Pre-cooling concrete production internet of things temperature control method and system

Abstract

The application discloses a pre-cooling concrete production Internet of Things temperature control system and method, which comprises a region demarcation module, a temperature control module, an environment data acquisition module, a data analysis and reasoning module, a device control module and a visual display module, and the modules cooperate with each other, realizes accurate temperature control of different pre-cooling concrete production regions, and the system demarcates different temperature control regions by using laser radar point cloud and hyperspectral image imaging technology, and fiber grating temperature sensing network array, model prediction MPC controller and magnetic suspension variable frequency refrigeration components ensure accurate temperature control, and meanwhile, the data analysis and reasoning module combines a deep learning time sequence prediction model and a reasoning rule mining algorithm to provide support for temperature control of production equipment; the overall temperature control scheme improves the intelligentization and accuracy of the pre-cooling concrete production Internet of Things temperature control, and is suitable for various pre-cooling concrete production scenes.

Description

Technical Field

[0001] This invention relates to the field of intelligent temperature control technology for pre-cooled concrete, and more specifically, to an IoT-based temperature control method and system for pre-cooled concrete production. Background Technology

[0002] With the continuous development of the construction industry, concrete has been widely used in various construction projects. However, temperature control has always been one of the key factors affecting concrete quality during the concrete production process. Especially for large-volume concrete, the temperature stress and shrinkage stress generated by the internal cement hydration heat may cause cracks in the concrete structure, thereby affecting the durability and safety of the structure. Traditional concrete temperature control methods mainly rely on manual monitoring and simple cooling water circulation systems, which have problems such as data acquisition lag, inaccurate monitoring data, and low temperature control efficiency.

[0003] In recent years, the rapid development of IoT technology has provided new solutions for concrete temperature control. IoT technology enables interconnection between devices, allowing sensors to collect temperature data in real time during the concrete production process, and using network transmission and data analysis technologies to control the concrete temperature.

[0004] However, most existing temperature control systems focus on temperature control throughout the entire production process. Temperature control throughout concrete production is extremely complex, making precise temperature control across the entire process difficult. Less attention is paid to temperature control during the pouring and curing stages of concrete production. During pouring and curing, the layout of the poured structure, environmental factors, and the settings of the pouring and curing equipment parameters all significantly impact the final quality of the concrete. Therefore, developing an IoT-based temperature control method and system for pre-cooled concrete production, capable of monitoring and controlling the temperature during the pouring and curing processes, is of great significance for improving concrete quality. Summary of the Invention

[0005] The purpose of this invention is to provide an IoT-based temperature control method and system for pre-cooled concrete production, in order to solve the aforementioned problems existing in the prior art.

[0006] The application is as follows:

[0007] An IoT-based temperature control system for pre-cooled concrete production includes:

[0008] The area delineation module is used to automatically delineate multiple separate pre-cooled concrete production temperature control zones based on the spatial layout of the cast-in-place structure. The area delineation module outputs the delineation information of each zone to the temperature control module, the environmental data acquisition module, and the data analysis and reasoning module.

[0009] The temperature control module, which is linked with the area delineation module, is used to control the temperature of each pre-cooled concrete production temperature control area, feeds the temperature back to the data analysis and inference module in real time, and receives the adjustment suggestions given by the data analysis and inference module in real time.

[0010] The environmental data acquisition module is used to monitor the environmental data of each temperature-controlled area in the pre-cooled concrete production process, including humidity and light intensity, and transmit the environmental data to the data analysis and inference module in real time.

[0011] The data analysis and reasoning module analyzes and reasons the received area delineation results, temperature and environmental data. It analyzes the temperature, characteristics of pre-cooled concrete and environmental data of each pre-cooled concrete production temperature control area, and then further reasons about the aforementioned data based on the data reasoning network model, and outputs control commands to the equipment control module and temperature control module.

[0012] The equipment control module dynamically adjusts the operating parameters of the pre-cooled concrete production equipment according to the control instructions of the data analysis and reasoning module, and pushes the operating parameter information of the pre-cooled concrete production equipment to the visualization display module. The pre-cooled concrete production equipment includes concrete pouring equipment and concrete curing equipment.

[0013] The visualization display module is used to display the temperature of each of the pre-cooled concrete production temperature control zones, the environmental data, and the operating parameter information of the pre-cooled concrete production equipment. It also allows the operator to manually set the temperature parameters and transmits the operator's manual setting instructions to the temperature control module.

[0014] Furthermore, the region delineation module includes:

[0015] The casting body spatial acquisition unit uses lidar point cloud and hyperspectral image imaging technology to acquire spatial distribution information of the casting body structure produced by pre-cooled concrete and transmits the acquired spatial distribution information to the area delineation decision unit.

[0016] The regional delineation decision unit has a built-in partitioning decision model based on particle swarm optimization algorithm and ant colony optimization algorithm. It is used to delineate the cast-in-place structure of pre-cooled concrete production into multiple regions according to the spatial distribution information, and output a regional delineation scheme that meets the needs of different pre-cooled concrete production scenarios. The multiple regions are uniformly named pre-cooled concrete production areas. Different pre-cooled concrete production scenarios include at least large-volume concrete, precast components, and core tubes of super high-rise buildings.

[0017] The casting body space acquisition unit further includes:

[0018] The LiDAR scanning subunit is equipped with multiple sets of adjustable azimuth and precision LiDARs, which are distributed in a matrix at key points in the pre-cooled concrete production area to scan the features of the internal space of the pre-cooled concrete production area and generate point cloud data.

[0019] The hyperspectral image acquisition subunit uses a hyperspectral camera with visible light to short-wave infrared imaging capabilities to capture concrete color and crack information on the surface of the pre-cooled concrete production area, and then uses an image fusion algorithm to match the point cloud data to output complete spatial distribution information of the pre-cooled concrete production area to the area delineation decision unit.

[0020] The regional delineation decision-making unit includes:

[0021] The initial scheme determines the sub-units, and has a built-in expert database of pre-cooled concrete production scenario cases. Based on the spatial distribution information, the type of concrete raw materials, and the information on the intended use of the concrete after production, the pre-cooled concrete production scenario is predicted. The corresponding sub-regional foundation templates are retrieved to form the initial area delineation scheme. The intended use information includes at least the distance information from the concrete production site to the construction site.

[0022] The algorithm optimization execution subunit uses the particle swarm optimization algorithm for fast convergence in high-dimensional parameter space and the discrete partition boundary optimization strategy of the ant colony optimization algorithm to iteratively optimize the initial region delineation scheme.

[0023] The specific implementation process of the algorithm optimization execution subunit is as follows:

[0024] The particle swarm optimization algorithm is used to optimize the partitioning scheme of the temperature control zone for pre-cooled concrete production in a high-dimensional parameter space. The velocity update formula satisfies:

[0025]

[0026] in, This represents the current velocity, reflecting the particle's search direction and step size in the parameter space. (pbest) id Gbest represents the individual best position, reflecting the particle's own historical best position. id Representing the global optimum, reflecting the optimal position found by all particles in the population, c1 and c2 represent learning factors, r1 and r2 represent random numbers, and k represents the current iteration number. max Indicates the maximum number of iterations the algorithm can run;

[0027] Combining ant colony optimization algorithm, topology optimization is performed on discrete partition boundaries, where the pheromone concentration update rule is as follows:

[0028]

[0029]

[0030] Where, π ij T represents the pheromone concentration, ρ represents the pheromone evaporation coefficient, and T represents the pheromone concentration. i T j L represents the temperature values ​​at nodes i and j, respectively. k1 Let Q represent the path length of the k1-th ant, where Q is a constant and m represents the number of ants.

[0031] Furthermore, the temperature control module includes:

[0032] The temperature data acquisition unit uses a fiber optic temperature sensor network array to measure the temperature of each of the pre-cooled concrete production temperature control zones and report the temperature data to the temperature control unit.

