A concrete temperature field intelligent monitoring and maintenance method and system

Abstract

The application discloses a kind of concrete temperature field intelligent monitoring and maintenance method and system, comprising: dividing initial boundary region, setting depth temperature measuring point, collecting multilayer temperature data and environmental and equipment data;Calculate temperature gradient and change rate, merge or split area, determine equivalent boundary thermal resistance parameter;Build multilayer virtual shell structure, generate virtual shell temperature data and temperature domain data;Input multi-source data to improved ContiFormer network, modeling continuous temperature propagation to obtain evolution result;Fusion temperature measurement and shell data, reconstruct three-dimensional temperature field and extract change characteristic parameter;Build future heat flow propagation cone, predict overlap risk to generate maintenance control instruction.The application realizes accurate monitoring and partition predictive maintenance control to concrete temperature field by dynamic boundary identification, multilayer virtual shell temperature domain construction and temperature evolution prediction.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring and curing control technology for concrete structures, and in particular to a method and system for intelligent monitoring and curing of concrete temperature field. Background Technology

[0002] During the large-volume pouring and early curing stages, concrete structures undergo significant hydration heat reactions, leading to a continuous rise in internal temperature. Meanwhile, the surface of the structure dissipates heat due to environmental factors such as temperature, wind speed, humidity, and sunlight, creating a distinct temperature gradient within the structure. When this temperature gradient and the internal-external temperature difference exceed a certain range, thermal stress can easily develop within the concrete, leading to temperature cracks. Therefore, real-time monitoring of the concrete temperature field and implementing appropriate curing controls based on temperature changes are crucial technical aspects for ensuring the durability and safety of concrete structures. Currently, in engineering practice, temperature sensors are typically embedded within the concrete structure to acquire temperature data, and curing measures such as spraying, covering, or ventilation are implemented in conjunction with empirical thresholds or simple temperature control strategies to reduce the risk of temperature cracks.

[0003] Existing methods for concrete temperature monitoring and curing are mostly based on temperature analysis within fixed monitoring areas and single-layer boundary conditions. This involves setting up several temperature measuring points within the structure, collecting temperature data to estimate internal concrete temperature changes, and then manually or semi-automatically adjusting curing measures based on the monitoring results. While these methods can reflect concrete temperature changes to some extent, they typically rely on data from only a small number of monitoring points, making it difficult to accurately describe the true temperature distribution of the concrete structure within a spatial range. Current technologies generally ignore the dynamic differences in heat dissipation from the structural surface caused by environmental changes, failing to effectively identify boundary heat transfer characteristics, leading to discrepancies between temperature field analysis results and actual conditions.

[0004] Existing technologies for temperature prediction and curing control mostly employ a delayed response approach, meaning that curing and control measures are only initiated after an abnormal temperature or excessive temperature difference is detected. This lack of ability to predict future temperature trends and the difficulty in analyzing the heat propagation path within the structure prevents the implementation of targeted curing measures in advance. When environmental disturbances are significant or local heat dissipation conditions change, sudden temperature changes or temperature stress concentrations can easily occur, affecting the overall quality of the concrete structure.

[0005] Therefore, how to provide a method and system for intelligent monitoring and curing of concrete temperature field is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose an intelligent monitoring and curing method and system for concrete temperature fields. This invention constructs dynamic boundary partitions, establishes a virtual shell temperature domain, and introduces an improved ContiFormer network to jointly model the internal temperature of concrete and the boundary heat dissipation state. Based on the temperature field change characteristics, it constructs a future heat flow propagation cone, enabling predictive analysis of the concrete temperature field and zoned pulse curing control. This invention can dynamically identify changes in the heat dissipation state of the concrete structure surface, accurately reconstruct the three-dimensional temperature field of concrete, and predict the heat propagation trend within the structure in advance, thereby achieving targeted curing control. It has the advantages of high accuracy in temperature field analysis, strong timeliness in curing control, and good effect in preventing temperature cracks.

[0007] A method for intelligent monitoring and curing of concrete temperature field according to an embodiment of the present invention includes: Multiple initial boundary regions are defined on the surface of the concrete structure to be monitored. Multiple depth temperature measurement points are set along the thickness direction in each initial boundary region to obtain multi-layer temperature measurement data. At the same time, environmental disturbance parameters and curing equipment operation status data are collected. Based on multi-layer temperature measurement data, calculate the temperature gradient and temperature change rate of each initial boundary region, merge or split the initial boundary regions to obtain dynamic boundary partitions and determine the equivalent boundary thermal resistance parameters. Based on the equivalent boundary thermal resistance parameters and the surface temperature measurement points, instantaneous boundary response shells, boundary delay response shells, and disturbance reflection shells are constructed for each dynamic boundary zone outside the concrete structure boundary. The shell temperature values ​​and shell spacing corresponding to each virtual shell are calculated, virtual shell temperature data is generated, and multi-layer virtual shell temperature domain data is formed. Multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and curing equipment operating status data are input into the improved ContiFormer network to model the continuous-time temperature propagation process between the internal temperature of concrete and the virtual shell temperature, and obtain the temperature evolution results at multiple future moments. The three-dimensional temperature field of concrete is reconstructed based on multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and the characteristic parameters of temperature field change in each dynamic boundary zone are extracted. Based on the characteristic parameters of temperature field change and the equivalent boundary thermal resistance parameters, a future heat flow propagation cone is constructed. When the predicted future heat flow propagation cone overlaps with the deep region of concrete within the future control period, a curing control command is generated to implement zoned pulse curing control on the concrete structure.

[0008] Optionally, the plurality of depth temperature measurement points include surface temperature measurement points, near-surface temperature measurement points, middle layer temperature measurement points, and core layer temperature measurement points arranged sequentially along the concrete thickness direction. The environmental disturbance parameters include ambient temperature, ambient humidity, wind speed, and solar radiation intensity. The maintenance equipment operation status data include the operation status of the spray device, the operation status of the covering and insulation device, and the operation status of the ventilation device.

[0009] Optionally, obtaining the dynamic boundary partition and determining the equivalent boundary thermal resistance parameters includes: The temperature values ​​of each initial boundary region at different depths at the current sampling time are extracted from the multi-layer temperature measurement data, and the corresponding temperature sequences are formed according to the depth order. The temperature gradient along the concrete thickness direction of each initial boundary region is calculated based on the temperature sequence. The temperature gradient is the ratio of the temperature difference between adjacent depth temperature measurement points to the corresponding depth spacing. The temperature change rate is calculated based on the temperature change at adjacent sampling times of the same depth temperature measurement point. The temperature change rate is the ratio of the difference between the current temperature and the previous sampling temperature to the sampling time interval. The boundary heat dissipation consistency index is determined based on the temperature gradient difference and temperature change rate difference between adjacent initial boundary regions. The boundary heat dissipation consistency index is compared with a preset consistency threshold. When the boundary heat dissipation consistency index is less than or equal to the preset consistency threshold, the corresponding initial boundary regions are merged. When the boundary heat dissipation consistency index is greater than the preset consistency threshold, the regions are kept independent or split, resulting in multiple dynamic boundary partitions. The corresponding boundary heat flow state is determined based on the surface temperature, internal temperature, and environmental disturbance parameters of each dynamic boundary zone, and the equivalent boundary thermal resistance parameter of each dynamic boundary zone is determined based on the temperature difference between the surface temperature and the internal temperature and the boundary heat flow state.

