Pile foundation compactness evaluation method and system based on thermal conductivity inversion, and medium

Abstract

The pile foundation compactness evaluation method and system based on thermal conductivity inversion, and the medium, comprise: collecting three-dimensional space-time temperature data of measuring points on the pile body through the distributed optical fiber fixedly wound in a spiral manner along the pile foundation reinforcement cage main reinforcement; using the three-dimensional space-time temperature data, a composite loss function of a physical information neural network is constructed, the network is trained by minimizing the composite loss function to invert the thermal conductivity distribution of the pile body concrete; the thermal conductivity distribution is substituted into a pre-established quantitative mapping model to obtain a three-dimensional compactness index of the pile body, and the quality grade of each spatial region is determined based on the compactness index and a preset grading threshold value. The present application can perceive the temperature distribution of the pile body cross section (ring direction); the physical information neural network is used for inversion, the temperature data is automatically and accurately converted into the thermal conductivity distribution representing the internal state of the material; the quantitative mapping model is used to quantify the compactness into an index, the pile foundation compactness is accurately evaluated, and an intelligent closed loop is formed.

Description

Technical Field

[0001] This application relates to the fields of power transmission and transformation engineering technology and non-destructive testing technology, specifically to a method, system and medium for evaluating the compactness of pile foundations based on thermal conductivity inversion. Background Technology

[0002] How can we use an integrated technology to automatically, efficiently, and quantitatively complete the entire process from three-dimensional temperature field monitoring and thermophysical parameter inversion to final density assessment, starting from the source of concrete pile foundation pouring? This would transform the three-dimensional distribution field of the thermal conductivity of the pile concrete into a quantitative result of density that can be directly understood in engineering and used to accurately assess the quality of concrete pouring. This would overcome the shortcomings of existing technologies, such as single monitoring dimensions, low inversion efficiency, qualitative assessment results, and reliance on human experience, which lead to incomplete, inaccurate, and untimely pile foundation quality assessments, thus supporting refined engineering management.

[0003] The following are common general knowledge in this field: 1) Distributed fiber optic temperature measurement principle: Based on optical time-domain reflectometry, continuous and real-time temperature measurement along the fiber path is achieved by analyzing the intensity ratio of backscattered Raman light in the fiber.

[0004] 2) Forward Problem of Heat Conduction: This involves calculating the temperature field by solving the partial differential equations (PDEs) of heat conduction, given the thermophysical parameters, boundary conditions, and initial conditions of the material. For concrete pile foundations, the governing equations are typically:

[0005] in: Temperature field is a spatial location A function of time t; Density (kg / m³); Specific heat capacity (J / (kg·K)); Thermal conductivity (W / (m·K)) is a spatial distribution function and is directly related to the density of concrete pouring. The heat release from concrete hydration (W / m³) is usually represented by an exponential decay model. ,in For the final release of heat, For the heat release rate parameter, For gradient.

[0006] 3) Inverse problem of heat conduction: Given partial temperature observation data and boundary / initial conditions, inversely deduce the internal thermophysical parameters of the material (e.g., k The distribution of (x)). This is a typical ill-posed problem, the solution may not be unique and it is sensitive to data noise.

[0007] 4) Physical Information Neural Network Concept: A novel framework that embeds physical laws into the training process of deep learning models. By using the residuals of physical equations as part of the loss function, the network can both fit the data and obey physical laws after training.

[0008] 5) Relationship between thermal conductivity and concrete density: The thermal conductivity of concrete is closely related to its composition, porosity, and water content. Generally speaking, the denser the concrete and the lower its porosity, the higher its thermal conductivity. Conversely, the presence of voids, cracks, or segregation significantly reduces the thermal conductivity. Porosity and density are highly related but physically opposite concepts; lower porosity corresponds to higher density. This is the physical basis for using thermal conductivity to assess density.

[0009] The following is an introduction to existing technologies for monitoring the three-dimensional temperature field of concrete pile foundations: The existing technology relates to the invention application with application number 202511167271.0, entitled "A Distributed Temperature Measuring Optical Cable and a Method for Accurate Evaluation of the Integrity of Cast-in-Place Piles." This method includes deploying a temperature measuring optical cable along the reinforcing cage of the cast-in-place pile; performing refined temperature measurements of the cast-in-place pile at different depths before and after pouring using the optical cable; correcting the measured temperatures to obtain temperature data at different depths of the cast-in-place pile; converting the temperature data at the corresponding depth into an equivalent concrete cover thickness, and combining this with the radius of the reinforcing cage to obtain the equivalent radius of the cast-in-place pile at the current depth; using spline interpolation based on the equivalent radius to draw the cross-section corresponding to the cast-in-place pile at the current depth; constructing the 3D shape of the cast-in-place pile from the cross-sections at different depths for visualization. Through refined temperature measurement, the spatial resolution of temperature measurement is effectively improved, not only reducing or even avoiding missed detections of defects and errors in severity evaluation of the cast-in-place pile, but also achieving visual transparency of the cast-in-place pile. The overall operation is simple, convenient, intuitive, and accurate, with broad application prospects.

[0010] The prior art includes an invention patent with application number 202111074861.0 entitled "An Integrity Detection Device for Cast-in-Place Piles Based on Fiber Bragg Grating Temperature Measurement". This device comprises a cast-in-place pile (1) and a reinforcing cage (2). A capillary seamless steel tube fiber Bragg grating temperature sensor (3) and an actively heated fiber Bragg grating temperature sensor (8) are fixed at equal intervals along the axial direction of the cast-in-place pile (1) onto the reinforcing cage (2), and both are connected to a signal demodulator (4). The actively heated fiber Bragg grating temperature sensor (8) is then connected to a pulse heating power supply (9). The signal demodulator (4) is connected to an Internet of Things (IoT) gateway module (5), which is then connected to a data visualization processing module (7). The signal demodulator (4), IoT gateway module (5), and pulse heating power supply (9) are then connected to a DC power supply module (6) via interface electricity. The advantages of this invention are improved intelligence and real-time capability of pile foundation detection, increased accuracy of pile foundation detection data, simplified detection process, and reduced cost.

