Construction method of anti-floating anchor rod, terminal equipment and storage medium
By screening construction points, fitting soil data and identifying soil layers using deep learning technology, the problem of low accuracy in determining the load layer by drilling pressure is solved, and the accurate placement and construction stability of anti-floating anchors are achieved.
Patent Information
- Application Number
- CN202411017698.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-22
AI Technical Summary
During building construction, due to the complex underground soil environment when applying anti-floating anchors, the accuracy of the drilling pressure determination layer is low, which affects the construction effect.
By screening construction points, querying soil data, fitting the elevation map of the holding layer, collecting drilling pressure and vibration signals, using reconstruction networks and classifiers to identify the soil layer, verify the holding layer, and ensuring the accurate placement of the anti-floating anchor rods put.
It improves the accuracy of the judgment and holding layer, ensures effective construction work of anti-floating anchors, and improves the stability and safety of construction.
Smart Images

Figure CN118997137A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application belong to the technical field of deep learning, and in particular, relate to a construction method, terminal equipment, and storage medium for an anti-floating anchor. Background Art
[0002] When constructing some buildings, such as building a low-rise building next to a high-rise building, in addition to driving engineering piles to support the building and provide compressive resistance, anti-floating anchor rods must also be driven in to prevent the building from having an unstable foundation due to the buoyancy underneath.
[0003] During the process of drilling anti-floating anchors, the elevation of the bearing layer will be measured and drilling will be carried out according to the elevation. During drilling, it will be determined whether the bearing layer has been reached based on the pressure.
[0004] However, the underground soil environment is relatively complex, and there are many factors that affect the pressure of the borehole, resulting in low accuracy in determining the bearing layer. Summary of the invention
[0005] In view of this, the embodiments of the present application provide a construction method of an anti-floating anchor, a terminal device and a storage medium, so as to improve the accuracy of determining the bearing layer during the construction of the anti-floating anchor.
[0006] A first aspect of an embodiment of the present application provides a construction method for an anti-floating anchor rod, comprising:
[0007] Screening sampling points for construction engineering piles and anchor points for construction anti-floating anchors at the construction site of the building;
[0008] When the advance drilling operation is completed for the sampling point, querying soil data generated by the advance drilling operation at the sampling point;
[0009] Fitting the elevation map data of the bearing layer of the construction site according to the soil data; recording the elevation data of the bearing layer for the anchor point in the elevation map data;
[0010] Fitting the overall soil type of the soil at the construction site according to the soil data;
[0011] When the drilling machine drills the anchor point according to the elevation data, the drilling pressure and the original vibration signal of the drilling machine are collected;
[0012] Inputting the original vibration signal into a reconstruction network trained for the soil type, converting the original vibration signal into a target vibration signal representing the friction between the drilling machine and the soil in the construction site;
[0013] Inputting the deviation vibration signal into the classifier trained for the soil type to identify the soil layer and confidence; the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal; the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site;
[0014] Verifying that the soil layer is a bearing layer based on the drilling pressure and the confidence level when the soil layer is a bearing layer;
[0015] If the soil layer is successfully verified to be a bearing layer, it is determined that the anti-floating anchor rod is placed in the anchor rod point to carry out construction work.
[0016] A second aspect of the embodiment of the present application provides a construction device for an anti-floating anchor rod, comprising:
[0017] A point determination module is used to select sampling points for construction engineering piles and anchor points for construction anti-floating anchors in the construction site of the building;
[0018] A soil data query module, used for querying soil data generated by the advance drilling operation at the sampling point when the advance drilling operation is completed on the sampling point;
[0019] An elevation map data fitting module is used to fit the elevation map data of the bearing layer of the construction site according to the soil data; the elevation data of the bearing layer is recorded for the anchor point in the elevation map data;
[0020] A soil type fitting module, used for fitting the overall soil type of the soil at the construction site according to the soil data;
[0021] A data acquisition module, used for collecting the drilling pressure and original vibration signal of the drilling machine when the drilling machine drills the anchor point according to the elevation data;
[0022] a vibration signal reconstruction module, configured to input the original vibration signal into a reconstruction network trained for the soil type, and convert the original vibration signal into a target vibration signal representing the friction between the drilling machine and the soil in the construction site;
[0023] A vibration signal classification module, used for inputting the deviation vibration signal into a classifier trained on the soil type to identify the soil layer and confidence; the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal; the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site;
[0024] A bearing layer verification module, used for verifying that the soil layer is a bearing layer according to the drilling pressure and the confidence level when the soil layer is a bearing layer;
[0025] The construction operation determination module is used to determine that the anti-floating anchor rod will be placed in the anchor rod point to carry out the construction operation if the soil layer is successfully verified to be a bearing layer.
[0026] The third aspect of an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the construction method of the anti-floating anchor rod as described in the first aspect above is implemented.
[0027] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the construction method of the anti-floating anchor rod as described in the first aspect above is implemented.
[0028] A fifth aspect of the embodiments of the present application provides a computer program product. When the computer program product is run on a computer, the computer is enabled to execute the construction method of the anti-floating anchor described in the first aspect.
[0029] In this embodiment, sampling points for construction engineering piles and anchor points for construction anti-floating anchors are selected in the construction site of the building; when the advance drilling operation is completed on the sampling points, the soil data generated by the advance drilling operation at the sampling points are queried; the elevation map data of the bearing layer is fitted to the construction site based on the soil data; the elevation data of the bearing layer is recorded for the anchor points in the elevation map data; the soil type of the construction site is fitted as a whole based on the soil data; when the drilling machine drills holes at the anchor points according to the elevation data, the drilling pressure and the original vibration signal of the drilling machine are collected; the original vibration signal is input into the soil type In the reconstruction network trained on the soil type, the original vibration signal is converted into a target vibration signal that characterizes the friction between the drilling machine and the soil in the construction site; the deviation vibration signal is input into the classifier trained on the soil type to identify the soil layer and confidence; the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal; the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site; the soil layer is verified as a bearing layer based on the drilling pressure and the confidence when the soil layer is a bearing layer; if the soil layer is successfully verified as a bearing layer, it is determined to place the anti-floating anchor into the anchor point for construction operations. This embodiment reconstructs and classifies the vibration signal based on the overall soil type of the soil at the construction site, which can weaken the influence of the complex soil environment and obtain more accurate soil layers and confidences, thereby assisting the drilling pressure in verifying the bearing layer, and can effectively improve the accuracy of judging the bearing layer during the anti-floating anchor. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or prior art descriptions. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0031] Figure 1 It is a schematic diagram of a construction method of an anti-floating anchor provided in an embodiment of the present application;
[0032] Figure 2 This is an example diagram of an extended construction area provided in an embodiment of the present application;
[0033] Figure 3 is a schematic diagram of a reconstructed network provided in an embodiment of the present application;
[0034] Figure 4 is a schematic diagram of a classifier provided in an embodiment of the present application;
[0035] Figure 5 It is a schematic diagram of a construction device for an anti-floating anchor provided in an embodiment of the present application;
[0036] Figure 6 It is a schematic diagram of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0037] In the following description, specific details such as specific system structures, technologies, etc. are proposed for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from hindering the description of the present application.
