Shield soil bin pressure prediction method and system based on multi-source heterogeneous data fusion
By fusing multi-source heterogeneous data and using neural network models, the problems of low accuracy and parameter migration in shield tunnel earth pressure prediction were solved, achieving higher accuracy earth pressure prediction and ensuring construction stability.
Patent Information
- Application Number
- CN202411126329.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods for predicting pressure in tunnel boring machine (TBM) soil chambers lack the integration of multi-source heterogeneous data, resulting in low prediction accuracy and a lack of scientific reference for parameter settings, making them unsuitable for application across different projects.
A multi-source heterogeneous data fusion method is adopted. By acquiring and processing the tunneling parameters, shield burial depth and soil physical parameters of the shield tunnel project, a wavelet coefficient preprocessing function and correlation analysis model are established. A convolutional neural network is constructed to predict the soil pressure in the tunnel embankment. The model structure is optimized by a genetic algorithm to achieve data transfer learning.
It improves the accuracy of shield tunneling earth pressure prediction, reduces the risk of instability at the construction excavation face, and provides a scientific reference for earth pressure setting, especially effective in tunneling through sand and gravel strata.
Smart Images

Figure CN119025918A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of shield tunneling parameter prediction, and in particular to a shield soil bin pressure prediction method and system based on multi-source heterogeneous data fusion. Background Art
[0002] In the process of urban underground engineering construction, the shield method is widely used due to its fast construction speed, small impact on environmental disturbance and low cost. For shield tunnels passing through gravel strata, the gravel strata have obvious unevenness and discontinuity. The soil bin pressure is one of the most important control parameters in shield construction. Imbalance of soil bin pressure will cause instability of the face, missing strata, and then cause a series of adverse consequences such as surface subsidence. In the past, during shield excavation, drivers mainly roughly estimated the water and soil pressure of the excavation face based on the earth pressure balance theory of the shield excavation face. This method will lead to untimely adjustment of excavation parameters, and even cause problems such as abnormal wear of the tool and low efficiency. With the vigorous development of deep learning and the continuous advancement of data acquisition technology, new means have been brought to the setting of shield soil bin pressure.
[0003] At present, most scholars use machine learning algorithms to predict shield bunker pressure based on tunneling parameters such as shield advancement speed, cutter head torque, and penetration rate. This method has solved the problem of predicting shield bunker pressure to a certain extent, but the relevant method lacks consideration of other heterogeneous data features such as shield diameter and tunnel burial depth. At the same time, it is only applicable to a single project and cannot be migrated to other projects for application. Summary of the invention
[0004] In view of the above-mentioned defects, the technical problem solved by the present invention is to provide a method and system for predicting shield soil bin pressure based on multi-source heterogeneous data fusion, so as to solve the problems of low shield soil bin pressure prediction accuracy and lack of scientific reference for parameter setting.
[0005] A first aspect of the present invention provides a method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion, the method being used in a shield project, comprising:
[0006] S101: Acquire data related to the shield project and analyze the related data; the related data include: excavation parameters, shield burial depth and physical parameters of stratum soil;
[0007] S102: using the analyzed relevant data as input variables and the shield soil bin pressure as output variables, establishing a multi-source heterogeneous data fusion model to obtain a first shield soil bin pressure prediction model;
[0008] S103: Using the first shield soil bin pressure prediction model for other shield projects to obtain a second shield soil bin pressure prediction model, and completing the prediction of the shield soil bin pressure.
[0009] Preferably, the obtaining of data related to the shield project includes:
[0010] Use sensors to collect shield project related data and extract corresponding shield parameter signals;
[0011] Performing wavelet decomposition on the shield parameter signal to obtain wavelet coefficients at each decomposition layer;
[0012] According to the wavelet coefficients at each decomposition level, a processing threshold that changes with the decomposition level is constructed;
[0013] Constructing a wavelet coefficient preprocessing function according to the processing threshold;
[0014] Using the wavelet coefficient preprocessing function to process the wavelet coefficients at each decomposition level to obtain preprocessed wavelet coefficients;
[0015] The pre-processed wavelet coefficients are reconstructed to obtain shield project related data.
