A method and device for predicting full-field settlement of a foundation under a heavy rainfall scenario

Abstract

This invention discloses a method and apparatus for predicting full-field foundation settlement under heavy rainfall scenarios. A specific implementation of the method includes: acquiring rainfall parameter data and real-time settlement data; establishing a spatial coordinate system and a time-dimensional foundation computational domain, and building a physical information neural network model; determining mechanical equilibrium constraints and pore pressure diffusion constraints based on the foundation settlement and seepage consolidation mechanism, and determining physical constraint residuals based on physical collocation points; constructing a joint loss function based on the fitting error of settlement data from limited monitoring points and the physical constraint residuals, and training the model to obtain a trained model; responding to the received real-time settlement data and real-time rainfall parameters from limited monitoring points, calling the trained model for forward inference to obtain predicted settlement values, thus forming the full-field distribution of foundation settlement. This implementation can quickly predict full-field settlement under limited monitoring point conditions, reducing the safety assessment time for foundations and structures under heavy rainfall conditions.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of computer technology, specifically to a method and apparatus for predicting full-field foundation settlement under heavy rainfall scenarios. Background Technology

[0002] Currently, under conditions of heavy rainfall, rainwater infiltration significantly alters the pore water pressure distribution in the foundation soil, reducing the effective stress of the foundation and thus causing uneven settlement. In severe cases, this can lead to cracking, tilting, or even failure of the superstructure. For foundation settlement prediction, the common approach is to rely on settlement plates, settlement sensors, or GNSS equipment to obtain settlement information at a finite number of discrete points in the foundation, and then use numerical simulation and data-driven methods for prediction.

[0003] However, when using the above method to predict foundation settlement, the following technical problems often arise:

[0004] Detection equipment can only acquire settlement information at a limited number of discrete points in the foundation, making it difficult to directly obtain the continuous settlement field distribution across the entire foundation area. Numerical simulation methods involve complex model construction, high computational costs, and are highly sensitive to parameter accuracy, making them unsuitable for meeting the engineering requirements for real-time prediction under heavy rainfall conditions. Data-driven prediction methods lack physical constraints, and with a limited number of monitoring points, the prediction results are unstable and cannot meet the reliability requirements of engineering safety assessments, hindering project progress under heavy rainfall conditions and causing significant losses of engineering resources.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] This disclosure discloses a method and apparatus for predicting full-field foundation settlement under heavy rainfall scenarios, along with electronic equipment and computer-readable media, to solve one or more of the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide a method for predicting full-field foundation settlement under heavy rainfall scenarios, including: acquiring rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data during a heavy rainfall process, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation; establishing a foundation computational domain with a spatial coordinate system and a time dimension based on the rainfall parameter data and the real-time settlement data; establishing a physical information neural network model characterizing the foundation response based on the foundation computational domain, the rainfall parameter data, and the real-time settlement data; and consolidating the foundation settlement based on seepage. The mechanism involves determining the mechanical equilibrium constraints and pore pressure diffusion constraints, as well as the physical constraint residuals based on physical collocation points, where the physical collocation points are selected within the aforementioned foundation computational domain. A joint loss function is constructed based on the fitting error of the settlement data from the limited monitoring points and the aforementioned physical constraint residuals. The aforementioned physical information neural network model is then trained to obtain the trained model. In response to receiving real-time settlement data and real-time rainfall parameters from the limited monitoring points, the trained model is invoked for forward inference to obtain the predicted settlement values ​​for the aforementioned foundation computational domain, thus forming the overall foundation settlement distribution.

