches the opening degree set by the plan.

[0253] Step S604: According to the new aggregate characteristic data, match and analyze in the aggregate conveying resistance analysis model to obtain the predicted conveying resistance.

[0254] Aggregate conveying resistance analysis model: Developed by professionals in the fields of mechanical engineering, materials science, etc. using professional software and algorithms. The model is stored in the control system of the 3D printing device in the form of a computer program, and its development process is based on a large amount of experimental data, theoretical research, and actual engineering cases. For example, use computational fluid dynamics (CFD) software to construct the model in combination with specific aggregate characteristic parameters and pipeline geometric parameters, and verify and optimize it multiple times.

[0255] The predicted conveying resistance: The result output after the model runs and calculates by inputting the new aggregate characteristic data into the aggregate conveying resistance analysis model. This result is also stored in the data storage area of the 3D printing device control system for use in the subsequent step S605.

[0256] Step S605: Make real-time adjustments according to the predicted conveying resistance, determine the additional power that the conveyor pump located in the conveying pipeline needs to provide, and make corresponding power adjustments to the conveyor pump located in the conveying pipeline.

[0257] The predicted conveying resistance: Calculated in step S604 by the aggregate conveying resistance analysis model according to the new aggregate characteristic data, which is the resistance value that the new aggregate is expected to face when flowing in the conveying pipeline. This value reflects the difficulty level of conveying the new aggregate.

[0258] Real-time adjustment: Immediately make an adaptive change to the conveyor pump power according to the predicted conveying resistance situation to ensure that the new aggregate can be smoothly and stably conveyed in the conveying pipeline to the printing nozzle position.

[0259] The additional power that the conveyor pump needs to provide: Due to the change in conveying resistance caused by the new aggregate characteristics, in order to overcome this part of the increased resistance, the additional power value that the conveyor pump needs to output to maintain the normal conveying flow rate and flow of the aggregate. The additional power that the conveyor pump needs to provide is obtained as follows: Calculate by combining the predicted conveying resistance value with the calculation formula for the relationship between the conveyor pump power and the conveying resistance established in advance. This calculation formula is determined based on the principles of fluid mechanics and the performance parameters of the conveyor pump, etc., and is stored in the program algorithm of the 3D printing device control system.

[0260] Conveyor pump: A device installed in the conveying pipeline that is responsible for providing power for the flow of the aggregate in the pipeline, and its power can be adjusted to adapt to different aggregate conveying requirements.

[0261] Power adjustment: Change the output power of the delivery pump to match the additional power demand caused by the change in the conveying resistance of the new aggregate, ensuring the efficiency and stability of aggregate conveyance.

[0262] General process description: Assume that in step S604, the conveying resistance of the composite phase change aggregate is predicted to be 1000 Pa. After the 3D printing equipment control system reads this conveying resistance value, it calculates the additional power required by the delivery pump according to a pre-set calculation formula. Assume that according to the formula calculation, to overcome the resistance of 1000 Pa, the delivery pump needs to provide an additional 500 W of power. The control system generates a power adjustment instruction and sends this instruction to the drive device of the delivery pump. After receiving the instruction, the drive device of the delivery pump adjusts the power of the delivery pump to increase the output power of the delivery pump by 500 W on the basis of the original power. Through such power adjustment, the delivery pump can provide sufficient power to overcome the resistance of the composite phase change aggregate in the conveying pipeline, ensuring that the aggregate is conveyed to the printing nozzle at a stable flow rate and flow volume, and guaranteeing the normal printing process of the thermally regulated wall.

[0263] Furthermore, the intelligent printing method for the thermally regulated wall further includes steps of real-time particle size monitoring, feedback, and collaborative optimization based on machine vision, as follows:

[0264] Step S70a: During printing, real-time collect the discharge image of the mixture of the new aggregate and the radiation-cooling enhanced cementitious material, use a convolutional neural network to extract multi-dimensional features such as the texture, shape, and spatial distribution of the aggregate, and combine a particle size prediction model that integrates multiple factors to analyze the particle size distribution and generate a heat map.

[0265] Specifically, the real-time collection of the discharge image involves the following: During the intelligent printing of the thermally regulated wall, use a high-speed high-definition camera to perform real-time image collection on the discharge situation of the mixture of the new aggregate and the radiation-cooling enhanced cementitious material at the outlet of the special nozzle.

