Foundation pit settlement depth prediction method and device, equipment, storage medium and program product

Abstract

The invention relates to a foundation pit settlement depth prediction method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring monitoring information at a plurality of monitoring points in a foundation pit; the monitoring information comprises settlement depths at a plurality of historical moments, the settlement depth at the current moment and position information of the monitoring points; inputting the monitoring information of the plurality of monitoring points into an information deduction model to obtain the monitoring information of a plurality of non-monitoring points in the foundation pit; the information deduction model is obtained through sample monitoring information mask training at a plurality of monitoring points; and based on the monitoring information of the multiple monitoring points and the monitoring information of the multiple non-monitoring points, the settlement depth of each position in the foundation pit is obtained. By adopting the method, the prediction accuracy of the settlement depth of the foundation pit can be improved.

Description

Technical Field

[0001] This application relates to the field of underground engineering technology, and in particular to a method, apparatus, equipment, storage medium and program product for predicting the settlement depth of a foundation pit. Background Technology

[0002] With the acceleration of urbanization, underground space development has become an important part of modern urban development. In densely populated urban areas, surrounded by various underground pipelines, buildings, traffic arteries, and existing subway tunnels, monitoring ground settlement caused by excavation and using the collected data for real-time risk analysis is crucial to ensuring the safety of the surrounding area.

[0003] Currently, traditional methods for predicting foundation pit settlement rely on a limited number of physical monitoring points and analyze the time dependence of foundation pit settlement data to achieve short-term predictions of settlement at a limited number of physical monitoring points.

[0004] However, the aforementioned methods for predicting foundation pit settlement suffer from low accuracy. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, equipment, storage medium, and program product for predicting the settlement depth of a foundation pit, which can improve the accuracy of foundation pit settlement prediction, in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a method for predicting the settlement depth of a foundation pit, including:

[0007] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0008] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0009] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0010] In one embodiment, the settlement depth at various locations within the foundation pit is obtained based on monitoring information from multiple monitoring points and monitoring information from multiple unmonitored points, including:

[0011] The monitoring information from each monitoring point and the monitoring information from each unmonitored point are spliced ​​together to obtain the initial point features;

[0012] For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points;

[0013] Based on the initial point characteristics and the adjacency matrix, the settlement depth at each location in the foundation pit is obtained.

[0014] In one embodiment, the above-described process of concatenating the monitoring information at each monitoring point and the monitoring information at each unmonitored point to obtain initial point features includes:

[0015] Time normalization is performed on the time features in each monitoring information to obtain the normalized time features;

[0016] The location features in each monitoring information are normalized to obtain the normalized location features;

[0017] The normalized time features and the normalized position features are concatenated to obtain the initial point features.

[0018] In one embodiment, the above-described concatenation of normalized time features and normalized position features to obtain initial point features includes:

[0019] The normalized time features and the normalized position features are concatenated to obtain the concatenated features.

[0020] The splicing features are combined to obtain the initial point features.

[0021] In one embodiment, the settlement depth at each location in the foundation pit is obtained based on the initial point features and the adjacency matrix, including:

[0022] The adjacency matrix is ​​normalized to obtain the normalized adjacency matrix;

[0023] Determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix;

[0024] The degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix are input into a convolutional network to obtain the settlement depth at each location in the foundation pit.

[0025] In one embodiment, the method further includes:

[0026] The sample monitoring information at multiple monitoring points is masked to obtain the masked sample monitoring information at candidate monitoring points.

[0027] Based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points, the initial information inference model is trained to obtain the information inference model.

[0028] Secondly, this application also provides a device for predicting the settlement depth of a foundation pit, comprising:

[0029] The acquisition module is used to acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0030] The inference module is used to input monitoring information from multiple monitoring points into the information inference model to obtain monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0031] The determination module is used to obtain the settlement depth at various locations in the foundation pit based on monitoring information from multiple monitoring points and monitoring information from multiple unmonitored points.

