Foundation pit horizontal displacement prediction method, device, system, electronic equipment and medium

By performing abnormal checking and smooth optimization of the inclination measurement data of foundation pit depth changes, and using a multimodal neural network model based on attention mechanism for prediction, the problem of low accuracy of deep horizontal displacement monitoring of foundation pits in the prior art is solved. Achieve higher prediction accuracy and construction safety.

CN118114571BActive Publication Date: 2024-11-26广州广检建设工程检测中心有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410338512.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-22
Publication Date
2024-11-26
Estimated Expiration
2044-03-22

AI Technical Summary

Technical Problem

The existing deep horizontal displacement monitoring method of foundation pit relies on manual operation, and there are problems such as easy damage to the measurement points, large workload, and reaction lag. The prediction algorithm fails to effectively consider the type of support structure, soil layer conditions and construction conditions, resulting in prediction. Low accuracy.

Method used

By performing abnormal checks and smooth optimization of the pending inclination measurement data of foundation pit depth changes, identifying influencing factors, using a multimodal neural network model based on attention mechanism for prediction, considering time course data and influencing factors, outputting the next moment predicted horizontal displacement.

Benefits of technology

It improves the accuracy and interpretability of the prediction of horizontal displacement of foundation pits, enhances the early warning capability of foundation pit construction process, and ensures production safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118114571B_ABST
    Figure CN118114571B_ABST
Patent Text Reader

Abstract

The present application is applicable to the field of engineering monitoring technology, and provides a method, device, system, electronic device and medium for predicting the horizontal displacement of a foundation pit. The method for predicting the horizontal displacement of a foundation pit includes: obtaining target inclinometric data by performing abnormal inspection and smoothing optimization on the inclinometric data to be processed that changes with the depth of the foundation pit; performing data analysis on the target inclinometric data, identifying the factors affecting the change of the horizontal displacement of the foundation pit according to the change law of the target inclinometric data, and performing time-course processing on the inclinometric data corresponding to multiple influencing factors to obtain the time-course data of the target inclinometric data; inputting the target inclinometric data and the time-course data at the current moment into a multimodal neural network model based on an attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometric data corresponding to the target inclinometric data at the next moment, so as to predict the horizontal displacement of the foundation pit at the next moment. The above scheme can improve the accuracy of the prediction of the horizontal displacement of the foundation pit, and further prevent sudden accidents that may occur during the construction process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the field of engineering monitoring technology, and in particular, relates to a method, device, system, electronic equipment and medium for predicting horizontal displacement of a foundation pit. Background Art

[0002] With the rapid development of urban construction, the scale of foundation pit projects is constantly expanding. The requirements and pressures faced in ensuring the safe production of foundation pit projects are also gradually increasing. The excavation of deep foundation pits not only causes soil settlement and structural deformation, but also poses a threat to the stability of surrounding buildings. Among them, deep horizontal displacement, as a key parameter, is crucial to project safety.

[0003] At present, the monitoring of deep horizontal displacement of foundation pits mainly relies on manual operation, supplemented by automated monitoring methods. However, traditional monitoring methods have a series of problems, such as the monitoring points are easily damaged, the operation workload is large, and the response to subsequent changes and emergencies in construction is delayed. Therefore, in order to achieve early warning during foundation pit construction, deep horizontal displacement is usually predicted. However, most of the current prediction algorithms only consider time-history displacement data, lack of consideration of support structure type, soil condition, construction conditions, etc., resulting in low prediction accuracy. Summary of the invention

[0004] The embodiments of the present application provide a method, device, system, electronic equipment and medium for predicting horizontal displacement of a foundation pit, which can improve the accuracy of predicting horizontal displacement of a foundation pit and further prevent unexpected accidents that may occur during the construction process.

[0005] In a first aspect, an embodiment of the present application provides a method for predicting horizontal displacement of a foundation pit, and the method for predicting horizontal displacement of a foundation pit includes:

[0006] Obtain the inclinometric data to be processed that varies with the depth of the foundation pit;

[0007] Perform abnormality check and smooth optimization on the inclinometer data to be processed to obtain the target inclinometer data;

[0008] Perform data analysis on the target inclination measurement data to determine the change law of the target inclination measurement data, and identify the influencing factors that affect the change of the horizontal displacement of the foundation pit based on the change law of the target inclination measurement data, perform time-history processing on the inclination measurement data corresponding to multiple influencing factors, and obtain the time-history data of the target inclination measurement data;

[0009] The target inclinometer data and time-course data at the current moment are input into a multimodal neural network model based on the attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment to predict the horizontal displacement of the foundation pit at the next moment.

[0010] In a possible implementation, the inclinometric data to be processed includes an inclinometric curve of the foundation pit retaining structure changing with depth; abnormality inspection and smoothing optimization are performed on the inclinometric data to be processed to obtain target inclinometric data, including:

[0011] Fitting the inclinometer curve based on the least square method to determine the fitting value;

[0012] According to the fitting value and the true value of the inclinometric data to be processed, a mean absolute error and a root mean square error between the fitting value and the true value are determined;

[0013] According to the mean absolute error and root mean square error, the smoothness of the optimized inclinometric data and the abnormal data are determined;

[0014] According to the smoothness of the optimized inclinometer data and the abnormal data, the target inclinometer data is determined.