[0266] Multi-dimensional feature extraction refers to using a convolutional neural network to perform multi-dimensional feature extraction on the collected discharge image. CNN is a deep learning model specifically used to process data with a grid structure (such as images). It automatically learns meaningful feature representations from images through components such as convolutional layers, pooling layers, and fully connected layers.

[0267] Texture feature: Texture is the spatial distribution pattern of pixel gray values in an image, reflecting information such as the roughness and fineness of the aggregate surface. In CNN, through convolution operations between convolution kernels of different scales and directions and the image, the texture features of the aggregate are extracted. For example, using convolution kernels with different receptive fields can capture texture information of different thicknesses.

[0268] Shape features: Shape features describe the geometric shape of the aggregate, such as roundness, aspect ratio, etc. By detecting and analyzing the contours of the aggregates in the image, the CNN can learn the shape features of the aggregates. Specifically, an edge detection algorithm (such as Canny edge detection) is used to first extract the edge contours of the aggregates, and then the contour information is input into the CNN for further feature extraction and analysis.

[0269] Spatial distribution relationship: The spatial distribution relationship reflects the relative positions and arrangement patterns of the aggregates in the mixed discharge. The CNN can extract the spatial distribution features of the aggregates by learning information such as the distances and angles between the aggregates in the image. To better capture the spatial distribution relationship, position encoding information is introduced into the input layer of the CNN to incorporate the position information of each pixel into the feature extraction process.

[0270] The particle size prediction model that fuses multiple factors is a deep learning-based regression model that fuses multi-dimensional features of the aggregates, printing parameters (such as agitator paddle speed, aggregate delivery volume, etc.), environmental factors (such as temperature, humidity), etc., to predict the particle size distribution of the aggregates.

[0271] Model construction: This model uses a multi-layer perceptron as the basic architecture, takes multi-dimensional features, printing parameters, and environmental factors as inputs, and through non-linear transformations of multiple fully connected layers, outputs the prediction results of the particle size distribution of the aggregates.

[0272] Data training: To train this model, a large amount of actual printing data and laboratory simulation data are collected, including discharge images, corresponding multi-dimensional features, printing parameters, environmental factors, and the actually measured particle size distribution of the aggregates. These data are divided into a training set, a validation set, and a test set according to a certain ratio, and optimization algorithms such as stochastic gradient descent are used to train and optimize the model until the model achieves satisfactory performance on the validation set.

[0273] Particle size distribution analysis and heat map generation involve the following: The extracted multi-dimensional features are input into the particle size prediction model that fuses multiple factors to obtain the prediction results of the particle size distribution of the aggregates. To visually display the particle size distribution, a particle size distribution heat map is generated according to the prediction results. The heat map uses the depth of color to represent the distribution density of aggregates with different particle sizes, and the darker the color, the denser the distribution of aggregates with that particle size. Through the particle size distribution heat map, the particle size distribution of the aggregates in the mixed discharge can be quickly and visually understood, providing an important basis for formulating subsequent adjustment strategies.

[0274] Step S70b, compare the analyzed particle size distribution with the preset standard in real time. If it exceeds the range, combine multi-source data such as printing progress, environment, and wall thermal performance, and use a fuzzy decision tree and reinforcement learning algorithm to determine the optimal adjustment strategy.

[0275] Among them, the preset standards are determined according to the design requirements, material properties of the heat-regulating wall, and relevant industry standards, covering key indicators such as the proportion of aggregates in different particle size ranges and the average particle size.

[0276] The steps for comprehensively determining the optimal strategy are as follows:

[0277] Step S70b1: The fuzzy decision tree gives a recommended degree value of 0-1 for each adjustment action. For example, the recommended degree for reducing the stirring paddle speed is 0.8, and the recommended degree for increasing the aggregate delivery volume is 0.6. When initializing the reinforcement learning, set the fusion coefficient β (usually taken as 0.6). The recommended degree for the adjustment of the stirring paddle speed by the fuzzy decision tree is R n , then the initial selection probability P n = β×R n , for example, when R n = 0.8 and β = 0.6, P n = 0.48. Similarly, determine the initial selection probabilities of actions such as the aggregate delivery volume. In reinforcement learning, the agent updates the action value according to the Q-learning algorithm based on the reward feedback (learning rate 0.1, discount factor 0.9). Positive rewards (such as reducing the stirring paddle speed to make the particle size more standard gets 5 points) or negative rewards (such as increasing the aggregate delivery volume resulting in a progress delay gets -3 points) will both prompt the update of the action value, optimize the action selection probability, and refine the initial strategy.