[0033] The temperature control unit integrates a model prediction MPC controller and a magnetic levitation frequency conversion refrigeration component, and is used to adjust the temperature of each of the pre-cooled concrete production temperature control zones based on the measurement results of the temperature data acquisition unit.

[0034] Furthermore, the temperature data acquisition unit includes:

[0035] The fiber optic temperature sensor network array layout subunit uses a dynamic deployment algorithm based on geometric features. It calculates the first sensor density of all nodes in the pre-cooled concrete production temperature control area based on the volumetric characteristics, shape complexity index, and historical temperature fluctuation characteristics of the temperature control area. Through finite element thermal conduction simulation, areas with temperature gradients greater than a fourth threshold are designated as sensitive zones, and nodes in these sensitive zones are represented as sensitive nodes. A second sensor density is generated by arranging a bidirectional staggered array at the nodes at the boundaries of sensitive nodes and adjacent temperature control areas. If a pre-cooled concrete production temperature control area is detected to exceed the fifth threshold three times consecutively, a moving sensor is automatically added to that area, generating a third sensor density.

[0036] The data processing and transmission subunit integrates Kalman adaptive filtering algorithm and generative adversarial network to preprocess the field-collected data, remove outliers, reduce noise in the data, and transmit the preprocessed temperature data to the temperature control unit.

[0037] The temperature control unit includes:

[0038] The model predicts the MPC control subunit, which takes the temperature deviation and its changing trend as input, and dynamically generates the optimal control sequence through a rolling time-domain optimization algorithm, and outputs the control signal to drive the magnetic levitation variable frequency cooling component.

[0039] The magnetic levitation variable frequency refrigeration execution subunit is equipped with a magnetic levitation centrifugal compressor, a digital pulse electronic expansion valve, and a biomimetic airflow guiding duct mechanism. Based on the model prediction of the control signal of the MPC control subunit, it dynamically adjusts the cooling capacity, cold air speed, and refrigerant flow.

[0040] Furthermore, the environmental data acquisition module includes:

[0041] The humidity data acquisition and processing unit is equipped with a distributed fiber optic acoustic sensor to measure the humidity data of each of the pre-cooled concrete production temperature control zones and to preprocess the humidity data error in real time.

[0042] The light intensity acquisition and processing unit uses a quantum dot spectral sensor to measure the light intensity of each of the pre-cooled concrete production temperature control areas, constructs a real-time light intensity field model within the pre-cooled concrete production temperature control area, and updates it dynamically.

[0043] Furthermore, the humidity data acquisition and processing unit also includes:

[0044] The distributed fiber optic acoustic wave sensing subunit monitors the acoustic vibration signal in the temperature control area of ​​pre-cooled concrete production in real time, and inputs the acoustic vibration signal into the hydration reaction acoustic feature library to invert the humidity change.

[0045] The data correction subunit sets the start and end points of the time period for correction, and combines a multi-parameter decoupling algorithm to eliminate the cross-sensitivity of temperature and humidity, and outputs accurate humidity values ​​to the data analysis and inference module.

[0046] The light intensity acquisition and processing unit further includes:

[0047] The quantum dot spectral sensor subunit is used to output the band-wise irradiance data of the temperature control zone in the pre-cooled concrete production process.

[0048] The geometric correction module, based on astronomical algorithms, eliminates errors in band irradiance data caused by zenith angle.

[0049] The illumination intensity modeling and analysis module calculates the regional light field and generates light field data based on the irradiance data of different wavelengths and the Monte Carlo ray tracing algorithm. According to the building BIM model and the light trajectory, it predicts the shadow movement path, marks the data of the light intensity dead zone area, and sends the light field data and the light intensity dead zone area data to the data analysis and inference module.

[0050] Furthermore, the data analysis and reasoning module includes:

[0051] The pre-cooled concrete characteristic analysis unit integrates a big data pre-cooled concrete characteristic knowledge base to analyze the pouring and curing requirements of pre-cooled concrete and generate pouring and curing strategies for pre-cooled concrete characteristics.

[0052] The integrated data analysis unit uses a deep learning time series prediction model and inference rule mining algorithm to analyze the impact of temperature and environmental data on pouring and curing, predict temperature and environmental change trends and associated risks in advance, generate control suggestions, and provide the control suggestions to the temperature control module and the equipment control module.

[0053] The fusion data analysis unit includes:

[0054] The deep learning time series prediction subunit mines the patterns of historical temperature, humidity and light intensity time series data based on an improved radial basis function interpolation algorithm and an improved U-Net++ network model, predicts the future trend of temperature and environmental data, and generates trend prediction data.

[0055] The rule reasoning subunit inputs the predicted trend data of future temperature and environmental data into the data reasoning network model to obtain the reasoning result. The reasoning result is used to output early warning information and control suggestions to the equipment control module and temperature control module. The reasoning result is used to quantify the degree of impact of temperature and environmental data changes on the quality of pouring and curing.

[0056] The specific implementation process of the deep learning time series prediction subunit is as follows:

[0057] Multi-source data fusion is performed by acquiring discrete point temperatures and humidity within the temperature-controlled area of ​​the pre-cooled concrete production zone through the temperature control module and the environmental data acquisition module, generating discrete data; and acquiring surface temperature field, humidity field, and light intensity field of the temperature-controlled area of ​​the pre-cooled concrete production zone through the temperature control module and the environmental data acquisition module, generating surface data. Based on an improved radial basis function interpolation algorithm, the discrete data and surface data are fused to construct a three-dimensional temperature distribution voxel model.

[0058] Intelligent segmentation and feature extraction of temperature field: The three-dimensional temperature voxel model is sliced ​​into XY, XZ, and YZ planes. The sliced ​​data is input into an improved U-Net++ network. The output of the improved U-Net++ network includes a binary mask of isothermal surface, semantic labels of abnormal high temperature region, and temperature gradient field tensor. The encoder of the improved U-Net++ network adopts ResNet34 skeleton and embeds coordinate attention module. The decoder of the improved U-Net++ network introduces dense skip connections and uses subpixel convolution.

[0059] Feature extraction by CP decomposition of the temperature gradient field tensor generates a spatiotemporal feature vector set {a r b r c r d r}, where a r Characterizing the spatial distribution pattern of the temperature field in the X direction, reflecting the temperature gradient of the concrete structure along the pouring length; br Characterizes the spatial distribution pattern of the temperature field in the Y direction, reflecting the interlayer temperature transfer characteristics during layered casting; c r Characterizing the temperature field distribution pattern in the Z direction, reflecting the attenuation relationship between the internal core temperature and the surface temperature; d r The weighting coefficients characterize the temperature field over time, reflecting the influence of environmental temperature changes and time-varying factors such as heat of hydration release on the overall temperature field; the CP decomposition represents a factorization method for a higher-order tensor.

[0060] A physical information neural network (PINN) is constructed to output trend prediction data for the next N minutes. The input of the PINN is a set of spatiotemporal feature vectors and environmental data. The trend prediction data includes temperature field evolution prediction data and confidence interval data. The environmental data includes ambient temperature, ambient humidity and ambient light intensity.

[0061] The rule-based reasoning subunit inputs the predicted trend data of future temperature and environmental data into the data reasoning network model to obtain the following reasoning results:

[0062] Based on the first-layer perceptron and the core multilayer perceptron in the data inference network model, the representation vector is extracted from the future temperature change trend prediction data to obtain the first temperature vector and the first temperature base vector. The representation vector extraction includes the representation vectors whose correlation is not less than a preset first correlation threshold.

[0063] Based on the second-layer perceptron and the core multilayer perceptron in the data inference network model, the representation vector is extracted from the future environmental data change trend prediction data to obtain the first environmental vector and the first environmental base vector. The representation vector extraction includes the representation vectors whose correlation is not less than a preset second correlation threshold.

[0064] The first weight calculation unit in the data reasoning network model performs basic vector correction on the first temperature basic vector and the first environment basic vector respectively, so as to fuse the corrected basic vector with the first temperature vector and the first environment vector respectively to obtain the first fused vector. Finally, the reasoning unit in the data reasoning network model performs data reasoning on the first fused vector to obtain the reasoning result. The reasoning result includes at least a general reasoning result and an emergency risk reasoning result. The general reasoning result represents non-emergency handling matters, and the emergency risk reasoning result represents emergency handling matters.