[0016] Preferably, constructing a processing threshold value that changes with the number of decomposition levels according to the wavelet coefficients at each decomposition level includes:
[0017] Extract all high-frequency wavelet coefficients at each decomposition level;
[0018] Calculating the standard deviation of environmental noise according to the high-frequency wavelet coefficients;
[0019] A processing threshold is constructed according to the standard deviation of the environmental noise; wherein the processing threshold is:
[0020]
[0021] Among them, λ i represents the processing threshold at the i-th decomposition level, σ represents the standard deviation of environmental noise, n represents the shield parameter signal, median represents the median, and D represents the high-frequency wavelet coefficient.
[0022] Preferably, constructing a wavelet coefficient preprocessing function according to the processing threshold comprises:
[0023] Using the formula:
[0024]
[0025] Construct the wavelet coefficient preprocessing function; where tanh represents the tangent function, ω i represents the i-th wavelet coefficient, k represents the adjustment coefficient, Represents the wavelet coefficients after preprocessing.
[0026] Preferably, the reconstructing the preprocessed wavelet coefficients to obtain shield project related data includes:
[0027] The preprocessed wavelet coefficients are reconstructed to obtain the reconstructed shield parameter signal;
[0028] The quality of the reconstructed shield parameter signal is evaluated to obtain the quality index; the quality index calculation formula is:
[0029]
[0030] Among them, SNR represents the first quality index number, x i represents the amplitude of the original shield parameter signal at the i-th point, y i represents the amplitude of the shield parameter signal at the i-th point after reconstruction, CC represents the second quality index, represents the mean value of the original shield parameter signal, represents the mean value of the shield parameter signal after reconstruction;
[0031] Determine whether the quality index is within the preset range. If the quality index is not within the preset range, change the adjustment coefficient of the wavelet coefficient preprocessing function and re-preprocess the wavelet coefficient until the quality index enters the preset range, and output the corresponding shield project related data.
[0032] Preferably, the analysis of the relevant data includes: processing the null value of the relevant data, denoising and standardizing the relevant data, and analyzing the processed relevant data using a correlation analysis method to determine whether the relevant data is redundant; wherein the denoising includes: setting the total length of the relevant data parameter column to n, a data collection frequency to f, a tolerance time span to Tmax, assuming that the first data collection time is t1, and the Nth data collection time is t N ,but:
[0033] T N =t N -t1
[0034] Among them, T N The time span for collecting data for Article N;
[0035] If, T N ≤Tmax, the collected relevant data are available;
[0036] If, T N >Tmax, the collected relevant data are not available.
[0037] Preferably, the denoising process further includes: increasing the number of fault-tolerant data items r, and collecting data over a time span of r / f, then:
[0038] Tmax=(n+r) / f
[0039] Among them, Tmax is the tolerance time span.
[0040] Preferably, the processing related data is analyzed by using a correlation analysis method to determine whether the related data is redundant, including:
[0041] The soil bin pressure is taken as the dependent variable, and the related data is taken as the independent variable, and they are sorted from large to small to form the dependent variable sequence and the independent variable sequence;
[0042] Differences are made between the rankings of the dependent variable and independent variable data pairs collected at the same time to form rank differences;
[0043] The correlation coefficient between the soil bin pressure and each relevant data is calculated according to the grade difference; wherein the correlation coefficient calculation formula is:
[0044]
[0045] Among them, r s represents the correlation coefficient, n represents the number of dependent variable and independent variable data pairs, d r It represents the rank difference between the dependent variable and the independent variable data collected at the same time;
[0046] The relevant data corresponding to the correlation coefficient that is not within the preset range are removed.
[0047] The second aspect of the present invention discloses a shield soil bin pressure prediction system based on multi-source heterogeneous data fusion, the system comprising:
[0048] An acquisition unit, used for acquiring data related to the shield project and analyzing the related data;
[0049] The acquisition unit includes: an acquisition subunit for acquiring excavation parameters, shield burial depth and physical parameters of stratum soil;
[0050] An analysis unit, configured to perform denoising on the relevant data using a wavelet analysis threshold method and to analyze the relevant data using a correlation analysis method to determine whether the relevant data is redundant;
[0051] A fusion unit is used to use the analyzed related data as input variables and the shield soil bin pressure as output variables to establish a multi-source heterogeneous data fusion model to obtain a first shield soil bin pressure prediction model;
[0052] The learning unit applies the first shield soil bin pressure prediction model to other shield projects to obtain the second shield soil bin pressure prediction model.