[0278] Step S70b2 is as follows: Regarding the aggregate particle size, calculate the sum of the absolute values of the differences in the proportion of each interval between the actual and the preset particle size distributions. The smaller the sum, the closer the particle size distribution is to the standard. Printing progress: Measured by the ratio of the actual to the expected printing progress. The closer the ratio is to 1, the more in line with the expectation the progress is. Environmental conditions: Fit the functional relationship between the temperature and humidity (T, H) and the change value M of the material properties through experiments. The smaller M is, the smaller the impact of the environment on the material properties.

[0279] Regarding the thermal performance of the wall, it is quantified by the absolute value of the difference between the actual and the designed thermal conductivity. The smaller the difference, the closer the thermal performance is to the design requirements. Determine the weights of each factor according to the printing requirements. For example, for the thermal performance of the heavy wall, its weight is set to 0.4; for the printing progress, the weight is set to 0.3; the weight for the adjustment of the aggregate particle size is 0.2; the weight for the environmental conditions is 0.1, and the sum of the weights is 1. For each strategy combination (different value combinations of parameters such as the stirring paddle speed and the aggregate delivery volume), calculate the sum after operating the weights of each factor with the corresponding quantified evaluation values according to the rules, obtain the comprehensive influence index, and compare the index values to weigh the adjustment effect.

[0280] Step S70b3: Select the strategy combination with the largest comprehensive influence index value as the candidate. Verify the feasibility of the candidate strategy to ensure that the adjustment value of the stirring paddle speed is within the rated speed range of the equipment, the adjustment value of the aggregate conveying volume is within the capacity range of the conveying equipment, the adjustment value of the valve opening degree is within the adjustable angle range of the valve, and the adjustment value of the conveying pipeline pressure is within the pressure range that the pipeline and the conveying pump can withstand. If the candidate strategy is feasible, determine it as the optimal strategy, and clearly give the adjustment values or ranges of parameters such as the stirring paddle speed (revolutions per minute), aggregate conveying volume (kilograms per minute), valve opening degree (degrees), and conveying pipeline pressure (Pascals) to guide the printing work.

[0281] Step S70c, according to the optimal adjustment strategy and combined with the multi-parameter coupling model, coordinately adjust the stirring paddle speed, aggregate conveying volume, valve opening degree, and pipeline pressure.

[0282] The public content of Step S70c is as follows:

[0283] Step S70c1, use deep learning to build a multi-parameter coupling dynamic model, collect printing parameters such as stirring paddle speed, aggregate conveying volume, valve opening degree, and pipeline pressure, as well as data such as wall thermal performance, structural strength, and material mixing effect for training. The model can accurately master the complex non-linear coupling relationship between parameters. Compared with traditional models, it can capture high-order coupling effects and achieve more accurate multi-parameter coordinated adjustment prediction.

[0284] Step S70c2, after receiving the optimal adjustment strategy instruction, input the target parameter values (such as the desired stirring paddle speed, aggregate conveying volume, etc.) into the multi-parameter coupling model. The model accurately calculates the specific values of the coordinated adjustment of each parameter according to the learned coupling relationship. For example, to improve the wall thermal performance, the model may give a plan to reduce the stirring paddle speed, increase the aggregate conveying volume, and synchronously adjust the valve opening degree and pipeline pressure, and the adjustment amounts are all calculated based on the coupling relationship, rather than relying on experience.

[0285] Step S70c3, during the adjustment process, use high-precision sensors to real-time monitor key parameters such as stirring paddle speed, aggregate conveying volume, valve opening degree, pipeline pressure, as well as wall performance and material mixing state. The sensor data is real-time transmitted to the data processing system and compared with the model prediction results. Once the actual parameters deviate from the predicted values, the system feeds back the deviation information to the model. The model uses the online learning algorithm to fine-tune its own parameters and structure according to the new data, recalculate a more accurate coordinated adjustment plan, and achieve dynamic optimization.

[0286] Step S70c4: The multi-parameter coupling model combines real-time and historical monitoring data and uses machine learning prediction algorithms to predict future printing conditions. For example, based on the current changes in environmental temperature and humidity and the printing progress, it predicts the changes in material properties and the requirements for parameter adjustment. If it is predicted that the increase in environmental temperature leads to an increase in the bonding property of the aggregate, affecting the mixing effect and particle size distribution, the model calculates in advance the values of the required increase in the stirring paddle speed, adjustment of the valve opening degree, and pipeline pressure, and takes preventive measures in advance to ensure stable printing and reliable wall quality.

