Safety status assessment method and system of retaining structure based on deep foundation pit inclinometric data

Abstract

The present application relates to a method and system for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data, including obtaining historical deep foundation pit inclinometer data of a construction site, constructing a semi-supervised neural network model, training the semi-supervised neural network model based on the historical deep foundation pit inclinometer data to obtain a trained smooth optimization model; based on the trained smooth optimization model, smooth optimization is performed on the deep foundation pit inclinometer data obtained from the construction site to obtain a curve of the internal force of the retaining structure to be evaluated varying with depth; based on the curve of the internal force of the retaining structure to be evaluated varying with depth, the actual bending moment of the retaining structure to be evaluated is calculated, and the bending moment limit value of the retaining structure to be evaluated is obtained, and based on the bending moment limit value and the actual bending moment, the safety status evaluation result of the retaining structure to be evaluated is obtained. The present application has the effect of improving the reliability of inclinometer data and the safety of retaining structures.

Description

Technical Field

[0001] The present invention relates to the technical field of safety assessment of deep foundation pit retaining structures, and in particular to a method and system for assessing the safety status of retaining structures based on deep foundation pit inclinometer data. Background Art

[0002] At present, with the continuous development of urban infrastructure construction and the country's emphasis on safety production, the requirements for deep foundation pit monitoring are gradually increasing. In actual foundation pit excavation, the inclinometer is used to detect the deep horizontal displacement of the side retaining structure and its rate of change to prevent excessive deformation of the soil and cause damage to the support structure.

[0003] At the same time, with the development of automated monitoring technology, more and more foundation pit projects use automated monitoring equipment. The increase in monitoring frequency has led to a sharp increase in the amount of monitoring data, and the reliability of automated monitoring data has also been challenged. At present, the optimization method for the inclinometer data automatically collected by the automatic inclinometer is achieved by using supervised learning algorithms such as support vector machines and recurrent neural networks. It can easily adapt to different types of foundation pit structures and complex soil layer changes, but it requires huge labeled data sets to support it. The labeling process is time-consuming and inefficient. There is an urgent need for a method that can detect the reliability of the currently acquired inclinometer data in real time and analyze the potential risks of the retaining structure during deep foundation pit excavation, so as to ensure the safety of foundation pit engineering production. Summary of the invention

[0004] In order to improve the reliability of inclinometric data and the safety of retaining structures, the present application provides a method and system for assessing the safety status of retaining structures based on inclinometric data of deep foundation pits.

[0005] The above-mentioned invention objective of the present application is achieved through the following technical solutions:

[0006] A method, the method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data comprises:

[0007] Acquire historical deep foundation pit inclinometer data of the construction site, and generate inspection and smoothing optimization information according to the historical deep foundation pit inclinometer data;

[0008] According to the inspection and smooth optimization information, a semi-supervised neural network model is constructed, and the semi-supervised neural network model is trained based on the historical deep foundation pit inclination measurement data to obtain a trained smooth optimization model;

[0009] Based on the trained smooth optimization model, the deep foundation pit inclinometric data obtained from the construction site are smoothly optimized to obtain a curve of the internal force of the retaining structure to be evaluated varying with depth;

[0010] According to the curve of the internal force variation with depth of the enclosure structure to be evaluated, the actual bending moment of the enclosure structure to be evaluated is calculated, and the bending moment limit value of the enclosure structure to be evaluated is obtained; based on the bending moment limit value and the actual bending moment, the safety status assessment result of the enclosure structure to be evaluated is obtained.

[0011] By adopting the above-mentioned technical scheme, since the current optimization method for the inclinometer data automatically collected by the automatic inclinometer is realized by using supervised learning algorithms such as support vector machines and recurrent neural networks, it can easily adapt to different foundation pit structure types and complex soil layer changes, but it requires a huge labeled data set to support it, and the labeling process is time-consuming and inefficient. Therefore, this application obtains the historical deep foundation pit inclinometer data at the construction site, generates inspection and smoothing optimization information based on the historical deep foundation pit inclinometer data, and then constructs a semi-supervised neural network model based on the inspection and smoothing optimization information. The semi-supervised neural network model is trained based on the historical deep foundation pit inclinometer data to obtain a trained smoothing optimization model, which helps to improve the adaptability of the semi-supervised neural network model, so that the model can perform data inspection and smoothing optimization on the deep foundation pit inclinometer data, and then through Based on the trained smooth optimization model, the deep foundation pit inclinometer data obtained from the construction site are smoothly optimized to obtain the model-optimized curve of the internal force of the retaining structure to be evaluated varying with depth, which helps to eliminate the error caused by directly using the deep foundation pit inclinometer data and improve the accuracy of the curve of the internal force of the retaining structure to be evaluated varying with depth. The actual bending moment of the retaining structure to be evaluated is calculated according to the curve of the internal force of the retaining structure to be evaluated varying with depth, and the bending moment limit value of the retaining structure to be evaluated is obtained. Based on the bending moment limit value and the actual bending moment, the safety status assessment result of the retaining structure to be evaluated is obtained, which helps to intuitively understand the safety status of the retaining structure to be evaluated, and according to the safety assessment result of the retaining structure to be evaluated, the reliability of the retaining structure is improved, so as to achieve the purpose of improving the accuracy and credibility of the results of monitoring the safety status of the retaining structure, thereby ensuring the safety of the foundation pit.

[0012] In a preferred example, the present application can be further configured as follows: constructing a semi-supervised neural network model based on the inspection and smoothing optimization information, training the semi-supervised neural network model based on the historical deep foundation pit inclinometer data, and obtaining a trained smoothing optimization model, specifically including:

[0013] Based on the fitting algorithm, part of the inclination measurement data in the historical deep foundation pit inclination measurement data is labeled to obtain labeled data; and an inclination measurement data set is constructed according to the labeled data and the unlabeled data in the historical deep foundation pit inclination measurement data;

[0014] The semi-supervised neural network model is trained according to the inclinometer data set to obtain the trained smoothing optimization model.

[0015] By adopting the above technical scheme, part of the inclinometer data in the historical deep foundation pit inclinometer data is labeled based on the fitting algorithm to obtain labeled data, and an inclinometer data set is constructed based on the labeled data and the unlabeled data in the historical deep foundation pit inclinometer data. This helps to provide necessary samples and labels for model training, so that the semi-supervised neural network model is trained according to the inclinometer data set to obtain a trained smooth optimization model, so that the model can achieve better results in the current application scenario, reduce the probability of overfitting, and easily adapt to different foundation pit structure types and complex soil layer changes.