[0015] In a possible implementation, the smoothing degree and abnormal data of the optimized inclinometric data are determined according to the mean absolute error and the root mean square error, including:

[0016] If the mean absolute error and the root mean square error are less than the first error threshold, it is determined that the smoothing degree of the optimized inclinometric data meets the optimization smoothing requirement;

[0017] If the mean absolute error and the root mean square error are greater than the second error threshold, the corresponding inclinometer data is determined to be abnormal data.

[0018] In a possible implementation, the time history data is a one-dimensional vector and includes at least: the depth of the target inclinometer data, the monitoring interval, the current depth inclinometer data, and the excavation depth.

[0019] In one possible implementation, the training process of the multimodal neural network model based on the attention mechanism includes:

[0020] Obtain the inclinometric curve image and time history data of the foundation pit retaining structure changing with depth at any time, where the time history data is obtained by optimizing and smoothing the inclinometric curve;

[0021] The inclinometer curve images and time course data are input into the multimodal neural network model based on the attention mechanism, and the network parameters of the multimodal neural network model based on the attention mechanism are optimized through back propagation and model evaluation indicators;

[0022] When the model evaluation index reaches the preset threshold, the training of the multimodal neural network model based on the attention mechanism is completed.

[0023] In a possible implementation, the method for predicting horizontal displacement of a foundation pit further includes:

[0024] Obtain the bending moment function of the enclosure structure that varies with depth and the internal force function of the enclosure structure that varies with depth;

[0025] According to the excavation depth of the foundation pit and the bending moment function, determine the bending moment value corresponding to the excavation depth;

[0026] According to the excavation depth of the foundation pit and the internal force function, determine the internal force value corresponding to the excavation depth;

[0027] The predicted safety state of the retaining structure is determined based on the bending moment value corresponding to the excavation depth and the internal force value corresponding to the excavation depth.

[0028] In a second aspect, an embodiment of the present application provides a device for predicting horizontal displacement of a foundation pit, wherein the device for predicting horizontal displacement of a foundation pit comprises:

[0029] A data acquisition module is used to acquire inclination data to be processed that varies with the depth of the foundation pit;

[0030] A smoothing optimization module is used to perform abnormality inspection and smoothing optimization on the inclination measurement data to be processed to obtain the target inclination measurement data;

[0031] The time history data determination module is used to perform data analysis on the target inclination measurement data, determine the change law of the target inclination measurement data, and identify the influencing factors that affect the change of the horizontal displacement of the foundation pit according to the change law of the target inclination measurement data, and perform time history processing on the inclination measurement data corresponding to multiple influencing factors to obtain the time history data of the target inclination measurement data;

[0032] The prediction module is used to input the target inclinometer data and time-course data at the current moment into a multimodal neural network model based on the attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment to predict the horizontal displacement of the foundation pit at the next moment.

[0033] In a third aspect, an embodiment of the present application provides a foundation pit horizontal displacement prediction system, wherein the foundation pit horizontal displacement prediction system comprises an inclinometer and a foundation pit horizontal displacement prediction device, wherein the inclinometer comprises a plurality of fixed inclinometer units;

[0034] The inclinometer unit is installed and buried in the inclinometer tube, and is used to collect the inclinometer data to be processed that changes with the depth of the deep pit using the inclinometer sensor, and transmit the inclinometer data to be processed to the foundation pit horizontal displacement prediction device;

[0035] The device for predicting horizontal displacement of a foundation pit is used to execute the method for predicting horizontal displacement of a foundation pit as described in the first aspect above.

[0036] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the aforementioned method for predicting horizontal displacement of a foundation pit when executing the computer program.

[0037] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for predicting horizontal displacement of a foundation pit.

[0038] In a sixth aspect, an embodiment of the present application provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute the above-mentioned method for predicting horizontal displacement of a foundation pit.

[0039] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0040] The embodiment of the present application obtains target inclinometer data by performing abnormality check and smoothing optimization on the acquired inclinometer data to be processed that changes with the depth of the foundation pit. The prediction based on the target inclinometer data improves the prediction progress compared with the prediction based on the inclinometer data to be processed. Secondly, data analysis is performed on the target inclinometer data to determine the changing law of the target inclinometer data. According to the changing law of the target inclinometer data, the influencing factors affecting the change of the horizontal displacement of the foundation pit are identified. The inclinometer data corresponding to multiple influencing factors are time-processed to obtain the time-course data of the target inclinometer data. Finally, the target inclinometer data and the time-course data are input into a multimodal neural network model of an attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment. When making predictions, the multimodal neural network model based on the attention mechanism takes into account the displacement influencing factors that affect the horizontal displacement of the foundation pit, so that the method has strong interpretability and high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 It is a structural schematic diagram of a foundation pit horizontal displacement prediction system provided by an embodiment of the present application;

[0043] Figure 2 It is a flow chart of a method for predicting horizontal displacement of a foundation pit provided in one embodiment of the present application;

[0044] Figure 3 It is a schematic diagram of abnormal data;

[0045] Figure 4 This is a normal data diagram;

[0046] Figure 5 It is a schematic flow chart of a method for predicting horizontal displacement of a foundation pit provided in another embodiment of the present application;

[0047] Figure 6 It is a structural schematic diagram of a foundation pit horizontal displacement prediction device provided in an embodiment of the present application;

[0048] Figure 7 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application; DETAILED DESCRIPTION