[0038] The technical solution of the present application is described below through specific embodiments.
[0039] Reference Figure 1 , shows a schematic diagram of a construction method of an anti-floating anchor provided in an embodiment of the present application, which may specifically include the following steps:
[0040] Step 101: Select sampling points for construction engineering piles and anchor points for construction anti-floating anchors at a construction site of a building.
[0041] In this embodiment, the designer of the building can set multiple sampling points for construction engineering piles and anchor points for construction anti-floating anchor rods according to the conditions of the construction site of the building, and construct a real-world coordinate system for the construction site, such as a Cartesian coordinate system, in which the sampling points (coordinate points) for construction engineering piles and anchor points (coordinate points) for construction anti-floating anchor rods are screened out.
[0042] To facilitate subsequent operations, the coordinate information of the sampling points and anchor points in this coordinate system can be converted into actual coordinate information such as longitude and latitude.
[0043] Step 102: When the advance drilling operation is completed for the sampling point, query the soil data generated by the advance drilling operation at the sampling point.
[0044] Generally, during the construction of engineering piles, advance drilling operations are performed on the sampling points (i.e., the actual locations in the construction site), and the soil data collected during the advance drilling operations are recorded in a database. During the construction of anti-floating anchor rods, the soil data of each sampling point can be queried from the database.
[0045] Among them, soil data includes longitude and latitude, elevation data of each soil layer (including bearing layer), etc.
[0046] Step 103: Fitting the elevation map data of the bearing layer of the construction site according to the soil data.
[0047] In this embodiment, the soil data of each sampling point can be input into the industrial control platform, and three-dimensional elevation map data representing the bearing layer can be fitted to the construction site.
[0048] In practical applications, the elevation data of the bearing layer for the anchor points are recorded in the elevation map data. Then, the longitude and latitude of each anchor point can be input into the elevation map data to query the elevation data of the bearing layer for each anchor point.
[0049] Step 104: Fit the overall soil type of the soil at the construction site based on the soil data.
[0050] In practical applications, the soil structures in different regions are diverse. In order to improve the processing efficiency, multiple soil types can be abstracted from the soils in different regions as a whole. Then, for the current construction site, the soil data of each sampling point is used to fit the overall soil type of the soil at the construction site.
[0051] In one embodiment of the present application, step 104 may include the following steps:
[0052] Step 1041: Set multiple soil points in the soil data.
[0053] In this embodiment, dense sampling can be performed in the soil data in an equidistant manner to obtain a plurality of soil points, and the elevation data of the plurality of soil points are spaced at fixed lengths.
[0054] Step 1042: Map the soil layers represented by the multiple soil points to specified vectors respectively to obtain an original soil vector sequence.
[0055] In practical applications, the soil layer represented by each soil point can be identified, and the soil layer represented by each soil point can be mapped to a specified vector using a table lookup method, and the vectors of each soil point are combined into a sequence, which is recorded as the original soil vector sequence.
[0056] Step 1043: perform a dimensionality reduction operation on the original soil vector sequence to obtain a target soil vector sequence.
[0057] In practical applications, the depths of advance drilling operations at different construction sites are different, which makes the length of the original soil vector sequence long and of varying lengths. To facilitate calculations, methods such as PCA (Principal Component Analysis) can be used to perform dimensionality reduction operations on the original soil vector sequence to reduce the dimension to a specified length to obtain a target soil vector sequence.
[0058] Step 1044: Calculate the distance between the target soil vector sequence and multiple standard vector sequences.
[0059] In this embodiment, a large number of historical target soil vector sequences can be collected and clustered using K-means (K-means) and other methods to obtain multiple clusters, where the multiple clusters represent multiple point types, and the point type represents the type of soil at a certain point (i.e., a small range).
[0060] The center points in multiple clusters are respectively set as standard vector sequences, and then each standard vector sequence is configured with a point type.
[0061] For the current target soil vector sequence, the distances (such as Euclidean distance, etc.) between the current target soil vector sequence and multiple standard vector sequences are calculated respectively.
[0062] Step 1045: assign the point type corresponding to the standard vector sequence with the shortest distance to the point type of the sampling point corresponding to the soil data.
[0063] The distances between the target soil vector sequence and multiple standard vector sequences (such as Euclidean distance, etc.) are compared, and the current sampling point (represented by the target soil vector sequence of soil data) is divided into the cluster where the standard vector sequence with the shortest distance is located, so that the point type represented by the cluster is assigned to the point type of the sampling point.
[0064] Step 1046: Use the point types of the sampling points to fit a construction map to the construction site.
[0065] In practical applications, the point type of each sampling point is used as a sample, and a two-dimensional construction map can be fitted to the construction site based on the point type of each sampling point, so that the two-dimensional construction map is represented.
[0066] In one embodiment of the present application, step 1046 may further include the following steps:
[0067] Step 10461, grid the construction site network to obtain multiple construction areas.
[0068] In this embodiment, the construction site grid may be gridded on a two-dimensional plane, and the construction site may be divided into a plurality of construction areas, wherein the size, shape (such as a square) and other attributes of the plurality of construction areas are the same.
[0069] Step 10462: Map the sampling points to the construction area to obtain the central area.
[0070] According to the coordinate information of the sampling points, the sampling points are mapped to a certain construction area. At this time, the construction area is recorded as the central area.
[0071] Step 10463, starting from the central area, expand outward layer by layer to assign the expanded construction area to the point type corresponding to the sampling point in the central area.
[0072] Generally speaking, risky areas such as faults have been excluded from construction sites, and the underground soil layers change linearly. Therefore, starting from the central area, we can expand outward layer by layer, and assign the expanded construction area to the point type corresponding to the sampling point in the central area.
[0073] In the specific implementation, the expansion is performed in an iterative manner, and each iteration can be regarded as expanding outward by one layer.
[0074] In each iteration, each construction area is traversed to determine whether there are adjacent blank areas in each positive direction of the filling area.
[0075] Among them, the filled area is a construction area with a point type, the filled area is initially the center area, the blank area is a construction area without a point type, and the positive directions are up, down, left and right.
[0076] If there is a blank area, a bucket is created for the blank area. A bucket is a structure for storing data in the cache.