[0016] In a preferred example, the present application can be further configured as follows: labeling part of the inclination measurement data in the historical deep foundation pit inclination measurement data based on the fitting algorithm to obtain labeled data specifically includes:

[0017] Selecting a portion of the inclination measurement data from the historical deep foundation pit inclination measurement data and marking the portion of the inclination measurement data as the portion of the inclination measurement data;

[0018] The polynomial function fitting algorithm of the linear least squares method, the arbitrary function fitting algorithm of the nonlinear least squares method, the quintic spline function fitting algorithm, the quintic smoothing spline fitting function and the polynomial fitting algorithm and the weighted least squares fitting algorithm are respectively used to perform curve fitting on the part of the inclinometric data to obtain corresponding fitting curves;

[0019] Based on the weighted formula set by the system, the fitting curves are weighted, and the weight coefficients that meet the preset conditions are set to obtain the fitting function of the inclinometer curve changing with depth. The weighted formula is f(x)=a1f1(x)+a2f2(x)+…+a n f n (x), where n is the number of fitting algorithms used, n is 5; a n is the fitting algorithm weight of the nth fitting algorithm; f n (x) is the fitting function corresponding to the nth fitting algorithm; f(x) is the fitting function of the inclination curve changing with depth;

[0020] The labeled data is obtained according to a fitting function of the inclinometer curve changing with depth.

[0021] By adopting the above technical scheme, by selecting a part of the inclinometer data from the historical deep foundation pit inclinometer data and marking it as partial inclinometer data, it is helpful to reduce the cost of making labeled data sets. By adopting five fitting algorithms to perform curve fitting on the partial inclinometer data, the corresponding fitting curves are obtained, which is helpful to understand the distribution and rules of the data. Based on the weighted formula set by the system, each fitting curve is weighted, and the weight coefficient that meets the preset conditions is set to obtain the fitting function of the inclinometer curve changing with depth, which is helpful to eliminate the error caused by partial inclinometer data. By fitting the function according to the change of inclinometer curve with depth, labeled data is obtained, which helps to reduce the time spent on manual annotation, reduce the links of manual intervention, and improve the credibility of the data.

[0022] In a preferred example, the present application may be further configured as follows: the semi-supervised neural network model is trained according to the inclination measurement data set to obtain the trained smooth optimization model, specifically including:

[0023] Obtain a fully connected neural network, input the labeled data into the fully connected neural network, and perform calculations based on the first prediction result and the manually annotated labels to obtain a supervised loss function value;

[0024] Performing enhancement processing on the unlabeled data to obtain enhanced data, inputting the enhanced data into the fully connected neural network, and calculating the obtained second prediction result to obtain an unsupervised loss function value;

[0025] Iteratively training the semi-supervised neural network model according to the supervised loss function value and the unsupervised loss function value, and determining the number of iterations;

[0026] When the number of iterations meets a preset iteration number threshold, the trained semi-supervised neural network model is output and used as the smoothing optimization model.

[0027] By adopting the above technical solution, a semi-supervised neural network is trained according to the inclinometer data set, and a data enhancement method suitable for unlabeled data is used to train a smoothing optimization model, so that the model can perform data inspection and smooth optimization on only a small number of labeled inclinometer curve data, thereby reducing the cost of manual labeling and being able to cope with the problem of scarce labels in real scenarios. Therefore, the model trained based on historical deep foundation pit inclinometer data can achieve better results in the current application scenario, and can easily adapt to different foundation pit structure types and complex soil layer changes, so as to obtain a smooth curve of the internal force of the retaining structure to be evaluated changing with depth.

[0028] In a preferred example, the present application can be further configured as follows: according to the curve of the internal force of the enclosure structure to be evaluated changing with depth, the actual bending moment of the enclosure structure to be evaluated is calculated, and the bending moment limit value of the enclosure structure to be evaluated is obtained; based on the bending moment limit value and the actual bending moment, the safety status assessment result of the enclosure structure to be evaluated is obtained, which specifically includes:

[0029] Based on the curvature solution formula, the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth is obtained according to the curve of the internal force of the enclosure structure to be evaluated changing with depth. The curvature solution formula is Φ(x)=d 2 y / dx 2 , where dx 2 is the second-order differential of the depth x of the enclosure to be assessed; d 2 y is the second-order differential of the curve y(x) of the internal force of the enclosure structure to be evaluated changing with depth; Φ(x) is the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth;

[0030] Based on the bending moment calculation formula, the actual bending moment of the enclosure structure to be evaluated is obtained according to the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth. The bending moment calculation formula is M(x)=EIΦ(x)=EI·d 2 y / dx 2 , where M(x) is the actual bending moment of the enclosure structure to be evaluated; EI is the bending stiffness of the enclosure structure to be evaluated;

[0031] Based on the bending moment limit calculation formula, the bending moment limit value of the enclosure structure to be evaluated is calculated. The bending moment limit calculation formula is M = (0.85 × f ck ×b×h) / γ m , where M is the bending moment limit value of the enclosure structure to be evaluated, f ck is the design compressive strength of the enclosure structure to be evaluated, b is the width of the enclosure structure to be evaluated, h is the effective height of the enclosure structure to be evaluated, γ m is the safety factor;

[0032] Based on the bending moment limit value and the actual bending moment, a bending coefficient is calculated, and according to the bending coefficient, a safety status assessment result of the enclosure structure to be assessed is obtained.

[0033] By adopting the above technical scheme, based on the curvature solving formula, according to the curve of the internal force of the enclosure structure to be evaluated varying with depth, the curvature of the curve of the internal force of the enclosure structure to be evaluated varying with depth is obtained, which is helpful to analyze the smoothness of the curve and the trend of the curvature change, and based on the bending moment calculation formula, according to the curvature of the curve of the internal force of the enclosure structure to be evaluated varying with depth, the actual bending moment of the enclosure structure to be evaluated is obtained, which is helpful to analyze according to the current bearing condition of the enclosure structure to be evaluated, and can effectively evaluate the safety status of the enclosure structure to be evaluated, so as to calculate the bending moment limit value of the enclosure structure to be evaluated based on the bending moment limit calculation formula, which is helpful to timely discover the deformation of the enclosure structure to be evaluated according to the bending moment limit value, and calculate the bending coefficient based on the bending moment limit value and the actual bending moment, and obtain the safety status assessment result of the enclosure structure to be evaluated according to the bending coefficient, which is helpful to monitor the enclosure structure according to the bending coefficient and reduce the risk of collapse of the enclosure structure, thereby ensuring the safety and quality of the construction process.

[0034] In a preferred example, the present application may be further configured as follows: obtaining the safety status assessment result of the enclosure structure to be assessed according to the bending coefficient, specifically including:

[0035] Compare the bending coefficient with a preset safety factor threshold, and determine the safety status assessment result of the enclosure structure to be assessed according to the comparison result;

[0036] According to the safety status assessment result, risk warning information is generated, and a warning prompt message is issued according to the risk warning information.

[0037] By adopting the above technical scheme, by obtaining a preset safety factor threshold as a safety limit, comparing the bending coefficient with the preset safety factor threshold, and determining the safety status assessment result of the retaining structure to be assessed based on the comparison result, the safety status assessment of the retaining structure is achieved, and the safety status of the retaining structure can be intuitively understood based on the safety status assessment result. By generating risk warning information based on the safety status assessment result, and issuing a warning prompt message based on the risk warning information, it is helpful to attract the attention of construction technicians, avoid illegal operations by construction technicians during the excavation process, improve safety protection for construction technicians, ensure the safety and quality of the construction process by adjusting construction methods, strengthening support or stopping construction, ensure the safety of construction technicians during the construction process, and improve the reliability of the retaining structure, which is conducive to ensuring the safety of the foundation pit.