[0065] Furthermore, the device control module includes:

[0066] A general algorithm processing unit is embedded with a deep reinforcement learning model. The deep reinforcement learning model is used to dynamically adjust the operating parameters of the pre-cooled concrete production equipment according to the general inference results, and push the operating parameters and energy consumption parameters to the visualization display module.

[0067] The emergency handling unit, based on the emergency risk reasoning result, adjusts the heating mode of the pre-cooling concrete production equipment to increase the heating power and accelerate the concrete forming speed when the concrete surface temperature is lower than the first emergency threshold for concrete pouring and curing. When the emergency risk reasoning result indicates that the concrete surface temperature exceeds the second emergency threshold for concrete pouring and curing, the unit activates the cooling mode of the pre-cooling concrete production equipment to increase the cooling power, thereby rapidly reducing the concrete surface temperature and preventing cracks from forming due to overheating.

[0068] The energy consumption equipment operation optimization unit increases the fan speed of the pre-cooling concrete production equipment to promote cooling when the surface temperature of the concrete exceeds the optimal temperature threshold for concrete pouring; conversely, it decreases the fan speed of the pre-cooling concrete production equipment to promote heat retention when the surface temperature of the concrete is below the optimal temperature threshold for concrete pouring. The formula used is as follows:

[0069]

[0070] If(T cur (t)andT inner (t)>T set ), thus obtaining F fan rise,

[0071] If(T cur (t)andT inner (t) <T set ), thus obtaining F fan decline,

[0072] Among them, F fan F represents the adjusted energy consumption operating rate of the pre-cooled concrete production equipment. base Based on the wind turbine speed, b f T is the fan speed coefficient. cur (t) represents the average outer surface temperature of the temperature-controlled area for pre-cooled concrete production at time t, T inner (t) represents the average inner surface temperature of the pre-cooled concrete production temperature control zone at time t, T set This indicates the temperature optimization threshold.

[0073] A method for IoT-based temperature control in pre-cooled concrete production, used to implement any of the IoT-based temperature control systems for pre-cooled concrete production, includes the following steps:

[0074] S1. The area delineation module automatically delineates multiple separate pre-cooled concrete production temperature control zones based on the spatial layout of the cast-in-place structure. The area delineation module outputs the delineation information of each zone to the temperature control module, the environmental data acquisition module, and the data analysis and reasoning module.

[0075] S2. Use the temperature control module to work in conjunction with the area delineation module to control the temperature of each pre-cooled concrete production temperature control area, feed the temperature back to the data analysis and inference module in real time, and receive the control suggestions given by the data analysis and inference module in real time.

[0076] S3. Use the environmental data acquisition module to monitor the environmental data of each pre-cooled concrete production temperature control area, including humidity and light intensity, and transmit the environmental data to the data analysis and inference module in real time.

[0077] S4. The data analysis and reasoning module is used to analyze and reason about the received area delineation results, temperature and environmental data. The temperature, characteristics of pre-cooled concrete and environmental data of each pre-cooled concrete production temperature control area are analyzed. Then, the data is further reasoned based on the data reasoning network model, and control commands are output to the equipment control module and temperature control module.

[0078] S5. The equipment control module dynamically adjusts the operating parameters of the pre-cooled concrete production equipment according to the control instructions of the data analysis and reasoning module, and pushes the operating parameter information of the pre-cooled concrete production equipment to the visualization display module. The pre-cooled concrete production equipment includes concrete pouring production equipment and concrete curing production equipment.

[0079] S6. Use a visualization display module to display the temperature of each of the pre-cooled concrete production temperature control zones, the environmental data, and the operating parameter information of the pre-cooled concrete production equipment, and allow the operator to manually set the temperature parameters, and transmit the operator's manual setting command to the temperature control module.

[0080] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects:

[0081] This invention provides a region delineation module, a temperature control module, an environmental data acquisition module, a data analysis and inference module, an equipment control module, and a visualization display module. These modules work collaboratively to achieve precise temperature control of different pre-cooled concrete production areas. The system utilizes lidar point cloud and hyperspectral image imaging technology to delineate different temperature control areas. A fiber optic temperature sensor network array, a model prediction MPC controller, and a magnetic levitation frequency conversion cooling component ensure accurate temperature control. Simultaneously, the data analysis and inference module combines a deep learning time series prediction model and inference rule mining algorithms to support temperature control of production equipment. The overall temperature control solution of this invention improves the intelligence and accuracy of IoT-based temperature control in pre-cooled concrete production and is applicable to various pre-cooled concrete production scenarios. Attached Figure Description

[0082] Figure 1 This is an architecture diagram of an IoT temperature control system for pre-cooled concrete production provided in an embodiment of the present invention;

[0083] Figure 2 This is a schematic diagram of a method for controlling the temperature of pre-cooled concrete production using the Internet of Things, provided in an embodiment of the present invention. Detailed Implementation

[0084] The present invention will now be described in detail with reference to the accompanying drawings.

[0085] Example 1

[0086] This invention provides an IoT-based temperature control system for pre-cooled concrete production, such as... Figure 1 ,include:

[0087] The area delineation module is used to automatically delineate multiple separate pre-cooled concrete production temperature control zones based on the spatial layout of the cast-in-place structure. The area delineation module outputs the delineation information of each zone to the temperature control module, the environmental data acquisition module, and the data analysis and reasoning module.

[0088] The temperature control module, which is linked with the area delineation module, is used to control the temperature of each pre-cooled concrete production temperature control area, feeds the temperature back to the data analysis and inference module in real time, and receives the adjustment suggestions given by the data analysis and inference module in real time.

[0089] The environmental data acquisition module is used to monitor the environmental data of each temperature-controlled area in the pre-cooled concrete production process, including humidity and light intensity, and transmit the environmental data to the data analysis and inference module in real time.

[0090] The data analysis and reasoning module analyzes and reasons the received area delineation results, temperature and environmental data. It analyzes the temperature, characteristics of pre-cooled concrete and environmental data of each pre-cooled concrete production temperature control area, and then further reasons about the aforementioned data based on the data reasoning network model, and outputs control commands to the equipment control module and temperature control module.

[0091] The equipment control module dynamically adjusts the operating parameters of the pre-cooled concrete production equipment according to the control instructions of the data analysis and reasoning module, and pushes the operating parameter information of the pre-cooled concrete production equipment to the visualization display module. The pre-cooled concrete production equipment includes concrete pouring equipment and concrete curing equipment.

[0092] The visualization display module is used to display the temperature of each of the pre-cooled concrete production temperature control zones, the environmental data, and the operating parameter information of the pre-cooled concrete production equipment. It also allows the operator to manually set the temperature parameters and transmits the operator's manual setting instructions to the temperature control module.

[0093] Specifically, the system in this embodiment includes a region delineation module, a temperature control module, an environmental data acquisition module, a data analysis and inference module, an equipment control module, and a visualization display module. These modules work collaboratively to achieve precise temperature control of different pre-cooled concrete production areas. The system utilizes lidar point cloud and hyperspectral image imaging technology to delineate different temperature control areas. A fiber optic temperature sensor network array, a model prediction MPC controller, and a magnetic levitation frequency conversion cooling component ensure precise temperature control. Simultaneously, the data analysis and inference module combines a deep learning time series prediction model and inference rule mining algorithms to support temperature control of production equipment. The overall temperature control solution of this invention improves the intelligence and accuracy of IoT-based temperature control in pre-cooled concrete production and is applicable to various outdoor pre-cooled concrete production scenarios.

[0094] In the above embodiments, specifically, the region delineation module includes:

[0095] The casting body spatial acquisition unit uses lidar point cloud and hyperspectral image imaging technology to acquire spatial distribution information of the casting body structure produced by pre-cooled concrete and transmits the acquired spatial distribution information to the area delineation decision unit.