[0032] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0033] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0034] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0035] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0036] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0037] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0038] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0039] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0040] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0041] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0042] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0043] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0044] The aforementioned method, apparatus, equipment, storage medium, and program product for predicting the settlement depth of foundation pits first deduce the monitoring information of multiple unmonitored points in the foundation pit based on the monitoring information at multiple actual monitoring points. Then, based on the monitoring information at the multiple actual monitoring points and the monitoring information at the multiple unmonitored points, the settlement depth of the foundation pit is predicted. Compared with methods that predict the settlement depth of foundation pits based solely on the monitoring information at actual monitoring points, this application also uses the monitoring information of multiple unmonitored points in the foundation pit, which greatly improves the accuracy of the prediction of the settlement depth of the foundation pit. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is an application environment diagram of the method for predicting the settlement depth of a foundation pit in one embodiment;

[0047] Figure 2 This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in one embodiment;

[0048] Figure 3 This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in another embodiment;

[0049] Figure 4 This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in another embodiment;

[0050] Figure 5This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in another embodiment;

[0051] Figure 6 This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in another embodiment;

[0052] Figure 7 This is a schematic diagram of the structure of a convolutional network in one embodiment;

[0053] Figure 8 This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in another embodiment;

[0054] Figure 9 This is a schematic diagram of the training process of an information inference model in one embodiment;

[0055] Figure 10 This is a flowchart illustrating a method for predicting the settlement depth of a foundation pit in another embodiment;

[0056] Figure 11 This is a schematic cloud diagram illustrating the overall settlement of the foundation pit in one embodiment;

[0057] Figure 12 This is a structural block diagram of a device for predicting the settlement depth of a foundation pit in one embodiment. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0059] With the acceleration of urbanization, underground space development has become an important part of modern urban development. In densely populated urban areas, surrounded by various underground pipelines, buildings, traffic arteries, and existing subway tunnels, monitoring ground settlement caused by excavation and using the collected data for real-time risk analysis is crucial to ensuring the safety of the surrounding area.

[0060] Currently, traditional methods for predicting foundation pit settlement rely on a limited number of physical monitoring points and the analysis of the time dependence of settlement data to achieve short-term predictions of settlement at these limited monitoring points. For example, researchers have developed deep learning models based on Long Short-Term Memory (LSTM) networks and Gate Recurrent Units (GRUs), which can effectively capture the temporal evolution of settlement data and provide important support for the safety assessment of foundation pit construction.

[0061] However, the aforementioned methods for predicting foundation pit settlement suffer from low accuracy. This application aims to provide a method for predicting foundation pit settlement depth to address this problem.

[0062] Having described the background technology of the method for predicting the settlement depth of foundation pits provided in the embodiments of this application, the implementation environment involved in the method for predicting the settlement depth of foundation pits provided in the embodiments of this application will be briefly described below. The method for predicting the settlement depth of foundation pits provided in the embodiments of this application can be applied to, for example... Figure 1 The computer device shown includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for predicting the settlement depth of a foundation pit. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0063] Those skilled in the art will understand that Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0064] Having described the application scenarios of the foundation pit settlement depth prediction method provided in the embodiments of this application above, the following focuses on the foundation pit settlement depth prediction method described in this application.

[0065] In one embodiment, such as Figure 2 As shown, a method for predicting the settlement depth of a foundation pit is provided, and this method is applied to... Figure 1Taking a computer device as an example, the explanation includes the following steps:

[0066] S201. Obtain monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points.

[0067] Among them, the monitoring point refers to the physical entity monitoring point set up in the foundation pit in advance. The monitoring point can monitor the settlement depth at the location of the monitoring point, including the settlement depth at the location of the monitoring point at multiple historical times and the settlement depth at the location of the monitoring point at the current time.

[0068] In this embodiment, multiple monitoring points can be pre-set in the foundation pit, and detectors can be installed at each monitoring point to detect the location information and settlement depth of multiple monitoring points in the foundation pit in real time. The location information and settlement depth of multiple monitoring points in the foundation pit are stored in a database. When it is necessary to determine the overall settlement depth of the foundation pit, the monitoring information of multiple monitoring points in the foundation pit can be obtained from the database.