[0038] The second object of the invention is achieved by the following technical solutions:

[0039] A system for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data, the system comprising:

[0040] A data acquisition module, used to acquire historical deep foundation pit inclinometric data of the construction site, and generate inspection and smoothing optimization information based on the historical deep foundation pit inclinometric data;

[0041] A model training module is used to construct a semi-supervised neural network model based on the inspection and smoothing optimization information, and train the semi-supervised neural network model based on the historical deep foundation pit inclination measurement data to obtain a trained smoothing optimization model;

[0042] A model output module is used to perform smooth optimization on the deep foundation pit inclinometric data obtained from the construction site based on the trained smooth optimization model to obtain a curve of the internal force of the retaining structure to be evaluated varying with depth;

[0043] The safety assessment module is used to calculate the actual bending moment of the retaining structure to be assessed according to the curve of the internal force variation with depth of the retaining structure to be assessed, and obtain the bending moment limit value of the retaining structure to be assessed, and obtain the safety status assessment result of the retaining structure to be assessed based on the bending moment limit value and the actual bending moment.

[0044] By adopting the above-mentioned technical scheme, since the current optimization method for the inclinometer data automatically collected by the automatic inclinometer is realized by using supervised learning algorithms such as support vector machines and recurrent neural networks, it can easily adapt to different foundation pit structure types and complex soil layer changes, but it requires a huge labeled data set to support it, and the labeling process is time-consuming and inefficient. Therefore, this application obtains the historical deep foundation pit inclinometer data at the construction site, generates inspection and smoothing optimization information based on the historical deep foundation pit inclinometer data, and then constructs a semi-supervised neural network model based on the inspection and smoothing optimization information. The semi-supervised neural network model is trained based on the historical deep foundation pit inclinometer data to obtain a trained smoothing optimization model, which helps to improve the adaptability of the semi-supervised neural network model, so that the model can perform data inspection and smoothing optimization on the deep foundation pit inclinometer data, and then through Based on the trained smooth optimization model, the deep foundation pit inclinometer data obtained from the construction site are smoothly optimized to obtain the model-optimized curve of the internal force of the retaining structure to be evaluated varying with depth, which helps to eliminate the error caused by directly using the deep foundation pit inclinometer data and improve the accuracy of the curve of the internal force of the retaining structure to be evaluated varying with depth. The actual bending moment of the retaining structure to be evaluated is calculated according to the curve of the internal force of the retaining structure to be evaluated varying with depth, and the bending moment limit value of the retaining structure to be evaluated is obtained. Based on the bending moment limit value and the actual bending moment, the safety status assessment result of the retaining structure to be evaluated is obtained, which helps to intuitively understand the safety status of the retaining structure to be evaluated, and according to the safety assessment result of the retaining structure to be evaluated, the reliability of the retaining structure is improved, so as to achieve the purpose of improving the accuracy and credibility of the results of monitoring the safety status of the retaining structure, thereby ensuring the safety of the foundation pit.

[0045] The third objective of the present application is achieved through the following technical solutions:

[0046] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method for assessing the safety status of a retaining structure based on deep foundation pit inclinometer data are implemented.

[0047] The fourth objective of the present application is achieved through the following technical solutions:

[0048] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for assessing the safety status of a retaining structure based on deep foundation pit inclinometer data.

[0049] In summary, the present application includes at least one of the following beneficial technical effects:

[0050] 1. By selecting a part of the inclination data from the historical deep foundation pit inclination data and marking it as partial inclination data, it helps to reduce the cost of making label data sets. By judging the linear relationship of the partial inclination data, the corresponding fitting algorithm is selected according to the linear type of the partial inclination data for fitting and labeling processing, so as to obtain the corresponding fitting weight, which helps to eliminate the error of the partial inclination data and improve the smoothness of the data. By weighting the fitting weights, the labeled data is obtained according to the weighted results, which helps to reduce the time spent on manual labeling, reduce the links of manual intervention, and improve the credibility of the data;

[0051] 2. By training the semi-supervised neural network based on the inclinometer data set, using the data enhancement method suitable for unlabeled data, and training the smoothing optimization model, the model can perform data inspection and smooth optimization on the inclinometer curve data with only a small number of labels, reducing the cost of manual labeling, and being able to deal with the problem of scarce labels in real scenarios. Therefore, the model obtained based on the inclinometer data can achieve better results in the current application scenario, and can easily adapt to different foundation pit structure types and complex soil layer changes, so as to obtain a smooth curve of the internal force of the retaining structure to be evaluated changing with depth;

[0052] 3. By solving the curvature-based formula and according to the curve of the internal force of the enclosure structure to be evaluated changing with depth, the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth is obtained, which is helpful to analyze the smoothness of the curve and the trend of the curvature change. By using the bending moment calculation formula and according to the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth, the actual bending moment of the enclosure structure to be evaluated is obtained, which is helpful to analyze according to the current load-bearing condition of the enclosure structure to be evaluated, and can effectively evaluate the safety status of the enclosure structure to be evaluated. Therefore, based on the bending moment limit calculation formula, the bending moment limit value of the enclosure structure to be evaluated is calculated, which is helpful to timely discover the deformation of the enclosure structure to be evaluated according to the bending moment limit value. By calculating the bending coefficient based on the bending moment limit value and the actual bending moment, the safety status assessment result of the enclosure structure to be evaluated is obtained according to the bending coefficient, which is helpful to monitor the enclosure structure according to the bending coefficient and reduce the enclosure. The risk of collapse of the retaining structure is reduced, thereby ensuring the safety and quality of the construction process; 4. By obtaining the preset safety factor threshold as the safety limit, the bending coefficient is compared with the preset safety factor threshold, and the safety status assessment result of the retaining structure to be assessed is determined according to the comparison result, so as to realize the safety status assessment of the retaining structure, and the safety status of the retaining structure can be intuitively understood according to the safety status assessment result. By generating risk warning information according to the safety status assessment result, and issuing warning prompt messages according to the risk warning information, it is helpful to attract the attention of construction technicians, avoid construction technicians from operating in violation of regulations during the excavation process, improve the safety protection of construction technicians, ensure the safety and quality of the construction process by adjusting the construction method, strengthening support or stopping construction, ensure the safety of construction technicians during the construction process, and improve the reliability of the retaining structure, which is conducive to ensuring the safety of the foundation pit. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart for realizing a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0054] Figure 2 It is a flowchart for implementing step S20 in a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0055] Figure 3 It is a flowchart for implementing step S21 in a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0056] Figure 4 It is a flowchart for implementing step S22 in a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0057] Figure 5It is a flowchart for implementing step S40 in a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0058] Figure 6 It is a flowchart for implementing step S44 in a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0059] Figure 7 It is a partial schematic diagram of a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0060] Figure 8 It is another partial schematic diagram of the method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data in one embodiment of the present application;

[0061] Fig. 9 This is a principle block diagram of a system for evaluating the safety status of a retaining structure based on deep foundation pit inclinometric data in one embodiment of the present application;

[0062] Fig.10 It is a schematic diagram of a device in an embodiment of the present application. DETAILED DESCRIPTION

[0063] The present application is further described in detail below in conjunction with the accompanying drawings.