[0010] Optionally, generating virtual shell temperature data and forming multi-layer virtual shell temperature domain data includes: Read the equivalent boundary thermal resistance parameters, surface temperature, internal temperature and environmental disturbance parameters, and use the structural surface of each dynamic boundary partition as the reference boundary to establish a virtual outward expansion path outside the reference boundary along the boundary normal direction. An instant boundary response shell is constructed along a virtual outward expansion path. The instant boundary response shell is formed by offsetting the surface boundary of the dynamic boundary partition outward by a first preset distance along the normal direction. The heat transfer difference between the surface temperature and the internal temperature, the current environmental disturbance parameters, and the operating status of the maintenance equipment are jointly mapped to the instant boundary response shell to generate the first layer shell temperature data that characterizes the instantaneous heat dissipation state of the current boundary. A boundary delay response shell is constructed outside the instantaneous boundary response shell. The boundary delay response shell is formed by offsetting the instantaneous boundary response shell outward by a second preset distance along the same normal direction. The surface temperature change, internal temperature change and environmental disturbance change at the current time and the previous time are superimposed on the boundary delay response shell to generate the temperature data of the second shell that characterizes the boundary heat exchange hysteresis effect. A disturbance reflection shell is constructed outside the boundary delay response shell. The disturbance reflection shell is formed by offsetting the boundary delay response shell outward by a third preset distance along the same normal direction. The boundary heat flow offset caused by environmental disturbance changes, the thermal resistance change trend of dynamic boundary partitions, and the changes in the operating status of maintenance equipment are all mapped to the disturbance reflection shell to generate the third shell temperature data characterizing the influence of external disturbances transmitted back to the boundary. The temperature data of the first, second, and third shell layers, along with the hierarchical order, normal spacing, and partition position of each shell relative to the reference boundary, are correlated and encoded to generate virtual shell temperature data. The virtual shell temperature data is then combined with the spatial position data of each shell to form multi-layer virtual shell temperature domain data corresponding to the dynamic boundary partition.

[0011] Optionally, obtaining the temperature evolution results at multiple future time points includes: The multi-layer temperature measurement data, virtual shell temperature domain data, environmental disturbance parameters and maintenance equipment operation status data of each dynamic boundary partition are aligned according to the timestamp. Missing sampling points are filled by linear interpolation and arranged in sequence as sensing temperature sequence, virtual shell temperature sequence, and environmental and equipment status sequence to construct the input tensor. The input tensor is fed into an improved ContiFormer network, which consists of a continuous-time coding layer, a first coding block, a boundary-aware gating layer, a second coding block, a multi-scale dilation aggregation layer, a third coding block, a thermal residual coupling layer, and a prediction decoding head, which are sequentially connected in series. The continuous-time coding layer adopts adaptive time coding based on cumulative heat flux. The boundary-aware gating layer is inserted after the output of the first coding block and adjusts the attention channel weights in real time according to the temperature difference between the surface and the interior. A multi-scale dilatation aggregation layer is inserted between the second and third coding blocks. The multi-scale dilatation aggregation layer uses minute-level and day-night-level dilated convolution kernels to capture multi-timescale temperature patterns and feeds them back to the output of the second coding block in a residual manner. The output of the third coding block is fused with the operating status vector of the maintenance equipment through the thermal residual coupling layer. The thermal residual coupling layer adopts a gated residual structure to superimpose the heat input or heat extraction generated by the equipment operation onto the temperature feature flow. The predictive decoder outputs the temperature sequences of temperature measurement points at different depths in each dynamic boundary partition and the corresponding virtual shell temperature sequences within the next three control cycles, forming the temperature evolution results at multiple future moments; The improved ContiFormer network was obtained through offline supervised training. The training samples consisted of historical temperature sequences and corresponding future temperature labels. The training process employed a joint optimization of temperature mean square error loss and boundary thermal equilibrium consistency loss.

[0012] Optionally, the extraction of temperature field change characteristic parameters for each dynamic boundary partition includes: Read multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and associate the temperature values ​​of temperature measurement points at different depths with the virtual shell temperature at the corresponding time according to dynamic boundary partitioning to form a temperature data set that includes spatial location, depth location, and time series. Based on the spatial relationship of each temperature measurement point in the temperature data set, the internal temperature measurement node and the corresponding virtual shell node are spatially mapped using dynamic boundary partitioning as the basic calculation unit, so that the virtual shell node, as a boundary constraint, together with the internal temperature measurement node, constitutes the temperature field reconstruction node set. Based on the spatial coordinates and corresponding temperature values ​​of each node in the temperature field reconstruction node set, the continuous temperature distribution of the internal space is calculated within the dynamic boundary partition range to obtain the three-dimensional temperature field of the concrete structure at different spatial locations and at different times. In a three-dimensional temperature field, the temperature changes between different spatial locations are extracted according to the spatial range of dynamic boundary partitions, and the spatial temperature gradient of each dynamic boundary partition is determined based on the temperature difference between adjacent spatial locations and the corresponding spatial distance. The rate of temperature change is determined based on the temperature changes at the same spatial location during continuous sampling, and the internal and external temperature differences are determined based on the temperature difference between the interior and surface of the dynamic boundary partition, thus forming the characteristic parameters of temperature field changes for each dynamic boundary partition.

[0013] Optionally, the generation of curing control instructions, which implements zoned pulse curing control for the concrete structure, includes: By reading the characteristic parameters of temperature field changes, equivalent boundary thermal resistance parameters, and temperature evolution results, the locations of high surface heat flux response, internal temperature sensitive locations, and deep key regions are determined in each dynamic boundary partition. The initial opening degree of the cone is determined by taking the high heat flux response position on the surface as the starting point of the cone, the dominant temperature gradient direction from the surface to the interior as the main propagation direction, the equivalent boundary thermal resistance parameter of the corresponding dynamic boundary partition as the initial opening degree of the cone, and the direction of cone axis deflection is determined by the temperature advancement direction at multiple consecutive moments in the temperature evolution result, thus forming the initial heat flux propagation cone skeleton of the corresponding dynamic boundary partition. The cone skeleton propagates along the initial heat flow, and the cone cross-section is extrapolated layer by layer according to each future control moment. The advancing distance of the cone along the main propagation direction is determined according to the temperature change rate at the corresponding moment. The expansion depth of the cone in the thickness direction is determined according to the internal and external temperature difference. The expansion width of the cone in the lateral direction is determined according to the spatial temperature gradient distribution. The cone cross-section sequence corresponding to each future control moment is constructed. The sequence of cone sections corresponding to each future control moment is stacked sequentially along time to form a future heat flow propagation cone with time-progression properties, wherein the future heat flow propagation cone includes a main propagation cone and a lateral disturbance branch cone; Determine whether the main propagation cone or the lateral disturbance branch cone in the future heat flow propagation cone will spatially overlap with the deep critical area during the future control period. When spatial overlap occurs, determine the maintenance control intensity, maintenance control start time and maintenance control duration of the dynamic boundary zone based on the overlap time, overlap depth, overlap range and the equivalent boundary thermal resistance parameters of the corresponding dynamic boundary zone. Based on the curing control intensity, curing control start time, and curing control duration, corresponding curing control instructions for dynamic boundary zones are generated. Curing control instructions are output separately for each dynamic boundary zone to implement zoned pulse curing control for concrete structures.