[0069] S202. Input the monitoring information from multiple monitoring points into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by masking the sample monitoring information from multiple monitoring points.

[0070] Among them, unmonitored points refer to virtual points in the foundation pit other than the monitoring points.

[0071] In this embodiment, after obtaining the monitoring information at multiple monitoring points, the monitoring information at multiple monitoring points can be input into the information inference model to infer the monitoring information at multiple unmonitored points in the foundation pit.

[0072] It should be noted that the sample monitoring information at multiple monitoring points can be masked, and the initial information inference model can be trained based on the masked sample monitoring information at multiple monitoring points to obtain the information inference model.

[0073] S203. Based on the monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit is obtained.

[0074] In this embodiment, after obtaining monitoring information from multiple monitoring points and multiple non-monitoring points, the monitoring information from multiple monitoring points and multiple non-monitoring points is input into the prediction model to predict the settlement depth of the foundation pit.

[0075] In this embodiment, monitoring information at multiple unmonitored points in the foundation pit is first deduced based on monitoring information at multiple actual monitoring points in the foundation pit. Then, the settlement depth of the foundation pit is predicted based on the monitoring information at multiple actual monitoring points and the monitoring information at multiple unmonitored points. Compared with the method of predicting the settlement depth of the foundation pit based solely on the monitoring information at actual monitoring points, this application also uses the monitoring information at multiple unmonitored points in the foundation pit, which greatly improves the accuracy of the prediction of the settlement depth of the foundation pit.

[0076] In one embodiment, in Figure 2 Based on the illustrated embodiment, the process of obtaining the settlement depth at various locations within the foundation pit can be described, such as... Figure 3 As shown, the above-mentioned S203 "based on monitoring information from multiple monitoring points and monitoring information from multiple unmonitored points, the settlement depth at each location in the foundation pit is obtained" includes:

[0077] S301. The monitoring information at each monitoring point and the monitoring information at each unmonitored point are spliced ​​together to obtain the initial point features.

[0078] In this embodiment, after acquiring monitoring information from multiple monitoring points and multiple unmonitored points, the monitoring information from the multiple monitoring points and the multiple unmonitored points can be spliced ​​together to obtain initial point features. Optionally, the monitoring information from multiple monitoring points and the monitoring information from multiple unmonitored points can be spliced ​​together to obtain initial point features.

[0079] Optionally, a method for obtaining initial point features is provided below; see [link to relevant documentation]. Figure 4 The aforementioned S301, "Based on monitoring information from multiple monitoring points and monitoring information from multiple unmonitored points, the settlement depth at each location in the foundation pit is obtained," includes:

[0080] S401. Perform time normalization processing on the time characteristics in each monitoring information to obtain the normalized time characteristics.

[0081] In this embodiment, in order to enhance the model's representation ability and take into account the temporal and spatial correlation of the monitoring data, the node feature matrix is ​​composed of features in multiple dimensions, including the location information of multiple monitoring points and multiple unmonitored points, the settlement depth at multiple historical times, and the settlement depth at the current time.

[0082] For example, any monitoring point Both include settlement depths at multiple historical moments and settlement depths at the current moment, i.e., temporal characteristics. ,in, This refers to the settlement depth at monitoring point (or unmonitored point) i. This refers to the historical moment when monitoring point (or unmonitored point) i is located. Settlement depth information, This refers to the historical moment when monitoring point (or unmonitored point) i is located. Settlement depth information. It should be noted that any monitoring point... It also includes location information. The above. This represents the time series characteristics of monitoring data from monitoring point i over the most recent h days. The value of h needs to be determined manually. For data with long-term time dependence, a larger value should be selected. However, the time dependence of foundation pit settlement monitoring data is relatively weak, so a smaller value can be selected. The dimension 'a' includes not only the location information of the monitored points (or unmonitored points), but also other relevant nearby information. For example, in engineering practice, settlement is usually related to the nearest distance to the edge of the foundation pit, so additional information can be added... middle.