[0064] In one embodiment, if Figure 1-Figure 8 As shown, the present application discloses a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data, which specifically includes the following steps:

[0065] S10: Obtain historical deep foundation pit inclinometric data of the construction site, and generate inspection and smoothing optimization information based on the historical deep foundation pit inclinometric data.

[0066] In this embodiment, the principle of inclinometer is to bury an inclinometer tube in the deep foundation pit, and then monitor the lateral deformation of the deep foundation pit by an automatic inclinometer. The monitoring is performed regularly according to the monitoring cycle to obtain a single lateral deformation measurement result. The automatic inclinometer stops once every 0.5m change in depth in the inclinometer tube and records the data once. The automatic inclinometer will have a lateral deformation measurement result every 0.5m depth. The lateral deformation measurement result is the cumulative displacement, and then the deep foundation pit inclinometer data is obtained. Historical deep foundation pit data refers to the inclinometer data collected by the instrument in the past period of time. Inspection and smoothing information is due to the fact that the inclinometer curve measured by the instrument has systematic errors and random errors, and even gross errors, so it is necessary to perform data inspection and smoothing optimization on the deep foundation pit inclinometer data.

[0067] Specifically, by setting the working parameters and working mode of the automatic inclinometer located at the construction site, the deep foundation pit inclinometer data collected by the automatic inclinometer is used, and the deep foundation pit data collected by the automatic inclinometer is placed into a database for storing data through a monitoring host connected to the automatic inclinometer. Then, historical deep foundation pit inclinometer data can be obtained from the database, and information for data inspection and smoothing optimization of the deep foundation pit inclinometer data can be generated based on the historical deep foundation pit inclinometer data, i.e., inspection and smoothing optimization information.

[0068] S20: Based on the inspection and smoothing optimization information, a semi-supervised neural network model is constructed, and the semi-supervised neural network model is trained based on historical deep foundation pit inclinometric data to obtain a trained smoothing optimization model.

[0069] At present, it is popular to use supervised learning algorithms such as support vector machines and recurrent neural networks, which have excellent robustness and can easily adapt to different types of foundation pit structures and complex soil layer changes, but require huge labeled data sets to support. In this embodiment, a semi-supervised learning algorithm is used, and only labeled data and a large amount of unlabeled data are used to train the neural network model. It not only has nonlinear capabilities, but also can reduce the cost of making labeled data sets. A semi-supervised neural network model refers to a neural network model that combines unsupervised learning and supervised learning. The smoothing optimization model is used to eliminate the errors caused by directly using deep foundation pit inclinometer data, and output a smooth curve of the internal force of the retaining structure changing with depth.

[0070] Specifically, based on the inspection and smoothing optimization information obtained from the data inspection and smoothing optimization of the deep foundation pit inclinometer data, a neural network model combining unsupervised learning and supervised learning, namely a semi-supervised neural network model, is constructed. Therefore, the semi-supervised neural network model is trained based on the acquired historical deep foundation pit inclinometer data as a training set, so that the trained smoothing optimization model receives the deep foundation pit inclinometer data as input, judges the optimization status of the deep foundation pit inclinometer data based on the deep foundation pit inclinometer data, and then outputs a smooth curve of the internal force of the retaining structure changing with depth.

[0071] S30: Based on the trained smooth optimization model, the deep foundation pit inclinometric data obtained from the construction site is smoothed and optimized to obtain the curve of the internal force of the retaining structure to be evaluated varying with depth.

[0072] In this embodiment, the deep foundation pit inclinometer data refers to the data obtained from the construction site for monitoring and evaluating the deformation and stability of the deep foundation pit structure. The curve of the internal force variation with depth of the enclosure structure to be evaluated refers to the curve used to describe the variation of the internal force of the enclosure structure to be evaluated at different depths.

[0073] Specifically, by using the trained smoothing optimization model, the deep foundation pit inclinometer data collected and calculated by the automatic inclinometer from the construction site is checked and smoothed and optimized to eliminate the errors caused by directly using the deep foundation pit inclinometer data. The trained smoothing optimization model can receive the deep foundation pit inclinometer data as input, and judge the optimization status of the deep foundation pit inclinometer data based on the deep foundation pit inclinometer data, and then output a smooth curve of the internal force of the retaining structure to be evaluated changing with depth.

[0074] S40: According to the curve of the internal force of the enclosure structure to be evaluated changing with depth, the actual bending moment of the enclosure structure to be evaluated is calculated, and the bending moment limit value of the enclosure structure to be evaluated is obtained. Based on the bending moment limit value and the actual bending moment, a safety status assessment result of the enclosure structure to be evaluated is obtained.

[0075] In this embodiment, the actual bending moment refers to the actual value of the bending moment borne by the enclosure structure to be evaluated when it is subjected to external loads. The bending moment limit value refers to the maximum value of the bending moment borne by the enclosure structure to be evaluated when it is subjected to external loads. The safety status assessment result refers to the result obtained from the safety status assessment of the enclosure structure to be evaluated.

[0076] Specifically, according to the curve of the internal force variation of the enclosing structure to be evaluated at different depths, that is, the curve of the internal force variation with depth of the enclosing structure to be evaluated, and with the help of the curvature definition, the curvature of the curve of the internal force variation with depth of the enclosing structure to be evaluated is calculated, so as to calculate the actual bending moment of the enclosing structure to be evaluated according to the curvature of the curve, and obtain the bending moment limit value of the enclosing structure to be evaluated, that is, the maximum value of the bending moment that the enclosing structure to be evaluated can withstand, and based on the bending moment limit value and the actual bending moment of the enclosing structure to be evaluated, the bending moment limit value is compared with the actual bending moment, so as to conduct a safety assessment of the enclosing structure, so as to obtain the safety status assessment result of the enclosing structure to be evaluated.

[0077] In one embodiment, if Figure 2 as well as Figure 7 As shown, in step S20, a semi-supervised neural network model is constructed based on the inspection and smoothing optimization information, and the semi-supervised neural network model is trained based on the historical deep foundation pit inclination data to obtain a trained smoothing optimization model, which specifically includes:

[0078] S21: Label some of the historical deep foundation pit inclinometer data based on the fitting algorithm to obtain labeled data.

[0079] In this embodiment, the labeled data refers to a set of labeled data obtained by labeling part of the inclination measurement data in the historical deep foundation pit inclination measurement data through a fitting algorithm.

[0080] Specifically, the fitting algorithm is used to fit and manually annotate some of the historical deep foundation pit inclinometer data, and the final label is obtained by manually designing weighted parameters, that is, the inclinometer curve changes with depth fitting function, so as to obtain a small batch of labeled data with labels. For example, the labeled data accounts for 1%-10% of the entire inclinometer data set, and the other inclinometer data are all unlabeled data, which reduces the sample production cost.

[0081] S22: Construct an inclinometer dataset based on the labeled data and the unlabeled data in the historical deep foundation pit inclinometer data.

[0082] Specifically, based on the labeled data obtained by fitting the fitting algorithm and manual annotation and the unlabeled data in the historical deep foundation pit inclinometer data, an inclinometer dataset for training the model is constructed.

[0083] S23: Perform model training on the semi-supervised neural network model according to the inclinometer data set to obtain a trained smoothing optimization model.