[0096] The regional delineation decision unit has a built-in partitioning decision model based on particle swarm optimization algorithm and ant colony optimization algorithm. It is used to delineate the cast-in-place structure of pre-cooled concrete production into multiple regions according to the spatial distribution information, and output a regional delineation scheme that meets the needs of different pre-cooled concrete production scenarios. The multiple regions are uniformly named pre-cooled concrete production areas. Different pre-cooled concrete production scenarios include at least large-volume concrete, precast components, and core tubes of super high-rise buildings.

[0097] The casting body space acquisition unit further includes:

[0098] The LiDAR scanning subunit is equipped with multiple sets of adjustable azimuth and precision LiDARs, which are distributed in a matrix at key points in the pre-cooled concrete production area to scan the features of the internal space of the pre-cooled concrete production area and generate point cloud data.

[0099] The hyperspectral image acquisition subunit uses a hyperspectral camera with visible light to short-wave infrared imaging capabilities to capture concrete color and crack information on the surface of the pre-cooled concrete production area, and then uses an image fusion algorithm to match the point cloud data to output complete spatial distribution information of the pre-cooled concrete production area to the area delineation decision unit.

[0100] The regional delineation decision-making unit includes:

[0101] The initial scheme determines the sub-units, and has a built-in expert database of pre-cooled concrete production scenario cases. Based on the spatial distribution information, the type of concrete raw materials, and the information on the intended use of the concrete after production, the pre-cooled concrete production scenario is predicted. The corresponding sub-regional foundation templates are retrieved to form the initial area delineation scheme. The intended use information includes at least the distance information from the concrete production site to the construction site.

[0102] It should be noted that the regional basic templates are built into the expert database of pre-cooled concrete production scenario cases. Multiple regional basic templates are pre-determined through the expert database of pre-cooled concrete production scenario cases. Multiple regional basic templates are generated after clustering through massive case analysis in the expert database. These multiple regional basic templates are subsequently used to determine the initial regional delineation scheme.

[0103] Specifically, the initial scheme determines the sub-units including:

[0104] Scene feature extraction, calculation of scene geometric complexity, raw material uniformity and transportation distance coefficient;

[0105] The expert case matching engine retrieves the scenario with the most similar historical solution using Euclidean distance;

[0106] The algorithm optimization execution subunit uses the particle swarm optimization algorithm for fast convergence in high-dimensional parameter space and the discrete partition boundary optimization strategy of the ant colony optimization algorithm to iteratively optimize the initial region delineation scheme, balancing the independence of the temperature control area of ​​the pre-cooled concrete production, the process connection and the synergy of multi-modal sensors.

[0107] Specifically, the implementation process of the algorithm optimization execution subunit is as follows:

[0108] The particle swarm optimization algorithm is used to optimize the partitioning scheme of the temperature control zone for pre-cooled concrete production in a high-dimensional parameter space. The velocity update formula satisfies:

[0109]

[0110] in, This represents the current velocity, reflecting the particle's search direction and step size in the parameter space. (pbest) id Gbest represents the individual best position, reflecting the particle's own historical best position. id Representing the global optimum, reflecting the optimal position found by all particles in the population, c1 and c2 represent learning factors, r1 and r2 represent random numbers, and k represents the current iteration number. max Indicates the maximum number of iterations the algorithm can run;

[0111] Combining ant colony optimization algorithm, topology optimization is performed on discrete partition boundaries, where the pheromone concentration update rule is as follows:

[0112]

[0113]

[0114] Where, π ij T represents the pheromone concentration, ρ represents the pheromone evaporation coefficient, and T represents the pheromone concentration. i T j L represents the temperature values ​​at nodes i and j, respectively. k1 Let Q represent the path length of the k1-th ant, where Q is a constant and m represents the number of ants.

[0115] In the above embodiments, specifically, the temperature control module includes:

[0116] The temperature data acquisition unit uses a fiber optic temperature sensor network array to measure the temperature of each of the pre-cooled concrete production temperature control zones and report the temperature data to the temperature control unit.

[0117] The temperature control unit integrates a model prediction MPC controller and a magnetic levitation frequency conversion refrigeration component, and is used to adjust the temperature of each of the pre-cooled concrete production temperature control zones based on the measurement results of the temperature data acquisition unit.

[0118] In the above embodiments, specifically, the temperature data acquisition unit includes:

[0119] The fiber optic temperature sensor network array layout subunit uses a dynamic deployment algorithm based on geometric features. It calculates the first sensor density of all nodes in the pre-cooled concrete production temperature control area based on the volumetric characteristics, shape complexity index, and historical temperature fluctuation characteristics of the temperature control area. Through finite element thermal conduction simulation, areas with temperature gradients greater than a fourth threshold are designated as sensitive zones, and nodes in these sensitive zones are represented as sensitive nodes. A second sensor density is generated by arranging a bidirectional staggered array at the nodes at the boundaries of sensitive nodes and adjacent temperature control areas. If a pre-cooled concrete production temperature control area is detected to exceed the fifth threshold three times consecutively, a moving sensor is automatically added to that area, generating a third sensor density.

[0120] The data processing and transmission subunit integrates Kalman adaptive filtering algorithm and generative adversarial network to preprocess the field-collected data, remove outliers, reduce noise in the data, and transmit the preprocessed temperature data to the temperature control unit.

[0121] The temperature control unit includes:

[0122] The model predicts the MPC control subunit, which takes the temperature deviation and its changing trend as input, and dynamically generates the optimal control sequence through a rolling time-domain optimization algorithm, and outputs the control signal to drive the magnetic levitation variable frequency cooling component.

[0123] The magnetic levitation variable frequency refrigeration execution subunit is equipped with a magnetic levitation centrifugal compressor, a digital pulse electronic expansion valve, and a biomimetic airflow guiding duct mechanism. Based on the model prediction of the control signal of the MPC control subunit, it dynamically adjusts the cooling capacity, cold air speed, and refrigerant flow.

[0124] It should be noted that the prediction model for the MPC control subunit is as follows:

[0125] A state-space model is constructed based on the concrete heat conduction equation:

[0126] T(t+1)=AT(t)+Bu(t)+Cd(t),

[0127] Where A represents the state matrix, describing the self-evolution characteristics of the internal temperature field of concrete and reflecting the diffusion and accumulation of heat; B represents the control matrix, quantifying the ability of refrigeration equipment (such as cooling pipes and spray systems) to regulate the temperature field; C represents the disturbance matrix, reflecting the impact of external disturbances such as changes in ambient temperature and light intensity on the system; T(t) represents the state vector, a column vector describing the temperature values ​​of each temperature control zone; u(t) represents the control input, consisting of cooling water flow rate, refrigeration compressor frequency, and spray valve opening; d(t) represents the disturbance input, consisting of ambient temperature, solar radiation intensity, and humidity.

[0128] In the above embodiments, specifically, the environmental data acquisition module includes:

[0129] The humidity data acquisition and processing unit is equipped with a distributed fiber optic acoustic sensor to measure the humidity data of each of the pre-cooled concrete production temperature control zones and to preprocess the humidity data error in real time.

[0130] The light intensity acquisition and processing unit uses a quantum dot spectral sensor to measure the light intensity of each of the pre-cooled concrete production temperature control areas, constructs a real-time light intensity field model within the pre-cooled concrete production temperature control area, and updates it dynamically.

[0131] In the above embodiments, specifically, the humidity data acquisition and processing unit further includes:

[0132] The distributed fiber optic acoustic wave sensing subunit monitors the acoustic vibration signal in the temperature control area of ​​pre-cooled concrete production in real time, and inputs the acoustic vibration signal into the hydration reaction acoustic feature library to invert the humidity change.

[0133] The data correction subunit sets the start and end points of the time period for correction, and combines a multi-parameter decoupling algorithm to eliminate the cross-sensitivity of temperature and humidity, and outputs accurate humidity values ​​to the data analysis and inference module.