[0014] A concrete temperature field intelligent monitoring and curing system according to an embodiment of the present invention includes: The temperature acquisition module is used to acquire multi-layer temperature data, and at the same time, it collects environmental disturbance parameters and maintenance equipment operating status data. The boundary identification module is used to calculate the temperature gradient and temperature change rate of each initial boundary region based on multi-layer temperature measurement data, and to determine the equivalent boundary thermal resistance parameters. The virtual shell module is used to calculate the virtual shell temperature value and spacing based on the equivalent boundary thermal resistance parameters and the surface temperature measurement point temperature, generate virtual shell temperature data, and form multi-layer virtual shell temperature domain data. The temperature prediction module is used to input multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and maintenance equipment operating status data into the improved ContiFormer network to obtain temperature evolution results. The temperature field reconstruction module is used to reconstruct the three-dimensional temperature field of concrete based on multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and to extract characteristic parameters of temperature field changes. The maintenance control module is used to construct the future heat flow propagation cone based on the temperature field change characteristic parameters and equivalent boundary thermal resistance parameters, and generate maintenance control commands to implement zoned pulse maintenance control.

[0015] The beneficial effects of this invention are: Compared with existing technologies, this invention divides the concrete structure surface into initial boundary regions and combines multi-layer temperature measurement data to dynamically merge and split the boundary regions. This enables real-time identification of heat dissipation differences in different regions, thereby forming dynamic boundary partitions and determining the corresponding equivalent boundary thermal resistance parameters. This avoids the boundary condition distortion problem caused by fixed monitoring areas in traditional technologies, allowing the heat dissipation state of the concrete surface to be dynamically updated with changes in environmental disturbances. This improves the accuracy of temperature monitoring and boundary heat transfer identification, providing more realistic boundary conditions for subsequent temperature field analysis.

[0016] This invention constructs an instantaneous boundary response shell, a boundary delayed response shell, and a disturbance reflection shell outside the boundary of a concrete structure, forming multi-layered virtual shell temperature domain data. This data is then combined with multi-layered temperature measurement data and temperature evolution results to reconstruct the three-dimensional temperature field of the concrete. By jointly constraining the internal temperature information with the virtual shell temperature information, the influence of environmental disturbances on the heat dissipation process of the concrete surface is more accurately reflected, resulting in a temperature field distribution closer to the real state and improving the spatial integrity and temperature distribution accuracy of the temperature field analysis.

[0017] This invention utilizes an improved ContiFormer network to model the continuous-time temperature propagation process between the internal temperature of concrete and the temperature of a virtual shell, obtaining temperature evolution results at multiple future time points. Based on temperature field change characteristic parameters and equivalent boundary thermal resistance parameters, a future heat flow propagation cone is constructed. When it is predicted that the future heat flow propagation cone may overlap with the deep regions of the concrete, the system can generate curing control commands in advance and implement zoned pulse curing control. This achieves predictive regulation of the concrete temperature field, reduces local temperature abrupt changes and temperature stress concentration, and improves the temperature crack prevention capability and overall construction quality of the concrete structure. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for intelligent monitoring and curing of concrete temperature field proposed in this invention. Figure 2 This is a schematic diagram of the structure of an intelligent monitoring and curing system for concrete temperature field proposed in this invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0020] refer to Figure 1A method for intelligent monitoring and curing of concrete temperature field, comprising: Multiple initial boundary regions are defined on the surface of the concrete structure to be monitored. Multiple depth temperature measurement points are set along the thickness direction in each initial boundary region to obtain multi-layer temperature measurement data. At the same time, environmental disturbance parameters and curing equipment operation status data are collected. Based on multi-layer temperature measurement data, calculate the temperature gradient and temperature change rate of each initial boundary region, merge or split the initial boundary regions to obtain dynamic boundary partitions and determine the equivalent boundary thermal resistance parameters. Based on the equivalent boundary thermal resistance parameters and the surface temperature measurement points, instantaneous boundary response shells, boundary delay response shells, and disturbance reflection shells are constructed for each dynamic boundary zone outside the concrete structure boundary. The shell temperature values ​​and shell spacing corresponding to each virtual shell are calculated, virtual shell temperature data is generated, and multi-layer virtual shell temperature domain data is formed. Multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and curing equipment operating status data are input into the improved ContiFormer network to model the continuous-time temperature propagation process between the internal temperature of concrete and the virtual shell temperature, and obtain the temperature evolution results at multiple future moments. The three-dimensional temperature field of concrete is reconstructed based on multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and the characteristic parameters of temperature field change in each dynamic boundary zone are extracted. Based on the characteristic parameters of temperature field change and the equivalent boundary thermal resistance parameters, a future heat flow propagation cone is constructed. When the predicted future heat flow propagation cone overlaps with the deep region of concrete within the future control period, a curing control command is generated to implement zoned pulse curing control on the concrete structure.

[0021] In this embodiment, the plurality of depth temperature measurement points include surface temperature measurement points, near-surface temperature measurement points, middle layer temperature measurement points, and core layer temperature measurement points arranged sequentially along the concrete thickness direction. Environmental disturbance parameters include ambient temperature, ambient humidity, wind speed, and solar radiation intensity. The operating status data of the curing equipment include the operating status of the spray device, the operating status of the covering and insulation device, and the operating status of the ventilation device. Among them, the surface temperature measurement points are set in the depth range of 0.05 meters to 0.2 meters from the outer surface of the concrete, the near-surface temperature measurement points are set in the depth range of 0.2 meters to 0.8 meters from the outer surface of the concrete, the middle layer temperature measurement points are set in the depth range of 0.8 meters from the outer surface of the concrete to half the thickness of the concrete, and the core layer temperature measurement points are set in the depth range of more than half the thickness of the concrete and close to the center area of ​​the concrete thickness.