[0084] Specifically, the semi-supervised neural network model is trained according to the inclinometer data set to obtain a trained smooth optimization model, which can be used to receive the deep foundation pit inclinometer data as input, judge the optimization of the deep foundation pit inclinometer data, and then output a smooth curve of the internal force of the retaining structure changing with depth. It should be understood that the process of training the training model using training samples belongs to the existing method and will not be repeated here.

[0085] In one embodiment, if Figure 3 As shown, in step S21, part of the inclination measurement data in the historical deep foundation pit inclination measurement data is labeled based on the fitting algorithm to obtain labeled data, which specifically includes:

[0086] S211: Select a portion of the inclinometer data from the historical deep foundation pit inclinometer data and mark it as partial inclinometer data.

[0087] Specifically, a part of the inclinometer data is selected from the historical deep foundation pit inclinometer data, marked as partial inclinometer data, and labeled.

[0088] S212: curve fitting is performed on part of the inclinometric data using a polynomial function fitting algorithm of a linear least squares method, an arbitrary function fitting algorithm of a nonlinear least squares method, a quintic spline function fitting algorithm, a quintic smoothing spline fitting function and a polynomial fitting algorithm, and a weighted least squares fitting algorithm to obtain corresponding fitting curves.

[0089] Specifically, the linear least squares polynomial function fitting algorithm, the nonlinear least squares arbitrary function fitting algorithm, the quintic spline function fitting algorithm, the quintic smoothing spline fitting function and the polynomial fitting algorithm and the weighted least squares fitting algorithm are used to fit the curve of some inclinometric data. The linear least squares polynomial function fitting algorithm is to satisfy the formula Thus, the fitting curve is obtained, and by setting the number of polynomial terms n to 6, the data points are (x i ,y i ), i = 0, 1, ..., m, Φ is the function class consisting of all polynomials with degree not exceeding n ≦ m: P n (x i )∈Φ; The arbitrary function fitting algorithm of the nonlinear least squares method is to obtain the nonlinear function formula with the help of the open source fitting toolbox xymath, and use the formula for fitting. There are three nonlinear function formulas, which are: f(x)=c0+c1x 2 +c2x 3 +a3e x +a4x5+a5x 6 +a6(log(x)) 3 , one of them is selected for fitting, so as to obtain a fitting curve; the quintic spline function fitting algorithm uses the quintic spline function to fit part of the inclinometer data, where the smoothing spline parameter s is 5, so as to obtain a fitting curve; the quintic smoothing spline fitting function and the polynomial fitting algorithm first use the quintic smoothing spline fitting function to perform a certain smoothing and noise reduction on part of the inclinometer data, and then further smooth and reduce the noise of the smoothed part of the inclinometer data based on the linear least squares polynomial fitting algorithm, so as to obtain a fitting curve; the weighted least squares fitting algorithm weights the fitting residuals at the maximum value of part of the inclinometer data, where the weight of the maximum value is set to 10 and the rest is set to 1, so as to obtain a fitting curve. Therefore, according to each fitting algorithm, the corresponding fitting curve is obtained.

[0090] S213: Based on the weighting formula set by the system, weight each fitting curve and set the weight coefficient that meets the preset conditions to obtain the fitting function of the inclinometer curve changing with depth. The weighting formula is f(x)=a1f1(x)+a2f2(x)+…+a n f n (x), where n is the number of fitting algorithms used, n is 5; a n is the fitting algorithm weight of the nth fitting algorithm; f n (x) is the fitting function corresponding to the nth fitting algorithm; f(x) is the fitting function of the inclinometer curve changing with depth.

[0091] Specifically, the corresponding fitting curves are obtained by respectively using the above-mentioned different fitting algorithms, and then the obtained fitting curves are weighted based on the weighting formula set by the system. The weighting formula is f(x)=a1f1(x)+a2f2(x)+…+a n f n (x), by manually setting the weight coefficient that meets the preset conditions, the fitting function of the inclinometer curve changing with depth is obtained.

[0092] S214: fitting a function according to the variation of the inclinometer curve with depth to obtain labeled data.

[0093] Specifically, by selecting the fitting curves obtained by the corresponding fitting algorithm, weighted processing is performed on each fitting curve to obtain a fitting function of the inclinometer curve changing with depth, thereby obtaining labeled data, so that the labeled data can be substituted into the smoothing optimization model for model training, so that the model can easily adapt to different foundation pit structure types and complex soil layer changes.

[0094] In one embodiment, if Figure 4 as well as Figure 8 As shown, in step S22, the semi-supervised neural network model is trained according to the inclinometer data set to obtain a trained smooth optimization model, which specifically includes:

[0095] S221: Obtain a fully connected neural network, input the labeled data into the fully connected neural network, and calculate according to the first prediction result and the manually annotated labels to obtain a supervised loss function value.

[0096] Specifically, the mean square error is selected as the loss function of the semi-supervised neural network model, a fully connected neural network is obtained, and the number of hidden layers in the fully connected neural network is set to three, and the number of neurons in the hidden layer is set to [128, 256, 128]. After the labeled data is input into the fully connected neural network, the first prediction result output by the fully connected neural network is obtained, and the MSE values ​​of the two are calculated based on the first prediction result and the manually labeled labels. MSE refers to the mean of the sum of squares of the differences between the predicted value and the target value, that is, the supervised loss function value is obtained by calculating the mean square error of the labeled data and the manually labeled labels.

[0097] S222: Perform enhancement processing on the unlabeled data to obtain enhanced data, input the enhanced data into the fully connected neural network, and calculate the obtained second prediction result to obtain an unsupervised loss function value.

[0098] Specifically, enhanced data is obtained by enhancing the unlabeled data, i.e., the unlabeled data in the historical deep foundation pit inclinometer data. For example, the enhanced processing method uses length scaling, noise addition, horizontal flipping around the y-axis, and scaling according to the inclinometer data characteristics of the inclinometer data set, and the processing probabilities are set to [0.5, 0.6, 0.5, 0.4] respectively. Two different output data are obtained according to the enhanced processing probabilities. The enhanced data at this time is composed of these two output data, so that after the enhanced data is input into the fully connected neural network, the second prediction result output by the fully connected neural network is obtained, and the MSE value is calculated for the second prediction result, that is, the mean square error of the two is calculated to obtain the unsupervised loss function value.

[0099] S223: Perform iterative training on the semi-supervised neural network model according to the supervised loss function value and the unsupervised loss function value, and determine the number of iterations.

[0100] Specifically, according to the supervised loss function value and the unsupervised loss function value obtained by the above calculation, the semi-supervised neural network model is iteratively trained to improve the performance of the semi-supervised neural network model so that it can adapt to the rules and characteristics in the data, and therefore the number of iterations of model training is judged.

[0101] S224: When the number of iterations meets a preset iteration number threshold, the trained semi-supervised neural network model is output and used as a smoothing optimization model.

[0102] Specifically, when the number of iterations meets the preset iteration threshold, it means that the performance of the semi-supervised neural network model has been adjusted to the best, and its accuracy, generalization ability and stability have achieved good results. Therefore, the trained semi-supervised neural network model is output as a smoothing optimization model that can be used for deep foundation pit inclinometer data.