[0049] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0050] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0051] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0052] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0053] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0054] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0055] It should be understood that the size of the serial numbers of the steps in this embodiment does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0056] With the rapid development of urban construction, the scale of foundation pit projects is constantly expanding. The requirements and pressures faced in ensuring the safe production of foundation pit projects are also gradually increasing. The excavation of deep foundation pits not only causes soil settlement and structural deformation, but also poses a threat to the stability of surrounding buildings. Among them, deep horizontal displacement, as a key parameter, is crucial to engineering safety. In order to achieve early warning during foundation pit construction and predict deep horizontal displacement, researchers tend to use soft computing methods instead of cumbersome theoretical analysis methods and finite element numerical simulation methods that need to simulate actual conditions. Commonly used soft computing methods for displacement prediction include gray fuzzy, time series regression, Kalman filtering, support vector machine, random forest, artificial neural network, and neural network based on optimization algorithms such as genetic algorithms. However, the current monitoring method for deep horizontal displacement of foundation pits is still mainly based on manual monitoring, supplemented by automatic monitoring. Traditional monitoring methods have problems such as easy destruction of measuring points, large workload, and delayed response to subsequent changes and emergencies in construction, resulting in low quality of deep horizontal displacement data in actual foundation pit projects. The accuracy and reliability of early warning using the above-mentioned common prediction algorithms face great challenges and cannot be put into practical use. On the other hand, most of the current prediction algorithms only consider time-history displacement data, and lack consideration of support structure type, soil conditions, construction conditions, etc., resulting in low accuracy in the prediction of horizontal displacement of foundation pits.

[0057] In order to solve the above problems, the present application provides a method for predicting the horizontal displacement of a foundation pit. The target inclinometer data can be obtained by performing abnormality inspection and smoothing optimization on the acquired inclinometer data to be processed that changes with the depth of the foundation pit. The prediction based on the target inclinometer data improves the prediction progress compared with the prediction based on the inclinometer data to be processed. Secondly, the target inclinometer data is analyzed to determine the changing law of the target inclinometer data. According to the changing law of the target inclinometer data, the influencing factors affecting the change of the horizontal displacement of the foundation pit are identified. The inclinometer data corresponding to multiple influencing factors are time-processed to obtain the time-course data of the target inclinometer data. Finally, the target inclinometer data and the time-course data are input into a multimodal neural network model with an attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment. When making predictions, the multimodal neural network model based on the attention mechanism takes into account the displacement influencing factors that affect the horizontal displacement of the foundation pit, so that the method has strong interpretability and high accuracy.

[0058] The following is a detailed description of the foundation pit horizontal displacement prediction method, device, system, electronic equipment, medium and computer program provided in the present application with reference to the accompanying drawings.

[0059] Figure 1 The schematic diagram of the structure of a foundation pit horizontal displacement prediction system provided by an embodiment of the present application is shown. The foundation pit horizontal displacement prediction system includes an inclinometer and a foundation pit horizontal displacement prediction device; the inclinometer is connected to the foundation pit horizontal displacement prediction device, and the inclinometer includes a plurality of fixed inclinometer units.

[0060] The foundation pit horizontal displacement prediction system of the embodiment of the present application can be applied to application scenarios such as foundation pit engineering monitoring and management, geological disaster monitoring and early warning. Among them, the inclinometer in the foundation pit horizontal displacement is composed of multiple fixed inclinometer units, each of which is installed in the inclinometer tube around the foundation pit. Each inclinometer unit is equipped with an inclinometer sensor for collecting inclinometer data to be processed under the change of foundation pit depth. The inclinometer data to be processed may cover multiple aspects, such as inclinometer angle, deformation, inclinometer curve, etc. When the inclinometer collects the inclinometer data to be processed, it transmits the inclinometer data to be processed to the foundation pit horizontal displacement prediction device, which is responsible for processing the inclinometer data to be processed to predict the horizontal displacement of the foundation pit at a future moment. First, the device performs anomaly inspection and smoothing optimization on the inclinometer data to be processed. Anomaly inspection can identify outliers in the inclinometer data to ensure data quality. Smoothing optimization can use filters or other techniques to eliminate noise in the inclinometer data and reduce the volatility of the data. The processed data is called target inclinometer data, which represents the actual deformation of the foundation pit. Next, the foundation pit horizontal displacement prediction device determines the time-course data based on the target inclinometer data. The time-course data is determined based on the target inclinometer data, and is used to represent the possible influencing factors of the foundation pit horizontal displacement. Specifically, the target inclinometer data is analyzed to determine the change law of the target inclinometer data, and the influencing factors that affect the change of the foundation pit horizontal displacement are identified according to the change law of the target inclinometer data. The inclinometer data corresponding to multiple influencing factors are processed in a time-course manner to obtain the time-course data of the target inclinometer data. Secondly, the target inclinometer data at the current moment is input into a multimodal neural network model based on the attention mechanism together with the time-course data. The neural network model combines information of different types of data and uses the attention mechanism to determine the importance of various information. Finally, based on the output of the neural network model, the prediction device generates the inclinometer data prediction value of the target inclinometer data corresponding to the next moment.

[0061] Optionally, the time course data includes at least: the depth of the target inclinometer data, the monitoring interval, the current depth inclinometer data and the excavation depth.