[0083] Optionally, to improve computational efficiency and accuracy, the time features in each monitoring information can be normalized to obtain the normalized time features, as shown in the following formula (1):

[0084]

[0085] in, This represents the maximum absolute value of all historical monitoring data from all monitoring points (or unmonitored points). It refers to the normalized time characteristics of each monitoring point (or unmonitored point).

[0086] S402. Perform position normalization processing on the position features in each monitoring information to obtain the normalized position features.

[0087] In this embodiment, in order to improve computational efficiency and accuracy, the location features in each monitoring information can be normalized to obtain the normalized location features, as shown in the following formula (2):

[0088]

[0089] in, Indicates the location characteristics of monitoring point (or unmonitored point) i; and These represent the maximum and minimum values ​​of each element in the location feature set, respectively. It refers to the normalized location characteristics of each monitoring point (or unmonitored point).

[0090] S403. The normalized time features and the normalized position features are concatenated to obtain the initial point features.

[0091] In this embodiment, after obtaining the normalized time features and the normalized position features, the normalized time features and the normalized position features can be spliced ​​together to obtain the initial point features.

[0092] Optionally, the method for obtaining the initial point features can be further described, such as... Figure 5 As shown, the above S403 "concatenates the normalized time features and the normalized position features to obtain the initial point features" includes:

[0093] S501. The normalized time features and the normalized position features are concatenated to obtain the concatenated features.

[0094] In this embodiment, after obtaining the normalized time features and the normalized position features, the normalized time features and the normalized position features can be spliced ​​together to obtain spliced ​​features.

[0095] Optionally, the normalized time features and the normalized location features are spliced ​​together to obtain the spliced ​​features of each monitoring point (or unmonitored point), as shown in the following formula (3):

[0096]

[0097] in, This refers to the splicing characteristics of each monitoring point (or unmonitored point). This indicates a vector concatenation operation.

[0098] S502. Combine the various splicing features to obtain the initial point features.

[0099] In this embodiment, after obtaining each splicing feature, the splicing features can be combined to obtain the initial point features. Optionally, the splicing features can be combined to form a matrix to obtain the initial node features. See formula (4) below:

[0100]

[0101] It should be noted that after obtaining the initial node features... Then, the initial node features can be... The data is input into a convolutional network to obtain the settlement depth at various locations within the foundation pit.

[0102] This provides a method for obtaining initial point features.

[0103] S302. For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points.

[0104] In this embodiment, after obtaining multiple monitoring points and multiple unmonitored points, the connection weight between any two points can be determined for any two points among the multiple monitoring points and multiple unmonitored points. Based on the obtained connection weight between any two points among the multiple monitoring points and multiple unmonitored points, the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points is obtained.

[0105] Optionally, a graph data structure can be constructed first based on multiple monitored points and multiple unmonitored points, treating these points as nodes in the graph data structure, with the relationships between nodes described by an adjacency matrix. Since the monitored and unmonitored points are not always arranged according to rules, it is difficult to determine the number of edges connecting each point, and the connection strength of each edge is difficult to define. Setting the connection strength of all edges to 1 would result in the loss of valuable similarity information. A more general approach to defining edges is to use a Gaussian kernel function combined with a threshold. This method uses a distance metric (such as Euclidean distance or other metrics) to reflect the similarity between nodes. In this case, the distance between nodes serves as the inverse measure of similarity: the smaller the distance, the more similar the nodes. This aligns with practical intuition in foundation pit engineering. Specifically, the weights between nodes are calculated based on Euclidean distance or other similarity metrics, and the connection weights between any two points from the multiple monitored and unmonitored points are determined. The connection weight between any two points The calculation method is shown in the following formula (5):

[0106]

[0107] in, This represents the Euclidean distance between monitoring point (or unmonitored point) i and j. Let be the standard deviation of the Gaussian kernel function. This is the connection threshold. When... When the threshold is exceeded, the connection weight between the two nodes is set to 0 to reduce computation and improve model efficiency.

[0108] S303. Based on the initial point characteristics and the adjacency matrix, the settlement depth at each location in the foundation pit is obtained.