[0103] In one embodiment, if Figure 5 As shown, in step S40, the actual bending moment of the enclosure structure to be evaluated is calculated according to the curve of the internal force of the enclosure structure to be evaluated changing with depth, and the bending moment limit value of the enclosure structure to be evaluated is obtained. Based on the bending moment limit value and the actual bending moment, the safety status assessment result of the enclosure structure to be evaluated is obtained, which specifically includes:

[0104] S41: Based on the curvature solution formula, according to the curve of the internal force of the enclosure structure to be evaluated changing with depth, the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth is obtained. The curvature solution formula is Φ(x)=d 2 y / dx 2 , where dx 2 is the second-order differential of the depth x of the enclosure to be assessed; d 2y is the second-order differential of the curve y(x) of the internal force of the enclosure structure to be evaluated changing with depth; Φ(x) is the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth.

[0105] In this embodiment, the curvature of a curve refers to the degree of curvature or curvature of the curve at a certain point.

[0106] Specifically, according to the curve of the internal force of the enclosure structure to be evaluated varying with depth, the curvature of the curve used for the degree of curvature of the curve of the internal force of the enclosure structure to be evaluated varying with depth is obtained using the curvature solution formula, that is, the curvature of the curve of the internal force of the enclosure structure to be evaluated varying with depth.

[0107] S42: Based on the bending moment calculation formula, according to the curvature of the curve of the internal force of the enclosure structure to be evaluated with depth, the actual bending moment of the enclosure structure to be evaluated is obtained. The bending moment calculation formula is M(x)=EIΦ(x)=EI·d 2 y / dx 2 , where M(x) is the actual bending moment of the enclosure structure to be evaluated; EI is the bending stiffness of the enclosure structure to be evaluated.

[0108] Specifically, according to the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth, the actual bending moment of the enclosure structure to be evaluated is calculated by using the bending moment calculation formula. Since the actual bending moment of the enclosure structure is proportional to the curvature of the curve, when the absolute value of the curvature of the curve is larger, it means that the curvature of the curve is greater, and the bending moment of the enclosure structure under the action of external load is also greater, thereby obtaining the internal force situation of the enclosure structure to be evaluated under the action of external load, that is, the actual bending moment of the enclosure structure to be evaluated.

[0109] S43: Based on the moment limit calculation formula, calculate the moment limit value of the enclosure structure to be evaluated. The moment limit calculation formula is M = (0.85 × f ck ×b×h) / γ m , where M is the bending moment limit, f ck is the design compressive strength of the enclosure structure to be evaluated, b is the width of the enclosure structure to be evaluated, h is the effective height of the enclosure structure to be evaluated, γ m is the safety factor.

[0110] Specifically, by obtaining the design data of the enclosure structure to be evaluated, as well as the width and effective height of the enclosure structure to be evaluated, and using the bending moment limit calculation formula, the bending moment limit value of the enclosure structure to be evaluated can be calculated.

[0111] S44: Based on the bending moment limit value and the actual bending moment, the bending coefficient is calculated, and according to the bending coefficient, the safety status assessment result of the enclosure structure to be assessed is obtained.

[0112] In this embodiment, the bending coefficient refers to a coefficient used to judge the safety of the enclosure structure under normal use or load conditions.

[0113] Specifically, based on the bending moment limit value and the actual bending moment of the enclosing structure to be evaluated, the ratio of the two (actual bending moment / bending moment limit value) is calculated to obtain a coefficient for judging the safety of the enclosing structure to be evaluated under normal use or load conditions, namely, the bending coefficient. Then, based on the bending coefficient, the safety status of the enclosing structure to be evaluated is judged, thereby obtaining a safety status assessment result of the enclosing structure to be evaluated.

[0114] In one embodiment, if Figure 6 As shown, in step S44, the safety status assessment result of the enclosure structure to be assessed is obtained according to the bending coefficient, which specifically includes:

[0115] S441: Compare the bending coefficient with a preset safety factor threshold, and determine a safety status assessment result of the enclosure structure to be assessed based on the comparison result.

[0116] In this embodiment, the preset safety factor threshold refers to a numerical value that ensures the limit of the safety level of the enclosure structure.

[0117] Specifically, the preset safety factor threshold provides quantitative information about the safety status of the enclosing structure, that is, the preset safety factor threshold is subsequently used as a benchmark, and the bending coefficient is compared with the preset safety factor threshold to obtain a comparison result. For example, when the comparison result is that the bending coefficient is greater than the preset safety factor threshold, the safety status of the enclosing structure to be evaluated under the bending coefficient is determined to be dangerous, that is, the safety status assessment result is unsafe. When the comparison result is that the bending coefficient is less than or equal to the preset safety factor threshold, the safety status of the enclosing structure to be evaluated under the bending coefficient is determined to be normal, that is, the safety status assessment result is safe. For example, the system defines the preset safety factor threshold as 0.7, and the bending coefficient is 0.8 after solving the above data, that is, the safety status assessment result of the enclosing structure to be evaluated is dangerous. In this way, the safety status assessment result of the enclosing structure to be evaluated is obtained.

[0118] S442: Generate risk warning information based on the safety status assessment result, and issue a warning prompt message based on the risk warning information.

[0119] In this embodiment, the risk warning information refers to information indicating that there is a risk in the enclosure structure. The warning prompt message refers to a message used to draw the attention of construction technicians to the enclosure structure.

[0120] Specifically, the safety status assessment result of the retaining structure to be assessed is obtained by comparing the bending coefficient with the preset safety factor threshold, and risk warning information is generated based on the safety status assessment result, and a warning prompt message is issued based on the risk warning information. By issuing the warning prompt message, construction technicians are prevented from operating in violation of regulations during the excavation process, and the safety protection of construction technicians is improved. The safety and quality of the construction process are ensured by adjusting the construction method, strengthening support or stopping construction, and the safety of construction technicians during the construction process is guaranteed.

[0121] It should be understood that the size of the serial numbers of the steps in the above embodiments 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 embodiments of the present application.

[0122] In one embodiment, a retaining structure safety status assessment system based on deep foundation pit inclinometer data is provided, and the retaining structure safety status assessment system based on deep foundation pit inclinometer data corresponds one-to-one to the retaining structure safety status assessment method based on deep foundation pit inclinometer data in the above embodiment. Fig. 9 As shown, the retaining structure safety status assessment system based on deep foundation pit inclinometer data includes a data acquisition module, a model training module, a model output module and a safety assessment module. The detailed description of each functional module is as follows:

[0123] A data acquisition module is used to obtain historical deep foundation pit inclination measurement data of the construction site, and generate inspection and smoothing optimization information based on the historical deep foundation pit inclination measurement data;

[0124] The model training module is used to construct a semi-supervised neural network model based on the inspection and smoothing optimization information, and train the semi-supervised neural network model based on the historical deep foundation pit inclination data to obtain a trained smoothing optimization model;

[0125] The model output module is used to perform smooth optimization on the deep foundation pit inclinometric data obtained from the construction site based on the trained smooth optimization model, and obtain the curve of the internal force of the retaining structure to be evaluated changing with depth;

[0126] The safety assessment module is used to calculate the actual bending moment of the enclosure structure to be assessed according to the curve of the internal force variation with depth of the enclosure structure to be assessed, and obtain the bending moment limit value of the enclosure structure to be assessed, and obtain the safety status assessment result of the enclosure structure to be assessed based on the bending moment limit value and the actual bending moment.