[0134] The light intensity acquisition and processing unit further includes:

[0135] The quantum dot spectral sensor subunit is used to output the band-wise irradiance data of the temperature control zone in the pre-cooled concrete production process.

[0136] The geometric correction module, based on astronomical algorithms, eliminates errors in band irradiance data caused by zenith angle.

[0137] The illumination intensity modeling and analysis module calculates the regional light field and generates light field data based on the irradiance data of different wavelengths and the Monte Carlo ray tracing algorithm. According to the building BIM model and the light trajectory, it predicts the shadow movement path, marks the data of the light intensity dead zone area, and sends the light field data and the light intensity dead zone area data to the data analysis and inference module.

[0138] It should be noted that quantum dot spectral sensors:

[0139] Hardware configuration:

[0140] Visible light band (400-700nm): Employs a silicon-based photodiode array with a resolution of ±5W / m. 2 ;

[0141] Near-infrared band (700-1100nm): InGaAs sensor, sensitivity 0.1mV / (W / m 2 );

[0142] Ultraviolet band (300-400nm): used to monitor photoaging effects;

[0143] Data calibration:

[0144] Daily calibration using a standard light source (such as AM1.5G spectrum) eliminates sensor attenuation;

[0145] Monte Carlo ray tracing algorithm implementation:

[0146] 1. Modeling process:

[0147] Light emission:

[0148] Generate 10 based on the sun's position (declination angle δ, hour angle ω). 6 A ray of light;

[0149] Wavelength weighting: Visible light: Near-infrared: Ultraviolet = 0.6:0.3:0.1;

[0150] 2. Collision Detection:

[0151] BIM model meshing (accuracy ≤ 0.1m), using BVH (Bounding Volume Hierarchy) to accelerate intersection calculations;

[0152] Modeling of concrete surface roughness (micro-plane normals follow a GGX distribution);

[0153] 3. Energy accumulation:

[0154] Each temperature-controlled zone is divided into grid cells, and each grid cell receives irradiance as follows:

[0155]

[0156] Among them, E cell I represents the irradiance received by a grid cell. i Cosθ represents the intensity of light. i Let d represent the cosine of the incident angle, a represent the attenuation coefficient, and d represent the attenuation i N represents the propagation distance, reflecting the straight-line distance of light rays from the light source (sun or reflector) to the grid cell; N represents the total number of light rays.

[0157] Shadow path prediction and blind spot marking:

[0158] Dynamic shadow calculation:

[0159] Input: BIM model geometric data + solar trajectory calculated by astronomical algorithms;

[0160] Output: Time-varying curve of shadow coverage;

[0161] Marking rules: If the slope of the time-varying curve of shadow coverage is greater than the preset threshold within 10 minutes, it is marked as a potential maintenance risk area and judged as a dead corner;

[0162] The specific implementation process of the astronomical algorithm is as follows:

[0163] Step 1: Spacetime Reference Synchronization

[0164] High-precision spatiotemporal calibration:

[0165] The latitude and longitude of the station are obtained through GPS / BeiDou modules;

[0166] UTC time is synchronized using atomic clocks;

[0167] Terrain data loading:

[0168] Read the elevation and slope / aspect of the measurement points from the DEM database;

[0169] Calculate the terrain roughness factor;

[0170] Step 2: Solving the Sun's Position

[0171] The declination angle was calculated using an improved Jean Meeus algorithm.

[0172] The solar zenith angle formula is used to calculate the latitude, longitude, declination angle, and solar hour angle of the station, while the mean error formula is used to eliminate the error caused by the zenith angle calculation.

[0173] It should be noted that the expert database of pre-cooled concrete production scenario cases includes a built-in hydration reaction acoustic feature database, which outputs a humidity-sound wave vibration signal feature mapping model through experimental simulation data.

[0174] In the above embodiments, specifically, the data analysis and reasoning module includes:

[0175] The pre-cooled concrete characteristic analysis unit integrates a big data pre-cooled concrete characteristic knowledge base to analyze the pouring and curing requirements of pre-cooled concrete and generate pouring and curing strategies for pre-cooled concrete characteristics.

[0176] The integrated data analysis unit uses a deep learning time series prediction model and inference rule mining algorithm to analyze the impact of temperature and environmental data on pouring and curing, predict temperature and environmental change trends and associated risks in advance, generate control suggestions, and provide the control suggestions to the temperature control module and the equipment control module.

[0177] The fusion data analysis unit includes:

[0178] The deep learning time series prediction subunit mines the patterns of historical temperature, humidity and light intensity time series data based on an improved radial basis function interpolation algorithm and an improved U-Net++ network model, predicts the future trend of temperature and environmental data, and generates trend prediction data.

[0179] The rule reasoning subunit inputs the predicted trend data of future temperature and environmental data into the data reasoning network model to obtain the reasoning result. The reasoning result is used to output early warning information and control suggestions to the equipment control module and temperature control module. The reasoning result is used to quantify the degree of impact of temperature and environmental data changes on the quality of pouring and curing.

[0180] The specific implementation process of the deep learning time series prediction subunit is as follows:

[0181] Multi-source data fusion is performed by acquiring discrete point temperatures and humidity within the temperature-controlled area of ​​the pre-cooled concrete production zone through the temperature control module and the environmental data acquisition module, generating discrete data; and acquiring surface temperature field, humidity field, and light intensity field of the temperature-controlled area of ​​the pre-cooled concrete production zone through the temperature control module and the environmental data acquisition module, generating surface data. Based on an improved radial basis function interpolation algorithm, the discrete data and surface data are fused to construct a three-dimensional temperature distribution voxel model.

[0182] Intelligent segmentation and feature extraction of temperature field: The three-dimensional temperature voxel model is sliced ​​into XY, XZ, and YZ planes. The sliced ​​data is input into an improved U-Net++ network. The output of the improved U-Net++ network includes a binary mask of isothermal surface, semantic labels of abnormal high temperature region, and temperature gradient field tensor. The encoder of the improved U-Net++ network adopts ResNet34 skeleton and embeds coordinate attention module. The decoder of the improved U-Net++ network introduces dense skip connections and uses subpixel convolution.

[0183] Feature extraction by CP decomposition of the temperature gradient field tensor generates a spatiotemporal feature vector set {a r b r c r d r}, where a r Characterizing the spatial distribution pattern of the temperature field in the X direction, reflecting the temperature gradient of the concrete structure along the pouring length; b r Characterizes the spatial distribution pattern of the temperature field in the Y direction, reflecting the interlayer temperature transfer characteristics during layered casting; c r Characterizing the temperature field distribution pattern in the Z direction, reflecting the attenuation relationship between the internal core temperature and the surface temperature; d r The weighting coefficients characterize the temperature field over time, reflecting the influence of environmental temperature changes and time-varying factors such as heat of hydration release on the overall temperature field; the CP decomposition represents a factorization method for a higher-order tensor.

[0184] A physical information neural network (PINN) is constructed to output trend prediction data for the next N minutes. The input of the PINN is a set of spatiotemporal feature vectors and environmental data. The trend prediction data includes temperature field evolution prediction data and confidence interval data. The environmental data includes ambient temperature, ambient humidity and ambient light intensity.

[0185] The rule-based reasoning subunit inputs the predicted trend data of future temperature and environmental data into the data reasoning network model to obtain the following reasoning results:

[0186] Based on the first-layer perceptron and the core multilayer perceptron in the data inference network model, the representation vector is extracted from the future temperature change trend prediction data to obtain the first temperature vector and the first temperature base vector. The representation vector extraction includes the representation vectors whose correlation is not less than a preset first correlation threshold.

[0187] Based on the second-layer perceptron and the core multilayer perceptron in the data inference network model, the representation vector is extracted from the future environmental data change trend prediction data to obtain the first environmental vector and the first environmental base vector. The representation vector extraction includes the representation vectors whose correlation is not less than a preset second correlation threshold.