[0022] In this embodiment, obtaining the dynamic boundary partition and determining the equivalent boundary thermal resistance parameters includes: The temperature values ​​of each initial boundary region at different depths at the current sampling time are extracted from the multi-layer temperature measurement data, and the corresponding temperature sequences are formed according to the depth order. The temperature gradient along the concrete thickness direction of each initial boundary region is calculated based on the temperature sequence. The temperature gradient is the ratio of the temperature difference between adjacent depth temperature measurement points to the corresponding depth spacing. The temperature change rate is calculated based on the temperature change at adjacent sampling times of the same depth temperature measurement point. The temperature change rate is the ratio of the difference between the current temperature and the previous sampling temperature to the sampling time interval. The boundary heat dissipation consistency index is determined based on the temperature gradient difference and temperature change rate difference between adjacent initial boundary regions. This index is then compared to a preset consistency threshold. When the boundary heat dissipation consistency index is less than or equal to the preset threshold, the corresponding initial boundary regions are merged. When the boundary heat dissipation consistency index is greater than the preset threshold, the regions remain independent or are split, resulting in multiple dynamic boundary partitions. The determination of the boundary heat dissipation consistency index is specifically as follows: First, extract the temperature gradient of adjacent initial boundary regions at the same sampling time, and calculate the gradient difference between them to characterize the difference in heat dissipation intensity of adjacent initial boundary regions along the thickness direction. The temperature change rate of adjacent initial boundary regions at the same sampling time is extracted, and the difference in the rate of change between the two is calculated to characterize the consistency of temperature evolution of adjacent initial boundary regions in the current time period. The gradient difference value and the rate of change difference value are weighted and fused to obtain the boundary heat dissipation consistency index of the corresponding adjacent initial boundary regions. The corresponding boundary heat flow state is determined based on the surface temperature, internal temperature, and environmental disturbance parameters of each dynamic boundary zone, and the equivalent boundary thermal resistance parameter of each dynamic boundary zone is determined based on the temperature difference between the surface temperature and the internal temperature and the boundary heat flow state.

[0023] In this embodiment, generating virtual shell temperature data and forming multi-layer virtual shell temperature domain data includes: Read the equivalent boundary thermal resistance parameters, surface temperature, internal temperature and environmental disturbance parameters, and use the structural surface of each dynamic boundary partition as the reference boundary to establish a virtual outward expansion path outside the reference boundary along the boundary normal direction. An instantaneous boundary response shell is constructed along a virtual outward expansion path. The instantaneous boundary response shell is formed by offsetting the surface boundary of the dynamic boundary partition outward along the normal direction by a first preset distance. The heat transfer difference between the surface temperature and the internal temperature, the current environmental disturbance parameters, and the operating status of the maintenance equipment are jointly mapped to the instantaneous boundary response shell to generate the first layer shell temperature data characterizing the instantaneous heat dissipation state of the current boundary. The first preset distance is determined according to the equivalent boundary thermal resistance parameters of the corresponding dynamic boundary partition. A boundary delay response shell is constructed outside the instantaneous boundary response shell. The boundary delay response shell is formed by offsetting the instantaneous boundary response shell outward by a second preset distance along the same normal direction. The surface temperature change, internal temperature change and environmental disturbance change at the current time and the previous time are time-series superimposed and mapped to the boundary delay response shell to generate the temperature data of the second shell that characterizes the boundary heat exchange hysteresis effect. The second preset distance is determined according to the equivalent boundary thermal resistance change amplitude of the corresponding dynamic boundary partition at multiple consecutive sampling times. A disturbance reflection shell is constructed outside the boundary delay response shell. The disturbance reflection shell is formed by offsetting the boundary delay response shell outward by a third preset distance along the same normal direction. The boundary heat flow offset caused by environmental disturbance changes, the thermal resistance change trend of dynamic boundary partitions, and the changes in the operating status of maintenance equipment are all mapped to the disturbance reflection shell to generate the temperature data of the third shell layer characterizing the influence of external disturbances transmitted back to the boundary. The third preset distance is determined according to the intensity and direction of change of environmental disturbance parameters. The temperature data of the first, second, and third shell layers, along with the hierarchical order, normal spacing, and partition position of each shell relative to the reference boundary, are correlated and encoded to generate virtual shell temperature data. The virtual shell temperature data is then combined with the spatial position data of each shell to form multi-layer virtual shell temperature domain data corresponding to the dynamic boundary partition.

[0024] In this embodiment, obtaining the temperature evolution results at multiple future times includes: The multi-layer temperature measurement data, virtual shell temperature domain data, environmental disturbance parameters and maintenance equipment operation status data of each dynamic boundary partition are aligned according to the timestamp. Missing sampling points are filled by linear interpolation and arranged in sequence as sensing temperature sequence, virtual shell temperature sequence, and environmental and equipment status sequence to construct the input tensor. The input tensor is fed into an improved ContiFormer network, which consists of a continuous-time coding layer, a first coding block, a boundary-aware gating layer, a second coding block, a multi-scale dilation aggregation layer, a third coding block, a thermal residual coupling layer, and a prediction decoder, sequentially connected in series. The continuous-time coding layer employs adaptive time coding based on cumulative heat flux. The boundary-aware gating layer is inserted after the output of the first coding block and adjusts the attention channel weights in real time according to the temperature difference between the surface and the interior. Specifically, the continuous-time coding layer employs adaptive time coding based on cumulative heat flux. Based on the sampling time interval between adjacent sampling times, multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and maintenance equipment operation status data of each dynamic boundary zone are read, and the temperature difference change information between the surface temperature and the internal temperature is extracted. Based on the heat exchange effects caused by the changes in surface and internal temperature difference, environmental disturbance parameters, and operating status of maintenance equipment in each time step, the heat flow changes in each time step are gradually accumulated to obtain the cumulative heat flow characterization value at the corresponding sampling time. The cumulative heat flow characterization value is jointly mapped with the sampling time interval of the corresponding sampling time to generate an adaptive time code for the corresponding sampling time. The adaptive time code is then appended to the corresponding temperature feature sequence in the input tensor. A multi-scale dilation aggregation layer is inserted between the second and third coding blocks. This layer uses minute-level and day-night dilated convolutional kernels to capture multi-timescale temperature patterns and feeds them back to the output of the second coding block as residuals. Specifically, the multi-scale dilation aggregation layer uses minute-level and day-night dilated convolutional kernels to capture multi-timescale temperature patterns. The temperature feature sequence output from the second coding block is input in parallel into multiple dilated convolution branches with different expansion rates. The dilated convolution branch with a smaller expansion rate is used to extract minute-level temperature fluctuation features to characterize short-term temperature changes caused by the start-up and shutdown of the spray device, the switching of the ventilation device, and local environmental disturbances. The dilated convolution branches corresponding to the diurnal time scale are used to extract temperature change features within the diurnal cycle to characterize the long-term temperature evolution trend caused by changes in ambient temperature, solar radiation intensity and overall heat dissipation. The temperature features output by each dilated convolution branch are spliced ​​and aggregated to form multi-scale temperature features that include short-term and long-term temperature patterns. The multi-scale temperature features are residually superimposed with the original temperature features output by the second coding block, so that the aggregated temperature features enhance the ability to represent temperature patterns at different time scales while retaining the original continuous time propagation information, and are then output to the third coding block. The output of the third coding block is fused with the operating status vector of the maintenance equipment through the thermal residual coupling layer. The thermal residual coupling layer adopts a gated residual structure to superimpose the heat input or heat extraction generated by the equipment operation onto the temperature feature flow. The predictive decoder outputs temperature sequences at different depths of temperature measurement points in each dynamic boundary partition and the corresponding virtual shell temperature sequences for the next three control cycles, forming the temperature evolution results at multiple future time points. Specifically, the formation of these temperature evolution results at multiple future time points is as follows: Based on the temperature sequence of temperature measurement points at different depths in each dynamic boundary partition output by the prediction decoder, the prediction time range is divided into multiple consecutive time periods according to the preset control cycle, and the prediction time corresponding to each time period is determined. At each prediction time, the temperature values ​​of temperature measurement points at different depths of the corresponding dynamic boundary partition and the temperature values ​​of the corresponding virtual shell are extracted and combined according to the depth location and time order to form the temperature status data at the corresponding time. Arrange the temperature state data at each predicted time in chronological order to form a temperature sequence that reflects the changes in the internal temperature of concrete and the heat dissipation state at the boundary over time, thus forming the temperature evolution results for multiple future times. The improved ContiFormer network was obtained through offline supervised training. The training samples consisted of historical temperature sequences and corresponding future temperature labels. The training process employed a joint optimization of temperature mean square error loss and boundary thermal equilibrium consistency loss.