[0053] The third aspect of the present invention provides an intelligent device, including a transmitter, a receiver, a memory and a processor; the memory is used to store computer instructions; the processor is used to run the computer instructions stored in the memory to implement the above-mentioned method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion.
[0054] The fourth aspect of the present invention provides a storage medium, comprising: a readable storage medium and computer instructions, wherein the computer instructions are stored in the readable storage medium; the computer instructions are used to implement the above-mentioned method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion.
[0055] The beneficial effects provided by the present invention are as follows: it breaks through the singleness and limitations of the training set data considered in the existing shield soil bin pressure prediction methods, can overcome the prediction limitations caused by the large differences in information from different data sources, improve the prediction accuracy of the shield soil bin pressure, and reduce the risk of instability faced by shield excavation construction. The prediction of the shield soil bin pressure provides a scientific reference for the setting of the soil bin pressure, especially during the excavation process in sandy and gravel formations. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[0057] Figure 1 A flow chart of a method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion disclosed in an embodiment of the present invention;
[0058] Figure 2 This is an example diagram of mean filtering and denoising of a shield soil bin pressure prediction method based on multi-source heterogeneous data fusion disclosed in an embodiment of the present invention;
[0059] Figure 3 It is a schematic diagram of the correlation analysis of tunneling parameters of the shield soil bin pressure prediction method based on multi-source heterogeneous data fusion disclosed in an embodiment of the present invention;
[0060] Figure 4 A comparison diagram of predicted and true values of soil bin pressure in a shield soil bin pressure prediction method based on multi-source heterogeneous data fusion disclosed in an embodiment of the present invention;
[0061] Figure 5 The present invention is a block diagram of a shield soil bin pressure prediction system based on multi-source heterogeneous data fusion disclosed in an embodiment of the present invention.
[0062] The above drawings show clear embodiments of the present disclosure, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present disclosure in any way, but to illustrate the concepts of the present disclosure to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION
[0063] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.
[0064] Prediction method of shield soil bin pressure based on multi-source heterogeneous data fusion. Figure 1 As shown, the method is used in a shield project.
[0065] S101: Acquire data related to the shield project and analyze the data.
[0066] Data collection and preprocessing: The shield machine often collects and stores the time series data of various parameters at a frequency of 1Hz through the PLC system. The data can be obtained through the interface provided by the project. In addition to irrelevant parameters, there are 18 characteristic parameter columns that affect the tunneling speed in shield construction, including cutter head torque, total thrust, penetration, cutter head speed, cutter head total extrusion pressure, total grouting volume, etc.
[0067] Furthermore, in practical applications, the present invention needs to connect the sensor to the PLC system to collect shield project related data and extract the corresponding shield parameter signal;
[0068] Performing wavelet decomposition on the shield parameter signal to obtain wavelet coefficients at each decomposition layer;
[0069] According to the wavelet coefficients at each decomposition level, a processing threshold that changes with the decomposition level is constructed;
[0070] Furthermore, a processing threshold that changes with the number of decomposition levels is constructed according to the wavelet coefficients at each decomposition level, including:
[0071] Extract all high-frequency wavelet coefficients at each decomposition level;
[0072] Calculating the standard deviation of environmental noise according to the high-frequency wavelet coefficients;
[0073] A processing threshold is constructed according to the standard deviation of the environmental noise; wherein the processing threshold is:
[0074]
[0075] Among them, λ irepresents the processing threshold at the i-th decomposition level, σ represents the standard deviation of environmental noise, n represents the shield parameter signal, median represents the median, and D represents the high-frequency wavelet coefficient.
[0076] In the process of wavelet processing, the selection of threshold directly affects the acquisition quality and effect of the target signal. If the threshold is too large, the effective signal will be mistakenly removed. If the threshold is too small, some noise will be identified as effective signal, thereby reducing the quality of the signal. The traditional threshold is usually a fixed value, and its denoising effect is usually not suitable for the current application scenario. Therefore, the characteristics and noise levels of different signals are different. Therefore, the adaptive threshold constructed in this application can be adjusted according to the characteristics of the current signal, which solves the limitation that the fixed threshold has different effects in different wavelet transform layers, and greatly improves the acquisition quality of the signal.