[0062] In an embodiment of the present application, a multimodal neural network algorithm based on an attention mechanism is used to consider multiple displacement influencing factors (such as the depth of the target inclinometer data, the monitoring interval, the current depth inclinometer data, and the excavation depth, etc.) to predict the deep horizontal displacement of the deep foundation pit, prevent sudden accidents that may occur during the construction process, and ensure the production safety of deep foundation pit excavation operations, thereby effectively preventing production safety accidents.

[0063] Figure 2 A schematic flow chart of a method for predicting horizontal displacement of a foundation pit provided in an embodiment of the present application is shown.

[0064] Step 201, obtaining inclinometric data to be processed that varies with the depth of the foundation pit.

[0065] The inclinometric data to be processed that changes with the depth of the foundation pit may refer to the inclinometric data collected as the depth of the foundation pit construction changes, which is usually used to monitor the deformation of the foundation pit project. The inclinometric data to be processed may include information such as inclinometric angle and horizontal displacement, and its specific form depends on the type of inclinometric instrument and sensor selected. For example, the inclinometric data to be processed may refer to the collected inclinometric curve.

[0066] In the embodiments of the present application, data collection is usually completed by professional technicians or automated monitoring equipment. For example, in foundation pit projects, inclinometers are usually installed on inclinometer tubes or other fixed measuring points to ensure accurate data collection. As the construction depth of the foundation pit changes, the inclinometer will regularly record relevant inclinometer data, which will accumulate over time, and the data collection process should be strictly monitored and recorded to ensure the accuracy and completeness of the data.

[0067] It should be understood that the collected inclination data usually needs to be processed to ensure the availability and quality of the data. The processing process of the inclination data to be processed is described in detail below.

[0068] Step 202: performing anomaly check and smoothing optimization on the inclinometer data to be processed to obtain target inclinometer data.

[0069] In the embodiment of the present application, abnormal data that may be caused by equipment failure or external interference can be identified and marked through abnormal inspection, the abnormal data can be eliminated, and the inclinometric data to be processed can be smoothly optimized to ensure the availability and quality of the target inclinometric data obtained. The embodiment of the present application does not limit the specific method of performing abnormal inspection and smooth optimization.

[0070] In a possible implementation, the inclinometric data to be processed includes an inclinometric curve of the foundation pit retaining structure changing with depth; the above step 202 may specifically include:

[0071] Fitting the inclinometer curve based on the least square method to determine the fitting value;

[0072] According to the fitting value and the true value of the inclinometric data to be processed, a mean absolute error and a root mean square error between the fitting value and the true value are determined;

[0073] According to the mean absolute error and root mean square error, the smoothness of the optimized inclinometric data and the abnormal data are determined;

[0074] According to the smoothness of the optimized inclinometer data and the abnormal data, the target inclinometer data is determined.

[0075] In the embodiment of the present application, the inclinometer curve measured by the inclinometer has systematic errors, random errors, and even gross errors, and cannot be used to produce a prediction network model data set. Directly inputting it into the prediction network model will cause the output prediction results to have large deviations. Therefore, it is necessary to perform abnormality inspection and smoothing optimization on the inclinometer data.

[0076] For example, during the excavation of the foundation pit, the retaining structure such as the underground continuous wall is only subject to its own gravity in the vertical direction, so it can be simplified as a pure bending member without axial force. According to the basic properties of bending members under the assumption of small deformation and plane section in material mechanics, the slope of any point on the plane is It can be obtained by differentiating the inclinometer curve y(x) that changes with depth:

[0077]

[0078] Where x is the depth; y(x) is the equation of the incline curve of the enclosure structure that changes with depth; The curvature of the inclinometric curve of the retaining structure varies with depth.

[0079] Since the inclinometer curve is relatively flat, it can be simplified as shown below.

[0080]

[0081] Based on the elastic foundation beam theory assumption, according to the plane strain problem, the enclosure structure is simplified to a pure curved beam on the elastic foundation, and the following equation is obtained:

[0082]

[0083] According to Rankine's earth pressure theory, p(x) and q(x) are first-order slopes that vary with depth x. Therefore, the equation of the slope curve y(x) of the retaining structure varying with depth must be at least fifth-order differentiable to satisfy the above equation. In addition, considering that there are many factors affecting earth pressure, the measured earth pressure data on site is not first-order linear, but parabolic. Therefore, the sixth-order polynomial based on the least squares method can be used to fit the slope curve equation y(x) of the retaining structure varying with depth, so as to check and smooth the slope data. The inspection and optimization effects can be seen in Figure 3 and Figure 4 The abnormal data diagram and normal data diagram are shown respectively. Figure 3 and Figure 4In the figure, the horizontal axis Depth is the depth, the vertical axis Derivation represents the derivative value, and the vertical axis Horizontal displacement represents the horizontal displacement. It can be concluded that the curve in the normal data is relatively smooth and the abnormal data that are too large or too small in the original curve are deleted; by calculating the mean absolute error and the root mean square error between the fitting value and the actual value, the degree of optimization smoothness of the inclinometer data can be obtained and abnormal inclinometer data can be detected.

[0084] In a possible implementation, the smoothing degree and abnormal data of the optimized inclinometric data are determined according to the mean absolute error and the root mean square error, including:

[0085] If the mean absolute error and the root mean square error are less than the first error threshold, it is determined that the smoothness of the optimized oblique measurement data meets the optimization smoothing requirement;

[0086] If the mean absolute error and the root mean square error are greater than the second error threshold, the corresponding inclinometer data is determined to be abnormal data.