[0077] For the same blank area, there may be one or more adjacent filled areas. In this case, the point types of all adjacent filled areas are written into the bucket.
[0078] A point type is randomly selected in the bucket and assigned to the blank area to simulate the linear change of soil layers and the existence of random changes.
[0079] If there are no blank areas, it means that all building areas have been configured with point types, and you can determine that the expansion is complete.
[0080] For example, Figure 2 For the partial building area shown, in the first iteration, there is a blank area A2 above the filled area (i.e., the central area) A0, a blank area A7 below, a blank area A4 on the left, and a blank area A5 on the right; there is a blank area B2 above the filled area (i.e., the central area) B0, a blank area B7 below, a blank area B4 on the left, and a blank area B5 on the right.
[0081] Since all blank areas at this time have an adjacent filled area, the point type of the filled area (i.e., the central area) A0 is assigned to the blank areas A2, A7, A4, and A5, turning them into filled areas; the class type of the filled area (i.e., the central area) B0 is assigned to the blank areas B2, B7, B4, and B5, turning them into filled areas.
[0082] In the second iteration, there is a blank area A3 on the right side of the filled area A2, there is a blank area A3 above the filled area A5, and there is a blank area A3 on the left side of the filled area B4. At this time, the bucket created for the blank area A3 can store the point type of the filled area A2, the point type of the filled area A5, and the point type of the filled area B4. A point type is randomly selected from the bucket and assigned to the blank area A3. At this time, the point type of the blank area A3 may be the same as that of the central area A0 or the same as that of the central area B0.
[0083] Step 10464: When the expansion is completed, each construction area and point type is output as a construction map of the construction site.
[0084] When the expansion is complete, the various construction areas and their point types can be organized into an overall structure and output as a construction map of the construction site.
[0085] Step 1047: Calculate the similarity between the construction map and multiple standard maps.
[0086] In this embodiment, a large number of historical construction maps can be collected, and the construction maps under the same soil type can be processed using statistical methods (such as histograms, linear superposition, etc.) to generate a representative standard map. Then, each standard map is configured with a soil type.
[0087] For the current construction map, the similarity between the current construction map and multiple standard maps can be calculated using methods such as histogram, cosine angle, hash algorithm, etc.
[0088] Step 1048: Assign the soil type corresponding to the standard map with the highest similarity as the overall soil type of the soil at the construction site.
[0089] The similarity between the current construction map and multiple standard maps is compared, and the standard map with the highest similarity is screened out, so that the soil type corresponding to the standard map with the highest similarity is assigned as the overall soil type of the soil at the construction site.
[0090] Step 105: When the drilling machine drills holes at the anchor points according to the elevation data, the drilling pressure and original vibration signal of the drilling machine are collected.
[0091] In this embodiment, a pressure sensor and a vibration sensor can be set on the drill bit of the drilling machine. When the drilling machine is drilling the anchor point according to the elevation data of the bearing layer in the anchor point, the pressure sensor can be called in real time to collect the drilling pressure of the drilling machine and the vibration sensor can be called to collect the original vibration signal of the drilling machine.
[0092] Step 106: Input the original vibration signal into the reconstruction network trained for soil types, and convert the original vibration signal into a target vibration signal that represents the friction between the drilling machine and the soil in the construction site.
[0093] In actual application, there are certain differences between the soil structures of different soil types, and there are certain differences in the vibrations caused by friction of the drilling machine in different soil structures. If an anchor drilling rig (anchor drilling rig) is used for dry drilling, although artificial watering is used at the construction site, it only serves for earth excavation and slope support, and the water level is only reduced to 30cm below the bottom of the pit. The local thick sand layer is due to the high groundwater level, and the self-stabilizing ability of the hole wall is weakened, which is easy to cause hole collapse. Moreover, dry drilling does not have the process of washing the hole, so it is difficult to ensure the normal functioning of friction resistance.
[0094] Therefore, the reconstruction network can be trained in advance for each soil type, that is, one soil type corresponds to one reconstruction network.
[0095] The reconstruction networks corresponding to various soil types have the same structure but different parameters. Therefore, when determining the soil type, the reconstruction network of the general structure can be loaded, and the parameters corresponding to the soil type can be loaded into the reconstruction network for operation, thereby improving the efficiency of managing the reconstruction network.
[0096] Furthermore, the reconstruction network can perform a reconstruction operation on the original vibration signal. During the reconstruction operation, the interfering signals are filtered out, the effective signals are enhanced, and the original vibration signal is converted into a relatively pure target vibration signal that characterizes the friction between the drilling machine and the soil in the construction site.
[0097] In one embodiment of the present application, Figure 3 As shown, the reconstructed network for soil type training includes a first long short-term memory network LSTM_1, a second long short-term memory network LSTM_2, a third long short-term memory network LSTM_3, a first dynamic convolution module DCM_1, a second dynamic convolution module DCM_2, a first deconvolution layer Deconv_1, a second deconvolution layer Deconv_2, a spatial feature module (Spatial Feature Module, SFM) and a first fully connected layer FC_1.
[0098] Among them, the first long short-term memory network LSTM_1, the second long short-term memory network LSTM_2 and the third long short-term memory network LSTM_3 all belong to the long short-term memory network (LSTM). LSTM can be divided into the following key variables:
[0099] Input: h t-1 (the hidden layer at time t-1) and x t (eigenvector at time t);
[0100] Output: h t (Adding activation functions such as softmax can be used as the real output, otherwise it is used as a hidden layer);
[0101] Memory: c t-1 and c t .
[0102] Then, there are three main stages inside LSTM:
[0103] 1. Forgetting stage
[0104] The forgetting stage is mainly to selectively forget the input from the previous node, which is calculated by z f (f stands for forget) as the forget gate to control the c of the previous state t-1 Which ones to keep and which ones to forget.
[0105] 2. Select the memory stage
[0106] The selective memory stage selectively "memorizes" the input, mainly the input x tSelect memory, focus on recording the important ones and remember less of the unimportant ones. The current input content is represented by the z calculated previously. The selected gating signal is represented by z i (i stands for information) to control.
[0107] Add the results of the above two stages to get c transmitted to the next state t .
[0108] 3. Output stage
[0109] The output stage will decide what will be considered as the output of the current state. Mainly through z o To control. And also the c obtained in the previous stage o The image is scaled (transformed via a tanh activation function).
[0110] The first dynamic convolution module DCM_1 and the second dynamic convolution module DCM_2 are both dynamic convolution modules (Dynamic Convolutional Module, DCM). DCM has two branches, one of which is responsible for compressing the feature channel (reducing the amount of calculation), and the other branch obtains the weight of the depthwise convolution through adaptive pooling, and then the output of the two branches is depthwise convolution. Because the weight is dynamically generated, it is more adaptable to different inputs and can adaptively change its own parameters.