[0025] In this embodiment, the extraction of temperature field change characteristic parameters for each dynamic boundary partition includes: Read multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and associate the temperature values ​​of temperature measurement points at different depths with the virtual shell temperature at the corresponding time according to dynamic boundary partitioning to form a temperature data set that includes spatial location, depth location, and time series. Based on the spatial relationships of temperature measurement points in the temperature dataset, and using dynamic boundary partitioning as the basic computational unit, the internal temperature measurement nodes are spatially mapped to their corresponding virtual shell nodes. This allows the virtual shell nodes, acting as boundary constraints, to jointly form a temperature field reconstruction node set with the internal temperature measurement nodes. Specifically, the spatial mapping between the internal temperature measurement nodes and their corresponding virtual shell nodes is as follows: Based on the embedment depth of each temperature measuring point in the concrete structure and its planar position within the dynamic boundary partition, the position of the internal temperature measuring node in the three-dimensional coordinate space is determined, and the structural surface position of the corresponding dynamic boundary partition is used as the boundary reference position. Based on the temperature domain data of the multi-layer virtual shell, the spatial distance and hierarchical order of each virtual shell node relative to the boundary reference position are determined, and the position of each virtual shell node in three-dimensional space is determined according to the same coordinate reference system as the internal temperature measuring node. Based on the spatial distance relationship between the internal temperature measuring nodes and the virtual shell nodes and the range of the dynamic boundary partition, each virtual shell node is associated with the internal temperature measuring node along the corresponding boundary normal direction to establish a spatial mapping relationship between the internal temperature measuring nodes and the virtual shell nodes. Based on the spatial coordinates and corresponding temperature values ​​of each node in the temperature field reconstruction node set, continuous temperature distribution calculation is performed on the internal space within the dynamic boundary partition range to obtain the three-dimensional temperature field of the concrete structure at different spatial locations and time points. Specifically, the continuous temperature distribution calculation within the dynamic boundary partition range is as follows: Based on the spatial coordinates of each internal temperature measurement node and the virtual shell node in the temperature field reconstruction node set, a three-dimensional spatial mesh is established within the dynamic boundary partition range, and the temperature value of each node is used as the initial temperature data of the corresponding mesh node. Based on the spatial distance relationship between adjacent grid nodes, the temperature of the internal grid nodes is spatially propagated and calculated, so that the temperature of the internal grid nodes is simultaneously constrained by the temperatures of the surrounding temperature measuring nodes and the virtual shell nodes, thus obtaining the spatial temperature value of each grid node. Based on the time sequence and temperature evolution results, the temperature of the grid nodes at each time point is updated, and the temperature of each grid node is combined according to its spatial location to form the three-dimensional temperature field distribution of concrete at the corresponding time point. In a three-dimensional temperature field, the temperature changes between different spatial locations are extracted according to the spatial range of dynamic boundary partitions, and the spatial temperature gradient of each dynamic boundary partition is determined based on the temperature difference between adjacent spatial locations and the corresponding spatial distance. The rate of temperature change is determined based on the temperature changes at the same spatial location during continuous sampling, and the internal and external temperature differences are determined based on the temperature difference between the interior and surface of the dynamic boundary partition, thus forming the characteristic parameters of temperature field changes for each dynamic boundary partition.

[0026] In this embodiment, the generation of curing control instructions to implement zoned pulse curing control of the concrete structure includes: By reading the characteristic parameters of temperature field changes, equivalent boundary thermal resistance parameters, and temperature evolution results, the locations of high surface heat flux response, internal temperature sensitive locations, and deep key regions are determined in each dynamic boundary partition. The initial opening degree of the cone is determined by taking the high heat flux response position on the surface as the starting point of the cone, the dominant temperature gradient direction from the surface to the interior as the main propagation direction, the equivalent boundary thermal resistance parameter of the corresponding dynamic boundary partition as the initial opening degree of the cone, and the direction of cone axis deflection is determined by the temperature advancement direction at multiple consecutive moments in the temperature evolution result, thus forming the initial heat flux propagation cone skeleton of the corresponding dynamic boundary partition. The cone skeleton propagates along the initial heat flow, and the cone cross-section is extrapolated layer by layer according to each future control moment. The advancing distance of the cone along the main propagation direction is determined according to the temperature change rate at the corresponding moment. The expansion depth of the cone in the thickness direction is determined according to the internal and external temperature difference. The expansion width of the cone in the lateral direction is determined according to the spatial temperature gradient distribution. The cone cross-section sequence corresponding to each future control moment is constructed. The sequence of cone sections corresponding to each future control moment is stacked sequentially along time to form a future heat flow propagation cone with time-progression properties. The future heat flow propagation cone includes a main propagation cone and lateral disturbance branch cones. The formation of the main propagation cone is as follows: The location with the most significant surface temperature change in each dynamic boundary partition is taken as the starting point of heat flow propagation, and the main direction of the temperature gradient at the starting point of heat flow propagation is taken as the initial propagation direction of heat flow propagation, thus determining the axial direction of the main propagation cone. Based on the temperature change rate and equivalent boundary thermal resistance parameters at different time points in the temperature evolution results, the propagation distance of heat along the axial direction inside the structure is determined, and the cone length is gradually extended along the axial direction at each predicted time. Based on the internal and external temperature difference and spatial temperature gradient distribution in the characteristic parameters of temperature field change, the expansion range of the cone in the lateral direction is determined, and the corresponding cone cross-section is formed at each time point, thereby constructing the main propagation cone in chronological order; The formation of the lateral disturbance branch cone is specifically as follows: As the main propagation cone advances along time, the spatial temperature gradient distribution between adjacent dynamic boundary partitions is read in real time. When the temperature gradient direction between adjacent partitions deflects relative to the main propagation direction, the corresponding position is determined as the lateral disturbance trigger position. Based on the deflection direction of the temperature gradient at the trigger location of the lateral disturbance and the wind speed direction and solar radiation change direction in the environmental disturbance parameters, the deflection and propagation direction of the heat flow in the lateral space is determined, with the trigger location as the starting point of the lateral branch. Following the same time-progression method as the main propagation cone, the cone cross-section is gradually expanded in the deflection propagation direction. The expansion range of the branch cone is determined according to the temperature change rate and spatial temperature gradient at the corresponding time, forming a lateral disturbance branch cone. Determine whether the main propagation cone or lateral disturbance branch cone in the future heat flow propagation cone spatially overlaps with the deep critical region during the future control period. If spatial overlap occurs, determine the maintenance control intensity, start time, and duration of the dynamic boundary zone based on the overlap time, overlap depth, overlap range, and the equivalent boundary thermal resistance parameters of the corresponding dynamic boundary zone. Specifically, the determination of spatial overlap involves: Based on the spatial range of the main propagation cone and the lateral disturbance branch cone in the future heat flow propagation cone at each future control moment, the three-dimensional coverage area of ​​each cone inside the concrete structure is determined, and the spatial location range of the deep key area at the corresponding moment is determined simultaneously. The spatial location range of the cone's three-dimensional coverage area and the deep key area at each future control time is compared hourly. When the cone's coverage area enters the deep key area, or when the cone's boundary comes into contact with the deep key area's boundary, it is determined that there is spatial overlap at the corresponding time. The overlap time is determined by the earliest time when the spatial overlap occurs, the overlap depth is determined by the depth corresponding to the overlap position, and the overlap range is determined by the overlapping spatial range of the cone-covered area and the deep key area, thus completing the spatial overlap judgment. Based on the curing control intensity, curing control start time, and curing control duration, corresponding curing control instructions for dynamic boundary zones are generated. Curing control instructions are output separately for each dynamic boundary zone to implement zoned pulse curing control for concrete structures. The curing control instructions include spray cooling control instructions, covering insulation control instructions, or ventilation and heat dissipation control instructions.