[0009] Secondly, some embodiments of this disclosure provide a foundation settlement prediction device under heavy rainfall scenarios, comprising: a data acquisition unit configured to acquire rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data during a heavy rainfall process, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation; a computational domain establishment unit configured to establish a foundation computational domain with a spatial coordinate system and a time dimension based on the rainfall parameter data and the real-time settlement data; a model establishment unit configured to establish a physical information neural network model characterizing the foundation response based on the foundation computational domain, the rainfall parameter data, and the real-time settlement data; and a constraint determination unit configured to... Based on the mechanism of foundation settlement and seepage consolidation, mechanical equilibrium constraints and pore pressure diffusion constraints are determined, and physical constraint residuals are determined based on physical collocation points, wherein the aforementioned physical collocation points are selected in the aforementioned foundation calculation domain; the model training unit is configured to construct a joint loss function based on the fitting error of the settlement data of the aforementioned limited monitoring points and the aforementioned physical constraint residuals, and to train the aforementioned physical information neural network model to obtain the trained model; the settlement prediction unit is configured to, in response to receiving the real-time settlement data and real-time rainfall parameters of the aforementioned limited monitoring points, call the aforementioned trained model to perform forward inference to obtain the settlement prediction value of the aforementioned foundation calculation domain, so as to form the full field distribution of foundation settlement.

[0010] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0011] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0012] The above-described embodiments of this disclosure have the following beneficial effects: The method for predicting full-field settlement of foundations under heavy rainfall scenarios, as described in some embodiments of this disclosure, can quickly predict full-field settlement under limited monitoring point conditions, reducing the safety assessment time for foundations and structures under heavy rainfall conditions. Specifically, common detection equipment can only acquire settlement information at a limited number of discrete points in the foundation, making it difficult to directly obtain the continuous settlement field distribution of the entire foundation area. Numerical simulation methods have high computational costs, and data-driven methods lack physical constraints, resulting in long prediction times. Based on this, the method for predicting full-field settlement of foundations under heavy rainfall scenarios, as described in some embodiments of this disclosure, firstly acquires rainfall parameter data and real-time settlement data. The rainfall parameter data is data collected during heavy rainfall, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation. This provides a real-time input data foundation, ensuring that in actual heavy rainfall scenarios, such as urban floods or mountain torrential rain events, changes in rainfall intensity and local settlement signals can be captured in a timely manner, avoiding structural risk assessment errors caused by delays. Secondly, based on the rainfall parameter data and the real-time settlement data, a foundation computational domain with a spatial coordinate system and a time dimension is established. Therefore, a complete spatiotemporal prediction range is defined, enabling continuous coverage of the entire area in practical engineering projects such as high-rise building foundations or bridge foundation monitoring. This compensates for the deficiency of limited monitoring points in reflecting the overall distribution and improves the comprehensiveness of flood season safety early warning. Next, based on the foundation calculation domain, the rainfall parameter data, and the real-time settlement data, a physical information neural network model characterizing the foundation response is established. This achieves intelligent mapping from data to the settlement field, efficiently integrating rainfall and settlement information, reducing the dependence of traditional methods on geological parameters, and improving prediction adaptability under complex geological conditions. Then, based on the foundation settlement and seepage consolidation mechanism, mechanical equilibrium constraints and pore pressure diffusion constraints are determined, and physical constraint residuals are determined based on physical points selected within the foundation calculation domain. This converts the settlement prediction results into actual dimensions, ensuring that the prediction results conform to physical laws. This avoids non-physical biases generated by purely data-driven methods, improves the accuracy of predicting uneven settlement, and prevents cracking or tilting of the foundation structure. Next, based on the fitting error of the settlement data from the limited monitoring points and the residuals of the physical constraints, a joint loss function is constructed, and the aforementioned physical information neural network model is trained to obtain the trained model. This improves generalization ability under limited data, ensures prediction stability under different rainfall intensities, and reduces uncertainty in engineering decisions. Finally, in response to receiving real-time settlement data and real-time rainfall parameters from the limited monitoring points, the trained model is invoked for forward inference to obtain the predicted settlement values ​​for the foundation computational domain, thus forming the overall distribution of foundation settlement.In summary, by fusing monitoring data with physical constraints through a physical information neural network, rapid prediction without iteration can be achieved in actual heavy rainfall scenarios. This reduces the safety assessment time for foundations and structures under heavy rainfall conditions, thereby improving the real-time performance and reliability of engineering safety assessments and avoiding sudden losses caused by disasters. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a flowchart of some embodiments of the method for predicting full-field settlement of foundations under heavy rainfall scenarios according to this disclosure;

[0015] Figure 2 This is a structural schematic diagram of some embodiments of the foundation settlement prediction device under heavy rainfall scenarios according to the present disclosure;

[0016] Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure.