[0087] In the embodiment of the present application, the mean absolute error refers to the average of the absolute values ​​of the differences between the actual values ​​and the fitted values, and the root mean square error refers to the square root of the average of the squares of the differences between the actual values ​​and the fitted values.

[0088] Exemplarily, if the mean absolute error and the root mean square error are greater than the set threshold 3, the corresponding inclinometric data is determined to be abnormal data.

[0089] Exemplarily, the calculation formula of mean absolute error is:

[0090]

[0091] Exemplarily, the calculation formula of the root mean square error is:

[0092]

[0093] Where n is the total number of inclination data at a certain moment; y i is the measured value of the inclinometer data; is the fitted value of the inclinometer data.

[0094] Step 203, performing data analysis on the target inclinometer data, determining the changing pattern of the target inclinometer data, and identifying the influencing factors that affect the horizontal displacement change of the foundation pit based on the changing pattern of the target inclinometer data, performing time-course processing on the inclinometer data corresponding to multiple influencing factors, and obtaining the time-course data of the target inclinometer data.

[0095] Among them, the time history data is used to represent the displacement influencing factors of the horizontal displacement of the foundation pit.

[0096] In the embodiment of the present application, the target inclinometer data is processed and optimized inclinometer data, which represents the actual horizontal displacement of the foundation pit. These data may include information such as the horizontal displacement and inclinometer angle at each depth point of the foundation pit. The accuracy of the target inclinometer data is crucial for subsequent prediction and analysis because they are based on actual observation data.

[0097] In the embodiment of the present application, the time history data is determined based on the target inclinometric data and is used to represent the displacement influencing factors of the horizontal displacement of the foundation pit. The time history data reflects the changes over time of various factors affecting the horizontal displacement of the foundation pit, such as the changing monitoring depth, monitoring interval, and excavation depth. The determination of the time history data is to establish a complete prediction model to more accurately predict the future horizontal displacement of the foundation pit.

[0098] In an embodiment of the present application, the specific process of determining the time-course data may be: analyzing the target inclinometer data to understand the actual changes in the horizontal displacement of the foundation pit. The analysis may involve statistical methods, time series analysis and other techniques to determine the trend and change pattern of the target inclinometer data. Then, identify the various factors that affect the horizontal displacement of the foundation pit, such as the changing monitoring depth, monitoring interval, and excavation depth. These factors may change over time, so they need to be time-coursed and the identified factors converted into time-course data to correspond to the target inclinometer data.

[0099] It should be understood that the target inclinometric data and time history data are integrated to form a complete data set. This data set can be used to train a multimodal neural network model based on an attention mechanism to perform predictive analysis of the horizontal displacement of the foundation pit. It can also be used to input into the trained multimodal neural network model based on an attention mechanism to predict the horizontal displacement of the foundation pit at the next moment.

[0100] Step 204, input the target inclinometer data and time course data at the current moment into the multimodal neural network model based on the attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment to predict the horizontal displacement of the foundation pit at the next moment.

[0101] In the embodiment of the present application, in order to improve the accuracy and reliability of the prediction of the inclinometric data, it is urgently necessary to adopt an advanced algorithm that comprehensively considers multiple key factors to achieve a more comprehensive and accurate prediction of the deep horizontal displacement of the foundation pit engineering. Therefore, a multimodal neural network model based on the attention mechanism is designed to consider the displacement influencing factors such as the overall foundation pit inclinometric deformation, monitoring interval time, excavation depth, etc., so as to improve the prediction accuracy and reliability and effectively respond to the challenges of foundation pit engineering safety.

[0102] The input of the multimodal neural network model based on the attention mechanism is the image of the inclinometer curve and time-course data that changes with depth at a certain moment. The time-course data is a one-dimensional vector, which contains four components: the depth of the inclinometer data, the monitoring interval, the inclinometer data at a certain depth, and the excavation depth. The final output of the multimodal neural network model based on the attention mechanism is the predicted value of the inclinometer data corresponding to the input depth at the next moment.

[0103] It should be understood that the network model data set is produced using measured data, and the inclinometer curve image and time-course data that change with depth at a certain moment are produced from the inclinometer data that have been checked and smoothed and optimized.

[0104] In one possible implementation, the training process of the multimodal neural network model based on the attention mechanism includes:

[0105] Obtain the inclinometric curve image and time history data of the foundation pit retaining structure changing with depth at any time, where the time history data is obtained by optimizing and smoothing the inclinometric curve;

[0106] The inclinometer curve images and time course data are input into the multimodal neural network model based on the attention mechanism, and the network parameters of the multimodal neural network model based on the attention mechanism are optimized through back propagation and model evaluation indicators;

[0107] When the model evaluation index reaches the preset threshold, the training of the multimodal neural network model based on the attention mechanism is completed.

[0108] In an embodiment of the present application, first, during the training process, it is necessary to first obtain the inclinometer curve image and corresponding time-history data of the foundation pit retaining structure changing with depth at any moment. The time-history data is obtained by optimizing and smoothing the inclinometer curve, and is used to represent the influencing factors of the horizontal displacement of the foundation pit. Secondly, the obtained inclinometer curve image and time-history data are used as training data, the inclinometer curve image is used as the input of the visual modality, and the time-history data is used as the input of other modalities. The inclinometer curve image and time-history data are passed as input to a multimodal neural network model based on the attention mechanism. In the model, the attention mechanism can help the network effectively integrate information from different modalities and determine the importance of each modality to the prediction results. Then, the network parameters of the model are optimized through the back-propagation algorithm.