[0111] The first deconvolution layer Deconv_1 and the second deconvolution layer Deconv_2 are both deconvolution layers (also called transposed convolution layers), which can provide deconvolution operations and adjust the size of features. They restore the features obtained through the deconvolution operation back to the size and details of the original input.
[0112] The spatial feature module FSM can use the mountain and context information to generate convolution kernels of different sizes and adaptively learn the semantic information of the features.
[0113] The first fully connected layer FC_1 belongs to the fully connected layer. Each of its nodes is connected to all the nodes in the previous layer, which can capture all the previous features. After calculation, all the features can be more effectively and completely integrated to obtain a global feature.
[0114] Then, in this embodiment, step 106 may include the following steps:
[0115] Step 10611: Input the original vibration signal into the first long short-term memory network to extract the first trunk vibration feature.
[0116] The original vibration signal is input into the first long short-term memory network LSTM_1, and the first long short-term memory network LSTM_1 extracts the time series features from the original vibration signal, which are recorded as the first trunk vibration features.
[0117] Step 10612: input the first trunk vibration feature into the first dynamic convolution module to extract the second trunk vibration feature.
[0118] The first trunk vibration feature is input into the first dynamic convolution module DCM_1, and the first dynamic convolution module DCM_1 extracts spatial features from the first trunk vibration feature, which are recorded as second trunk vibration features.
[0119] Step 10613: Input the second trunk vibration feature into the spatial feature module, and apply the attention mechanism at multiple scales to extract the third trunk vibration feature.
[0120] When the drilling machine rubs against different soils, signals of different frequency bands are generated. The energy strength of these signals is different and the range is wide. The second backbone vibration feature is input into the spatial feature module DCM. The spatial feature module DCM applies the attention mechanism from multiple parallel scales to extract features from the second backbone vibration feature, which is recorded as the third backbone vibration feature, so that the third backbone vibration feature can cover signals of different frequency bands and generate higher quality features.
[0121] Among them, multiple parallel scales enable the space to obtain a larger receptive field and extract more contextual information.
[0122] The attention mechanism can make the spatial feature module DCM pay more attention to the effective signals of vibration caused by friction, thereby improving the accuracy and robustness of feature expression.
[0123] In a structure such as Figure 3 As shown, the spatial feature module DCM includes a first convolutional layer Conv_1, a second convolutional layer Conv_2, a third convolutional layer Conv_3, a fourth convolutional layer Conv_4, a fourth long short-term memory network LSTM_4, a fifth long short-term memory network LSTM_5, a first multi-head attention layer MHA_1 and a second multi-head attention layer MHA_2.
[0124] Among them, the first convolution layer Conv_1, the second convolution layer Conv_2, the third convolution layer Conv_3 and the fourth convolution layer Conv_4 are all convolution layers, which can provide convolution operations and extract shallow features.
[0125] The fourth long short-term memory network LSTM_4 and the fifth long short-term memory network LSTM_5 both belong to long short-term memory networks.
[0126] The first multi-head attention layer MHA_1 and the second multi-head attention layer MHA_2 both belong to the multi-head attention mechanism (Multi-Head Attention, MHA). In the multi-head attention mechanism, multiple sets of attention weights can be used, each set of weights can learn different semantic information, and each set of weights will generate a context vector. Finally, these context vectors will be concatenated and then a linear transformation will be performed to obtain the final output.
[0127] In this structure, the spatial feature module DCM can be divided into four branches:
[0128] In the first branch, the second trunk vibration feature is input into the first convolution layer Conv_1 to perform a convolution operation and extract the first branch vibration feature.
[0129] In the second branch, the second trunk vibration feature is input into the second convolutional layer Conv_2 to perform a convolution operation and extract the second branch vibration feature.
[0130] In the third branch, the second trunk vibration feature is input into the fourth long short-term memory network LSTM_4 to extract the time series feature, which is recorded as the third branch vibration feature.
[0131] The vibration feature of the third branch is input into the first multi-head attention layer MHA_1 and the multi-head attention mechanism is applied to extract the feature, which is recorded as the vibration feature of the fourth branch.
[0132] In the fourth branch, the second trunk vibration feature is input into the fifth long short-term memory network LSTM_5 to extract the time series features, which are recorded as the fifth branch vibration features.
[0133] The fifth branch vibration feature is input into the second multi-head attention layer MHA_2 and the multi-head attention mechanism is applied to extract the feature, which is recorded as the sixth branch vibration feature.
[0134] Use functions such as Concat to merge the fourth branch vibration feature and the sixth branch vibration feature into the seventh branch vibration feature.
[0135] The seventh branch vibration feature is input into the third convolutional layer Conv_3 to perform a convolution operation and extract the eighth branch vibration feature.
[0136] Use functions such as Concat to merge the first branch vibration feature, the second branch vibration feature and the eighth branch vibration feature into the ninth branch vibration feature.
[0137] The ninth branch vibration feature is input into the fourth convolutional layer Conv_4 to perform a convolution operation and extract the third trunk vibration feature.
[0138] Step 10614: Input the third trunk vibration feature into the second dynamic convolution module to extract the fourth trunk vibration feature.
[0139] The third trunk vibration feature is input into the second dynamic convolution module DCM_2, and the second dynamic convolution module DCM_2 extracts spatial features from the third trunk vibration feature, which is recorded as the fourth trunk vibration feature.
[0140] Step 10615: Input the second trunk vibration feature into the first deconvolution layer to extract the fifth trunk vibration feature.
[0141] The second main-trunk vibration feature is input into the first deconvolution layer Deconv_1, and the first deconvolution layer Deconv_1 performs a deconvolution operation on the second main-trunk vibration feature to extract the fifth main-trunk vibration feature.
[0142] Step 10616: Input the fifth trunk vibration feature into the second long short-term memory network to extract the sixth trunk vibration feature.
[0143] The fifth trunk vibration feature is input into the second long short-term memory network LSTM_2. The second long short-term memory network LSTM_2 extracts the time series feature from the fifth trunk vibration feature and records it as the sixth trunk vibration feature.
[0144] Step 10617: merge the fourth main-trunk vibration feature and the sixth main-trunk vibration feature into a seventh main-trunk vibration feature.
[0145] Use the Concat function to merge the fourth main-trunk vibration feature and the sixth main-trunk vibration feature into the seventh main-trunk vibration feature.
[0146] Step 10618: Input the seventh main-trunk vibration feature into the second deconvolution layer to extract the eighth main-trunk vibration feature.