[0077] A wavelet coefficient preprocessing function is constructed according to the processing threshold; wherein the wavelet coefficient preprocessing function is:
[0078]
[0079] Construct the wavelet coefficient preprocessing function; where tanh represents the tangent function, ω i represents the i-th wavelet coefficient, k represents the adjustment coefficient, Represents the wavelet coefficients after preprocessing.
[0080] The existing commonly used threshold functions all use the sgn function to transform the wavelet coefficients. However, the sgn function has an oscillation phenomenon at the discontinuity point. The present invention uses the tanh function to replace the sgn function, which is smoother at the discontinuity point position, so it can more accurately and meticulously screen out the wavelet coefficients represented by the target signal.
[0081] Using the wavelet coefficient preprocessing function to process the wavelet coefficients at each decomposition level to obtain preprocessed wavelet coefficients;
[0082] The pre-processed wavelet coefficients are reconstructed to obtain shield project related data.
[0083] Furthermore, the pre-processed wavelet coefficients are reconstructed to obtain shield project related data, including:
[0084] The preprocessed wavelet coefficients are reconstructed to obtain the reconstructed shield parameter signal;
[0085] The quality of the reconstructed shield parameter signal is evaluated to obtain the quality index; the quality index calculation formula is:
[0086]
[0087] where SNR represents the first quality index number, and x i represents the amplitude of the original shield parameter signal at the i-th point, and y i represents the amplitude of the reconstructed shield parameter signal at the i-th point, and CC represents the second quality index number. represents the mean value of the original shield parameter signal, represents the mean value of the reconstructed shield parameter signal;
[0088] Judge whether the quality index is within the preset range. If the quality index is not within the preset range, then change the adjustment coefficient of the wavelet coefficient preprocessing function and preprocess the wavelet coefficient again until the quality index enters the preset range, and output the relevant data of the corresponding shield project.
[0089] By screening out the quality indexes within the preset range, the present invention can adjust the relevant parameters for different types of signals, thereby improving the signal acquisition quality.
[0090] The preprocessing process of the data includes null value processing of the data, denoising of time series data based on mean filtering, and standardization of the data.
[0091] Null value processing: The acquisition of the tunneling parameter data in time series is continuous and large in quantity. Therefore, for the null values in the data, the direct deletion method is adopted in this paper, and the null values and the characteristic values collected at the same time point are directly deleted. Let the original data set be D, and the total number of characteristic parameter columns be l. For a certain parameter column Xi, if at time t, the value Xit of the parameter column is null, then all the parameter values collected at time t are deleted.
[0092] Mean filtering denoising: During the tunneling of the tunnel shield, manual operations and geological mutations often cause data fluctuations and anomalies. Especially in the case of complex and changeable geological environments, the fluctuations of the time series parameter data are more unstable. These abnormal fluctuation data are the noise data in the tunneling parameter sequence. The mean filtering method is an effective method for denoising time series data in the time domain. For a certain parameter column assuming that its sequence total length is n, when using the mean filtering method to denoise it, a window size w (w < n, w represents the number of data in the window) needs to be set. During the forward sliding of the window, calculate the mean value of the parameter values in the window to form a new sequence, such as Figure 2As shown. In actual construction, data may be missed due to machine failure or acquisition equipment failure, so the time span of the data in the window needs to be judged during calculation. Suppose the data acquisition frequency is f, the maximum tolerance time span is Tmax, the acquisition time of the first data in the window is t1, and the acquisition time of the last data is tw, then the current window time span is Tw=tw-t1. If Tw≤Tmax, the current window is available, otherwise ignore the current window and continue to slide forward. For Tmax, its value can be set by increasing the number of fault-tolerant data. When the data in the window is continuous, the window time span is w / f. Suppose the number of fault-tolerant data (that is, the number of data allowed to be missed) is r, then Tmax=(w+r) / f.