[0011] The prior art relates to an invention patent with application number 201310006961.9 entitled "Integrity Detection System for Cast-in-Place Piles Based on Distributed Fiber Optic Temperature Sensing," which relates to a technology for detecting the integrity of cast-in-place piles. The system structure is as follows: a sensing fiber (20) is placed in the cast-in-place pile (00), and a heating device (30), a distributed fiber optic temperature measuring instrument (10), and a cast-in-place pile integrity evaluation device (40) are connected sequentially. The heating device (30) heats the sensing fiber (20), the sensing fiber (20) senses the pile foundation temperature, the distributed fiber optic temperature measuring instrument (10) measures the temperature, and the cast-in-place pile integrity evaluation device (40) converts the temperature value into an integrity value. This invention can effectively and quickly measure the integrity of cast-in-place piles; it can perform long-term, real-time, online monitoring of the integrity of cast-in-place piles; the detection is simple and can meet various requirements for cast-in-place piles; it is suitable for detecting the integrity of various cast-in-place piles.

[0012] The above-mentioned existing technologies have the following technical disadvantages: (1) Insufficient monitoring data dimensions, resulting in blind spots in defect identification. Conventional straight-line fiber optic cables can only acquire one-dimensional axial temperature information and cannot sense the temperature distribution of the pile cross-section (circumferential direction). This results in insensitivity to defects such as eccentric voids and circumferential segregation, leading to a high risk of missed detection and making it impossible to achieve true three-dimensional quality diagnosis.

[0013] (2) The chain from data to evaluation is broken, the process is inefficient and non-quantitative. There is a lack of an automatic and accurate physical model and algorithm bridge between temperature data and the final quality conclusions. It relies on manual interpretation, and the interpretation only stays at the level of temperature anomalies. This results in a slow evaluation process, strong subjectivity, and the inability to quantify temperature anomalies into indicators directly needed for engineering, such as "density" and "porosity". The evaluation results are crude and have weak support for engineering decision-making.

[0014] (3) The technical links are fragmented and no intelligent closed loop has been formed. Temperature measurement, simple analysis, and manual evaluation are separate processes. The failure to build an integrated intelligent system from sensing and inversion to evaluation results in low overall efficiency and makes it difficult to achieve large-scale, standardized engineering applications. Summary of the Invention

[0015] To address the problems of insufficient dimensions of existing monitoring data, blind spots in defect identification, broken data-to-assessment chain, inefficient and non-quantitative process, and fragmented technical links that fail to form an intelligent closed loop, this application proposes: 1. A pile foundation compaction assessment method based on thermal conductivity inversion, characterized by comprising: Three-dimensional spatiotemporal temperature data of measuring points on the pile body are collected by distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage. Using three-dimensional spatiotemporal temperature data, a composite loss function of a physical information neural network is constructed. The physical information neural network is trained by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete. The composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term. The thermal conductivity distribution is substituted into a pre-established quantitative mapping model to obtain the three-dimensional compactness index of the pile body, and the quality level of each spatial region is determined based on the compactness index and the preset grading threshold. Among them, the quantitative mapping model is constructed based on the quantitative mapping between thermal conductivity and concrete density established by the theory of materials physics.

[0016] Preferably, the step of constructing a composite loss function for a physical information neural network using three-dimensional spatiotemporal temperature data, and training the physical information neural network by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete, includes: By inputting the coordinates from the three-dimensional spatiotemporal temperature data into a physical neural network model, the predicted temperature field and thermal conductivity distribution can be obtained. The data driving loss is calculated by combining the temperature in the three-dimensional spatiotemporal temperature data and the predicted temperature value in the predicted temperature field with the driving loss equation. The physical law loss is calculated based on the combination of the predicted temperature field and thermal conductivity distribution, and the thermal conductivity in the thermal conductivity distribution combined with the physical equation of heat conduction. Based on the predicted thermal conductivity and temperature field, combined with the pre-set reasonableness constraints of the thermal conductivity distribution law, the loss of prior knowledge is determined. The composite loss is calculated from data-driven loss, physical law loss, and prior knowledge loss. The gradient descent algorithm is used to iteratively update the neural network parameters, minimize the composite loss until convergence, and obtain the thermal conductivity distribution of the pile concrete.

[0017] Preferably, the driving loss equation is as follows:

[0018] in The total number of data points. To predict temperature, For observation data, For the spatial coordinates of the data points, Time coordinates of data points Preferably, the physical equation for heat conduction is as follows:

[0019] in, Loss due to physical laws For the total number of configuration points, To predict the temperature field, To predict thermal conductivity, For temperature gradient, The heat released during hydration is j, where j is the placement point number. For gradient.

[0020] Preferably, the prior knowledge loss is calculated using the following formula:

[0021] In the formula, For prior knowledge loss, For smoothness constraint loss, For range-constrained losses, The loss is determined by boundary conditions.

[0022] Preferably, the composite loss is calculated using the following formula:

[0023] in, For composite loss function, Data-driven loss weights Weight loss due to physical laws Weights for prior knowledge loss. For data-driven loss, Loss due to physical laws This is due to the loss of prior knowledge.

[0024] Preferably, the construction of the quantization mapping model includes: Based on the theory of effective media in composite materials, a parameterized physical constitutive relationship between thermal conductivity and density is established. Measured thermal conductivity and density data of standard concrete pile foundations are obtained. The parameters in the constitutive relation are determined by fitting the measured thermal conductivity and density data. The parameters are then substituted into the constitutive relation to obtain a quantitative mapping model.