[0109] During the optimization process, the model will adjust the network parameters according to the feedback of the loss function, so that the model can more accurately predict the horizontal displacement of the foundation pit. During the training process, the performance of the model needs to be evaluated regularly to ensure the effectiveness and generalization ability of the model. Some evaluation indicators (such as mean square error, mean absolute error, etc.) can be used to monitor the training progress of the model and determine whether the preset threshold is reached.

[0110] When the model evaluation index reaches the preset threshold, the training process ends, indicating that the model has converged and achieved satisfactory performance. The training stop condition can be that the index reaches the set threshold, or there is no further improvement after a certain number of training rounds.

[0111] Through the above steps, the multimodal neural network model based on the attention mechanism can be effectively trained and used to predict the horizontal displacement of the foundation pit after training.

[0112] In an embodiment of the present application, target inclinometer data are obtained by performing abnormality check and smoothing optimization on the acquired inclinometer data to be processed that changes with the depth of the foundation pit. The prediction based on the target inclinometer data improves the prediction progress compared with the prediction based on the inclinometer data to be processed. Secondly, according to the target inclinometer data, the displacement influencing factors that affect the horizontal displacement of the foundation pit, that is, the time-course data, are determined. Finally, the target inclinometer data and the time-course data are input into a multimodal neural network model with an attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the inclinometer data prediction value corresponding to the target inclinometer data at the next moment. The multimodal neural network model based on the attention mechanism takes into account the displacement influencing factors that affect the horizontal displacement of the foundation pit when making predictions, so that the method has strong interpretability and high accuracy.

[0113] Figure 5 A schematic flow chart of a method for predicting horizontal displacement of a foundation pit provided in another embodiment of the present application is shown.

[0114] Step 501, obtaining inclinometric data to be processed that varies with the depth of the foundation pit.

[0115] Step 502: Perform an abnormality check and smoothing optimization on the inclinometer data to be processed to obtain target inclinometer data.

[0116] Step 503, perform data analysis on the target inclinometer data, determine the change pattern of the target inclinometer data, and identify the influencing factors that affect the change of the horizontal displacement of the foundation pit based on the change pattern of the target inclinometer data, perform time-course processing on the inclinometer data corresponding to multiple influencing factors, and obtain the time-course data of the target inclinometer data.

[0117] Step 504, input the target inclinometer data and time course data at the current moment into the multimodal neural network model based on the attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment to predict the horizontal displacement of the foundation pit at the next moment.

[0118] The specific implementation process and principle of the above steps 501-504 can refer to the detailed description of steps 201-204 of the above embodiment, which will not be repeated here.

[0119] Step 505, obtaining the bending moment function of the enclosure structure that varies with depth and the internal force function of the enclosure structure that varies with depth.

[0120] Step 505: Determine the bending moment value corresponding to the excavation depth according to the excavation depth of the foundation pit and the bending moment function.

[0121] Step 507, determining the internal force value corresponding to the excavation depth according to the excavation depth of the foundation pit and the internal force function.

[0122] Step 508, determining the predicted safety state of the enclosure structure according to the bending moment value corresponding to the excavation depth and the internal force value corresponding to the excavation depth.

[0123] In the embodiments of the present application, according to the foundation pit damage mode, in addition to the longitudinal landslide, the deformation shape of the retaining wall before the foundation pit damage can be divided into wall overturning, wall kicking, and wall breaking. The fracture of the retaining wall may directly lead to local damage or even overall damage and collapse of the foundation pit, and the process of damage to the retaining structure to collapse is often instantaneous. In order to control the safety of the foundation pit and avoid the possible fracture of the retaining wall during the excavation process, the key is to control the internal force during the excavation process not to exceed the bending moment limit value, and continuously monitor the internal force of the retaining structure during the excavation process.

[0124] In the embodiment of the present application, it is necessary to monitor in real time that the internal force of the retaining structure does not exceed the bending moment limit value, so as to avoid accidents such as fracture of the retaining structure during the excavation process.

[0125] For example, according to the theory of structural mechanics, the calculation formula for the bending moment of an elastic beam in pure bending (i.e., the bending moment function) is:

[0126]

[0127] Where φ is the curvature of the cross-section deformation, R is the radius of curvature; M(x) is the bending moment of the beam that changes with depth; EI is the cross-sectional bending stiffness of the elastic beam (its calculation can refer to the "Concrete Structure Design Code" and "Technical Code for Building Foundation Pit Engineering"); x is the depth of the retaining structure.

[0128] Combining the above bending moment function and the simplified formula of the inclinometer curve in the above embodiment, it can be obtained that the internal force value of the enclosure structure can be represented by the inclinometer curve y(x) function that changes with depth:

[0129]

[0130] The inclinometric data obtained through prediction is brought into the inclinometric curve, and the internal force of the enclosure structure is back-calculated and compared with the bending moment limit value to obtain the predicted safety state of the enclosure structure.

[0131] Based on the previous embodiment, the embodiment of the present application continuously monitors the predicted safety status of the enclosure structure by calculating and comparing the internal force value and the bending moment limit value of the enclosure structure, so as to facilitate timely engineering intervention when the predicted safety status is found to be abnormal.