[0287] An embodiment of the present application also provides a thermally regulated wall. The thermally regulated wall is a layered structure, which from the outside to the inside is successively an outer layer, a middle layer, and an inner layer; each layer plays a different function to achieve good thermal regulation performance.

[0288] Outer layer: It is composed of a radiation cooling enhanced cement-based material mixed with polystyrene particle aggregate. Among them, the radiation cooling enhanced cement-based material plays a key role. It is composed of white cement and a radiation cooling material. The radiation cooling material can be one or more of barium sulfate, titanium dioxide, and aluminum oxide. White cement itself has a relatively high solar radiation reflectivity, usually between 0.7 and 0.8, which can reflect a large amount of solar radiation and reduce the entry of external heat into the room. The addition of radiation cooling materials such as barium sulfate, titanium dioxide, and aluminum oxide significantly increases the emissivity of the material, enabling it to dissipate the heat absorbed by the wall to the external environment through infrared radiation when there is no solar radiation, enhancing the radiation cooling effect. The polystyrene particle aggregate, with its low thermal conductivity characteristics, effectively reduces heat conduction, prevents the loss of indoor temperature and the entry of external temperature, acts as a heat insulation layer in the wall, and maintains the stability of the indoor temperature. For example, in tropical regions, the thickness of the outer layer is 10 cm, and this structure can effectively block the external high temperature and reduce the demand for air conditioning.

[0289] Middle layer: It is composed of a radiation cooling enhanced cement-based material mixed with 35°C phase change aggregate. The radiation cooling enhanced cement-based material provides the structural strength and a certain radiation cooling ability for the wall. The 35°C phase change aggregate has a special thermal regulation function. When the external temperature rises and approaches or exceeds 35°C, the phase change material absorbs heat and undergoes a phase change, changing from a solid state to a liquid state, storing heat, and slowing down the transfer of heat to the inner layer or the room, preventing the rapid penetration of external high temperature. In hot summer weather, the middle layer can effectively relieve heat accumulation, reduce the direct conduction of high-temperature air to the inner layer, and help the wall maintain a relatively stable temperature during the day.

[0290] Inner layer: It is composed of a radiation cooling enhanced cement-based material mixed with 25°C phase change aggregates. While the radiation cooling enhanced cement-based material safeguards the wall structure, it also participates in certain thermal regulation. The 25°C phase change aggregates are related to the most comfortable temperature for the human body. When the wall temperature drops and approaches or is lower than 25°C, the phase change material changes from a liquid state to a solid state, releasing the stored heat, helping the wall maintain warmth, and slowing down the impact of cold air on the interior. In winter, the inner layer can effectively prevent the indoor temperature from being too low, especially at night or in cold seasons, ensuring that the interior remains within a comfortable temperature range.

[0291] The embodiments of this specific implementation manner are all preferred embodiments of this application, and do not limit the protection scope of this application accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A thermal temperature control wall intelligent printing method, characterized in that: include: The 3D printing device is started, and according to the preset printing path, the mixed material of the radiation cooling enhanced cement-based material and the aggregate is extruded and printed layer by layer through a special nozzle according to the preliminarily set printing parameters of each layer, wherein the special nozzle is a printing nozzle with two independent inlets, one of which is connected to the radiation cooling enhanced cement-based material, and the other is connected to the aggregate, and a stirring paddle is arranged inside the special nozzle; During the printing process, a sensor matrix is ​​used to collect multi-source data, including environmental data, printing progress data, and wall thermal performance related data; Collect historical climate data of the printing area, and perform data preprocessing on the historical climate data and the collected multi-source data; Input the multi-source data and historical climate data that have completed data preprocessing into the pre-built multi-objective optimization model, and output the optimal parameter combination for printing each layer of the wall, and replace the initially set printing parameters of each layer according to the optimal parameter combination for printing each layer of the wall; During the printing process, based on the multi-source data collected in real time and pre-processed, it is determined whether the pre-set trigger conditions for aggregate switching are met; If yes, an aggregate switching instruction is issued to control the aggregate channel of the special nozzle to switch. During the aggregate switching process, the valve opening and closing degree and the conveying pipeline pressure are regulated by a preset control method to allow new aggregate to enter the printing nozzle. During the aggregate switching process, an adaptive fuzzy PID algorithm is used to synchronously adjust the stirring blade speed and the aggregate conveying amount, so that the aggregate particle size distribution after the new aggregate and the radiation cooling enhanced cement-based material are mixed is within the preset aggregate particle size distribution standard range; If not, keep the original setting.