[0025] Preferably, the quantization mapping model is as follows:

[0026] in, and These represent the thermal conductivity of fully poured concrete and concrete with pouring defects, respectively. This is a density index, where 1 indicates density and 0 indicates severe defects.

[0027] Preferably, determining the quality level of each spatial region based on the density index and a preset grading threshold includes: The density level range is constructed by multiple preset grading thresholds; Determine the density level range to which the density index belongs, and use the quality level corresponding to the density level range as the quality level of the spatial area. The quality levels are: I. Healthy, II. Minor Defects, III. Moderate Defects, and IV. Major Defects.

[0028] Preferably, after determining the quality level of each spatial region based on the density index and a preset grading threshold, the method further includes: Based on the quality level of each spatial region, a defect statistics table is generated, and a classification conclusion is drawn. A three-dimensional density cloud map of the pile foundation is constructed based on the density index.

[0029] Furthermore, this application also provides a pile foundation compaction assessment system based on thermal conductivity inversion, including: The temperature acquisition module is used to collect three-dimensional spatiotemporal temperature data of measuring points on the pile body by using distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage. The inversion module is used to construct a composite loss function of a physical information neural network using three-dimensional spatiotemporal temperature data. The physical information neural network is trained by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete. The composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term. The compactness assessment module is used to substitute the thermal conductivity distribution into a pre-established quantitative mapping model to obtain the compactness index of the three-dimensional pile body, and to determine the quality level of each spatial region based on the compactness index and the preset grading threshold. Among them, the quantitative mapping model is constructed based on the quantitative mapping between thermal conductivity and concrete density established by the theory of materials physics.

[0030] Preferably, the inversion module is specifically used for: By inputting the coordinates from the three-dimensional spatiotemporal temperature data into a physical neural network model, the predicted temperature field and thermal conductivity distribution can be obtained. The data driving loss is calculated by combining the temperature in the three-dimensional spatiotemporal temperature data and the predicted temperature value in the predicted temperature field with the driving loss equation. The physical law loss is calculated based on the combination of the predicted temperature field and thermal conductivity distribution, and the thermal conductivity in the thermal conductivity distribution combined with the physical equation of heat conduction. Based on the predicted thermal conductivity and temperature field, combined with the pre-set reasonableness constraints of the thermal conductivity distribution law, the loss of prior knowledge is determined. The composite loss is calculated from data-driven loss, physical law loss, and prior knowledge loss. The gradient descent algorithm is used to iteratively update the neural network parameters, minimize the composite loss until convergence, and obtain the thermal conductivity distribution of the pile concrete.

[0031] Preferably, the driving loss equation is as follows:

[0032] in The total number of data points. To predict temperature, For observation data, For the spatial coordinates of the data points, For the time coordinates of the data points.

[0033] Preferably, the physical equation for heat conduction is as follows:

[0034] in, Loss due to physical laws For the total number of configuration points, To predict the temperature field, To predict thermal conductivity, For temperature gradient, The heat released during hydration is represented by j, which is the placement point number.

[0035] Preferably, the prior knowledge loss is calculated using the following formula:

[0036] In the formula, For prior knowledge loss, For smoothness constraint loss, For range-constrained losses, The loss is determined by boundary conditions.

[0037] Preferably, the composite loss is calculated using the following formula:

[0038] in, For composite loss function, Data-driven loss weights Weight loss due to physical laws Weights for prior knowledge loss. For data-driven loss, Loss due to physical laws This is due to the loss of prior knowledge.

[0039] Preferably, it further includes a model building module, used for: Based on the theory of effective media in composite materials, a parameterized physical constitutive relationship between thermal conductivity and density is established. Measured thermal conductivity and density data of standard concrete pile foundations are obtained. The parameters in the constitutive relation are determined by fitting the measured thermal conductivity and density data. The parameters are then substituted into the constitutive relation to obtain a quantitative mapping model.

[0040] Preferably, the quantization mapping model is as follows:

[0041] in, and These represent the thermal conductivity of fully poured concrete and concrete with pouring defects, respectively. This is a density index, where 1 indicates density and 0 indicates severe defects.

[0042] Preferably, the implementation steps in the density assessment module for determining the quality level of each spatial region based on the density index and a preset grading threshold include: The density level range is constructed by multiple preset grading thresholds; Determine the density level range to which the density index belongs, and use the quality level corresponding to the density level range as the quality level of the spatial area. The quality levels are: I. Healthy, II. Minor Defects, III. Moderate Defects, and IV. Major Defects.

[0043] Preferably, it further includes a report generation module, used for: Based on the quality level of each spatial region, a defect statistics table is generated, and a classification conclusion is drawn. A three-dimensional density cloud map of the pile foundation is constructed based on the density index.

[0044] In another aspect, this application also provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the pile compaction assessment method based on thermal conductivity inversion as described above is implemented.

[0045] In another aspect, this application also provides a computer-readable storage medium having an executable program stored thereon, which, when executed, implements the pile foundation compaction assessment method based on thermal conductivity inversion as described above.