[0147] The seventh main-trunk vibration feature is input into the second deconvolution layer Deconv_2, and the second deconvolution layer Deconv_2 performs a deconvolution operation on the seventh main-trunk vibration feature to extract the eighth main-trunk vibration feature.
[0148] Step 10619: Input the eighth trunk vibration feature into the third long short-term memory network to extract the ninth trunk vibration feature.
[0149] The eighth trunk vibration feature is input into the third long short-term memory network LSTM_3. The third long short-term memory network LSTM_3 extracts the time series feature from the eighth trunk vibration feature and records it as the ninth trunk vibration feature.
[0150] Step 10620: Input the ninth trunk vibration feature into the first fully connected layer and map it into a target vibration signal representing the friction between the drilling machine and the soil in the construction site.
[0151] The ninth main trunk vibration feature is input into the first fully connected layer FC_1, and the first fully connected layer FC_1 maps the ninth main trunk vibration feature into a target vibration signal representing the friction between the drilling machine and the soil in the construction site.
[0152] Step 107: Input the deviation vibration signal into a classifier trained on soil types to identify soil layers and confidence levels.
[0153] In practical applications, there are certain differences between the soil structures of different soil types, and there are certain differences in the vibrations generated by the drilling machine in different soil structures. Therefore, classifiers can be trained in advance for each soil type, that is, one soil type corresponds to one classifier, and each classifier can perform multiple classifications of the soil to obtain multiple soil layers and the confidence of the soil layer.
[0154] The classifiers corresponding to each soil type have the same structure but different parameters. Therefore, when determining the soil type, a classifier of a general structure can be loaded, and the parameters corresponding to the soil type can be loaded into the classifier for operation, thereby improving the efficiency of managing the classifier.
[0155] Furthermore, the reconstruction network and the classifier can form a cascade model. When training the reconstruction network and the classifier, the reconstruction network and the classifier can be regarded as a whole and trained simultaneously. Alternatively, the reconstruction network can be trained first, and when the training of the reconstruction network is completed, the classifier can be trained using the signal output by the reconstruction network. This embodiment does not impose any restrictions on this.
[0156] In this embodiment, a baseline vibration signal can be constructed based on the vibration signals generated by the friction between the drilling machine and different soil types in history, wherein the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site.
[0157] Taking the baseline vibration signal as a reference, the target vibration signal is subtracted from the baseline vibration signal to obtain a deviation vibration signal, that is, the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal.
[0158] The deviation vibration signal is input into the corresponding classifier, and the classifier can identify the soil layer reached by the current drilling machine and the confidence level.
[0159] In a structure such as Figure 4 As shown, the classifier trained for soil type includes the fifth convolution layer Conv_5, the sixth convolution layer Conv_6, the seventh convolution layer Conv_7, the eighth convolution layer Conv_8, the sixth long short-term memory network LSTM_6, the pooling layer Pooling and the second fully connected layer FC_2.
[0160] Among them, the fifth convolutional layer Conv_5, the sixth convolutional layer Conv_6, the seventh convolutional layer Conv_7 and the eighth convolutional layer Conv_8 all belong to convolutional layers, the sixth long short-term memory network LSTM_6 belongs to the long short-term memory network, the pooling layer Pooling can provide pooling operations, such as maximum pooling operations, average pooling operations, etc., and the second fully connected layer FC_2 belongs to the fully connected layer.
[0161] In this structure, the deviation vibration signal is input into the fifth convolutional layer Conv_5 to perform a convolution operation and extract the first head vibration feature.
[0162] The first head vibration feature is input into the sixth convolutional layer Conv_6 to perform a convolution operation, and the second head vibration feature is extracted.
[0163] The second head vibration feature is input into the sixth long short-term memory network LSTM_6 to extract the time series feature, which is recorded as the third head vibration feature.
[0164] The third head vibration feature is input into the seventh convolutional layer Conv_7 to perform a convolution operation, and the fourth head vibration feature is extracted.
[0165] The fourth head vibration feature is input into the pooling layer to perform a pooling operation, and the fifth head vibration feature is extracted.
[0166] Use the Concat function to merge the second head vibration feature with the fifth head vibration feature into the sixth head vibration feature.
[0167] The sixth head vibration feature is input into the eighth convolutional layer Conv_8 to perform a convolution operation, and the seventh head vibration feature is extracted.
[0168] The seventh head vibration feature is input into the second fully connected layer FC_2 and mapped into soil layer and confidence.
[0169] Step 108: Verify that the soil layer is a bearing layer based on the drilling pressure and the confidence level when the soil layer is a bearing layer.
[0170] In this embodiment, the bearing layer is screened out from each soil layer, and the soil layer is verified to be the bearing layer by combining the drilling pressure and the confidence level when the soil layer is the bearing layer, that is, it is verified whether the soil layer currently reached by the drilling machine is the bearing layer.
[0171] In a specific implementation, the drilling pressure can be mapped to the probability that the soil layer is a bearing layer by looking up a table. The table records the mapping relationship between the drilling pressure and the probability that the soil layer is a bearing layer, which is calculated based on the drilling pressure recorded during historical drilling and the actual verification of whether it is a bearing layer.
[0172] Substitute the probability and the confidence level when the soil layer is the bearing layer into the following formula to calculate the comprehensive value:
[0173]
[0174] Among them, Score is the comprehensive value, Probability is the probability, Confidence is the confidence, Pexample is the first benchmark value set for the probability under the soil type, Cexample is the second benchmark value set for the confidence under the soil type, and α, β and γ are all weights.
[0175] The first reference value and the second reference value can be obtained by looking up a table, which has an adjustment effect.
[0176] The drilling pressure is compared with a preset first threshold value, and the integrated value is compared with a preset second threshold value.
[0177] If the drilling pressure is greater than the first threshold value, and the comprehensive value is greater than the second threshold value, it is determined that the verification of the soil layer as the bearing layer is successful.
[0178] If the borehole pressure is less than or equal to the first threshold value, and the comprehensive value is less than or equal to the second threshold value, it is determined that the verification of the soil layer as the bearing layer has failed.
[0179] Step 109: If the soil layer is successfully verified as a bearing layer, it is determined that an anti-floating anchor rod is placed in the anchor rod point to carry out construction work.
[0180] If the soil layer is successfully verified as a bearing layer, a prompt message may be generated to prompt construction personnel to place anti-floating anchors into anchor points for pressure grouting, anchor acceptance and other construction operations.