[0017] Figure 4 This is a schematic diagram of a physical information neural network model structure suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0018] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0019] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0020] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0021] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0022] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0023] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] refer to Figure 1 The flowchart 100 illustrates some embodiments of the method for predicting full-field foundation settlement under heavy rainfall scenarios according to the present disclosure. This method for predicting full-field foundation settlement under heavy rainfall scenarios includes the following steps:

[0025] Step 101: Obtain rainfall parameter data and real-time settlement data.

[0026] In some embodiments, the execution entity (e.g., electronic device) of the above-described method for predicting full-field foundation settlement under heavy rainfall scenarios can be hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster composed of multiple servers or terminal devices, or as a single server or a single terminal device. When the computing device is software, it can be installed in the hardware devices listed above. It can be implemented as multiple software programs or software modules to provide distributed services, or as a single software program or software module. No specific limitations are made here.

[0027] In some embodiments, the aforementioned implementing entity can acquire rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data from a heavy rainfall event, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation. These limited monitoring points can be locations where a limited number of sensors are deployed in the actual foundation engineering project. As an example, the rainfall parameter data can be quantitative data related to rainfall during a heavy rainfall event. In practice, the rainfall parameter data can be acquired through meteorological monitoring equipment (e.g., rain gauges, weather stations). The real-time settlement data can be settlement observations collected in real-time by monitoring devices deployed within the foundation. This real-time settlement data can be collected through settlement plates, settlement sensors, or GNSS equipment.

[0028] In some optional implementations of certain embodiments, the rainfall parameter data includes one or more of rainfall intensity, cumulative rainfall, and rainfall duration. The rainfall intensity can be the rainfall amount per unit time. The cumulative rainfall can be the total rainfall from the start of rainfall to a certain moment. The rainfall duration can be the length of time elapsed from the start of rainfall to the current moment. Optionally, in different embodiments, one or any combination of the above parameters can be selected as model input according to engineering needs and data availability.

[0029] In some optional implementations of certain embodiments, the aforementioned real-time settlement data is vertical settlement data, wherein the aforementioned vertical settlement data is acquired in real time by a settlement monitoring device.

[0030] Step 102: Based on rainfall parameter data and real-time settlement data, establish a spatial coordinate system and a ground calculation domain with a time dimension.

[0031] In some embodiments, the aforementioned executing entity can establish a foundation calculation domain with a spatial coordinate system and a time dimension based on the aforementioned rainfall parameter data and the aforementioned real-time settlement data. The foundation calculation domain can be a calculation area containing spatial coordinates (e.g., two-dimensional or three-dimensional coordinates) and a time dimension. As an example, a calculation domain can be constructed based on the spatial extent of the engineering foundation, and the rainfall parameter data and real-time settlement data can be mapped into this domain.

[0032] Step 103: Based on the foundation computational domain, rainfall parameter data, and real-time settlement data, establish a physical information neural network model characterizing the foundation response.

[0033] In some embodiments, the aforementioned execution entity may establish a physical information neural network model characterizing the foundation response based on the aforementioned foundation computing domain, the aforementioned rainfall parameter data, and the aforementioned real-time settlement data.

[0034] In some optional implementations of certain embodiments, the aforementioned physical information neural network model is a neural network model that takes data representing spatial coordinates, time, and rainfall parameters as input and outputs data representing the settlement field and pore water pressure field. The aforementioned physical information neural network model can employ a multi-layer feedforward structure (fully connected layers). The input to the aforementioned physical information neural network model can be data samples representing spatial coordinates, time, and rainfall parameters. The output of the aforementioned physical information neural network model can be data samples representing the vertical settlement of the foundation and pore water pressure.