2. The intelligent printing method for thermal temperature control wall according to claim 1, characterized in that: It also includes the following steps after preprocessing the historical climate data and the collected multi-source data: The multi-source data and historical climate data that have completed data preprocessing are unified in data format and standardized to form various data vectors; Input various data vectors into the preset fuzzy membership function, calculate the membership of various data vectors in different fuzzy sets, and perform weighted summation of multiple fuzzy membership functions to obtain the comprehensive membership; Analyze and calculate whether the comprehensive membership degree calculated exceeds the trigger threshold, where the calculation formula of the trigger threshold is as follows: ,in, is the trigger threshold, is the adjustment coefficient, which is used to adjust the weight of the comprehensive membership mean when calculating the trigger threshold. is the mean of the comprehensive membership, is the adjustment coefficient, which is used to adjust the weight of the comprehensive membership standard deviation when calculating the trigger threshold. is the standard deviation of comprehensive membership; If no, continue with the next steps; If yes, then based on the multi-source data that has completed data preprocessing and historical climate data, key state features are extracted, and principal component analysis is used to reduce the dimension to obtain the principal component feature vector representing the environment and equipment status, where the key state features include the environment change rate, printing progress deviation, and wall thermal performance deviation; Using a fuzzy logic system, based on the principal component eigenvector and preset rules, the adjustment coefficient of each performance target weight is calculated to update the weight of the multi-objective optimization model, where the performance targets include thermal regulation performance, structural stability, and material cost; Using the trained deep Q network, different principal component feature vectors are input to obtain corresponding printing parameter combinations, and this operation is continued until a preset number of printing parameter combinations under different states are generated; According to the updated multi-objective optimization model weights, the performance of each printing parameter combination generated is evaluated to obtain the scores of each performance target, and the scores of each performance target are weighted and summed according to the adjusted weights to obtain the comprehensive score of each parameter combination; The parameter combinations of all groups are sorted according to the comprehensive scores, and the parameter combination with the highest comprehensive score is selected as the optimal parameter combination for printing each layer of the wall. The initially set printing parameters of each layer are replaced according to the optimal parameter combination for printing each layer of the wall.

3. The intelligent printing method for thermal temperature control wall according to claim 2, characterized in that: The steps after obtaining the comprehensive score of each parameter combination are also included, which are as follows: Analyze whether the relative deviation between the comprehensive membership and the trigger threshold is always within the preset deviation range within a preset number of continuous monitoring cycles; If yes, continue with the next steps; If not, then select the parameter combination with the highest comprehensive score ranking as the candidate parameter combination from all the generated parameter combinations; Based on the preliminary candidate parameter combinations for printing each layer of the wall, determine the basic parameters related to wall printing, run a random algorithm to generate random parameters within the floating range of basic parameters related to different types of wall printing, and integrate the random parameters and input them into the multi-physics field coupling digital twin model; For each candidate parameter combination for printing each layer of the wall, the digital twin model is run a preset number of times in each set scenario. During each run, the newly generated random parameters are called, and the wall performance-related data output by each simulation is recorded; According to the wall performance related data output by simulation, the preset performance index evaluation scoring method is adopted to obtain the performance evaluation scores of different categories; Based on the obtained performance evaluation scores of different categories and the weight allocation schemes matched under different simulation conditions, the average comprehensive scores of different candidate parameter combinations under all simulation conditions are calculated; Calculate the coefficient of variation of the comprehensive score of each candidate parameter combination under different simulation conditions; The parameter combinations whose comprehensive scores under all simulation conditions are ranked before the preset positions and whose coefficient of variation is less than the preset value are selected as the optimal parameter combinations for printing each layer of the wall, and the initially set printing parameters for each layer are replaced according to the optimal parameter combinations for printing each layer of the wall.