[0188] The first weight calculation unit in the data reasoning network model performs basic vector correction on the first temperature basic vector and the first environment basic vector respectively, so as to fuse the corrected basic vector with the first temperature vector and the first environment vector respectively to obtain the first fused vector. Finally, the reasoning unit in the data reasoning network model performs data reasoning on the first fused vector to obtain the reasoning result. The reasoning result includes at least a general reasoning result and an emergency risk reasoning result. The general reasoning result represents non-emergency handling matters, and the emergency risk reasoning result represents emergency handling matters.

[0189] It is important to explain the implementation details of the key steps in the data inference network model:

[0190] 1. Representation Vector Extraction (Taking Temperature Data as an Example)

[0191] Input data:

[0192] Temperature trend forecast sequence for the next 6 hours (sampling interval 10 minutes): T = [T1, T2, ... T 36 ];

[0193] Environmental change trend data (ambient temperature, ambient humidity, and ambient light intensity): E = [H1, W1, S1, ... H 36W 36 S 36 ];

[0194] First-level sensor (temperature branch):

[0195] The first layer of the perceptron structure consists of a fully connected layer (36-dimensional input, 64-dimensional output) + ReLU activation;

[0196] First-layer perceptron output: First temperature vector (Higher-order features), First temperature fundamental vector (Low-dimensional features with clear physical meaning, such as slope and curvature);

[0197] Correlation threshold: A preset first correlation threshold is used, and the actual calculation is verified by the Pearson coefficient;

[0198] 2. Basic Vector Correction

[0199] Corrected algorithm:

[0200] V′ temp_base =V temp_base +a*MLP([V temp ||V env ]),

[0201] Where a represents the learnable parameter, || represents vector concatenation, MLP represents the perceptron function, and V env Represents the environment vector;

[0202] 3. Vector fusion and inference includes a fusion layer and inference units;

[0203] General reasoning result (non-urgent matters):

[0204] Output: Maintenance recommendations (e.g., "Start spraying after 6 hours");

[0205] Model: Softmax classifier (10 classes, standard operation);

[0206] Emergency Risk Reasoning Results:

[0207] Output: Risk type and level (e.g., "Core temperature rise exceeds limit within 2 hours: Level 3");

[0208] Model: Binary threshold judgment (triggered when Sigmoid output > 0.8);

[0209] It should be noted that the multilayer perceptron in the data inference network model is an expert neural network, and the weight calculation unit is a gated neural network. The data inference network model includes an input layer, a core multilayer perceptron, a first-layer perceptron, a second-layer perceptron, a first weight calculation unit, and an inference unit. Based on the data inference network model, it performs accurate inference on the predicted trend of future temperature and environmental data, infers the relationship between parameter adjustment and temperature control of concrete pouring equipment and concrete curing equipment, and then derives precise control commands, which are output to the equipment control module and temperature control module. The aforementioned technical features are one of the technical highlights of this embodiment. Most existing technologies only adjust the temperature control parameters in the temperature control module, without adjusting the parameters of the concrete pouring equipment and concrete curing equipment involved in the pouring and curing process to achieve further precise adjustment of the concrete temperature.

[0210] In the above embodiments, specifically, the device control module includes:

[0211] A general algorithm processing unit is embedded with a deep reinforcement learning model. The deep reinforcement learning model is used to dynamically adjust the operating parameters of the pre-cooled concrete production equipment according to the general inference results, and push the operating parameters and energy consumption parameters to the visualization display module.

[0212] The emergency handling unit, based on the emergency risk reasoning result, adjusts the heating mode of the pre-cooling concrete production equipment to increase the heating power and accelerate the concrete forming speed when the concrete surface temperature is lower than the first emergency threshold for concrete pouring and curing. When the emergency risk reasoning result indicates that the concrete surface temperature exceeds the second emergency threshold for concrete pouring and curing, the unit activates the cooling mode of the pre-cooling concrete production equipment to increase the cooling power, thereby rapidly reducing the concrete surface temperature and preventing cracks from forming due to overheating.

[0213] It should be noted that the pre-cooled concrete production equipment is equipped with both cooling and heating functions to assist in temperature intervention during the concrete pouring and curing process. In the existing technology, most pre-cooled concrete production equipment does not intervene in concrete temperature control. However, in the actual production process, the pre-cooled concrete production equipment can improve the efficiency of concrete temperature control by intervening in the concrete temperature in real time, which is also one of the technical highlights of this invention.

[0214] The energy consumption optimization unit increases the fan speed of the pre-cooling concrete production equipment to promote cooling when the surface temperature of the concrete exceeds the optimal temperature threshold for concrete pouring; conversely, it decreases the fan speed to promote heat retention when the surface temperature of the concrete is below the optimal temperature threshold for concrete pouring. The formula used is as follows:

[0215]

[0216] If(T cur (t)andT inner (t)>T set ), thus obtaining F fan rise,

[0217] If(T cur (t)andT inner (t) <T set ), thus obtaining F fan decline,

[0218] Among them, F fan F represents the adjusted energy consumption operating rate of the pre-cooled concrete production equipment. base Based on the wind turbine speed, b f T is the fan speed coefficient. cur (t) represents the average outer surface temperature of the temperature-controlled area for pre-cooled concrete production at time t, T inner (t) represents the average inner surface temperature of the pre-cooled concrete production temperature control zone at time t, T set This indicates the temperature optimization threshold.

[0219] It should be noted that the energy consumption optimization unit in this embodiment only considers the surface temperature of the concrete.

[0220] Example 2

[0221] A method for IoT-based temperature control in the production of pre-cooled concrete includes the following steps:

[0222] S1. The area delineation module automatically delineates multiple separate pre-cooled concrete production temperature control zones based on the spatial layout of the cast-in-place structure. The area delineation module outputs the delineation information of each zone to the temperature control module, the environmental data acquisition module, and the data analysis and reasoning module.

[0223] S2. Use the temperature control module to work in conjunction with the area delineation module to control the temperature of each pre-cooled concrete production temperature control area, feed the temperature back to the data analysis and inference module in real time, and receive the control suggestions given by the data analysis and inference module in real time.

[0224] S3. Use the environmental data acquisition module to monitor the environmental data of each pre-cooled concrete production temperature control area, including humidity and light intensity, and transmit the environmental data to the data analysis and inference module in real time.

[0225] S4. The data analysis and reasoning module is used to analyze and reason about the received area delineation results, temperature and environmental data. The temperature, characteristics of pre-cooled concrete and environmental data of each pre-cooled concrete production temperature control area are analyzed. Then, the data is further reasoned based on the data reasoning network model, and control commands are output to the equipment control module and temperature control module.

[0226] S5. The equipment control module dynamically adjusts the operating parameters of the pre-cooled concrete production equipment according to the control instructions of the data analysis and reasoning module, and pushes the operating parameter information of the pre-cooled concrete production equipment to the visualization display module. The pre-cooled concrete production equipment includes concrete pouring production equipment and concrete curing production equipment.

[0227] S6. Use a visualization display module to display the temperature of each of the pre-cooled concrete production temperature control zones, the environmental data, and the operating parameter information of the pre-cooled concrete production equipment, and allow the operator to manually set the temperature parameters, and transmit the operator's manual setting command to the temperature control module.

[0228] It should be understood that the above embodiments are one or more embodiments of the present invention, and there are many other embodiments and variations based on the present invention; any variations and modifications made by those skilled in the art through the present invention without making pioneering innovations are all within the protection scope of the present invention.