[0093] Data standardization: When using data for machine learning modeling, the dimensions and magnitudes of different features will affect the model training effect. Generally, higher eigenvalues will weaken the role of lower eigenvalues in model calculations. Data standardization is to normalize the data to the same data level to eliminate this effect. The z-score method is used to standardize the data. The processed data conforms to the standard normal distribution, that is, the mean is 0 and the standard deviation is 1. Suppose the length of the sequence after mean filtering is N, for a certain parameter sequence Calculate its mean μi and standard deviation σi, then the normalized parameter value is:
[0094]
[0095] The correlation analysis method is used to perform redundancy analysis on the relevant data, including: determining whether there is information overlap in the relevant data. Data redundancy and correlation analysis: The stability of the shield excavation face is affected by multiple factors, including but not limited to shield thrust, torque, excavation speed, shield burial depth, cohesion of the soil in the strata passed through, friction angle, etc. In order to set a suitable shield soil bin pressure to balance the water and soil pressure of the shield excavation face, it is necessary to determine whether there is redundancy between variables such as shield thrust, torque, excavation speed, cutterhead speed, total cutterhead extrusion pressure, total grouting volume, shield burial depth, cohesion of the soil in the strata passed through, friction angle, etc., and calculate the correlation coefficients between these variables.
[0096] In the embodiment of the present invention, the calculation process of the correlation coefficient of the present invention is as follows:
[0097] The soil bin pressure is taken as the dependent variable, and the related data is taken as the independent variable, and they are sorted from large to small to form the dependent variable sequence and the independent variable sequence;
[0098] Differences are made between the rankings of the dependent variable and independent variable data pairs collected at the same time to form rank differences;
[0099] The correlation coefficient between the soil bin pressure and each relevant data is calculated according to the grade difference; wherein the correlation coefficient calculation formula is:
[0100]
[0101] Among them, r s represents the correlation coefficient, n represents the number of dependent variable and independent variable data pairs, d r It represents the rank difference between the dependent variable and the independent variable data collected at the same time;
[0102] The relevant data corresponding to the correlation coefficient that is not within the preset range are removed.
[0103] In this example, the shield tunneling parameters that are closely related to the soil bin pressure include total thrust, cutter head torque and propulsion speed, such as Figure 3 As shown in the figure, the correlation coefficients of excavation parameters are 0.26, 0.47 and 0.27 respectively.
[0104] S102: Taking the analyzed data as input variables and the shield soil bin pressure as output variables, a multi-source heterogeneous data fusion model is established to obtain a first shield soil bin pressure prediction model. Multi-source heterogeneous data fusion: Taking shield thrust, torque, tunneling speed, shield burial depth, cohesion of the stratum soil, friction angle, etc. as input dependent variables and shield soil bin pressure as input independent variable, a multi-source heterogeneous data fusion model based on convolutional neural network is established to obtain a shield soil bin pressure prediction model.
[0105] By selecting three influencing factors, namely, shield thrust, cutterhead torque and tunneling speed, from S101 as input variables and taking the change of shield soil bin pressure as output, a neural network is constructed for training.
[0106] In the first fusion, a convolutional neural network architecture optimized by genetic algorithm was applied. The network design contains two hidden layers, and the ReLU function is selected as the activation function. The output layer uses the Log-sigmoid function as its activation function. In addition, L2 regularization is introduced to reduce the risk of overfitting. The model is used to predict the shield soil bin pressure and the first fusion result is obtained.
[0107] The first fusion result was fused with the shield depth data, and ReLU was used as the activation function and L2 regularization was introduced. After determining the structure of the convolutional neural network, the genetic algorithm (GA) was further introduced to enhance the search capability, especially the performance in global search, to improve the prediction accuracy. The model was used to predict the shield soil bin pressure and the second fusion result was obtained.
[0108] According to the above calculation method, the second fusion result is further integrated with the cohesion and friction angle of the soil in the stratum to obtain the final soil bin pressure prediction result.
[0109] S103: Use the first shield soil bin pressure prediction model for other shield projects to obtain a second shield soil bin pressure prediction model... and so on, to obtain a shield soil bin pressure prediction model. According to step S101 and step S102, obtain a soil bin pressure prediction model based on shield project A, and then perform transfer learning on the initial soil bin pressure prediction model; predict the shield soil bin pressure according to the parameters of shield project B, such as shield thrust, torque, excavation speed, shield burial depth, cohesion of the stratum soil, friction angle, etc., and adjust the weights and biases in the initial soil bin pressure prediction model according to the prediction results, to obtain a soil bin pressure prediction model suitable for shield project B.
[0110] Analysis on the prediction effect of shield soil bin pressure:
[0111] The coefficient of determination (R2) and mean absolute error (MAE) are used as indicators to evaluate the prediction accuracy, such as Figure 4 As shown, the predicted results of shield soil bin pressure are compared with the actual results.