[0027] refer to Figure 2 A concrete temperature field intelligent monitoring and curing system, comprising: The temperature acquisition module is used to acquire multi-layer temperature data, and at the same time, it collects environmental disturbance parameters and maintenance equipment operating status data. The boundary identification module is used to calculate the temperature gradient and temperature change rate of each initial boundary region based on multi-layer temperature measurement data, and to determine the equivalent boundary thermal resistance parameters. The virtual shell module is used to calculate the virtual shell temperature value and spacing based on the equivalent boundary thermal resistance parameters and the surface temperature measurement point temperature, generate virtual shell temperature data, and form multi-layer virtual shell temperature domain data. The temperature prediction module is used to input multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and maintenance equipment operating status data into the improved ContiFormer network to obtain temperature evolution results. The temperature field reconstruction module is used to reconstruct the three-dimensional temperature field of concrete based on multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and to extract characteristic parameters of temperature field changes. The maintenance control module is used to construct the future heat flow propagation cone based on the temperature field change characteristic parameters and equivalent boundary thermal resistance parameters, and generate maintenance control commands to implement zoned pulse maintenance control.

[0028] Example 1:

[0029] To verify the feasibility of this invention in practice, it was applied to the large-volume concrete foundation slab of a subway station in a city. The slab was 3.2m thick, and the volume of concrete poured in a single operation was approximately 360m³. Due to high summer temperatures and large diurnal temperature variations, the concrete is prone to significant internal temperature accumulation due to the heat of hydration. Simultaneously, the surface of the structure is affected by environmental factors such as wind speed, humidity, and sunlight, leading to heat dissipation and a clear temperature gradient between the interior and surface. In previous projects, construction units primarily relied on a small number of temperature sensors for monitoring and judged based on experience whether to activate spray systems or cover the surface for insulation. This method is insufficient for timely identification of changes in heat dissipation in different areas and cannot predict future temperature trends, easily leading to excessive localized temperature differences and temperature cracks.

[0030] In this embodiment, the intelligent monitoring and curing method for concrete temperature field proposed in this invention is used to monitor and control the temperature of the station's base slab. First, an initial boundary area is divided on the base slab surface using a grid of approximately 2m × 2m. Multiple depth temperature measurement points are then deployed along the concrete thickness direction in each area, at depths of 0.1m, 0.6m, 1.4m, 2.2m, and 3.0m from the surface, to acquire multi-layer temperature data. Simultaneously, environmental monitoring equipment is deployed at the construction site to collect environmental disturbance parameters such as ambient temperature, air humidity, wind speed, and solar radiation intensity, and to record the operational status of the spray system and insulation covering equipment. All monitoring data is transmitted in real-time to the control server via a wireless acquisition terminal.

[0031] The system calculates the temperature gradient and rate of change of each initial boundary region based on the collected multi-layer temperature measurement data. Regions are automatically merged when their temperature change trends are basically consistent, and split when the temperature change of a local area is significantly abnormal, thus forming dynamic boundary partitions. In this embodiment, nine dynamic boundary partitions were formed within 24 hours after the base slab was poured. The area near the south construction passage was further subdivided into three sub-regions due to higher wind speeds. Subsequently, based on the equivalent boundary thermal resistance parameters and surface temperature of each partition, the system constructs an instantaneous boundary response shell, a boundary delayed response shell, and a disturbance reflection shell outside the concrete structure boundary, generating multi-layer virtual shell temperature domain data. The multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and curing equipment operating status data are input into an improved ContiFormer network to model the continuous-time temperature propagation process between the internal concrete temperature and the virtual shell temperature, thereby obtaining the temperature evolution results for the next 6, 12, and 24 hours.

[0032] Based on the predicted temperature evolution, the system reconstructs the three-dimensional temperature field of the foundation slab concrete and extracts the characteristic parameters of temperature field changes in each dynamic boundary zone. During the monitoring process, 18 hours after pouring, the system predicted that the future heat flow propagation cone in the southern area would extend towards the central area of ​​the concrete within the next 5 hours, which would result in a large internal and external temperature difference if left uncontrolled. Therefore, the system automatically generated curing control instructions, initiating intermittent spray cooling and appropriately extending the insulation coverage time in this area. Actual monitoring results show that after control, the peak temperature in this area decreased from the originally predicted 69℃ to approximately 63℃, the temperature difference between the interior and the surface was controlled within 16℃, and no obvious temperature cracks appeared. Compared with adjacent construction sections using traditional monitoring methods, the method of this invention can identify abnormal temperature trends in advance and implement targeted curing control, resulting in a smoother temperature change process and a more uniform overall temperature distribution of the structure.

[0033] To further illustrate the technical effects of the present invention, temperature monitoring data at different depths were statistically analyzed and compiled within 48 hours after the base plate was poured.