[0035] As an example, the above rainfall parameter data can be expressed in the following form: Settlement monitoring points within the foundation can be... . No. The coordinates of each monitoring point can be .exist The observation value at time can be Then the real-time data sample can be .

[0036] In practice, the input to the above-mentioned physical information neural network model can be... The output of the aforementioned physical information neural network model can be the vertical settlement of the foundation. and pore pressure The aforementioned pore water pressure can be used for physical constraint calculations, allowing for calculations without direct monitoring.

[0037] Optionally, in response to the model entering training and real-time inference, the preprocessing of input and observation data can be carried out in one of the following ways: outlier handling, missing data handling, and variable normalization. The outlier handling method can be a threshold filtering or median filtering of the settlement observation sequence to remove obvious jumps. The missing data handling method can be a method that, in response to a missing data point at a certain time, uses linear interpolation between adjacent time points to fill in the missing data. The variable normalization method can be a method of normalizing the input variables.

[0038] Step 104: Based on the foundation settlement and seepage consolidation mechanism, determine the mechanical equilibrium constraints and pore pressure diffusion constraints, and based on the physical collocation, determine the physical constraint residuals.

[0039] In some embodiments, the aforementioned execution entity can determine mechanical equilibrium constraints and pore pressure diffusion constraints based on the foundation settlement and seepage consolidation mechanism, and determine physical constraint residuals based on physical collocation points, wherein the aforementioned physical collocation points are selected within the aforementioned foundation calculation domain. As an example, several physical collocation points are sampled within the calculation domain (e.g., uniform or random sampling), and the residuals are calculated using automatic differentiation. The aforementioned foundation settlement and seepage consolidation mechanism can be a fundamental principle in soil mechanics describing the vertical displacement (settlement) of foundation soil under load (e.g., self-weight, rainfall, etc.) and the volume change (consolidation) of soil due to pore water pressure dissipation. This fundamental principle can include the effective stress principle and Darcy's law. The aforementioned mechanical equilibrium constraints can be physical constraints imposed on the settlement field predicted by the neural network according to mechanical equilibrium equations (e.g., stress equilibrium conditions or the geometric relationship between strain and displacement). The aforementioned pore pressure diffusion constraints can be physical constraints imposed on the pore water pressure field predicted by the neural network by the governing equations of pore water pressure changing with time. The aforementioned physical collocation points can be discrete points randomly or uniformly selected within the foundation calculation domain. The aforementioned discrete points can include spatial coordinates and temporal coordinates. The aforementioned physical constraint residuals can be the difference between the left and right sides of the physical control equations after substituting the settlement and pore pressure output by the neural network into the physical control equations at the physical collocation points. For example, the aforementioned physical constraint residuals can be the non-zero value on the left side of the mechanical equilibrium equations or the difference between the left and right sides of the pore pressure diffusion equations.

[0040] In some optional implementations of certain embodiments, the aforementioned mechanical equilibrium constraints are used to constrain the coupling relationship between the settlement field and the pore water pressure field, and the aforementioned pore pressure diffusion constraints are used to constrain the diffusion evolution of the pore water pressure field over time, as well as the correlation between the pore water pressure field and rainfall parameters. The aforementioned mechanical equilibrium constraints can be... The aforementioned pore pressure diffusion constraint can be... Among them, the above These can be coefficients related to soil properties. (The above...) These can be parameters characterizing the foundation displacement field. (The above...) This can be a parameter characterizing pore pressure. (The above...) This could be the partial derivative of pore water pressure with respect to time. (The above...) This can be the diffusion coefficient. The diffusion coefficient mentioned above is related to the permeability and compressibility of the soil. The above... It can be the Laplace operator for pore water pressure. (The above...) This can be a coefficient characterizing the impact of rainfall. (The above...) This can be a time function of rainfall parameters (e.g., rainfall intensity or cumulative rainfall). The above. It can be a time parameter.

[0041] Step 105: Based on the fitting error and physical constraint residual of the settlement data from limited monitoring points, construct a joint loss function and train the physical information neural network model to obtain the trained model.