[0112] The prediction effect of shield soil bunker pressure is analyzed, and the determination coefficient (R2), mean absolute error (MAE) and root mean square error (RMSE) are used as indicators to evaluate the prediction accuracy, and the predicted results of shield soil bunker pressure are compared with the actual results. The present invention can overcome the prediction limitations caused by the large difference in information from different data sources, improve the prediction accuracy of shield soil bunker pressure, and reduce the risk of instability faced by shield excavation construction. It provides a scientific reference for the setting of soil bunker pressure during shield excavation in sandy and gravel strata.
[0113] The second aspect of the present invention discloses a shield soil bin pressure prediction system 50 based on multi-source heterogeneous data fusion, such as Figure 5 As shown, it includes: an acquisition unit 501, which is used to acquire data related to the shield project and analyze the related data.
[0114] The fusion unit 502 is used to use the analyzed related data as input variables and the shield soil bin pressure as output variables to establish a multi-source heterogeneous data fusion model to obtain a first shield soil bin pressure prediction model.
[0115] The learning unit 503 applies the first shield soil bin pressure prediction model to other shield projects to obtain a second shield soil bin pressure prediction model.
[0116] According to an embodiment of the present invention, the acquisition unit includes: an acquisition subunit for acquiring excavation parameters, shield burial depth and physical parameters of stratum soil.
[0117] The analysis unit is used to use a wavelet analysis threshold method to denoise the relevant data and use a correlation analysis method to analyze the relevant data to determine whether the relevant data is redundant.
[0118] The third aspect of the present invention provides an intelligent device, including a transmitter, a receiver, a memory and a processor; the memory is used to store computer instructions; the processor is used to run the computer instructions stored in the memory to implement the above-mentioned method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion.
[0119] The fourth aspect of the present invention provides a storage medium, comprising: a readable storage medium and computer instructions, wherein the computer instructions are stored in the readable storage medium; the computer instructions are used to implement the above-mentioned method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion.
[0120] The beneficial effects provided by the present invention are as follows: it breaks through the singleness and limitations of the training set data considered in the existing shield soil bin pressure prediction methods, can overcome the prediction limitations caused by the large differences in information from different data sources, improve the prediction accuracy of the shield soil bin pressure, and reduce the risk of instability faced by shield excavation construction. The prediction of the shield soil bin pressure provides a scientific reference for the setting of the soil bin pressure, especially during the excavation process in sandy and gravel formations.
[0121] Obviously, the above specific implementation cases are only examples for illustrating the application of the present method, rather than limiting the implementation methods. For those skilled in the art, other different forms of changes and modifications can be made on the basis of the above description to study other related issues. Therefore, the protection scope of the present invention should be the protection scope of the claims.
[0122] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.
[0123] The electronic device and other embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.
[0124] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, rather than to limit them. Although the embodiments of the present invention have been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
[0126] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be regarded as exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.
[0127] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. A method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion, characterized in that: The method is used in shield projects, including: S101: Acquire data related to the shield project and analyze the related data; the related data include: excavation parameters, shield burial depth and physical parameters of stratum soil; S102: using the analyzed relevant data as input variables and the shield soil bin pressure as output variables, establishing a multi-source heterogeneous data fusion model to obtain a first shield soil bin pressure prediction model; S103: Using the first shield soil bin pressure prediction model for other shield projects to obtain a second shield soil bin pressure prediction model, and completing the prediction of the shield soil bin pressure.
2. The method according to claim 1, characterized in that The obtaining of data related to the shield project includes: Use sensors to collect shield project related data and extract corresponding shield parameter signals; Performing wavelet decomposition on the shield parameter signal to obtain wavelet coefficients at each decomposition layer; According to the wavelet coefficients at each decomposition level, a processing threshold that changes with the decomposition level is constructed; Constructing a wavelet coefficient preprocessing function according to the processing threshold; Using the wavelet coefficient preprocessing function to process the wavelet coefficients at each decomposition level to obtain preprocessed wavelet coefficients; The pre-processed wavelet coefficients are reconstructed to obtain shield project related data.