[0046] Compared with the prior art, the beneficial effects of this application are as follows: This application provides a method for assessing the compactness of pile foundations based on thermal conductivity inversion, comprising: acquiring three-dimensional spatiotemporal temperature data of measuring points on the pile body by using distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage; constructing a composite loss function of a physical information neural network using the three-dimensional spatiotemporal temperature data, training the physical information neural network by minimizing the composite loss function to invert the thermal conductivity distribution of the pile body concrete, wherein the composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term; substituting the thermal conductivity distribution into a pre-established quantization mapping model to obtain a three-dimensional compactness index of the pile body, and determining the quality level of each spatial region based on the compactness index and a preset grading threshold; wherein the quantization mapping model is constructed based on the quantization mapping between thermal conductivity and concrete compactness established by the theory of materials physics. The distributed optical fiber in this application is spirally wound and fixed along the main reinforcement of the pile foundation cage. It can acquire three-dimensional spatiotemporal temperature data and sense the temperature distribution of the pile body cross section (circumferential direction), providing a data foundation for three-dimensional quality diagnosis and overcoming the problems of insufficient monitoring data dimensions and blind spots in defect identification. This application uses a physical information neural network for inversion, automatically and accurately converting temperature data into thermal conductivity distribution that characterizes the internal state of the material, overcoming the broken chain from data to evaluation, inefficient process and non-quantitative nature. The quantitative mapping model of this application quantifies compaction into an index, accurately assesses the compaction of the pile foundation, and forms an intelligent closed loop. Attached Figure Description

[0047] Figure 1 This is a flowchart of the pile foundation compaction assessment method based on thermal conductivity inversion in this application; Figure 2 This is a schematic diagram of the spiral fiber optic sensor network of this application; Figure 3 This is a schematic diagram of the non-destructive testing system structure of this application; Figure 4 This is a schematic diagram illustrating the specific implementation of the pile foundation compaction assessment method based on thermal conductivity inversion in this application; Figure 5 This is a schematic diagram of an electronic device structure according to this application; Among them, 1-distributed optical fiber, 11-longitudinal distributed optical fiber, 12-spiral distributed optical fiber, 2-steel skeleton, 3-temperature acquisition module, 4-inversion module, and 5-density assessment module. Detailed Implementation

[0048] The purpose of this invention is to provide a systematic, quantitative, and automated evaluation method to achieve the following effects: (1) By using distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation to collect three-dimensional spatiotemporal temperature data of the measuring points on the pile body, the complete and efficient acquisition of the three-dimensional temperature field inside the pile foundation is realized. Through the innovative sensor deployment method, three-dimensional temperature data containing axial and circumferential information is obtained.

[0049] (2) Using advanced physical information neural network algorithms, a fast, stable and high-precision inversion channel from temperature field to thermophysical parameter field is established, and temperature data is automatically and accurately converted into thermal conductivity distribution that characterizes the internal state of the material.

[0050] (3) Define a clear quantitative mapping model and hierarchical threshold, and construct a quantitative, objective and automated evaluation system from thermophysical parameters to engineering quality indicators. The inverted parameter field is directly converted into quantitative density evaluation results and engineering suggestions.

[0051] To better understand this application, the content of this application will be further described below in conjunction with the accompanying drawings and embodiments.

[0052] Example 1: Methods for assessing pile foundation compaction based on thermal conductivity inversion, such as Figure 1 As shown, it includes: Step 1: Collect three-dimensional spatiotemporal temperature data of measuring points on the pile body by using distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage; Step 2: Using three-dimensional spatiotemporal temperature data, construct a composite loss function for the physical information neural network. Train the physical information neural network by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete. The composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term. Step 3: Substitute the thermal conductivity distribution into the pre-established quantitative mapping model to obtain the three-dimensional compactness index of the pile body, and determine the quality level of each spatial region based on the compactness index and the preset grading threshold. Among them, the quantitative mapping model is constructed based on the quantitative mapping between thermal conductivity and concrete density established by the theory of materials physics.

[0053] The data fitting term is the mean square error (MSE) between the predicted temperature field and the temperature data of the distributed optical fiber 1 in the same spatiotemporal coordinates.

[0054] This invention proposes a "spiral fiber optic three-dimensional temperature measurement - PINN intelligent inversion - " k - η The solution employs a four-in-one closed-loop technology approach: quantitative mapping, graded assessment report, and more. Its core lies in: first, acquiring the three-dimensional spatiotemporal temperature field of the pile body through a spirally deployed network of temperature-sensing optical cables; then, efficiently and stably retrieving the overall thermal conductivity distribution from the temperature field using a specially designed physical information neural network; next, converting the thermal conductivity field into a density field through an established quantitative mapping model; and finally, performing automated spatial statistics, defect identification, and quality grading based on preset standards, generating a visual report.

[0055] The following is combined with Figure 4 The technical solution of this application is further defined as follows: Methods for assessing pile foundation compaction based on thermal conductivity inversion include: Step 1: Collect three-dimensional spatiotemporal temperature data of measuring points on the pile body using distributed optical fibers that are spirally wound and fixed along the main reinforcement bars of the pile foundation steel cage. Specifically, this includes: Distributed optical fibers are spirally wound around the main reinforcement bars of the pile foundation to form a three-dimensional sensing network covering the entire surface of the pile. Three-dimensional spatiotemporal temperature data of various measuring points on the pile are collected by a three-dimensional sensor network; The measuring points here can be set manually or based on a set interval. This set interval can be determined based on empirical data or other statistical data. The three-dimensional spatiotemporal temperature data includes coordinates, temperature, and time.

[0056] Step 1 specifically includes: Spiral fiber optic sensor network layout: Abandoning the traditional single-axis linear layout, corrosion-resistant temperature sensing optical cables are spirally wound and fixed along the main reinforcement bars of the pile foundation cage. This allows the measurement path of a single optical fiber to simultaneously cover both the axial and circumferential directions, forming a three-dimensional sensing network covering the entire surface of the pile. Figure 2The longitudinally arranged distributed optical fibers 11 and the spirally arranged distributed optical fibers 12 form a three-dimensional sensing network covering the entire surface of the pile. The fiber pitch P is related to the pile diameter D, with a preferred range of 0.2D to 0.5D, but can also be dynamically adjusted according to the pile foundation bearing capacity design and spatial resolution requirements. In key stress-bearing sections such as the pile end and pile top, a denser pitch (e.g., P density = 0.1D to 0.2D) is set to form an enhanced monitoring section, thereby improving the accuracy of thermal conductivity inversion in key areas.