[0181] In this embodiment, sampling points for construction engineering piles and anchor points for construction anti-floating anchors are selected in the construction site of the building; when the advance drilling operation is completed on the sampling points, the soil data generated by the advance drilling operation at the sampling points are queried; the elevation map data of the bearing layer is fitted to the construction site based on the soil data; the elevation data of the bearing layer is recorded for the anchor points in the elevation map data; the soil type of the construction site is fitted as a whole based on the soil data; when the drilling machine drills holes at the anchor points according to the elevation data, the drilling pressure and the original vibration signal of the drilling machine are collected; the original vibration signal is input into the soil type In the reconstruction network trained on the soil type, the original vibration signal is converted into a target vibration signal that characterizes the friction between the drilling machine and the soil in the construction site; the deviation vibration signal is input into the classifier trained on the soil type to identify the soil layer and confidence; the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal; the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site; the soil layer is verified as a bearing layer based on the drilling pressure and the confidence when the soil layer is a bearing layer; if the soil layer is successfully verified as a bearing layer, it is determined to place the anti-floating anchor into the anchor point for construction operations. This embodiment reconstructs and classifies the vibration signal based on the overall soil type of the soil at the construction site, which can weaken the influence of the complex soil environment and obtain more accurate soil layers and confidences, thereby assisting the drilling pressure in verifying the bearing layer, and can effectively improve the accuracy of judging the bearing layer during the anti-floating anchor.
[0182] It should be noted 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.
[0183] Reference Figure 5 , shows a schematic diagram of a construction device for an anti-floating anchor provided in an embodiment of the present application, which may specifically include the following modules:
[0184] Point determination module 501, for selecting sampling points for construction engineering piles and anchor points for construction anti-floating anchors in the construction site of the building;
[0185] A soil data query module 502, for querying soil data generated by the advance drilling operation at the sampling point when the advance drilling operation is completed on the sampling point;
[0186] The elevation map data fitting module 503 is used to fit the elevation map data of the bearing layer of the construction site according to the soil data; the elevation map data records the elevation data of the bearing layer for the anchor point;
[0187] A soil type fitting module 504 is used to fit the overall soil type of the soil at the construction site according to the soil data;
[0188] The data acquisition module 505 is used to collect the drilling pressure and original vibration signal of the drilling machine when the drilling machine drills the anchor point according to the elevation data;
[0189] A vibration signal reconstruction module 506 is used to input the original vibration signal into a reconstruction network trained for the soil type, and convert the original vibration signal into a target vibration signal representing the friction between the drilling machine and the soil in the construction site;
[0190] The vibration signal classification module 507 is used to input the deviation vibration signal into the classifier trained for the soil type to identify the soil layer and confidence; the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal; the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site;
[0191] A bearing layer verification module 508, configured to verify that the soil layer is a bearing layer according to the drilling pressure and the confidence level when the soil layer is a bearing layer;
[0192] The construction operation determination module 509 is used to determine that the anti-floating anchor rod is placed in the anchor rod point to perform the construction operation if the soil layer is successfully verified to be a bearing layer.
[0193] In one embodiment of the present application, the soil type fitting module 504 is further used for:
[0194] Setting a plurality of soil points in the soil data;
[0195] Mapping the soil layers represented by the plurality of soil points into designated vectors respectively to obtain an original soil vector sequence;
[0196] Performing a dimensionality reduction operation on the original soil vector sequence to obtain a target soil vector sequence;
[0197] Calculating the distance between the target soil vector sequence and a plurality of standard vector sequences; each of the standard vector sequences is configured with a point type;
[0198] Assigning the point type corresponding to the standard vector sequence with the shortest distance to the point type of the sampling point corresponding to the soil data;
[0199] Fitting a construction map to the construction site using the point types of the sampling points;
[0200] Calculating the similarity between the construction map and a plurality of standard maps; each of the standard maps is configured with a soil type;
[0201] The soil type corresponding to the standard map with the highest similarity is assigned as the overall soil type of the soil at the construction site.
[0202] In one embodiment of the present application, the soil type fitting module 504 is further used to:
[0203] Gridding the construction site network to obtain multiple construction areas;
[0204] Mapping the sampling points to the construction area to obtain a central area;
[0205] Expand outward layer by layer from the central area, so as to assign the expanded construction area to the point type corresponding to the sampling point in the central area;
[0206] When the expansion is completed, each of the construction areas and the point types are output as a construction map of the construction site.
[0207] In one embodiment of the present application, the soil type fitting module 504 is further used to:
[0208] Traversing each of the construction areas, determining whether there is an adjacent blank area in each positive direction of the filling area; the filling area is the construction area with a point type, the filling area is initially the central area, the blank area is the construction area without a point type, and the positive directions are up, down, left and right;
[0209] If the blank area exists, create a bucket for the blank area;
[0210] Writing the point type of the fill area into the bucket;
[0211] Randomly select a point type in the bucket and assign it to the blank area;
[0212] If the blank area does not exist, it is determined that the expansion is completed.
[0213] In one embodiment of the present application, the reconstruction network trained for the soil type includes a first long short-term memory network, a second long short-term memory network, a third long short-term memory network, a first dynamic convolution module, a second dynamic convolution module, a first deconvolution layer, a second deconvolution layer, a spatial feature module and a first fully connected layer;
[0214] The vibration signal reconstruction module 506 is also used for:
[0215] Inputting the original vibration signal into the first long short-term memory network to extract the first trunk vibration feature;
[0216] Inputting the first trunk vibration feature into the first dynamic convolution module to extract the second trunk vibration feature;
[0217] Inputting the second trunk vibration feature into the spatial feature module, and applying an attention mechanism at multiple scales to extract the third trunk vibration feature;
[0218] Inputting the third trunk vibration feature into the second dynamic convolution module to extract the fourth trunk vibration feature;
[0219] Inputting the second main-trunk vibration feature into the first deconvolution layer to extract the fifth main-trunk vibration feature;
[0220] Inputting the fifth trunk vibration feature into the second long short-term memory network to extract the sixth trunk vibration feature;
[0221] fusing the fourth trunk vibration feature and the sixth trunk vibration feature into a seventh trunk vibration feature;
[0222] Inputting the seventh main-trunk vibration feature into the second deconvolution layer to extract the eighth main-trunk vibration feature;
[0223] Inputting the eighth trunk vibration feature into the third long short-term memory network to extract the ninth trunk vibration feature;
[0224] The ninth trunk vibration feature is input into the first fully connected layer and mapped into a target vibration signal representing the friction between the drilling machine and the soil in the construction site.