[0132] Corresponding to the foundation pit horizontal displacement prediction method in the above embodiment, Figure 6 The schematic diagram of the structure of a device for predicting horizontal displacement of a foundation pit provided by an embodiment of the present application is shown, and the device for predicting horizontal displacement of a foundation pit is configured in an electronic device. For the convenience of description, only the part related to the embodiment of the present application is shown.

[0133] See also Figure 6 , the foundation pit horizontal displacement prediction device 600 comprises:

[0134] The data acquisition module 601 is used to acquire the inclinometric data to be processed that varies with the depth of the foundation pit;

[0135] A smoothing optimization module 602 is used to perform abnormality inspection and smoothing optimization on the inclination measurement data to be processed to obtain target inclination measurement data;

[0136] The time history data determination module 603 is used to perform data analysis on the target inclination measurement data, determine the change law of the target inclination measurement data, and identify the influencing factors that affect the change of the horizontal displacement of the foundation pit according to the change law of the target inclination measurement data, and perform time history processing on the inclination measurement data corresponding to multiple influencing factors to obtain the time history data of the target inclination measurement data;

[0137] The prediction module 604 is used to input the target inclinometer data and time-course data at the current moment into the multimodal neural network model based on the attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment to predict the horizontal displacement of the foundation pit at the next moment.

[0138] In the embodiment of the present application, the inclinometric data to be processed includes an inclinometric curve of the foundation pit retaining structure changing with depth, and the smoothing optimization module 602 may specifically include the following units:

[0139] A fitting unit, used for fitting the inclinometric curve based on the least square method to determine a fitting value;

[0140] An error determination unit, used for determining a mean absolute error and a root mean square error between the fitting value and the true value according to the fitting value and the true value of the inclination measurement data to be processed;

[0141] An abnormality determination unit, used to determine the smoothness of the optimized inclinometric data and abnormal data according to the mean absolute error and the root mean square error;

[0142] The target determination unit is used to determine the target inclination measurement data according to the smoothness of the optimized inclination measurement data and the abnormal data.

[0143] In the embodiment of the present application, the abnormality determination unit may also be used to:

[0144] If the mean absolute error and the root mean square error are less than the first error threshold, it is determined that the smoothing degree of the optimized inclinometric data meets the optimization smoothing requirement;

[0145] If the mean absolute error and the root mean square error are greater than the second error threshold, the corresponding inclinometer data is determined to be abnormal data.

[0146] In an embodiment of the present application, the time history data is a one-dimensional vector and includes at least: the depth of the target inclinometer data, the monitoring interval, the current depth inclinometer data, and the excavation depth.

[0147] In the embodiment of the present application, the prediction module 604 may also be used for:

[0148] Obtain the inclinometric curve image and time history data of the foundation pit retaining structure changing with depth at any time, where the time history data is obtained by optimizing and smoothing the inclinometric curve;

[0149] The inclinometer curve images and time course data are input into the multimodal neural network model based on the attention mechanism, and the network parameters of the multimodal neural network model based on the attention mechanism are optimized through back propagation and model evaluation indicators;

[0150] When the model evaluation index reaches the preset threshold, the training of the multimodal neural network model based on the attention mechanism is completed.

[0151] In the embodiment of the present application, the foundation pit horizontal displacement prediction device 600 may further include the following modules:

[0152] Function acquisition module, used to obtain the bending moment function of the enclosure structure varying with depth and the internal force function of the enclosure structure varying with depth;

[0153] A bending moment value determination module is used to determine the bending moment value corresponding to the excavation depth according to the excavation depth of the foundation pit and the bending moment function;

[0154] An internal force value determination module is used to determine the internal force value corresponding to the excavation depth according to the excavation depth of the foundation pit and the internal force function;

[0155] The safety state prediction module is used to determine the predicted safety state of the enclosure structure according to the bending moment value corresponding to the excavation depth and the internal force value corresponding to the excavation depth.

[0156] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0157] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0158] See also Figure 7 , shows a schematic diagram of the structure of an electronic device provided by an embodiment of the present application, such as Figure 7 As shown, the electronic device 700 of this embodiment includes: at least one processor 710 ( Figure 7 Only one is shown), a memory 720, and a computer program 721 stored in the memory 720 and executable on the at least one processor 710, wherein the processor 710 implements the steps in the above-mentioned foundation pit horizontal displacement prediction method embodiment when executing the computer program 721.

[0159] It should be noted that the electronic device 700 may refer to a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The electronic device may include, but is not limited to, a processor 710 and a memory 720. Those skilled in the art will understand that Figure 7 This is merely an example of the electronic device 700 and does not constitute a limitation on the electronic device 700 . The electronic device 700 may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.

[0160] The processor 710 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0161] In some embodiments, the memory 720 may be an internal storage unit of the electronic device 700, such as a hard disk or memory of the electronic device 700. In other embodiments, the memory 720 may also be an external storage device of the electronic device 700, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 700. Further, the memory 720 may also include both an internal storage unit of the electronic device 700 and an external storage device. The memory 720 is used to store an operating system, an application program, a boot loader (Boot Loader), data and other programs, such as the program code of the computer program, etc. The memory 720 may also be used to temporarily store data that has been output or is to be output.

[0162] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0163] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0164] In the embodiments provided in the present application, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are merely schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0165] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0166] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0167] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0168] The present application implements all or part of the processes in the above-mentioned embodiment method, and may also be completed through a computer program product. When the computer program product runs on an electronic device, the electronic device can implement the steps in the above-mentioned method embodiments when executing.