[0127] Optionally, the model training module specifically includes:

[0128] The data labeling submodule is used to label some of the historical deep foundation pit inclinometer data based on the fitting algorithm to obtain labeled data;

[0129] Construct a training submodule to construct an inclination survey data set based on labeled data and unlabeled data in historical deep foundation pit inclination survey data;

[0130] The model training submodule is used to train the semi-supervised neural network model according to the inclinometer data set to obtain a trained smooth optimization model.

[0131] Optionally, the data annotation submodule specifically includes:

[0132] The data selection unit is used to select a part of the inclination measurement data from the historical deep foundation pit inclination measurement data and mark it as partial inclination measurement data; the algorithm fitting unit is used to respectively use the polynomial function fitting algorithm of the linear least squares method, the arbitrary function fitting algorithm of the nonlinear least squares method, the quintic spline function fitting algorithm, the quintic smoothing spline fitting function and the polynomial fitting algorithm and the weighted least squares fitting algorithm to perform curve fitting on the partial inclination measurement data to obtain corresponding fitting curves;

[0133] The curve output unit is used to weight each fitting curve based on the weighting formula set by the system, and set the weight coefficient that meets the preset conditions to obtain the fitting function of the inclinometer curve changing with depth. The weighting formula is f(x)=a1f1(x)+a2f2(x)+…+a n f n (x), where n is the number of fitting algorithms used, n is 5; a n is the fitting algorithm weight of the nth fitting algorithm; f n (x) is the fitting function corresponding to the nth fitting algorithm; f(x) is the fitting function of the inclination curve changing with depth;

[0134] The data annotation unit is used to obtain labeled data by fitting a function according to the variation of the inclinometer curve with depth.

[0135] Optionally, the model training submodule specifically includes:

[0136] The first branch computing unit is used to obtain a fully connected neural network, input the labeled data into the fully connected neural network, and then perform a calculation based on the first prediction result and the manually labeled label to obtain a supervised loss function value;

[0137] The second branch computing unit is used to perform enhancement processing on the unlabeled data to obtain enhanced data, input the enhanced data into the fully connected neural network, and then calculate the obtained second prediction result to obtain an unsupervised loss function value;

[0138] A model iterative training unit, used to iteratively train the semi-supervised neural network model according to the supervised loss function value and the unsupervised loss function value, and to determine the number of iterations;

[0139] The iterative output model unit is used to output the trained semi-supervised neural network model as a smoothing optimization model when the number of iterations meets a preset iteration number threshold.

[0140] Optionally, the security assessment module specifically includes:

[0141] The curvature calculation submodule is used to obtain the curvature of the curve of the internal force of the enclosure structure to be evaluated with depth based on the curvature solution formula. The curvature solution formula is Φ(x)=d 2 y / dx 2 , where dx 2 is the second-order differential of the depth x of the enclosure to be assessed; d 2 y is the second-order differential of the curve y(x) of the internal force of the enclosure structure to be evaluated changing with depth; Φ(x) is the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth;

[0142] The bending moment calculation submodule is used to obtain the actual bending moment of the enclosure structure to be evaluated based on the bending moment calculation formula and the curvature of the curve of the internal force of the enclosure structure to be evaluated with depth. The bending moment calculation formula is M(x)=EIΦ(x)=EI·d 2 y / dx 2 , where M(x) is the actual bending moment of the enclosure structure to be evaluated; EI is the bending stiffness of the enclosure structure to be evaluated;

[0143] The moment limit calculation submodule is used to calculate the moment limit value of the enclosure structure to be evaluated based on the moment limit calculation formula. The moment limit calculation formula is M = (0.85 × f ck ×b×h) / γ m , where M is the bending moment limit value of the enclosure structure to be evaluated, f ck is the design compressive strength of the enclosure structure to be evaluated, b is the width of the enclosure structure to be evaluated, h is the effective height of the enclosure structure to be evaluated, γ m is the safety factor;

[0144] The safety assessment submodule is used to calculate the bending coefficient based on the bending moment limit value and the actual bending moment, and obtain the safety status assessment result of the enclosure structure to be assessed according to the bending coefficient.

[0145] Optionally, the security assessment submodule specifically includes:

[0146] A comparison and evaluation unit, used for comparing the bending coefficient with a preset safety factor threshold, and determining a safety status evaluation result of the enclosure structure to be evaluated according to the comparison result;

[0147] The risk warning unit is used to generate risk warning information according to the safety status assessment results, and issue a warning prompt message according to the risk warning information.

[0148] Regarding the specific limitations of the retaining structure safety status assessment system based on deep foundation pit inclinometer data, please refer to the limitations of the retaining structure safety status assessment method based on deep foundation pit inclinometer data mentioned above, which will not be repeated here. Each module in the above-mentioned retaining structure safety status assessment system based on deep foundation pit inclinometer data can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0149] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig.10 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data information such as historical deep foundation pit inclinometer data, data information of smoothing optimization models and preset safety factor thresholds. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data is implemented.

[0150] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned method for assessing the safety status of a retaining structure based on deep foundation pit inclinometer data are implemented.

[0151] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0152] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0153] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for evaluating the safety status of retaining structures based on deep foundation pit inclinometer data, characterized in that: The method for evaluating the safety status of the retaining structure based on the deep foundation pit inclinometer data comprises: Acquire historical deep foundation pit inclinometer data of the construction site, and generate inspection and smoothing optimization information according to the historical deep foundation pit inclinometer data; According to the inspection and smooth optimization information, a semi-supervised neural network model is constructed, and the semi-supervised neural network model is trained based on the historical deep foundation pit inclination measurement data to obtain a trained smooth optimization model; Based on the trained smooth optimization model, the deep foundation pit inclinometric data obtained from the construction site are smoothly optimized to obtain a curve of the internal force of the retaining structure to be evaluated varying with depth; According to the curve of the internal force variation with depth of the enclosure structure to be evaluated, the actual bending moment of the enclosure structure to be evaluated is calculated, and the bending moment limit value of the enclosure structure to be evaluated is obtained, and based on the bending moment limit value and the actual bending moment, the safety status assessment result of the enclosure structure to be evaluated is obtained; The method of constructing a semi-supervised neural network model based on the inspection and smooth optimization information, and training the semi-supervised neural network model based on the historical deep foundation pit inclinometer data to obtain a trained smooth optimization model specifically includes: Labeling part of the inclination measurement data in the historical deep foundation pit inclination measurement data based on a fitting algorithm to obtain labeled data; Constructing an inclinometer data set according to the labeled data and the unlabeled data in the historical deep foundation pit inclinometer data; Performing model training on the semi-supervised neural network model according to the inclinometer data set to obtain the trained smoothing optimization model; The method of labeling part of the inclination measurement data in the historical deep foundation pit inclination measurement data based on the fitting algorithm to obtain labeled data specifically includes: Selecting a portion of the inclination measurement data from the historical deep foundation pit inclination measurement data and marking the portion of the inclination measurement data as the portion of the inclination measurement data; The polynomial function fitting algorithm of the linear least squares method, the arbitrary function fitting algorithm of the nonlinear least squares method, the quintic spline function fitting algorithm, the quintic smoothing spline fitting function and the polynomial fitting algorithm and the weighted least squares fitting algorithm are respectively used to perform curve fitting on the part of the inclinometric data to obtain corresponding fitting curves; Based on the weighted formula set by the system, the fitting curves are weighted, and the weight coefficients that meet the preset conditions are set to obtain the fitting function of the inclinometer curve changing with depth. The weighted formula is f(x)=a1f1(x)+a2f2(x)+…+a n f n (x), where n is the number of fitting algorithms used, n is 5; a n is the fitting algorithm weight of the nth fitting algorithm; f n (x) is the fitting function corresponding to the nth fitting algorithm; f(x) is the fitting function of the inclination curve changing with depth; The labeled data is obtained according to a fitting function of the inclinometer curve changing with depth.