[0109] In this embodiment, after obtaining the initial point features and the adjacency matrix, the settlement depth at each location in the foundation pit can be obtained based on the initial point features and the adjacency matrix.

[0110] Optionally, the method for obtaining the settlement depth at various locations within the foundation pit can be described; see [link to relevant documentation]. Figure 6 The aforementioned S303, "Based on the initial point characteristics and adjacency matrix, the settlement depth at each location in the foundation pit is obtained," includes:

[0111] S601. Normalize the adjacency matrix to obtain the normalized adjacency matrix.

[0112] In this embodiment, after obtaining the adjacency matrix, the critical matrix can be normalized based on the following formula (6) to obtain the normalized adjacency matrix:

[0113]

[0114] in, This represents the normalized adjacency matrix. This refers to the adjacency matrix obtained by formula (5) above. It refers to the identity matrix.

[0115] S602. Determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix.

[0116] In this embodiment, after obtaining the normalized adjacency matrix, the normalized adjacency matrix can be processed based on the following formula (7) to obtain the degree matrix of the adjacency matrix:

[0117]

[0118] in, The degree matrix representing the adjacency matrix is ​​the same as the degree matrix of the adjacency matrix. The values ​​on the diagonal are equal to the normalized adjacency matrix. The sum of each row.

[0119] S603. Input the degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix into the convolutional network to obtain the settlement depth at each location in the foundation pit.

[0120] In this embodiment, after obtaining the degree matrix, initial point features, and normalized adjacency matrix of the adjacency matrix, the degree matrix, initial point features, and normalized adjacency matrix of the adjacency matrix can be input into the convolutional network to obtain the settlement depth at each location in the foundation pit.

[0121] It should be noted that convolutional networks are shown in the attached diagram. Figure 7As shown, the model consists of three graph convolutional layers. The input to the model includes the degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix. The output includes the predicted settlement values ​​for all monitoring points. By inputting these spatiotemporal multidimensional features—the degree matrix, the initial point matrix, and the normalized adjacency matrix—into the convolutional network, the network's ability to capture settlement trends can be effectively improved.

[0122] Furthermore, the above convolutional network aggregates neighborhood information through graph convolution operations, and the core calculation process is shown in the following formula (8):

[0123]

[0124] in, Represents the node features of the l-th layer. For learnable parameters, The activation function is a non-linear function (such as ReLU). Through multi-layer graph convolution operations, the model can propagate information in the spatial domain and learn deep features between monitoring points, giving it a stronger ability to predict settlement. The vector output by the last graph convolutional layer is the settlement depth at each location in the foundation pit. It should be noted that each location refers to the locations of the monitoring points and the locations of the unmonitored points mentioned above.

[0125] In this embodiment, based on the initial point features at multiple monitoring points and multiple unmonitored points, and the connection weights between any two points among the multiple monitoring points and multiple unmonitored points, an adjacency matrix corresponding to multiple monitoring points and multiple unmonitored points is obtained. Considering the initial point features and the adjacency matrix as two spatiotemporal multidimensional features, the input to the convolutional network is used to obtain the settlement depth at each location in the foundation pit. This can effectively improve the convolutional network's ability to capture settlement trends and ensure the accuracy of determining the settlement depth at each location in the foundation pit.

[0126] In one embodiment, in Figure 2 Based on the illustrated embodiment, the process of acquiring the information inference model can be described, such as... Figure 8 As shown, the above method also includes:

[0127] S701. Mask the sample monitoring information at multiple monitoring points to obtain the masked sample monitoring information at candidate monitoring points.

[0128] In this embodiment, when it is necessary to train the information inference model, the sample monitoring information of a portion of the sample monitoring information at multiple monitoring points can be masked to obtain the masked sample monitoring information at the candidate monitoring points.

[0129] See Figure 9The sample monitoring information at some monitoring points in the training graph is masked to obtain a masked graph, which represents the sample monitoring information at the candidate monitoring points after masking. It should be noted that a specified number of monitoring points can be randomly selected from all monitoring points in the training graph to create the masked graph. In the masked graph, the temporal features of the masked measurement points are replaced with zeros, while the positional features remain unchanged.