[0034] Table 1. Temperature monitoring data of station floor concrete at different depths

[0035] As shown in Table 1, in the early stages after concrete pouring, the temperature at all depths generally showed an upward trend as the hydration heat reaction continued. Between 6 and 24 hours, the internal temperature of the concrete increased significantly, with the surface temperature rising from 30.5℃ to 39.1℃, while the temperature at a depth of 3.0m rose from 34.6℃ to 54.0℃. Due to the continuous accumulation of hydration heat and its slower heat dissipation, the temperature rise rate at deeper levels was higher than at the surface, thus creating a certain temperature gradient. Around 24 hours, the internal temperature reached a peak, with a temperature difference of 14.9℃ between the inside and outside, indicating that this stage is a critical period for temperature control.

[0036] After 24 hours, as the release of heat of hydration gradually weakened and the curing control measures took effect, the temperature at each depth began to gradually decrease. Between 30 and 48 hours, the temperature at a depth of 3.0m decreased from 54.0℃ to 42.2℃, the surface temperature decreased from 39.1℃ to 33.7℃, and the temperature difference between the inside and outside also decreased from 14.9℃ to 8.5℃, indicating that the temperature difference between the inside and the surface of the concrete gradually decreased and the temperature field tended to stabilize.

[0037] In summary, by using the method of this invention for dynamic monitoring and zoned curing control, the maximum internal temperature of concrete is controlled within a reasonable range, the temperature difference between the inside and outside is always kept within the allowable range of the project, the temperature change process at each depth is stable, the phenomenon of temperature stress concentration is effectively reduced, thereby reducing the risk of temperature cracks and improving the construction quality of concrete structures.

[0038] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent monitoring and curing of concrete temperature field, characterized in that, include: Multiple initial boundary regions are defined on the surface of the concrete structure to be monitored. Multiple depth temperature measurement points are set along the thickness direction in each initial boundary region to obtain multi-layer temperature measurement data. At the same time, environmental disturbance parameters and curing equipment operation status data are collected. Based on multi-layer temperature measurement data, calculate the temperature gradient and temperature change rate of each initial boundary region, merge or split the initial boundary regions to obtain dynamic boundary partitions and determine the equivalent boundary thermal resistance parameters. Based on the equivalent boundary thermal resistance parameters and the surface temperature measurement points, instantaneous boundary response shells, boundary delay response shells, and disturbance reflection shells are constructed for each dynamic boundary zone outside the concrete structure boundary. The shell temperature values ​​and shell spacing corresponding to each virtual shell are calculated, virtual shell temperature data is generated, and multi-layer virtual shell temperature domain data is formed. Multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and curing equipment operating status data are input into the improved ContiFormer network to model the continuous-time temperature propagation process between the internal temperature of concrete and the virtual shell temperature, and obtain the temperature evolution results at multiple future moments. The three-dimensional temperature field of concrete is reconstructed based on multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and the characteristic parameters of temperature field change in each dynamic boundary zone are extracted. Based on the characteristic parameters of temperature field change and the equivalent boundary thermal resistance parameters, a future heat flow propagation cone is constructed. When the predicted future heat flow propagation cone overlaps with the deep region of concrete within the future control period, a curing control command is generated to implement zoned pulse curing control on the concrete structure.

2. The method for intelligent monitoring and curing of concrete temperature field according to claim 1, characterized in that, The multiple depth temperature measurement points include surface temperature measurement points, near-surface temperature measurement points, middle layer temperature measurement points, and core layer temperature measurement points arranged sequentially along the concrete thickness direction. Environmental disturbance parameters include ambient temperature, ambient humidity, wind speed, and solar radiation intensity. The operating status data of the maintenance equipment include the operating status of the spray device, the operating status of the covering and insulation device, and the operating status of the ventilation device.

3. The method for intelligent monitoring and curing of concrete temperature field according to claim 1, characterized in that, The process of obtaining dynamic boundary partitioning and determining equivalent boundary thermal resistance parameters includes: The temperature values ​​of each initial boundary region at different depths at the current sampling time are extracted from the multi-layer temperature measurement data, and the corresponding temperature sequences are formed according to the depth order. The temperature gradient along the concrete thickness direction of each initial boundary region is calculated based on the temperature sequence. The temperature gradient is the ratio of the temperature difference between adjacent depth temperature measurement points to the corresponding depth spacing. The temperature change rate is calculated based on the temperature change at adjacent sampling times of the same depth temperature measurement point. The temperature change rate is the ratio of the difference between the current temperature and the previous sampling temperature to the sampling time interval. The boundary heat dissipation consistency index is determined based on the temperature gradient difference and temperature change rate difference between adjacent initial boundary regions. The boundary heat dissipation consistency index is compared with a preset consistency threshold. When the boundary heat dissipation consistency index is less than or equal to the preset consistency threshold, the corresponding initial boundary regions are merged. When the boundary heat dissipation consistency index is greater than the preset consistency threshold, the regions are kept independent or split, resulting in multiple dynamic boundary partitions. The corresponding boundary heat flow state is determined based on the surface temperature, internal temperature, and environmental disturbance parameters of each dynamic boundary zone, and the equivalent boundary thermal resistance parameter of each dynamic boundary zone is determined based on the temperature difference between the surface temperature and the internal temperature and the boundary heat flow state.

4. The method for intelligent monitoring and curing of concrete temperature field according to claim 1, characterized in that, The process of generating virtual shell temperature data and forming multi-layer virtual shell temperature domain data includes: Read the equivalent boundary thermal resistance parameters, surface temperature, internal temperature and environmental disturbance parameters, and use the structural surface of each dynamic boundary partition as the reference boundary to establish a virtual outward expansion path outside the reference boundary along the boundary normal direction. An instant boundary response shell is constructed along a virtual outward expansion path. The instant boundary response shell is formed by offsetting the surface boundary of the dynamic boundary partition outward by a first preset distance along the normal direction. The heat transfer difference between the surface temperature and the internal temperature, the current environmental disturbance parameters, and the operating status of the maintenance equipment are jointly mapped to the instant boundary response shell to generate the first layer shell temperature data that characterizes the instantaneous heat dissipation state of the current boundary. A boundary delay response shell is constructed outside the instantaneous boundary response shell. The boundary delay response shell is formed by offsetting the instantaneous boundary response shell outward by a second preset distance along the same normal direction. The surface temperature change, internal temperature change and environmental disturbance change at the current time and the previous time are superimposed on the boundary delay response shell to generate the temperature data of the second shell that characterizes the boundary heat exchange hysteresis effect. A disturbance reflection shell is constructed outside the boundary delay response shell. The disturbance reflection shell is formed by offsetting the boundary delay response shell outward by a third preset distance along the same normal direction. The boundary heat flow offset caused by environmental disturbance changes, the thermal resistance change trend of dynamic boundary partitions, and the changes in the operating status of maintenance equipment are all mapped to the disturbance reflection shell to generate the third shell temperature data characterizing the influence of external disturbances transmitted back to the boundary. The temperature data of the first, second, and third shell layers, along with the hierarchical order, normal spacing, and partition position of each shell relative to the reference boundary, are correlated and encoded to generate virtual shell temperature data. The virtual shell temperature data is then combined with the spatial position data of each shell to form multi-layer virtual shell temperature domain data corresponding to the dynamic boundary partition.