[0042] In some embodiments, the execution entity can construct a joint loss function based on the fitting error of the settlement data from the limited monitoring points and the physical constraint residuals, and train the physical information neural network model to obtain the trained model. The joint loss function can be a loss function consisting of a weighted sum of the monitoring data loss and the physical constraint residuals. The monitoring data loss can be the sum of squared errors between the settlement value predicted by the neural network and the measured settlement value.

[0043] As an example, the aforementioned monitoring data loss could be The aforementioned physical constraint residual loss can be... The total loss function can then be... Among them, the above This could be a loss function for monitoring data. (The above...) This can be used to measure the error between model-predicted settlement and measured settlement. (The above...) It can be the index value of the monitoring point (e.g., =1, 2, ..., , (This could be the total number of monitoring points). The above. It can be the index value of the time step (e.g., =1, 2, ..., , (This could be the total number of monitoring times). The above. It can be the first The spatial coordinates (horizontal and vertical) of each monitoring point. (The above...) It can be the first Each monitoring point. (The above) This could be a settlement value predicted by a neural network. (The above...) It can be the first Each monitoring point is at The measured settlement value at time [time]. The above. This could be the physical residual loss function. (The above...) This can be used to measure the degree to which the model output violates the governing equations at physical collocation points. The above... (Subscript) can be a physical coordinate index (e.g., =1, 2, ..., , (This could be the total number of monitoring points). The above. It can be the first The spatial coordinates and time of each physical point. (The above...) This could be a settlement value predicted by a neural network. (The above...) This could be the pore water pressure value predicted by a neural network. (The above...) These can be coefficients in mechanical equilibrium constraints. (The above...) It can be used to reflect the degree of influence of pore water pressure on the settlement field. The above... This could be the partial derivative of pore water pressure with respect to time. (The above...) This can be the pore pressure diffusion coefficient. The aforementioned diffusion coefficient is related to the permeability and compressibility of the soil. This can be a coefficient characterizing the impact of rainfall. (The above...) It can be in Rainfall parameters at any given time. (Above) This could be a joint total loss function. (The above...) It can be used to train neural networks. (The above...) It can be a weighting coefficient. (The above) It can be used to balance the fitting of monitoring data with physical consistency constraints.

[0044] In some optional implementations of certain embodiments, the coefficients in the aforementioned mechanical equilibrium constraints and pore pressure diffusion constraints are preset constants, wherein the preset constants are constants determined by normalization based on the foundation scale, time scale, and dimensions of rainfall parameters; or, the aforementioned coefficients are coefficients obtained by the joint inversion of the parameters to be estimated from monitoring data and physical constraints during the optimization process of the aforementioned joint loss function; the aforementioned physical constraint residuals are determined by automatic differentiation at the aforementioned physical collocation points, wherein the aforementioned physical constraint residuals include mechanical equilibrium constraint residuals and pore pressure diffusion constraint residuals.

[0045] Step 106: In response to receiving real-time settlement data and real-time rainfall parameters from a limited number of monitoring points, the trained model is invoked to perform forward inference to obtain the settlement prediction value of the foundation computational domain, so as to form the full-field distribution of foundation settlement.

[0046] In some embodiments, the execution entity may, in response to receiving real-time settlement data and real-time rainfall parameters from the limited monitoring points, invoke the trained model to perform forward inference to obtain the predicted settlement value of the foundation computational domain, thereby forming a full-field distribution of foundation settlement. The real-time rainfall parameters may be rainfall characteristic data acquired in real-time during heavy rainfall. The real-time settlement data may be vertical displacement data collected in real-time by settlement monitoring devices deployed within the foundation during heavy rainfall. The predicted settlement value may be the vertical settlement value output by the model. The full-field distribution of foundation settlement may be a continuous distribution field composed of settlement values ​​at various spatial locations (full field) within the foundation computational domain.

[0047] In some optional implementations of certain embodiments, the above-mentioned full-field distribution of foundation settlement includes a preset grid point settlement value table, settlement profile data, and settlement contour line data.