[0225] In one embodiment of the present application, the spatial feature module includes a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, a fourth long short-term memory network, a fifth long short-term memory network, a first multi-head attention layer and a second multi-head attention layer;
[0226] The vibration signal reconstruction module 506 is also used for:
[0227] Inputting the second trunk vibration feature into the first convolutional layer to extract the first branch vibration feature;
[0228] Inputting the second trunk vibration feature into the second convolutional layer to extract the second branch vibration feature;
[0229] Inputting the second trunk vibration feature into the fourth long short-term memory network to extract the third branch vibration feature;
[0230] Inputting the third branch vibration feature into the first multi-head attention layer to extract the fourth branch vibration feature;
[0231] Inputting the second trunk vibration feature into the fifth long short-term memory network to extract the fifth branch vibration feature;
[0232] Inputting the fifth branch vibration feature into the second multi-head attention layer to extract the sixth branch vibration feature;
[0233] Merging the fourth branch vibration feature and the sixth branch vibration feature into a seventh branch vibration feature;
[0234] Inputting the seventh branch vibration feature into the third convolutional layer to extract the eighth branch vibration feature;
[0235] Merging the first branch vibration feature, the second branch vibration feature and the eighth branch vibration feature into a ninth branch vibration feature;
[0236] The ninth branch vibration feature is input into the fourth convolutional layer to extract the third trunk vibration feature.
[0237] In one embodiment of the present application, the classifier trained for the soil type includes a fifth convolutional layer, a sixth convolutional layer, a seventh convolutional layer, an eighth convolutional layer, a sixth long short-term memory network, a pooling layer and a second fully connected layer;
[0238] The vibration signal classification module 507 is also used for:
[0239] Inputting the deviation vibration signal into the fifth convolutional layer to extract the first head vibration feature;
[0240] Inputting the first head vibration feature into the sixth convolutional layer to extract the second head vibration feature;
[0241] Inputting the second head vibration feature into the sixth long short-term memory network to extract the third head vibration feature;
[0242] Inputting the third head vibration feature into the seventh convolutional layer to extract the fourth head vibration feature;
[0243] Inputting the fourth head vibration feature into the pooling layer to extract the fifth head vibration feature;
[0244] fusing the second head vibration feature and the fifth head vibration feature into a sixth head vibration feature;
[0245] Inputting the sixth head vibration feature into the eighth convolutional layer to extract the seventh head vibration feature;
[0246] The seventh head vibration feature is input into the second fully connected layer and mapped into a soil layer and confidence.
[0247] In one embodiment of the present application, the bearing layer verification module 508 is further used to:
[0248] Mapping the borehole pressure to a probability that the soil layer is a bearing layer;
[0249] Substitute the probability and the confidence level when the soil layer is the bearing layer into the following formula to calculate the comprehensive value:
[0250]
[0251] Wherein, Score is the comprehensive value, Probability is the probability, CConfidence is the confidence, Pexample is the first reference value set for the probability under the soil type, CCexample is the second reference value set for the confidence under the soil type, and α, β and γ are all weights;
[0252] If the drilling pressure is greater than a first threshold value and the comprehensive value is greater than a second threshold value, it is determined that the soil layer is successfully verified as a bearing layer.
[0253] An embodiment of the present application provides a construction device for an anti-floating anchor rod. By using the construction device for an anti-floating anchor rod, each step in the aforementioned method embodiments can be implemented.
[0254] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment part.
[0255] Reference Figure 6 , shows a schematic diagram of a terminal device provided by an embodiment of the present application. Figure 6 As shown, the terminal device 400 in the embodiment of the present application includes: a processor 410, a memory 420, and a computer program 421 stored in the memory 420 and executable on the processor 410. When the processor 410 executes the computer program 421, the steps in each embodiment of the above-mentioned anti-floating anchor construction method are implemented, such as Figure 1 Alternatively, when the processor 410 executes the computer program 421, the functions of each module / unit in the above-mentioned device embodiments are realized, for example Figure 5 Functions of modules 501 to 509 are shown.
[0256] Exemplarily, the computer program 421 may be divided into one or more modules / units, which are stored in the memory 420 and executed by the processor 410 to complete the present application. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, which may be used to describe the execution process of the computer program 421 in the terminal device 400.
[0257] The terminal device 400 may include, but is not limited to, a processor 410 and a memory 420. Those skilled in the art will appreciate that Figure 6 It is only an example of the terminal device 400 and does not constitute a limitation on the terminal device 400. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device 400 may also include input and output devices, network access devices, buses, etc.
[0258] The processor 410 may be a central processing unit (CPU), or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
[0259] The memory 420 may be an internal storage unit of the terminal device 400, such as a hard disk or memory of the terminal device 400. The memory 420 may also be an external storage device of the terminal device 400, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 400. Further, the memory 420 may also include both an internal storage unit of the terminal device 400 and an external storage device. The memory 420 is used to store the computer program 421 and other programs and data required by the terminal device 400. The memory 420 may also be used to temporarily store data that has been output or is to be output.
[0260] An embodiment of the present application also discloses a terminal device, including 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 construction method of the anti-floating anchor rod as described in the aforementioned embodiments is implemented.
[0261] The embodiment of the present application further discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the construction method of the anti-floating anchor rod as described in the above-mentioned embodiments is implemented.
[0262] The embodiment of the present application further discloses a computer program product. When the computer program product is run on a computer, the computer is enabled to execute the construction method of the anti-floating anchor described in the above-mentioned embodiments.
[0263] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application is described in detail with reference to the above-mentioned embodiments, a person skilled in the art should understand that the technical solutions described in the above-mentioned embodiments can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.
Claims
1. A construction method for an anti-floating anchor rod, characterized in that: include: Screening sampling points for construction engineering piles and anchor points for construction anti-floating anchors at the construction site of the building; When the advance drilling operation is completed for the sampling point, querying soil data generated by the advance drilling operation at the sampling point; Fitting the elevation map data of the bearing layer of the construction site according to the soil data; recording the elevation data of the bearing layer for the anchor point in the elevation map data; Fitting the overall soil type of the soil at the construction site according to the soil data; When the drilling machine drills the anchor point according to the elevation data, the drilling pressure and the original vibration signal of the drilling machine are collected; Inputting the original vibration signal into a reconstruction network trained for the soil type, converting the original vibration signal into a target vibration signal representing the friction between the drilling machine and the soil in the construction site; Inputting the deviation vibration signal into the classifier trained for the soil type to identify the soil layer and confidence; the deviation vibration signal is the difference between the target vibration signal and the baseline vibration signal; the baseline vibration signal is a vibration signal that fits the soil type and characterizes the friction between the drilling machine and the soil in the construction site; Verifying that the soil layer is a bearing layer based on the drilling pressure and the confidence level when the soil layer is a bearing layer; If the soil layer is successfully verified to be a bearing layer, it is determined that the anti-floating anchor rod is placed in the anchor rod point to carry out construction work.