4. The intelligent printing method for thermal temperature control wall according to claim 3, characterized in that: Determining whether the preset triggering conditions for aggregate switching are met includes: Extract the key features of environment, printing progress, and wall thermal performance, and use them as matrix elements to construct a correlation matrix , where environmental data has m key features, printing progress has n key features, and wall thermal performance has p key features; At each monitoring moment, the state vector is composed according to the key characteristic states ,in, , is the status value of the i-th key feature. When the feature meets the trigger condition, the value is 1, and when it does not meet the trigger condition, the value is 0; The formula for multi-condition association score is as follows: ； in, is the multi-condition association score, is the state vector The transpose of , which converts a row vector into a column vector for matrix multiplication, is the incidence matrix The element in the i-th row and j-th column in represents the strength of association between the i-th feature and the j-th feature; Based on multiple condition correlation scores The comparison result with the associated score threshold determines whether the preset triggering condition for aggregate switching is met.

5. The intelligent printing method for thermal temperature control wall according to claim 4, characterized in that: Determining whether the preset triggering conditions for aggregate switching are met also includes building a correlation matrix The previous steps are as follows: Analyze whether the relative deviation between the comprehensive membership and the trigger threshold is always within the preset deviation range within a preset number of continuous monitoring cycles; If yes, continue to execute the original steps; If not, then according to the multi-source data that has completed data preprocessing, a preset cluster analysis method is used to determine the parameter combination category, wherein the parameter combination category includes a thermal insulation priority type wall and a structure priority type wall; According to the mapping relationship between the parameter combination category and the key feature, the key feature matching the determined parameter combination category is determined, and the determined parameter combination category and the key feature matching therewith are input into the trigger condition judgment model constructed based on the decision tree algorithm, and the trigger condition is output as the pre-set trigger condition for aggregate switching; Based on the multi-source data collected in real time and pre-processed, it is determined whether the pre-set triggering conditions for aggregate switching are met.

6. The intelligent printing method for thermal temperature control wall according to claim 1, characterized in that: The method also includes steps before determining whether a preset trigger condition for aggregate switching is met, which are as follows: Determine the thermal performance requirements of the wall according to the different functions of the building, and determine the quantitative indicators of the thermal performance requirements of the wall based on the preset energy-saving standards; Preprocess the real-time environmental data and the determined quantitative value of the wall thermal performance requirement; Match the collected real-time environmental data, the preset quantitative value of the wall thermal performance requirements, and the set aggregate type selection criteria to screen out the aggregate types that meet the conditions; Whether to execute the subsequent steps is determined based on the comparison results between the screened aggregate category and the currently used aggregate category.

7. The intelligent printing method for thermal temperature control wall according to claim 6, characterized in that: The analysis process of the triggering conditions for aggregate switching is as follows: According to the selected aggregate categories, determine the characteristics of the corresponding aggregate categories; According to the mapping relationship between the characteristics of the aggregate category and the trigger condition setting information of the aggregate switching, the trigger condition setting information of the aggregate switching is analyzed and determined.

8. The method for intelligent printing of thermal temperature-controlled walls according to claim 1, characterized in that: The valve opening and closing degree and pipeline pressure are regulated by the preset control methods, including: Control the valve that is currently conveying aggregate to close; Collect and obtain new aggregate property data; According to the new aggregate characteristic data, the valve opening and closing adjustment scheme is matched and obtained from the preset valve opening and closing adjustment scheme analysis model, and the valve opening and closing adjustment is performed; According to the new aggregate characteristic data, the predicted conveying resistance is obtained by matching and analyzing the aggregate conveying resistance analysis model; According to the predicted conveying resistance, real-time adjustment is made to determine the additional power that the conveying pump located in the conveying pipeline needs to provide, and the conveying pump located in the conveying pipeline is adjusted accordingly.

9. A thermal temperature control wall, characterized in that: The method of any one of claims 1 to 8 is used for printing, specifically as follows: The thermal temperature control wall is a layered structure, which consists of an outer layer, a middle layer and an inner layer from the outside to the inside; the outer layer is a mixture of radiation cooling enhanced cement-based materials and polystyrene particle aggregates, the middle layer is a mixture of radiation cooling enhanced cement-based materials and 35°C phase change aggregates, and the inner layer is a mixture of radiation cooling enhanced cement-based materials and 25°C phase change aggregates.

10. The thermal temperature control wall according to claim 9, characterized in that: The radiation cooling enhanced cement-based material consists of white cement and radiation cooling material, and the radiation cooling material is one or more of barium sulfate, titanium dioxide and aluminum oxide.

Images