Claims

1. An IoT-based temperature control system for pre-cooled concrete pouring and curing, characterized in that, include: The area delineation module is used to automatically delineate multiple separate pre-cooled concrete pouring and curing temperature control zones based on the spatial layout of the poured body structure. The region delineation module outputs the region delineation information to the temperature control module, the environmental data acquisition module, and the data analysis and reasoning module. The temperature control module, which is linked with the area delineation module, is used to control the temperature of each pre-cooled concrete pouring and curing temperature control area, feeds the temperature back to the data analysis and inference module in real time, and receives the adjustment suggestions given by the data analysis and inference module in real time. The environmental data acquisition module is used to monitor the environmental data of each pre-cooled concrete pouring and curing temperature control zone, including humidity and light intensity, and transmit the environmental data to the data analysis and inference module in real time. The data analysis and reasoning module analyzes and reasons the received area delineation results, temperature and environmental data. It analyzes the temperature, characteristics of pre-cooled concrete and environmental data of each pre-cooled concrete pouring and curing temperature control zone, and then further reasons about the aforementioned data based on the data reasoning network model, and outputs control commands to the equipment control module and temperature control module. The equipment control module dynamically adjusts the operating parameters of the pre-cooled concrete pouring and curing equipment according to the control instructions of the data analysis and reasoning module, and pushes the operating parameter information and energy consumption parameter information of the pre-cooled concrete pouring and curing equipment to the visualization display module. The pre-cooled concrete pouring and curing equipment includes concrete pouring equipment and concrete curing equipment. The visualization display module is used to display the temperature of each of the pre-cooled concrete pouring and curing temperature control zones, the environmental data, the operating parameter information of the pre-cooled concrete pouring and curing equipment, and allows the operator to manually set the temperature parameters and transmit the operator's manual setting instructions to the temperature control module. The data analysis and reasoning module includes a fusion data analysis unit that uses a deep learning time series prediction model and reasoning rule mining algorithm to analyze the impact of temperature and environmental data on pouring and curing, predict temperature and environmental change trends and associated risks in advance, generate control suggestions, and provide the control suggestions to the temperature control module and the equipment control module. The fusion data analysis unit includes: The deep learning time series prediction subunit mines the patterns of historical temperature, humidity and light intensity time series data based on an improved radial basis function interpolation algorithm and an improved U-Net++ network model, predicts the future trend of temperature and environmental data, and generates trend prediction data. The rule-based reasoning subunit inputs the predicted trend data of future temperature and environmental data into the data reasoning network model to obtain the reasoning results. The reasoning results are used to output early warning information and control suggestions to the equipment control module and temperature control module. The reasoning results are used to quantify the degree of impact of temperature and environmental data changes on the quality of pouring and curing.

2. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 1, characterized in that, The region delineation module includes: The casting body space acquisition unit uses lidar point cloud and hyperspectral image imaging technology to acquire spatial distribution information of the casting body structure during pre-cooled concrete casting and curing, and transmits the acquired spatial distribution information to the area delineation decision unit. The region delineation decision unit has a built-in partitioning decision model based on particle swarm optimization algorithm and ant colony optimization algorithm. It is used to delineate the pre-cooled concrete pouring and curing structure into multiple regions according to the spatial distribution information, and output a region delineation scheme that meets the needs of different pre-cooled concrete pouring and curing scenarios. The multiple regions are uniformly named pre-cooled concrete pouring and curing regions. Different pre-cooled concrete pouring and curing scenarios include at least large-volume concrete, precast components, and core tubes of super high-rise buildings. The casting body space acquisition unit further includes: The LiDAR scanning subunit is equipped with multiple sets of adjustable azimuth and precision LiDARs, which are distributed in a matrix at key points in the pre-cooled concrete pouring and curing area to scan the features of the internal space of the pre-cooled concrete pouring and curing area and generate point cloud data. The hyperspectral image acquisition subunit uses a hyperspectral camera with imaging capabilities from visible light to short-wave infrared to capture the surface concrete color and crack information of the pre-cooled concrete pouring and curing area. It then matches the image with the point cloud data through an image fusion algorithm and outputs complete spatial distribution information of the pre-cooled concrete pouring and curing area to the area delineation decision unit. The regional delineation decision-making unit includes: The initial scheme determines the sub-units, and has a built-in expert database of pre-cooled concrete pouring and curing scenario cases. Based on the spatial distribution information, the type of concrete raw materials, and the usage information after concrete pouring and curing, the pre-cooled concrete pouring and curing scenario is predicted. The corresponding sub-regional foundation templates are retrieved to form the initial area delineation scheme. The usage information includes at least the distance information from the concrete pouring and curing site to the construction site. The algorithm optimization execution subunit uses the particle swarm optimization algorithm for fast convergence in high-dimensional parameter space and the discrete partition boundary optimization strategy of the ant colony optimization algorithm to iteratively optimize the initial region delineation scheme. The specific implementation process of the algorithm optimization execution subunit is as follows: The particle swarm optimization algorithm is used to optimize the partitioning scheme of the temperature control zone for pre-cooled concrete pouring and curing in a high-dimensional parameter space. The velocity update formula satisfies: ,in, This indicates the current velocity, reflecting the particle's search direction and step size in the parameter space. This represents the individual optimal position, reflecting the particle's own historical optimal position. This represents the global optimum, reflecting the optimal position found by all particles in the population. They represent learning factors, Let represent random numbers, and k represent the current iteration number. This indicates the maximum number of iterations the algorithm can run. Indicates the current position; Combining ant colony optimization algorithm, topology optimization is performed on discrete partition boundaries, where the pheromone concentration update rule is as follows: ， ,in, Indicates pheromone concentration. Indicates the pheromone evaporation coefficient. These represent the temperature values ​​at nodes i and j, respectively. Let Q represent the path length of the k1-th ant, where Q is a constant and m represents the number of ants.

3. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 1, characterized in that, The temperature control module includes: The temperature data acquisition unit uses a fiber optic temperature sensor network array to measure the temperature of each of the pre-cooled concrete pouring and curing temperature control zones and report the temperature data to the temperature control unit. The temperature control unit integrates a model prediction MPC controller and a magnetic levitation frequency conversion refrigeration component, and is used to adjust the temperature of each of the pre-cooled concrete pouring and curing temperature control zones according to the measurement results of the temperature data acquisition unit.

4. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 3, characterized in that, The temperature data acquisition unit includes: The fiber optic temperature sensor network array layout subunit, based on a dynamic deployment algorithm of geometric features, calculates the first sensor density of all nodes in the pre-cooled concrete pouring and curing temperature control area according to the volume characteristic value, shape complexity index, and historical temperature fluctuation characteristics of the pre-cooled concrete pouring and curing temperature control area. Through finite element thermal conduction simulation, the temperature gradient greater than the fourth threshold is marked as a sensitive area, and the nodes in the sensitive area are represented as sensitive nodes. By arranging a bidirectional staggered array at the nodes at the boundary between the sensitive nodes and adjacent temperature control areas, a second sensor density is generated. If a pre-cooled concrete pouring and curing temperature control area is detected to exceed the fifth threshold three times consecutively, a moving sensor is automatically added to the area to generate a third sensor density. The data processing and transmission subunit integrates Kalman adaptive filtering algorithm and generative adversarial network to preprocess the field-collected data, remove outliers, reduce noise in the data, and transmit the preprocessed temperature data to the temperature control unit. The temperature control unit includes: The model predicts the MPC control subunit, which takes the temperature deviation and its changing trend as input, and dynamically generates the optimal control sequence through a rolling time-domain optimization algorithm, and outputs the control signal to drive the magnetic levitation variable frequency cooling component. The magnetic levitation variable frequency refrigeration execution subunit is equipped with a magnetic levitation centrifugal compressor, a digital pulse electronic expansion valve, and a biomimetic airflow guiding duct mechanism. Based on the model prediction of the control signal of the MPC control subunit, it dynamically adjusts the cooling capacity, cold air speed, and refrigerant flow.

5. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 1, characterized in that, The environmental data acquisition module includes: The humidity data acquisition and processing unit is equipped with a distributed fiber optic acoustic sensor to measure the humidity data of each of the pre-cooled concrete pouring and curing temperature control zones and to preprocess the humidity data error in real time. The light intensity acquisition and processing unit uses a quantum dot spectral sensor to measure the light intensity of each of the pre-cooled concrete pouring and curing temperature control areas, constructs a real-time light intensity field model within the pre-cooled concrete pouring and curing temperature control areas, and updates it dynamically.

6. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 5, characterized in that, The humidity data acquisition and processing unit further includes: The distributed fiber optic acoustic wave sensing subunit monitors the acoustic vibration signal in the temperature control zone of pre-cooled concrete pouring and curing in real time, and inputs the acoustic vibration signal into the hydration reaction acoustic feature library to invert the humidity change. The data correction subunit sets the start and end points of the time period for correction, and combines a multi-parameter decoupling algorithm to eliminate the cross-sensitivity of temperature and humidity, and outputs accurate humidity values ​​to the data analysis and inference module. The light intensity acquisition and processing unit further includes: The quantum dot spectral sensor subunit is used to output the band-wise irradiance data of the temperature-controlled area for pre-cooled concrete pouring and curing. The geometric correction module, based on astronomical algorithms, eliminates errors in band irradiance data caused by zenith angle. The illumination intensity modeling and analysis module calculates the regional light field and generates light field data based on the irradiance data of different wavelengths and the Monte Carlo ray tracing algorithm. According to the building BIM model and the light trajectory, it predicts the shadow movement path, marks the data of the light intensity dead zone area, and sends the light field data and the light intensity dead zone area data to the data analysis and inference module.

7. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 1, characterized in that, The data analysis and reasoning module also includes: The pre-cooled concrete characteristic analysis unit integrates a big data pre-cooled concrete characteristic knowledge base to analyze the pouring and curing requirements of pre-cooled concrete and generate pouring and curing strategies for pre-cooled concrete characteristics. The specific implementation process of the deep learning time series prediction subunit is as follows: Multi-source data fusion is performed by acquiring discrete point temperatures and humidity within the pre-cooled concrete pouring and curing temperature control zone through temperature control and environmental data acquisition modules to generate discrete data; surface temperature field, humidity field, and light intensity field of the pre-cooled concrete pouring and curing temperature control zone are acquired through temperature control and environmental data acquisition modules to generate surface data; based on an improved radial basis function interpolation algorithm, the discrete data and surface data are fused to construct a three-dimensional temperature distribution voxel model. Intelligent segmentation and feature extraction of temperature field: The three-dimensional temperature voxel model is sliced ​​into XY, XZ, and YZ planes. The sliced ​​data is input into an improved U-Net++ network. The output of the improved U-Net++ network includes a binary mask of isothermal surface, semantic labels of abnormal high temperature region, and temperature gradient field tensor. The encoder of the improved U-Net++ network adopts ResNet34 skeleton and embeds coordinate attention module. The decoder of the improved U-Net++ network introduces dense skip connections and uses subpixel convolution. Feature extraction by CP decomposition of the temperature gradient field tensor generates a set of spatiotemporal feature vectors. ,in, It characterizes the spatial distribution pattern of the temperature field in the X direction, reflecting the temperature gradient of the concrete structure along the pouring length. It characterizes the spatial distribution pattern of the temperature field in the Y direction and reflects the interlayer temperature transfer characteristics during layered casting. It characterizes the distribution pattern of the temperature field in the Z direction, reflecting the decay relationship between the internal core temperature and the surface temperature. The weighting coefficients characterize the temperature field over time, reflecting the influence of environmental temperature changes and time-varying factors such as heat of hydration release on the overall temperature field; the CP decomposition represents a factorization method for a higher-order tensor. A physical information neural network (PINN) is constructed to output trend prediction data for the next N minutes. The input of the PINN is a set of spatiotemporal feature vectors and environmental data. The trend prediction data includes temperature field evolution prediction data and confidence interval data. The environmental data includes ambient temperature, ambient humidity and ambient light intensity. The rule-based reasoning subunit inputs the predicted trend data of future temperature and environmental data into the data reasoning network model to obtain the following reasoning results: Based on the first-layer perceptron and the core multilayer perceptron in the data inference network model, the representation vector is extracted from the future temperature change trend prediction data to obtain the first temperature vector and the first temperature base vector. The representation vector extraction includes the representation vectors whose correlation is not less than a preset first correlation threshold. Based on the second-layer perceptron and the core multilayer perceptron in the data inference network model, the representation vector is extracted from the future environmental data change trend prediction data to obtain the first environmental vector and the first environmental base vector. The representation vector extraction includes the representation vectors whose correlation is not less than a preset second correlation threshold. The first weight calculation unit in the data reasoning network model performs basic vector correction on the first temperature basic vector and the first environment basic vector respectively, so as to fuse the corrected basic vector with the first temperature vector and the first environment vector respectively to obtain the first fused vector. Finally, the reasoning unit in the data reasoning network model performs data reasoning on the first fused vector to obtain the reasoning result. The reasoning result includes at least a general reasoning result and an emergency risk reasoning result. The general reasoning result represents non-emergency handling matters, and the emergency risk reasoning result represents emergency handling matters.

8. The IoT temperature control system for pre-cooled concrete pouring and curing according to claim 7, characterized in that, The device control module includes: A general algorithm processing unit is embedded with a deep reinforcement learning model. The deep reinforcement learning model is used to dynamically adjust the operating parameters of the pre-cooled concrete pouring and curing equipment according to the general inference results, and push the operating parameters and energy consumption parameters to the visualization display module. The emergency handling unit, based on the emergency risk reasoning result, determines that when the concrete surface temperature is lower than the first emergency threshold for concrete pouring and curing, it adjusts the heating mode of the pre-cooling concrete pouring and curing equipment, increases the heating power, and accelerates the concrete forming speed; based on the emergency risk reasoning result, determines that when the concrete surface temperature exceeds the second emergency threshold for concrete pouring and curing, it activates the cooling mode of the pre-cooling concrete pouring and curing equipment, increases the cooling power, and quickly reduces the concrete surface temperature to prevent cracks caused by overheating. The energy consumption equipment operation optimization unit increases the fan speed of the pre-cooling concrete pouring and curing equipment to promote cooling when the surface temperature of the concrete exceeds the temperature optimization threshold for concrete pouring; and decreases the fan speed of the pre-cooling concrete pouring and curing equipment to promote heat retention when the surface temperature of the concrete is below the temperature optimization threshold for concrete pouring.

9. A method for IoT-based temperature control in the pouring and curing of pre-cooled concrete, used to implement the IoT-based temperature control system for the pouring and curing of pre-cooled concrete as described in any one of claims 1-8, characterized in that, Includes the following steps: S1. The area delineation module automatically delineates multiple separate pre-cooled concrete pouring and curing temperature control zones based on the spatial layout of the poured body structure. The area delineation module outputs the delineation information of each zone to the temperature control module, the environmental data acquisition module, and the data analysis and reasoning module. S2. Use the temperature control module to work in conjunction with the area delineation module to control the temperature of each pre-cooled concrete pouring and curing temperature control area, feed the temperature back to the data analysis and inference module in real time, and receive the adjustment suggestions given by the data analysis and inference module in real time. S3. Use the environmental data acquisition module to monitor the environmental data of each pre-cooled concrete pouring and curing temperature control area, including humidity and light intensity, and transmit the environmental data to the data analysis and inference module in real time. S4. The data analysis and reasoning module is used to analyze and reason about the received area delineation results, temperature and environmental data. The temperature, characteristics of pre-cooled concrete and environmental data of each pre-cooled concrete pouring and curing temperature control area are analyzed. Then, the data is further reasoned based on the data reasoning network model, and control commands are output to the equipment control module and temperature control module. S5. The equipment control module dynamically adjusts the operating parameters of the pre-cooled concrete pouring and curing equipment according to the control instructions of the data analysis and reasoning module, and pushes the operating parameter information of the pre-cooled concrete pouring and curing equipment to the visualization display module. The pre-cooled concrete pouring and curing equipment includes concrete pouring equipment and concrete curing equipment. S6. Use a visualization display module to display the temperature of each of the pre-cooled concrete pouring and curing temperature control zones, the environmental data, and the operating parameter information of the pre-cooled concrete pouring and curing equipment, and allow the operator to manually set the temperature parameters, and transmit the operator's manual setting command to the temperature control module.

Images