2. The method for evaluating the safety status of retaining structures based on deep foundation pit inclinometer data according to claim 1 is characterized in that: The performing model training on the semi-supervised neural network model according to the inclinometer data set to obtain the trained smooth optimization model specifically includes: Obtain a fully connected neural network, input the labeled data into the fully connected neural network, and perform calculations based on the first prediction result and the manually annotated labels to obtain a supervised loss function value; Performing enhancement processing on the unlabeled data to obtain enhanced data, inputting the enhanced data into the fully connected neural network, and calculating the obtained second prediction result to obtain an unsupervised loss function value; Iteratively training the semi-supervised neural network model according to the supervised loss function value and the unsupervised loss function value, and determining the number of iterations; When the number of iterations meets a preset iteration number threshold, the trained semi-supervised neural network model is output and used as the smoothing optimization model.

3. The method for evaluating the safety status of retaining structures based on deep foundation pit inclinometer data according to claim 1 is characterized in that: The step of calculating the actual bending moment of the enclosure structure to be evaluated according to the curve of the variation of the internal force of the enclosure structure to be evaluated with depth, obtaining the bending moment limit value of the enclosure structure to be evaluated, and obtaining the safety status evaluation result of the enclosure structure to be evaluated based on the bending moment limit value and the actual bending moment, specifically includes: Based on the curvature solution formula, the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth is obtained according to the curve of the internal force of the enclosure structure to be evaluated changing with depth. The curvature solution formula is Φ(x)=d 2 y / dx 2 , where dx 2 is the second-order differential of the depth x of the enclosure to be assessed; d 2 y is the second-order differential of the curve y(x) of the internal force of the enclosure structure to be evaluated changing with depth; Φ(x) is the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth; Based on the bending moment calculation formula, the actual bending moment of the enclosure structure to be evaluated is obtained according to the curvature of the curve of the internal force of the enclosure structure to be evaluated changing with depth. The bending moment calculation formula is M(x)=EIΦ(x)=EI·d 2 y / dx 2 , where M(x) is the actual bending moment of the enclosure structure to be evaluated; EI is the bending stiffness of the enclosure structure to be evaluated; Based on the bending moment limit calculation formula, the bending moment limit value of the enclosure structure to be evaluated is calculated. The bending moment limit calculation formula is M=(0.85×f ck ×b×h) / γ m , where M is the bending moment limit value of the enclosure structure to be evaluated, f ck is the design compressive strength of the enclosure structure to be evaluated, b is the width of the enclosure structure to be evaluated, h is the effective height of the enclosure structure to be evaluated, γ m is the safety factor; Based on the bending moment limit value and the actual bending moment, a bending coefficient is calculated, and according to the bending coefficient, a safety status assessment result of the enclosure structure to be assessed is obtained.

4. The method for evaluating the safety status of retaining structures based on deep foundation pit inclinometric data according to claim 3 is characterized in that: The step of obtaining the safety status assessment result of the enclosure structure to be assessed according to the bending coefficient specifically includes: Compare the bending coefficient with a preset safety factor threshold, and determine the safety status assessment result of the enclosure structure to be assessed according to the comparison result; According to the safety status assessment result, risk warning information is generated, and a warning prompt message is issued according to the risk warning information.

5. A system for evaluating the safety status of retaining structures based on deep foundation pit inclinometer data, characterized in that: The retaining structure safety status assessment system based on deep foundation pit inclinometer data includes: A data acquisition module, used to acquire historical deep foundation pit inclinometric data of the construction site, and generate inspection and smoothing optimization information based on the historical deep foundation pit inclinometric data; A model training module is used to construct a semi-supervised neural network model based on the inspection and smoothing optimization information, and train the semi-supervised neural network model based on the historical deep foundation pit inclination measurement data to obtain a trained smoothing optimization model; A model output module is used to perform smooth optimization on the deep foundation pit inclinometric data obtained from the construction site based on the trained smooth optimization model to obtain a curve of the internal force of the retaining structure to be evaluated varying with depth; A safety assessment module, used to calculate the actual bending moment of the enclosure structure to be assessed according to the curve of the internal force of the enclosure structure to be assessed changing with depth, and obtain the bending moment limit value of the enclosure structure to be assessed, and obtain the safety status assessment result of the enclosure structure to be assessed based on the bending moment limit value and the actual bending moment; The model training module specifically includes: A data labeling submodule is used to label part of the inclination measurement data in the historical deep foundation pit inclination measurement data based on a fitting algorithm to obtain labeled data; Constructing a training submodule, which is used to construct an inclination measurement data set based on the labeled data and the unlabeled data in the historical deep foundation pit inclination measurement data; A model training submodule, used for performing model training on the semi-supervised neural network model according to the inclination measurement data set to obtain the trained smooth optimization model; The data annotation submodule specifically includes: A data selection unit, used for selecting a part of the inclination measurement data from the historical deep foundation pit inclination measurement data, and marking the part of the inclination measurement data; An algorithm fitting unit is used to respectively use a polynomial function fitting algorithm of a linear least squares method, an arbitrary function fitting algorithm of a nonlinear least squares method, a quintic spline function fitting algorithm, a quintic smoothing spline fitting function and a polynomial fitting algorithm, and a weighted least squares fitting algorithm to perform curve fitting on the part of the inclinometric data to obtain corresponding fitting curves; The curve output unit is used to weight the fitting curves based on the weighting formula set by the system, and set the weight coefficient that meets the preset conditions to obtain the fitting function of the inclinometer curve changing with depth. The weighting formula is f(x)=a1f1(x)+a2f2(x)+…+a n f n (x), where n is the number of fitting algorithms used, n is 5; a n is the fitting algorithm weight of the nth fitting algorithm; f n (x) is the fitting function corresponding to the nth fitting algorithm; f(x) is the fitting function of the inclination curve changing with depth; The data labeling unit is used to obtain the labeled data by fitting a function according to the variation of the inclinometer curve with depth.

6. A computer 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 steps of the method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data as described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for evaluating the safety status of a retaining structure based on deep foundation pit inclinometer data as described in any one of claims 1 to 4 are implemented.

Images