[0057] Four-dimensional temperature field data acquisition: During the hydration process of cementitious materials, distributed optical fiber temperature measurement technology is used to continuously acquire three-dimensional spatiotemporal temperature data of a limited number of measuring points on the pile body with a time resolution of minutes to hours.

[0058] Step 2: Using three-dimensional spatiotemporal temperature data, construct a composite loss function for the physical information neural network. Train the physical information neural network by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete, including: By inputting the coordinates from the three-dimensional spatiotemporal temperature data into a physical neural network model, the predicted temperature field and thermal conductivity distribution can be obtained. The data driving loss is calculated by combining the temperature in the three-dimensional spatiotemporal temperature data and the predicted temperature value in the predicted temperature field with the driving loss equation. The physical law loss is calculated based on the combination of the predicted temperature field and thermal conductivity distribution, and the thermal conductivity in the thermal conductivity distribution combined with the physical equation of heat conduction. Based on the predicted thermal conductivity and temperature field, combined with the pre-set reasonableness constraints of the thermal conductivity distribution law, the loss of prior knowledge is determined. The composite loss is calculated from data-driven loss, physical law loss, and prior knowledge loss. The gradient descent algorithm is used to iteratively update the neural network parameters, minimize the composite loss until convergence, and obtain the thermal conductivity distribution of the pile concrete.

[0059] Furthermore, the driving loss equation is shown below:

[0060] in The total number of data points. To predict temperature, For observation data, For the spatial coordinates of the data points, Time coordinates of data points Furthermore, the physical equation for heat conduction is shown below:

[0061] in, Loss due to physical laws For the total number of configuration points, To predict the temperature field, To predict thermal conductivity, For temperature gradient, The heat released during hydration is represented by j, which is the placement point number.

[0062] Furthermore, the prior knowledge loss is calculated using the following formula:

[0063]

[0064]

[0065]

[0066] In the formula, For prior knowledge loss, For smoothness constraint loss, For range-constrained losses, For boundary condition constraint loss, This represents the total number of sampling points in dense sampling. This represents the total number of sampling points at the pile-soil interface. The thermal conductivity at the j-th sampling point Spatial gradient at that location To correct the linear unit, This is the predicted value of the thermal conductivity at the j-th sampling point. , These are the upper and lower limits of thermal conductivity, respectively. The m-th sampling point at the pile-soil interface Predicted thermal conductivity value at that location. This represents the derivative of the predicted temperature value with respect to the direction outside the pile-soil boundary. Let be the convective heat transfer coefficient at the pile-soil interface. Boundary sampling points Predicted temperature value at time t. Boundary sampling points The ambient temperature at that location.

[0067] Step 2 specifically includes: Thermal conductivity inversion: Based on the collected sparse temperature field data, and with the physical equation of heat conduction as a strong constraint, a Physical Information Neural Network (PINN) is used to directly invert the three-dimensional distribution field k(x,y,z) of the thermal conductivity of the pile concrete. This network takes spatiotemporal coordinates as input, simultaneously predicts the temperature field and thermal conductivity field through dual output branches, and utilizes a triple loss function (data fitting loss, residual loss from the physical equation of heat conduction, and prior constraint loss for thermal conductivity) to ensure the physical rationality and numerical stability of the inversion results.

[0068] Quantitative Mapping of Thermal Conductivity and Concrete Density: Based on materials physics theory, a quantitative mapping model is established between thermal conductivity k and concrete density η. Substituting the inverted full-field thermal conductivity distribution into this model yields the three-dimensional density distribution field η(x,y,z) of the pile body. Based on the calculated density field, a preset grading threshold is used to automatically determine the quality level of each spatial region. Three-dimensional connected domain analysis identifies defective regions and outputs parameters such as their volume and equivalent center location.

[0069] 1. Governing equations The transient heat conduction during concrete hydration follows the following partial differential equation:

[0070] in: Temperature field is a function of spatial location x and time t; Density (kg / m³); Specific heat capacity (J / (kg·K)); Thermal conductivity (W / (m·K)) is a spatial distribution function and is directly related to the density of concrete pouring. The heat release from concrete hydration (W / m³) is usually represented by an exponential decay model. ,in For the final release of heat, This is the heat release rate parameter.

[0071] 2. Boundary and Initial Conditions Initial conditions: (Ambient temperature or mold temperature).

[0072] Boundary conditions: Third type of boundary conditions (Robin conditions): , denoted as the surface heat transfer coefficient.

[0073] 3. Physical Information Neural Network Model Architecture Input is spatial coordinates x and time t The output is the predicted temperature field. and thermal conductivity distribution .

[0074]

[0075] in For network parameters, For physical information neural network models, For spatial coordinates x The time is t Predicted temperature field at that time For spatial coordinates x The time is t The thermal conductivity distribution at that time.

[0076] 4. Loss Function Design The loss function consists of three parts: data-driven loss, physical law loss, and prior constraint loss.

[0077] in, For loss function, Data-driven loss weights Weight loss due to physical laws As a priori constraint loss weight, For data-driven loss, Loss due to physical laws The loss is due to prior constraints.

[0078] 1) Data-driven loss Ensure that the predicted temperature matches the observed data:

[0079] in The total number of data points. To predict temperature, For observation data, For the spatial coordinates of the data points, For the time coordinates of the data points.

[0080] 2) Loss of physical laws The forced prediction of the temperature field and thermal conductivity distribution satisfies the physical equations of heat conduction, and random sampling is performed within the computational domain. One configuration point:

[0081] in, Loss due to physical laws For the total number of configuration points, To predict the temperature field, To predict thermal conductivity, For temperature gradient, The heat released during hydration is represented by j, which is the placement point number.