[0169] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application is described in detail with reference to the above-mentioned embodiments, a person skilled in the art should understand that the technical solutions described in the above-mentioned embodiments can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for predicting horizontal displacement of a foundation pit, characterized in that: The method for predicting horizontal displacement of foundation pit comprises: Obtain the inclinometric data to be processed that varies with the depth of the foundation pit; Performing abnormality inspection and smoothing optimization on the inclination measurement data to be processed to obtain target inclination measurement data; Performing data analysis on the target inclination measurement data to determine the change law of the target inclination measurement data, and identifying the influencing factors that affect the change of the horizontal displacement of the foundation pit according to the change law of the target inclination measurement data, and performing time history processing on the inclination measurement data corresponding to the multiple influencing factors to obtain the time history data of the target inclination measurement data; The target inclinometer data and the time course data at the current moment are input into a multimodal neural network model based on an attention mechanism, and the multimodal neural network model based on the attention mechanism outputs a predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment, so as to predict the horizontal displacement of the foundation pit at the next moment.

2. The method for predicting horizontal displacement of a foundation pit according to claim 1, characterized in that: The inclinometric data to be processed includes an inclinometric curve of the foundation pit retaining structure changing with depth; the abnormality check and smooth optimization of the inclinometric data to be processed to obtain target inclinometric data includes: Fitting the inclinometer curve based on the least square method to determine a fitting value; Determine a mean absolute error and a root mean square error between the fitting value and the true value according to the fitting value and the true value of the inclination measurement data to be processed; Determine the smoothness of the optimized inclinometric data and abnormal data according to the mean absolute error and the root mean square error; The target inclination measurement data is determined according to the smoothness of the optimized inclination measurement data and the abnormal data.

3. The method for predicting horizontal displacement of a foundation pit according to claim 2, characterized in that: Determining the smoothness of the optimized inclinometric data and abnormal data according to the mean absolute error and the root mean square error includes: If the mean absolute error and the root mean square error are less than a first error threshold, it is determined that the smoothness of the optimized inclinometric data meets the optimization smoothness requirement; If the mean absolute error and the root mean square error are greater than a second error threshold, the corresponding inclinometric data is determined to be abnormal data.

4. The method for predicting horizontal displacement of a foundation pit according to claim 1, characterized in that: The time history data is a one-dimensional vector and includes at least: the depth of the target inclinometer data, the monitoring interval, the current depth inclinometer data and the excavation depth.

5. The method for predicting horizontal displacement of a foundation pit according to claim 1, characterized in that: The training process of the multimodal neural network model based on the attention mechanism includes: Obtaining an inclinometric curve image of the foundation pit retaining structure changing with depth at any time and the time history data, wherein the time history data is obtained by optimizing and smoothing the inclinometric curve; Inputting the inclinometer curve image and the time course data into the multimodal neural network model based on the attention mechanism, and optimizing the network parameters of the multimodal neural network model based on the attention mechanism through back propagation and model evaluation indicators; When the model evaluation index reaches a preset threshold, the training of the multimodal neural network model based on the attention mechanism is completed.

6. The method for predicting horizontal displacement of a foundation pit according to claim 1, characterized in that: The method for predicting horizontal displacement of foundation pit further comprises: Obtaining a bending moment function of a retaining structure that varies with depth and an internal force function of the retaining structure that varies with depth; Determining the bending moment value corresponding to the excavation depth according to the excavation depth of the foundation pit and the bending moment function; Determining an internal force value corresponding to the excavation depth according to the excavation depth of the foundation pit and the internal force function; The predicted safety state of the retaining structure is determined according to the bending moment value corresponding to the excavation depth and the internal force value corresponding to the excavation depth.

7. A device for predicting horizontal displacement of a foundation pit, characterized in that: The foundation pit horizontal displacement prediction device comprises: A data acquisition module is used to acquire inclination data to be processed that varies with the depth of the foundation pit; A smoothing optimization module, used for performing abnormality inspection and smoothing optimization on the inclination measurement data to be processed to obtain target inclination measurement data; A time history data determination module is used to perform data analysis on the target inclination measurement data, determine the change law of the target inclination measurement data, and identify the influencing factors that affect the change of the horizontal displacement of the foundation pit according to the change law of the target inclination measurement data, and perform time history processing on the inclination measurement data corresponding to the multiple influencing factors to obtain the time history data of the target inclination measurement data; A prediction module is used to input the target inclinometer data and the time course data at the current moment into a multimodal neural network model based on an attention mechanism, and the multimodal neural network model based on the attention mechanism outputs the predicted value of the inclinometer data corresponding to the target inclinometer data at the next moment, so as to predict the horizontal displacement of the foundation pit at the next moment.

8. A foundation pit horizontal displacement prediction system, characterized in that: The foundation pit horizontal displacement prediction system comprises an inclinometer and a foundation pit horizontal displacement prediction device, wherein the inclinometer comprises a plurality of fixed inclinometer units; The inclinometer unit is installed and buried in the inclinometer tube, and is used to collect the inclinometer data to be processed that changes with the depth of the deep pit using the inclinometer sensor, and transmit the inclinometer data to be processed to the foundation pit horizontal displacement prediction device; The foundation pit horizontal displacement prediction device is used to execute the foundation pit horizontal displacement prediction method as described in any one of claims 1 to 6.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.