[0130] S702. Based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points, the initial information inference model is trained to obtain the information inference model.

[0131] In this embodiment, after obtaining the sample monitoring information at the candidate monitoring points after masking, the initial information inference model is trained based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points to obtain the information inference model.

[0132] Optionally, the masked image is input into the initial information inference model for inference processing to obtain the initial monitoring information at the masked location. Based on the loss difference between the initial monitoring information at the masked location and the sample monitoring information at the masked location, the initial information inference model is trained to obtain the information inference model. During the above training process, the loss is calculated using the mean squared error formula, see formula (9) below:

[0133]

[0134] in, This indicates all the sensors that are blocked. and The measuring points are respectively The actual and predicted values.

[0135] In one embodiment, see Figure 10 It also provides a method for predicting the settlement depth of foundation pits, including:

[0136] T1. Obtain monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0137] T2. Input the monitoring information from multiple monitoring points into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0138] T3. Perform time normalization on the time characteristics of each monitoring information to obtain the normalized time characteristics.

[0139] T4. Perform position normalization processing on the position features in each monitoring information to obtain the normalized position features;

[0140] T5. The normalized time features and the normalized position features are concatenated to obtain the concatenated features.

[0141] T6. Combine the various splicing features to obtain the initial point features;

[0142] T7. For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points;

[0143] T8. Normalize the adjacency matrix to obtain the normalized adjacency matrix;

[0144] T9. Determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix.

[0145] T10. Input the degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix into the convolutional network to obtain the settlement depth at each location in the foundation pit.

[0146] It should be noted that the descriptions of T1-T10 above can be found in the relevant descriptions in the above embodiments, and their effects are similar, so they will not be repeated here.

[0147] In one embodiment, the method for predicting the settlement depth of the foundation pit provided in this application can also be used to predict the settlement depth at various locations throughout the entire foundation pit, generating a settlement cloud map of the entire foundation pit. (See [link to relevant documentation]). Figure 11 The process of generating a full-site settlement cloud map of the foundation pit includes: rasterizing adjacent areas potentially affected by the foundation pit, treating each grid cell as a location to be simulated, and then using the proposed method to generate a settlement cloud map of the foundation pit surface. The final generated settlement cloud map can visually display the settlement trends of different areas of the foundation pit, helping engineers conduct risk assessments and improving construction safety and transparency. Compared with traditional interpolation methods, such as inverse distance weighted interpolation and Kriging interpolation, this method can more accurately reflect settlement patterns under complex geological conditions, improving the accuracy and stability of full-site settlement spatial simulation.

[0148] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0149] Based on the same inventive concept, this application also provides a device for predicting the settlement depth of a foundation pit, which is used to implement the aforementioned method for predicting the settlement depth of a foundation pit. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the device for predicting the settlement depth of a foundation pit provided below can be found in the limitations of the method for predicting the settlement depth of a foundation pit described above, and will not be repeated here.

[0150] In one exemplary embodiment, such as Figure 12 As shown, a device for predicting the settlement depth of a foundation pit is provided, comprising: an acquisition module 10, a deduction module 11, and a determination module 12, wherein:

[0151] The acquisition module 10 is used to acquire monitoring information at multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points.

[0152] The inference module 11 is used to input the monitoring information from multiple monitoring points into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0153] The determination module 12 is used to obtain the settlement depth at each location in the foundation pit based on the monitoring information at multiple monitoring points and the monitoring information at multiple unmonitored points.

[0154] In an exemplary embodiment, the determining module 12 includes: a splicing unit, an acquisition unit, and a determining unit, wherein:

[0155] The splicing unit is specifically used to splice the monitoring information at each monitoring point and the monitoring information at each unmonitored point to obtain the initial point features;

[0156] The acquisition unit is specifically used to determine the connection weight between any two points among multiple monitoring points and multiple unmonitored points, and to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points.

[0157] The unit is determined specifically to obtain the settlement depth at each location in the foundation pit based on the initial point characteristics and the adjacency matrix.