[0082] 3) Prior constraint loss Introducing constraints on the rationality of the thermal conductivity distribution pattern:

[0083]

[0084]

[0085]

[0086] in, For prior knowledge loss, For smoothness constraint loss, For range-constrained losses, For boundary condition constraint loss, This represents the total number of sampling points in dense sampling. This represents the total number of sampling points at the pile-soil interface. The thermal conductivity at the j-th sampling point Spatial gradient at that location To correct the linear unit, This is the predicted value of the thermal conductivity at the j-th sampling point. , These are the upper and lower limits of thermal conductivity, respectively. The m-th sampling point at the pile-soil interface Predicted thermal conductivity value at that location. This represents the derivative of the predicted temperature value with respect to the direction outside the pile-soil boundary. Let be the convective heat transfer coefficient at the pile-soil interface. Boundary sampling points Predicted temperature value at time t. Boundary sampling points The ambient temperature at that location.

[0087] 5. Network Training and Inversion Process 1) Input: Three-dimensional spatiotemporal temperature data, which includes: observed temperature data and corresponding spatiotemporal coordinates; 2) Forward propagation: The neural network output predicts the temperature field and thermal conductivity distribution; 3) Loss Calculation: Calculate the composite loss; 4) Backpropagation and optimization: Use gradient descent algorithm to update network parameters and minimize the composite loss; 5) Output: After the training model converges, the output temperature field and thermal conductivity distribution are displayed.

[0088] Step 3 is preceded by: an introduction to the construction process of the geoquantization mapping model involved in this application, wherein the construction of the geoquantization mapping model includes: Based on the theory of effective media in composite materials, a parameterized physical constitutive relationship between thermal conductivity and density is established. Measured thermal conductivity and density data of standard concrete pile foundations are obtained. The parameters in the constitutive relation are determined by fitting the measured thermal conductivity and density data. The parameters are then substituted into the constitutive relation to obtain a quantitative mapping model.

[0089] Step 3: Substitute the thermal conductivity distribution into the pre-established quantization mapping model to obtain the three-dimensional compactness index of the pile body, and determine the quality level of each spatial region based on the compactness index and the preset grading threshold, including: Substituting the thermal conductivity distribution into a pre-established quantitative mapping model, the three-dimensional compactness index of the pile body is obtained. The density level range is constructed by multiple preset grading thresholds; Determine the density level range to which the density index belongs, and use the quality level corresponding to the density level range as the quality level of the spatial area. The quality levels are: I. Healthy, II. Minor Defects, III. Moderate Defects, and IV. Major Defects.

[0090] Furthermore, the density index is calculated using the following formula:

[0091] in and The thermal conductivity values ​​represent the thermal conductivity of fully poured concrete and concrete with pouring defects, respectively (reference values ​​determined by testing). A value of 1 indicates a dense structure, while a value of 0 indicates a serious defect.

[0092] Step 3 specifically includes: quantitative assessment of density.

[0093] 1) Mapping of density index Obtaining thermal conductivity distribution Then, it is transformed into a concrete pouring density index through a preset mapping relationship (linear model, or a nonlinear model can be used depending on the experimental conditions):

[0094] in and The thermal conductivity values ​​represent the thermal conductivity of fully poured concrete and concrete with pouring defects, respectively (reference values ​​determined by testing). A value of 1 indicates a dense structure, while a value of 0 indicates a serious defect.

[0095] 2) Spatial Analysis Calculate the average density of the pile foundation across the entire field. Standard deviation. Set a threshold. (Based on the type of concrete and experimental test results), all independent defect areas are automatically identified using three-dimensional connected domain analysis, and the volume, location, and minimum density of each area are calculated.

[0096] 3) Quantitative grading The following table automatically classifies the overall or local defect areas of the pile body:

[0097] 4) Report generation It automatically outputs a comprehensive report that includes a 3D density cloud map, a defect statistics table, and a classification conclusion.

[0098] This application addresses the inverse problem (i.e., given sparse temperature observation data, inversely determining the three-dimensional distribution of thermal conductivity, and then evaluating density). This inverse problem is ill-posed and requires unique constraint design. Therefore, this application makes the following adjustments: This application specifically adds prior knowledge loss constraints for thermal conductivity inversion and adopts an adaptive weight adjustment strategy (emphasizing physical loss in the early stage and data fitting in the later stage).

[0099] The technical solution of this application is directly aimed at pile foundation engineering. It combines a spirally deployed fiber optic sensing network to obtain the real temperature field, deeply couples PINN with non-destructive testing of the project, outputs the thermal conductivity field, and establishes a quantitative mapping model of thermal conductivity-density and a four-level classification standard (healthy / minor / moderate / severe). Finally, it outputs a quality assessment report that can be directly used in the project.

[0100] Compared with the prior art, the present invention has the following advantages: (1) It realizes full three-dimensional, blind-spot-free quality perception and diagnosis. Spiral fiber optic deployment fundamentally solves the problem of missing dimensions in traditional straight-line deployment. Combined with PINN's powerful inversion capability of three-dimensional temperature fields, it achieves true three-dimensional imaging of the internal state of pile foundations for the first time, and greatly improves the recognition rate and positioning accuracy of various complex defects.

[0101] (2) It opens up an intelligent channel from raw data to quantitative decision-making, achieving both high efficiency and high accuracy. The PINN algorithm achieves rapid and stable end-to-end inversion from the temperature field to the thermal conductivity field, avoiding the bottleneck of large computational load in traditional iterative methods. The subsequent quantitative evaluation system provides objective and unified engineering criteria. The entire process is automated, reducing what originally required several days and relied on expert experience to just a few hours to complete and output a quantitative report, resulting in a significant improvement in efficiency.