[0048] Optionally, after model training is complete, the system proceeds to the real-time prediction phase in response to a new time frame. Receive real-time settlement data and real-time rainfall parameters The entire settlement can be obtained by performing forward inference from a neural network without the need for iterative solutions using traditional numerical models. First, a prediction grid can be established. For example, preset grid points within the computational threshold could be... The above. This could be a set of predicted grid points. For example, a predefined set of spatial locations (grid points) within the foundation computational domain, used to output the predicted total settlement values. This could be the horizontal coordinates of the predicted grid points (e.g., coordinates along the transverse direction of the foundation). The above. This can be the predicted vertical coordinates (depth or elevation) of the grid points. In response to performing forward and backward inference, the inference formula can be... The above. It can be the trained model in At this point, the current time is The predicted settlement value. The above. (Superscript) can indicate a prediction result. The above It can be the forward inference of a trained model, with spatial coordinates as input. ,time and rainfall parameters The settlement value at that time. The above. It can represent the current moment, the first... At each point in time. The above. This could mean performing the aforementioned forward inference on each grid point of the predicted grid points.

[0049] The above-described embodiments of this disclosure have the following beneficial effects: The method for predicting full-field settlement of foundations under heavy rainfall scenarios, as described in some embodiments of this disclosure, can quickly predict full-field settlement under limited monitoring point conditions, reducing the safety assessment time for foundations and structures under heavy rainfall conditions. Specifically, common detection equipment can only acquire settlement information at a limited number of discrete points in the foundation, making it difficult to directly obtain the continuous settlement field distribution of the entire foundation area. Numerical simulation methods have high computational costs, and data-driven methods lack physical constraints, resulting in long prediction times. Based on this, the method for predicting full-field settlement of foundations under heavy rainfall scenarios, as described in some embodiments of this disclosure, firstly acquires rainfall parameter data and real-time settlement data. The rainfall parameter data is data collected during heavy rainfall, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation. This provides a real-time input data foundation, ensuring that in actual heavy rainfall scenarios, such as urban floods or mountain torrential rain events, changes in rainfall intensity and local settlement signals can be captured in a timely manner, avoiding structural risk assessment errors caused by delays. Secondly, based on the rainfall parameter data and the real-time settlement data, a foundation computational domain with a spatial coordinate system and a time dimension is established. Therefore, a complete spatiotemporal prediction range is defined, enabling continuous coverage of the entire area in practical engineering projects such as high-rise building foundations or bridge foundation monitoring. This compensates for the inability of limited monitoring points to reflect the overall distribution and improves the comprehensiveness of flood season safety early warnings. Next, based on the aforementioned foundation calculation domain, rainfall parameter data, and real-time settlement data, a physical information neural network model characterizing foundation response is established. This achieves intelligent mapping from data to the settlement field, efficiently integrating rainfall and settlement information, reducing the dependence of traditional methods on geological parameters, and improving prediction adaptability under complex geological conditions. Then, based on the foundation settlement and seepage consolidation mechanism, mechanical equilibrium constraints and pore pressure diffusion constraints are determined, and physical constraint residuals are determined based on physical points selected within the aforementioned foundation calculation domain. This converts the settlement prediction results into actual dimensions, ensuring that the prediction results conform to physical laws. This avoids non-physical biases generated by purely data-driven methods, improves the accuracy of predicting uneven settlement, and prevents cracking or tilting of foundation structures. Next, based on the fitting error of the settlement data from the limited monitoring points and the residuals of the physical constraints, a joint loss function is constructed, and the aforementioned physical information neural network model is trained to obtain the trained model. This improves generalization ability under limited data, ensures prediction stability under different rainfall intensities, and reduces uncertainty in engineering decisions. Finally, in response to receiving real-time settlement data and real-time rainfall parameters from the limited monitoring points, the trained model is invoked for forward inference to obtain the predicted settlement values ​​for the foundation computational domain, thus forming the overall distribution of foundation settlement.In summary, by fusing monitoring data with physical constraints through a physical information neural network, rapid prediction without iteration can be achieved in actual heavy rainfall scenarios. This reduces the safety assessment time for foundations and structures under heavy rainfall conditions, thereby improving the real-time performance and reliability of engineering safety assessments and avoiding sudden losses caused by disasters.