[0158] In an exemplary embodiment, the splicing unit is further configured to perform time normalization processing on the time features in each monitoring information to obtain normalized time features; perform position normalization processing on the position features in each monitoring information to obtain normalized position features; and splice the normalized time features and the normalized position features to obtain initial point features.

[0159] In an exemplary embodiment, the above-mentioned splicing unit is further configured to splice the normalized time features and the normalized position features to obtain spliced ​​features; and to combine the spliced ​​features to obtain initial point features.

[0160] In an exemplary embodiment, the aforementioned determining unit is further configured to normalize the adjacency matrix to obtain a normalized adjacency matrix; determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix; and input the degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix into a convolutional network to obtain the settlement depth at each location in the foundation pit.

[0161] In an exemplary embodiment, the above-described apparatus further includes: a masking module and a training module, wherein:

[0162] The masking module is used to mask the sample monitoring information at multiple monitoring points to obtain the masked sample monitoring information at candidate monitoring points.

[0163] The training module is used to train the initial information inference model based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points, so as to obtain the information inference model.

[0164] Each module in the aforementioned foundation pit settlement depth prediction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0165] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0166] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0167] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0168] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0169] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0170] The monitoring information from each monitoring point and the monitoring information from each unmonitored point are spliced ​​together to obtain the initial point features;

[0171] For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points;

[0172] Based on the initial point characteristics and the adjacency matrix, the settlement depth at each location in the foundation pit is obtained.

[0173] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0174] Time normalization is performed on the time features in each monitoring information to obtain the normalized time features;

[0175] The location features in each monitoring information are normalized to obtain the normalized location features;

[0176] The normalized time features and the normalized position features are concatenated to obtain the initial point features.

[0177] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0178] The normalized time features and the normalized position features are concatenated to obtain the concatenated features.

[0179] The splicing features are combined to obtain the initial point features.

[0180] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0181] The adjacency matrix is ​​normalized to obtain the normalized adjacency matrix;

[0182] Determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix;

[0183] The degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix are input into a convolutional network to obtain the settlement depth at each location in the foundation pit.

[0184] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0185] The sample monitoring information at multiple monitoring points is masked to obtain the masked sample monitoring information at candidate monitoring points.

[0186] Based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points, the initial information inference model is trained to obtain the information inference model.

[0187] 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 following steps are implemented:

[0188] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0189] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0190] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0191] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0192] The monitoring information from each monitoring point and the monitoring information from each unmonitored point are spliced ​​together to obtain the initial point features;

[0193] For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points;

[0194] Based on the initial point characteristics and the adjacency matrix, the settlement depth at each location in the foundation pit is obtained.

[0195] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0196] Time normalization is performed on the time features in each monitoring information to obtain the normalized time features;

[0197] The location features in each monitoring information are normalized to obtain the normalized location features;

[0198] The normalized time features and the normalized position features are concatenated to obtain the initial point features.

[0199] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0200] The normalized time features and the normalized position features are concatenated to obtain the concatenated features.

[0201] The splicing features are combined to obtain the initial point features.

[0202] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0203] The adjacency matrix is ​​normalized to obtain the normalized adjacency matrix;

[0204] Determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix;

[0205] The degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix are input into a convolutional network to obtain the settlement depth at each location in the foundation pit.

[0206] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0207] The sample monitoring information at multiple monitoring points is masked to obtain the masked sample monitoring information at candidate monitoring points.

[0208] Based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points, the initial information inference model is trained to obtain the information inference model.

[0209] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0210] Acquire monitoring information from multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points;

[0211] The monitoring information from multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points.

[0212] Based on monitoring information from multiple monitoring points and multiple non-monitoring points, the settlement depth at each location in the foundation pit was obtained.

[0213] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0214] The monitoring information from each monitoring point and the monitoring information from each unmonitored point are spliced ​​together to obtain the initial point features;

[0215] For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points;

[0216] Based on the initial point characteristics and the adjacency matrix, the settlement depth at each location in the foundation pit is obtained.