[0102] (3) It has formed a highly integrated technology closed loop, with strong practicality and scalability. This invention deeply integrates cutting-edge sensing technology, AI algorithms, and engineering knowledge to form a standardized and automated hardware and software system solution. This significantly lowers the technical barrier to entry, ensures the consistency of evaluation results, and lays a solid foundation for the large-scale, standardized application and promotion of this technology in pile foundation engineering.

[0103] Example 2 Based on the same inventive concept, this application also provides a pile foundation compaction assessment system based on thermal conductivity inversion, such as... Figure 3 As shown, it includes: The temperature acquisition module is used to collect three-dimensional spatiotemporal temperature data of measuring points on the pile body by using distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage. The inversion module is used to construct a composite loss function of a physical information neural network using three-dimensional spatiotemporal temperature data. The physical information neural network is trained by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete. The composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term. The compactness assessment module is used to substitute the thermal conductivity distribution into a pre-established quantitative mapping model to obtain the compactness index of the three-dimensional pile body, and to determine the quality level of each spatial region based on the compactness index and the preset grading threshold. Among them, the quantitative mapping model is constructed based on the quantitative mapping between thermal conductivity and concrete density established by the theory of materials physics.

[0104] Preferably, the temperature acquisition module 3 is used to acquire optical signals in the distributed optical fiber and convert the optical signals into temperature data. In this application, the conversion of optical signals into temperature data can be achieved by a temperature demodulator.

[0105] Preferably, the inversion module 4 is specifically used for: By inputting the coordinates from the three-dimensional spatiotemporal temperature data into a physical neural network model, the predicted temperature field and thermal conductivity distribution can be obtained. The data driving loss is calculated by combining the temperature in the three-dimensional spatiotemporal temperature data and the predicted temperature value in the predicted temperature field with the driving loss equation. The physical law loss is calculated based on the combination of the predicted temperature field and thermal conductivity distribution, and the thermal conductivity in the thermal conductivity distribution combined with the physical equation of heat conduction. Based on the predicted thermal conductivity and temperature field, combined with the pre-set reasonableness constraints of the thermal conductivity distribution law, the loss of prior knowledge is determined. The composite loss is calculated from data-driven loss, physical law loss, and prior knowledge loss. The gradient descent algorithm is used to iteratively update the neural network parameters, minimize the composite loss until convergence, and obtain the thermal conductivity distribution of the pile concrete.

[0106] Preferably, the driving loss equation is as follows:

[0107] in The total number of data points. To predict temperature, For observation data, For the spatial coordinates of the data points, For the time coordinates of the data points.

[0108] Preferably, the physical equation for heat conduction is as follows:

[0109] in, Loss due to physical laws For the total number of configuration points, To predict the temperature field, To predict thermal conductivity, For temperature gradient, The heat released during hydration is represented by j, which is the placement point number.

[0110] Preferably, the prior knowledge loss is calculated using the following formula:

[0111] In the formula, For prior knowledge loss, For smoothness constraint loss, For range-constrained losses, The loss is determined by boundary conditions.

[0112] Preferably, the composite loss is calculated using the following formula:

[0113] in, For composite loss function, Data-driven loss weights Weight loss due to physical laws Weights for prior knowledge loss. For data-driven loss, Loss due to physical laws This is due to the loss of prior knowledge.

[0114] Preferably, it further includes a model building module, used for: Based on the theory of effective media in composite materials, a parameterized physical constitutive relationship between thermal conductivity and density is established. Measured thermal conductivity and density data of standard concrete pile foundations are obtained. The parameters in the constitutive relation are determined by fitting the measured thermal conductivity and density data. The parameters are then substituted into the constitutive relation to obtain a quantitative mapping model.

[0115] Preferably, the quantization mapping model is as follows:

[0116] in, and These represent the thermal conductivity of fully poured concrete and concrete with pouring defects, respectively. This is a density index, where 1 indicates density and 0 indicates severe defects.

[0117] Preferably, the implementation steps of the density assessment module 5 in determining the quality level of each spatial region based on the density index and the preset grading threshold include: The density level range is constructed by multiple preset grading thresholds; Determine the density level range to which the density index belongs, and use the quality level corresponding to the density level range as the quality level of the spatial area. The quality levels are: I. Healthy, II. Minor Defects, III. Moderate Defects, and IV. Major Defects.

[0118] Preferably, it further includes a report generation module, used for: Based on the quality level of each spatial region, a defect statistics table is generated, and a classification conclusion is drawn. A three-dimensional density cloud map of the pile foundation is constructed based on the density index.

[0119] Example 3 like Figure 5 As shown, this application also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0120] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to realize the corresponding method flow or corresponding function, so as to realize the steps of the pile foundation compaction assessment method based on thermal conductivity inversion in the above embodiments.

[0121] Example 4 Based on the same inventive concept, this application also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the storage medium here can include both built-in storage media within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. Loading and executing one or more instructions stored in the storage medium by the processor can implement the steps of the pile foundation compaction assessment method based on thermal conductivity inversion in the above embodiments.

[0122] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0123] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0124] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0125] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0126] The above are merely embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of the claims of this application.

Claims

1. A method for evaluating the compactness of pile foundations based on thermal conductivity inversion, characterized in that, include: Three-dimensional spatiotemporal temperature data of measuring points on the pile body are collected by distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage. Using three-dimensional spatiotemporal temperature data, a composite loss function of a physical information neural network is constructed. The physical information neural network is trained by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete. The composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term. The thermal conductivity distribution is substituted into a pre-established quantitative mapping model to obtain the three-dimensional compactness index of the pile body, and the quality level of each spatial region is determined based on the compactness index and the preset grading threshold. Among them, the quantitative mapping model is constructed based on the quantitative mapping between thermal conductivity and concrete density established by the theory of materials physics.