[0050] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a foundation full-field settlement prediction device under heavy rainfall scenarios. These device embodiments are similar to... Figure 1 Corresponding to the method embodiments shown, this ground settlement prediction device under heavy rainfall scenarios can be specifically applied to various electronic devices.

[0051] like Figure 2 As shown, a foundation settlement prediction device 200 under heavy rainfall scenarios includes: a data acquisition unit 201, a computational domain establishment unit 202, a model establishment unit 203, a constraint determination unit 204, a model training unit 205, and a settlement prediction unit 206. The data acquisition unit 201 is configured to acquire rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data collected during heavy rainfall, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation. The computational domain establishment unit 202 is configured to establish a foundation computational domain with a spatial coordinate system and a time dimension based on the rainfall parameter data and the real-time settlement data. The model establishment unit 203 is configured to establish a physical information neural network model characterizing the foundation response based on the foundation computational domain, the rainfall parameter data, and the real-time settlement data. The constraint determination unit 204 is configured to: determine mechanical equilibrium constraints and pore pressure diffusion constraints based on the foundation settlement and seepage consolidation mechanism, and determine physical constraint residuals based on physical collocation points, wherein the aforementioned physical collocation points are selected within the aforementioned foundation calculation domain. The model training unit 205 is configured to: construct a joint loss function based on the fitting error of the settlement data from the aforementioned limited monitoring points and the aforementioned physical constraint residuals, and train the aforementioned physical information neural network model to obtain the trained model. The settlement prediction unit 206 is configured to: in response to receiving real-time settlement data and real-time rainfall parameters from the aforementioned limited monitoring points, call the trained model to perform forward inference to obtain the predicted settlement values ​​for the aforementioned foundation calculation domain, thereby forming the overall foundation settlement distribution.

[0052] It is understandable that the units recorded in the 200-meter foundation settlement prediction device under this heavy rainfall scenario are related to the reference... Figure 1 The steps described in the method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the foundation settlement prediction device 200 and its constituent units under heavy rainfall scenarios, and will not be repeated here.

[0053] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device (e.g., an electronic device) 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0054] like Figure 3 As shown, the electronic device 300 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0055] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0056] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0057] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0058] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0059] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data from a heavy rainfall process, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation; establish a foundation computational domain with a spatial coordinate system and a time dimension based on the rainfall parameter data and the real-time settlement data; establish a physical information neural network model characterizing the foundation response based on the foundation computational domain, the rainfall parameter data, and the real-time settlement data; and establish a physical information neural network model characterizing the foundation response based on the foundation settlement data. The settlement and seepage consolidation mechanism is used to determine the mechanical equilibrium constraints and pore pressure diffusion constraints, as well as the physical constraint residuals based on physical collocation points, which are selected in the aforementioned foundation calculation domain. Based on the fitting error of the settlement data from the aforementioned limited monitoring points and the aforementioned physical constraint residuals, a joint loss function is constructed, and the aforementioned physical information neural network model is trained to obtain the trained model. In response to receiving the real-time settlement data and real-time rainfall parameters from the aforementioned limited monitoring points, the trained model is invoked to perform forward inference to obtain the settlement prediction values ​​for the aforementioned foundation calculation domain, thereby forming the overall distribution of foundation settlement.

[0060] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0061] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0062] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a data acquisition unit, a computational domain establishment unit, a model establishment unit, a constraint determination unit, a model training unit, and a settlement prediction unit. The names of these units do not necessarily limit the unit itself; for example, the data acquisition unit may also be described as "a unit that acquires rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data from a heavy rainfall process, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation."