2. The method according to claim 1, characterized in that: The step of fitting the overall soil type of the soil at the construction site based on the soil data includes: Setting a plurality of soil points in the soil data; Mapping the soil layers represented by the plurality of soil points into designated vectors respectively to obtain an original soil vector sequence; Performing a dimensionality reduction operation on the original soil vector sequence to obtain a target soil vector sequence; Calculating the distance between the target soil vector sequence and a plurality of standard vector sequences; each of the standard vector sequences is configured with a point type; Assigning the point type corresponding to the standard vector sequence with the shortest distance to the point type of the sampling point corresponding to the soil data; Fitting a construction map to the construction site using the point types of the sampling points; Calculating the similarity between the construction map and a plurality of standard maps; each of the standard maps is configured with a soil type; The soil type corresponding to the standard map with the highest similarity is assigned as the overall soil type of the soil at the construction site.
3. The method according to claim 2, characterized in that The step of fitting a construction map to the construction site using the point type of the sampling point includes: Gridding the construction site network to obtain multiple construction areas; Mapping the sampling points to the construction area to obtain a central area; Expand outward layer by layer from the central area, so as to assign the expanded construction area to the point type corresponding to the sampling point in the central area; When the expansion is completed, each of the construction areas and the point types are output as a construction map of the construction site.
4. The method according to claim 3, characterized in that The step of expanding outwards layer by layer from the central area to assign the expanded construction area to the point type corresponding to the sampling point in the central area includes: Traversing each of the construction areas, determining whether there is an adjacent blank area in each positive direction of the filling area; the filling area is the construction area with a point type, the filling area is initially the central area, the blank area is the construction area without a point type, and the positive directions are up, down, left and right; If the blank area exists, create a bucket for the blank area; Writing the point type of the fill area into the bucket; Randomly select a point type in the bucket and assign it to the blank area; If the blank area does not exist, it is determined that the expansion is completed.
5. The method according to any one of claims 1 to 4, characterized in that The reconstruction network trained on the soil type includes a first long short-term memory network, a second long short-term memory network, a third long short-term memory network, a first dynamic convolution module, a second dynamic convolution module, a first deconvolution layer, a second deconvolution layer, a spatial feature module and a first fully connected layer; The step of inputting the original vibration signal into a reconstruction network trained for the soil type and converting the original vibration signal into a target vibration signal representing the friction between the drilling machine and the soil in the construction site includes: Inputting the original vibration signal into the first long short-term memory network to extract the first trunk vibration feature; Inputting the first trunk vibration feature into the first dynamic convolution module to extract the second trunk vibration feature; Inputting the second trunk vibration feature into the spatial feature module, and applying an attention mechanism at multiple scales to extract the third trunk vibration feature; Inputting the third trunk vibration feature into the second dynamic convolution module to extract the fourth trunk vibration feature; Inputting the second main-trunk vibration feature into the first deconvolution layer to extract the fifth main-trunk vibration feature; Inputting the fifth trunk vibration feature into the second long short-term memory network to extract the sixth trunk vibration feature; fusing the fourth trunk vibration feature and the sixth trunk vibration feature into a seventh trunk vibration feature; Inputting the seventh main-trunk vibration feature into the second deconvolution layer to extract the eighth main-trunk vibration feature; Inputting the eighth trunk vibration feature into the third long short-term memory network to extract the ninth trunk vibration feature; The ninth trunk vibration feature is input into the first fully connected layer and mapped into a target vibration signal representing the friction between the drilling machine and the soil in the construction site.
6. The method according to claim 5, characterized in that The spatial feature module includes a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, a fourth long short-term memory network, a fifth long short-term memory network, a first multi-head attention layer and a second multi-head attention layer; The step of inputting the second trunk vibration feature into the spatial feature module and applying an attention mechanism at multiple scales to extract the third trunk vibration feature comprises: Inputting the second trunk vibration feature into the first convolutional layer to extract the first branch vibration feature; Inputting the second trunk vibration feature into the second convolutional layer to extract the second branch vibration feature; Inputting the second trunk vibration feature into the fourth long short-term memory network to extract the third branch vibration feature; Inputting the third branch vibration feature into the first multi-head attention layer to extract the fourth branch vibration feature; Inputting the second trunk vibration feature into the fifth long short-term memory network to extract the fifth branch vibration feature; Inputting the fifth branch vibration feature into the second multi-head attention layer to extract the sixth branch vibration feature; Merging the fourth branch vibration feature and the sixth branch vibration feature into a seventh branch vibration feature; Inputting the seventh branch vibration feature into the third convolutional layer to extract the eighth branch vibration feature; fusing the first branch vibration feature, the second branch vibration feature and the eighth branch vibration feature into a ninth branch vibration feature; The ninth branch vibration feature is input into the fourth convolutional layer to extract the third trunk vibration feature.
7. The method according to claim 5, characterized in that The classifier trained for the soil type includes a fifth convolutional layer, a sixth convolutional layer, a seventh convolutional layer, an eighth convolutional layer, a sixth long short-term memory network, a pooling layer and a second fully connected layer; The step of inputting the deviation vibration signal into a classifier trained on the soil type to identify the soil layer and the confidence level includes: Inputting the deviation vibration signal into the fifth convolutional layer to extract the first head vibration feature; Inputting the first head vibration feature into the sixth convolutional layer to extract the second head vibration feature; Inputting the second head vibration feature into the sixth long short-term memory network to extract the third head vibration feature; Inputting the third head vibration feature into the seventh convolutional layer to extract the fourth head vibration feature; Inputting the fourth head vibration feature into the pooling layer to extract the fifth head vibration feature; fusing the second head vibration feature and the fifth head vibration feature into a sixth head vibration feature; Inputting the sixth head vibration feature into the eighth convolutional layer to extract the seventh head vibration feature; The seventh head vibration feature is input into the second fully connected layer and mapped into a soil layer and confidence.
8. The method according to claim 7, characterized in that The verifying that the soil layer is a bearing layer according to the drilling pressure and the confidence level when the soil layer is a bearing layer comprises: Mapping the borehole pressure to a probability that the soil layer is a bearing layer; Substitute the probability and the confidence level when the soil layer is the bearing layer into the following formula to calculate the comprehensive value: Wherein, Score is the comprehensive value, Probability is the probability, Confidence is the confidence, Pexample is the first reference value set for the probability under the soil type, Cexample is the second reference value set for the confidence under the soil type, and α, β and γ are all weights; If the drilling pressure is greater than a first threshold value and the comprehensive value is greater than a second threshold value, it is determined that the soil layer is successfully verified as a bearing layer.
9. A terminal 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 construction method of the anti-floating anchor rod according to any one of claims 1 to 8 is implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the construction method of the anti-floating anchor rod according to any one of claims 1 to 8 is implemented.