2. The method as described in claim 1, characterized in that, The process of constructing a composite loss function for a physical information neural network using three-dimensional spatiotemporal temperature data, and training the physical information neural network by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete includes: By inputting the coordinates from the three-dimensional spatiotemporal temperature data into a physical neural network model, the predicted temperature field and thermal conductivity distribution can be obtained. The data driving loss is calculated by combining the temperature in the three-dimensional spatiotemporal temperature data and the predicted temperature value in the predicted temperature field with the driving loss equation. The physical law loss is calculated based on the combination of the predicted temperature field and thermal conductivity distribution, and the thermal conductivity in the thermal conductivity distribution combined with the physical equation of heat conduction. Based on the predicted thermal conductivity and temperature field, combined with the pre-set reasonableness constraints of the thermal conductivity distribution law, the loss of prior knowledge is determined. The composite loss is calculated from data-driven loss, physical law loss, and prior knowledge loss. The gradient descent algorithm is used to iteratively update the neural network parameters, minimize the composite loss until convergence, and obtain the thermal conductivity distribution of the pile concrete.

3. The method as described in claim 1, characterized in that, The driving loss equation is shown below: in The total number of data points. To predict temperature, For observation data, For the spatial coordinates of the data points, For the time coordinates of the data points.

4. The method as described in claim 2, characterized in that, The physical equation for heat conduction is shown below: in, Loss due to physical laws For the total number of configuration points, To predict the temperature field, To predict thermal conductivity, For temperature gradient, The heat released during hydration is j, where j is the placement point number. For gradient.

5. The method as described in claim 2, characterized in that, The prior knowledge loss is calculated using the following formula: In the formula, For prior knowledge loss, For smoothness constraint loss, For range-constrained losses, The loss is determined by boundary conditions.

6. The method as described in claim 2, characterized in that, The composite loss is calculated using the following formula: in, For composite loss function, Data-driven loss weights Weight loss due to physical laws Weights for prior knowledge loss. For data-driven loss, Loss due to physical laws This is due to the loss of prior knowledge.

7. The method as described in claim 1, characterized in that, The construction of the quantization mapping model includes: Based on the theory of effective media in composite materials, a parameterized physical constitutive relationship between thermal conductivity and density is established. Measured thermal conductivity and density data of standard concrete pile foundations are obtained. The parameters in the constitutive relation are determined by fitting the measured thermal conductivity and density data. The parameters are then substituted into the constitutive relation to obtain a quantitative mapping model.

8. The method as described in claim 1, characterized in that, The quantization mapping model is shown in the following equation: in, and These represent the thermal conductivity of fully poured concrete and concrete with pouring defects, respectively. This is a density index, where 1 indicates density and 0 indicates severe defects.

9. The method as described in claim 1, characterized in that, The process of determining the quality level of each spatial region based on the density index and a preset grading threshold includes: The density level range is constructed by multiple preset grading thresholds; Determine the density level range to which the density index belongs, and use the quality level corresponding to the density level range as the quality level of the spatial area. The quality levels are: I. Healthy, II. Minor Defects, III. Moderate Defects, and IV. Major Defects.

10. The method as described in claim 1, characterized in that, After determining the quality level of each spatial region based on the density index and the preset grading threshold, the process further includes: Based on the quality level of each spatial region, a defect statistics table is generated, and a classification conclusion is drawn. A three-dimensional density cloud map of the pile foundation is constructed based on the density index.

11. A pile foundation compaction assessment system based on thermal conductivity inversion, characterized in that, include: The temperature acquisition module is used to collect three-dimensional spatiotemporal temperature data of measuring points on the pile body by using distributed optical fibers that are spirally wound and fixed along the main reinforcement of the pile foundation steel cage. The inversion module is used to construct a composite loss function of a physical information neural network using three-dimensional spatiotemporal temperature data. The physical information neural network is trained by minimizing the composite loss function to invert the thermal conductivity distribution of the pile concrete. The composite loss function includes a physical law loss term, a data-driven loss term, and a prior knowledge loss term. The compactness assessment module is used to substitute the thermal conductivity distribution into a pre-established quantitative mapping model to obtain the compactness index of the three-dimensional pile body, and to determine the quality level of each spatial region based on the compactness index and the preset grading threshold. Among them, the quantitative mapping model is constructed based on the quantitative mapping between thermal conductivity and concrete density established by the theory of materials physics.

12. The system as claimed in claim 11, characterized in that, The inversion module is specifically used for: By inputting the coordinates from the three-dimensional spatiotemporal temperature data into a physical neural network model, the predicted temperature field and thermal conductivity distribution can be obtained. The data driving loss is calculated by combining the temperature in the three-dimensional spatiotemporal temperature data and the predicted temperature value in the predicted temperature field with the driving loss equation. The physical law loss is calculated based on the combination of the predicted temperature field and thermal conductivity distribution, and the thermal conductivity in the thermal conductivity distribution combined with the physical equation of heat conduction. Based on the predicted thermal conductivity and temperature field, combined with the pre-set reasonableness constraints of the thermal conductivity distribution law, the loss of prior knowledge is determined. The composite loss is calculated from data-driven loss, physical law loss, and prior knowledge loss. The gradient descent algorithm is used to iteratively update the neural network parameters, minimize the composite loss until convergence, and obtain the thermal conductivity distribution of the pile concrete.

13. An electronic device, characterized in that, include: At least one processor and memory; The memory and processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the pile compaction assessment method based on thermal conductivity inversion as described in any one of claims 1 to 10 is implemented.

14. A readable storage medium, characterized in that, It contains an execution program, which, when executed, implements the pile foundation compaction assessment method based on thermal conductivity inversion as described in any one of claims 1 to 10.

Images