[0063] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0064] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for predicting full-field foundation settlement under heavy rainfall scenarios, comprising: Acquire rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data during a heavy rainfall process, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation; Based on the rainfall parameter data and the real-time settlement data, a ground calculation domain with a spatial coordinate system and a time dimension is established; Based on the foundation computing domain, the rainfall parameter data, and the real-time settlement data, a physical information neural network model characterizing the foundation response is established. Based on the foundation settlement and seepage consolidation mechanism, mechanical equilibrium constraints and pore pressure diffusion constraints are determined, and physical constraint residuals are determined based on physical collocation points, wherein the physical collocation points are selected in the foundation calculation domain; Based on the fitting error of the settlement data from the limited monitoring points and the physical constraint residuals, a joint loss function is constructed, and the physical information neural network model is trained to obtain the trained model. In response to receiving real-time settlement data and real-time rainfall parameters from the limited monitoring points, the trained model is invoked to perform forward inference to obtain the settlement prediction value of the foundation computation domain, thereby forming the full-field distribution of foundation settlement.

2. The method for predicting full-field foundation settlement under heavy rainfall scenarios according to claim 1, wherein, The rainfall parameter data includes one or more of rainfall intensity, cumulative rainfall, and rainfall duration.

3. The method for predicting full-field foundation settlement under heavy rainfall scenarios according to claim 1, wherein, The real-time settlement data is vertical settlement data, which is acquired in real time by the settlement monitoring device.

4. The method for predicting full-field foundation settlement under heavy rainfall scenarios according to claim 1, wherein, The physical information neural network model is a neural network model that takes data representing spatial coordinates, time and rainfall parameters as input and data representing settlement field and pore water pressure field as output.

5. The method for predicting full-field foundation settlement under heavy rainfall scenarios according to claim 1, wherein, The mechanical equilibrium constraint is used to constrain the coupling relationship between the settlement field and the pore water pressure field, and the pore pressure diffusion constraint is used to constrain the diffusion evolution of the pore water pressure field over time, as well as the correlation between the pore water pressure field and rainfall parameters.

6. The method for predicting full-field foundation settlement under heavy rainfall scenarios according to claim 1, wherein, The coefficients in the mechanical equilibrium constraint and the pore pressure diffusion constraint are preset constants, wherein the preset constants are constants determined by normalization based on the foundation scale, time scale and the dimensions of the rainfall parameter; or, the coefficients are obtained by the joint inversion of the parameters to be estimated by the monitoring data and physical constraints during the optimization process of the joint loss function; the physical constraint residuals are determined by automatic differentiation at the physical collocation points, wherein the physical constraint residuals include the mechanical equilibrium constraint residuals and the pore pressure diffusion constraint residuals.

7. The method for predicting full-field foundation settlement under heavy rainfall scenarios according to claim 1, wherein, The overall distribution of foundation settlement includes a pre-set grid point settlement value table, settlement profile data, and settlement contour line data.

8. A device for predicting full-field foundation settlement under heavy rainfall scenarios, comprising: The data acquisition unit is configured to acquire rainfall parameter data and real-time settlement data, wherein the rainfall parameter data is data during a heavy rainfall process, and the real-time settlement data is data collected from a limited number of monitoring points within the foundation. The computational domain establishment unit is configured to establish a ground-based computational domain with a spatial coordinate system and a time dimension based on the rainfall parameter data and the real-time settlement data. The model building unit is configured to build a physical information neural network model characterizing the foundation response based on the foundation computing domain, the rainfall parameter data, and the real-time settlement data. The constraint determination unit is configured to determine mechanical equilibrium constraints and pore pressure diffusion constraints based on the foundation settlement and seepage consolidation mechanism, and to determine physical constraint residuals based on physical collocation points, wherein the physical collocation points are selected in the foundation calculation domain; The model training unit is configured to construct a joint loss function based on the fitting error of the settlement data at the limited monitoring points and the physical constraint residuals, and to train the physical information neural network model to obtain the trained model. The settlement prediction unit is configured to, in response to receiving real-time settlement data and real-time rainfall parameters from the limited monitoring points, call the trained model to perform forward inference to obtain the settlement prediction value of the foundation computation domain, so as to form the full-field distribution of foundation settlement.

9. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images