[0217] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0218] Time normalization is performed on the time features in each monitoring information to obtain the normalized time features;

[0219] The location features in each monitoring information are normalized to obtain the normalized location features;

[0220] The normalized time features and the normalized position features are concatenated to obtain the initial point features.

[0221] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0222] The normalized time features and the normalized position features are concatenated to obtain the concatenated features.

[0223] The splicing features are combined to obtain the initial point features.

[0224] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0225] The adjacency matrix is ​​normalized to obtain the normalized adjacency matrix;

[0226] Determine the degree matrix of the adjacency matrix based on the normalized adjacency matrix;

[0227] The degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix are input into a convolutional network to obtain the settlement depth at each location in the foundation pit.

[0228] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0229] The sample monitoring information at multiple monitoring points is masked to obtain the masked sample monitoring information at candidate monitoring points.

[0230] Based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at multiple monitoring points, the initial information inference model is trained to obtain the information inference model.

[0231] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0232] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0233] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for predicting the settlement depth of a foundation pit, characterized in that, The method includes: Acquire monitoring information at multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points; The monitoring information from the multiple monitoring points is input into the information inference model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information inference model is trained by a mask of sample monitoring information from multiple monitoring points. Based on the monitoring information at the multiple monitoring points and the monitoring information at the multiple unmonitored points, the settlement depth at each location in the foundation pit is obtained.

2. The method according to claim 1, characterized in that, The settlement depth at each location in the foundation pit is obtained based on the monitoring information from the multiple monitoring points and the monitoring information from the multiple unmonitored points, including: The monitoring information at each of the monitoring points and the monitoring information at each of the unmonitored points are spliced ​​together to obtain the initial point features; For any two points among multiple monitoring points and multiple unmonitored points, determine the connection weight between the two points to obtain the adjacency matrix corresponding to the multiple monitoring points and multiple unmonitored points; Based on the initial point features and the adjacency matrix, the settlement depth at each location in the foundation pit is obtained.

3. The method according to claim 2, characterized in that, The process of concatenating the monitoring information at each of the monitoring points and the monitoring information at each of the unmonitored points to obtain the initial point features includes: The time features in each of the monitoring information are subjected to time normalization processing to obtain the normalized time features; The location features in each of the monitoring information are normalized to obtain the normalized location features. The normalized time features and the normalized position features are concatenated to obtain the initial point features.

4. The method according to claim 3, characterized in that, The step of concatenating the normalized time features and the normalized position features to obtain the initial point features includes: The normalized time features and the normalized position features are concatenated to obtain concatenated features. The initial point features are obtained by combining the splicing features.

5. The method according to claim 2, characterized in that, The step of obtaining the settlement depth at each location in the foundation pit based on the initial point features and the adjacency matrix includes: The adjacency matrix is ​​normalized to obtain a normalized adjacency matrix; Based on the normalized adjacency matrix, determine the degree matrix of the adjacency matrix; The degree matrix of the adjacency matrix, the initial point features, and the normalized adjacency matrix are input into a convolutional network to obtain the settlement depth at each location in the foundation pit.

6. The method according to claim 1, characterized in that, The method further includes: The sample monitoring information at the multiple monitoring points is masked to obtain the masked sample monitoring information at the candidate monitoring points. Based on the sample monitoring information at the candidate monitoring points after masking and the sample monitoring information at the multiple monitoring points, the initial information inference model is trained to obtain the information inference model.

7. A device for predicting the settlement depth of a foundation pit, characterized in that, The device includes: The acquisition module is used to acquire monitoring information at multiple monitoring points in the foundation pit; the monitoring information includes settlement depth at multiple historical times, settlement depth at the current time, and location information of the monitoring points; The deduction module is used to input the monitoring information from the multiple monitoring points into the information deduction model to obtain the monitoring information from multiple unmonitored points in the foundation pit; the information deduction model is trained by a mask of sample monitoring information from multiple monitoring points. The determination module is used to obtain the settlement depth at each location in the foundation pit based on the monitoring information at the multiple monitoring points and the monitoring information at the multiple unmonitored points.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images