5. The method for intelligent monitoring and curing of concrete temperature field according to claim 1, characterized in that, The obtained temperature evolution results at multiple future times include: The multi-layer temperature measurement data, virtual shell temperature domain data, environmental disturbance parameters and maintenance equipment operation status data of each dynamic boundary partition are aligned according to the timestamp. Missing sampling points are filled by linear interpolation and arranged in sequence as sensing temperature sequence, virtual shell temperature sequence, and environmental and equipment status sequence to construct the input tensor. The input tensor is fed into an improved ContiFormer network, which consists of a continuous-time coding layer, a first coding block, a boundary-aware gating layer, a second coding block, a multi-scale dilation aggregation layer, a third coding block, a thermal residual coupling layer, and a prediction decoding head, which are sequentially connected in series. The continuous-time coding layer adopts adaptive time coding based on cumulative heat flux. The boundary-aware gating layer is inserted after the output of the first coding block and adjusts the attention channel weights in real time according to the temperature difference between the surface and the interior. A multi-scale dilatation aggregation layer is inserted between the second and third coding blocks. The multi-scale dilatation aggregation layer uses minute-level and day-night-level dilated convolution kernels to capture multi-timescale temperature patterns and feeds them back to the output of the second coding block in a residual manner. The output of the third coding block is fused with the operating status vector of the maintenance equipment through the thermal residual coupling layer. The thermal residual coupling layer adopts a gated residual structure to superimpose the heat input or heat extraction generated by the equipment operation onto the temperature feature flow. The predictive decoder outputs the temperature sequences of temperature measurement points at different depths in each dynamic boundary partition and the corresponding virtual shell temperature sequences within the next three control cycles, forming the temperature evolution results at multiple future moments; The improved ContiFormer network was obtained through offline supervised training. The training samples consisted of historical temperature sequences and corresponding future temperature labels. The training process employed a joint optimization of temperature mean square error loss and boundary thermal equilibrium consistency loss.

6. The method for intelligent monitoring and curing of concrete temperature field according to claim 1, characterized in that, The extraction of temperature field change characteristic parameters for each dynamic boundary partition includes: Read multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and associate the temperature values ​​of temperature measurement points at different depths with the virtual shell temperature at the corresponding time according to dynamic boundary partitioning to form a temperature data set that includes spatial location, depth location, and time series. Based on the spatial relationship of each temperature measurement point in the temperature data set, the internal temperature measurement node and the corresponding virtual shell node are spatially mapped using dynamic boundary partitioning as the basic calculation unit, so that the virtual shell node, as a boundary constraint, together with the internal temperature measurement node, constitutes the temperature field reconstruction node set. Based on the spatial coordinates and corresponding temperature values ​​of each node in the temperature field reconstruction node set, the continuous temperature distribution of the internal space is calculated within the dynamic boundary partition range to obtain the three-dimensional temperature field of the concrete structure at different spatial locations and at different times. In a three-dimensional temperature field, the temperature changes between different spatial locations are extracted according to the spatial range of dynamic boundary partitions, and the spatial temperature gradient of each dynamic boundary partition is determined based on the temperature difference between adjacent spatial locations and the corresponding spatial distance. The rate of temperature change is determined based on the temperature changes at the same spatial location during continuous sampling, and the internal and external temperature differences are determined based on the temperature difference between the interior and surface of the dynamic boundary partition, thus forming the characteristic parameters of temperature field changes for each dynamic boundary partition.

7. The method for intelligent monitoring and curing of concrete temperature field according to claim 1, characterized in that, The generated curing control command implements zoned pulse curing control for the concrete structure, including: By reading the characteristic parameters of temperature field changes, equivalent boundary thermal resistance parameters, and temperature evolution results, the locations of high surface heat flux response, internal temperature sensitive locations, and deep key regions are determined in each dynamic boundary partition. The initial opening degree of the cone is determined by taking the high heat flux response position on the surface as the starting point of the cone, the dominant temperature gradient direction from the surface to the interior as the main propagation direction, the equivalent boundary thermal resistance parameter of the corresponding dynamic boundary partition as the initial opening degree of the cone, and the direction of cone axis deflection is determined by the temperature advancement direction at multiple consecutive moments in the temperature evolution result, thus forming the initial heat flux propagation cone skeleton of the corresponding dynamic boundary partition. The cone skeleton propagates along the initial heat flow, and the cone cross-section is extrapolated layer by layer according to each future control moment. The advancing distance of the cone along the main propagation direction is determined according to the temperature change rate at the corresponding moment. The expansion depth of the cone in the thickness direction is determined according to the internal and external temperature difference. The expansion width of the cone in the lateral direction is determined according to the spatial temperature gradient distribution. The cone cross-section sequence corresponding to each future control moment is constructed. The sequence of cone sections corresponding to each future control moment is stacked sequentially along time to form a future heat flow propagation cone with time-progression properties, wherein the future heat flow propagation cone includes a main propagation cone and a lateral disturbance branch cone; Determine whether the main propagation cone or the lateral disturbance branch cone in the future heat flow propagation cone will spatially overlap with the deep critical area during the future control period. When spatial overlap occurs, determine the maintenance control intensity, maintenance control start time and maintenance control duration of the dynamic boundary zone based on the overlap time, overlap depth, overlap range and the equivalent boundary thermal resistance parameters of the corresponding dynamic boundary zone. Based on the curing control intensity, curing control start time, and curing control duration, corresponding curing control instructions for dynamic boundary zones are generated. Curing control instructions are output separately for each dynamic boundary zone to implement zoned pulse curing control for concrete structures.

8. A concrete temperature field intelligent monitoring and curing system, comprising the concrete temperature field intelligent monitoring and curing method according to any one of claims 1 to 8, characterized in that, include: The temperature acquisition module is used to acquire multi-layer temperature data, and at the same time, it collects environmental disturbance parameters and maintenance equipment operating status data. The boundary identification module is used to calculate the temperature gradient and temperature change rate of each initial boundary region based on multi-layer temperature measurement data, and to determine the equivalent boundary thermal resistance parameters. The virtual shell module is used to calculate the virtual shell temperature value and spacing based on the equivalent boundary thermal resistance parameters and the surface temperature measurement point temperature, generate virtual shell temperature data, and form multi-layer virtual shell temperature domain data. The temperature prediction module is used to input multi-layer temperature measurement data, virtual shell temperature data, environmental disturbance parameters, and maintenance equipment operating status data into the improved ContiFormer network to obtain temperature evolution results. The temperature field reconstruction module is used to reconstruct the three-dimensional temperature field of concrete based on multi-layer temperature measurement data, multi-layer virtual shell temperature domain data, and temperature evolution results, and to extract characteristic parameters of temperature field changes. The maintenance control module is used to construct the future heat flow propagation cone based on the temperature field change characteristic parameters and equivalent boundary thermal resistance parameters, and generate maintenance control commands to implement zoned pulse maintenance control.

Images