3. The method according to claim 2, characterized in that The step of constructing a processing threshold value that changes with the number of decomposition levels according to the wavelet coefficients at each number of decomposition levels includes: Extract all high-frequency wavelet coefficients at each decomposition level; Calculating the standard deviation of environmental noise according to the high-frequency wavelet coefficients; A processing threshold is constructed according to the standard deviation of the environmental noise; wherein the processing threshold is: Among them, λ i represents the processing threshold at the i-th decomposition level, σ represents the standard deviation of environmental noise, n represents the shield parameter signal, median represents the median, and D represents the high-frequency wavelet coefficient.
4. The method according to claim 3, characterized in that The step of constructing a wavelet coefficient preprocessing function according to the processing threshold comprises: Using the formula: Construct the wavelet coefficient preprocessing function; where tanh represents the tangent function, ω i represents the i-th wavelet coefficient, k represents the adjustment coefficient, Represents the wavelet coefficients after preprocessing.
5. The method according to claim 4, characterized in that The reconstructing the pre-processed wavelet coefficients to obtain shield project related data includes: The preprocessed wavelet coefficients are reconstructed to obtain the reconstructed shield parameter signal; The quality of the reconstructed shield parameter signal is evaluated to obtain the quality index; the quality index calculation formula is: Among them, SNR represents the first quality index number, x i represents the amplitude of the original shield parameter signal at the i-th point, y i represents the amplitude of the shield parameter signal at the i-th point after reconstruction, CC represents the second quality index, represents the mean value of the original shield parameter signal, represents the mean value of shield parameter signal after reconstruction; Determine whether the quality index is within the preset range. If the quality index is not within the preset range, change the adjustment coefficient of the wavelet coefficient preprocessing function and re-preprocess the wavelet coefficient until the quality index enters the preset range, and output the corresponding shield project related data.
6. The method according to claim 5, characterized in that The analysis of the relevant data includes: processing the null value of the relevant data, denoising and standardizing the relevant data, and analyzing the processed relevant data using a correlation analysis method to determine whether the relevant data is redundant; wherein the denoising includes: setting the total length of the relevant data parameter column to n, a data collection frequency to f, a tolerance time span to Tmax, assuming that the first data collection time is t1, and the Nth data collection time is t N ,but: T N =t N -t1 Among them, T N The time span for collecting data for Article N; If, T N ≤Tmax, the collected relevant data are available; If, T N >Tmax, the collected relevant data are not available.
7. The method according to claim 6, characterized in that The denoising process also includes: increasing the number of fault-tolerant data items r, and collecting data over a time span of r / f, then: Tmax=(n+r) / f Among them, Tmax is the tolerance time span.
8. The method according to claim 6, characterized in that The processing related data is analyzed by using a correlation analysis method to determine whether the related data is redundant, including: The soil bin pressure is taken as the dependent variable, and the related data is taken as the independent variable, and they are sorted from large to small to form the dependent variable sequence and the independent variable sequence; Differences are made between the rankings of the dependent variable and independent variable data collected at the same time to form rank differences; The correlation coefficient between the soil bin pressure and each relevant data is calculated according to the grade difference; wherein the correlation coefficient calculation formula is: Among them, r s represents the correlation coefficient, n represents the number of dependent variable and independent variable data pairs, d r It represents the rank difference between the dependent variable and the independent variable data collected at the same time; The relevant data corresponding to the correlation coefficients that are not within the preset range are removed.
9. The shield soil bin pressure prediction system based on multi-source heterogeneous data fusion is characterized by: The system comprises: An acquisition unit, used for acquiring data related to the shield project and analyzing the related data; The acquisition unit includes: an acquisition subunit for acquiring excavation parameters, shield burial depth and physical parameters of stratum soil; An analysis unit, configured to perform denoising on the relevant data using a wavelet analysis threshold method and to analyze the relevant data using a correlation analysis method to determine whether the relevant data is redundant; A fusion unit is used to use the analyzed related data as input variables and the shield soil bin pressure as output variables to establish a multi-source heterogeneous data fusion model to obtain a first shield soil bin pressure prediction model; The learning unit applies the first shield soil bin pressure prediction model to other shield projects to obtain the second shield soil bin pressure prediction model.
10. A storage medium, characterized in that: include: A readable storage medium and computer instructions, wherein the computer instructions are stored in the readable storage medium; The computer instructions are used to implement the method for predicting shield soil bin pressure based on multi-source heterogeneous data fusion as described in any one of claims 1 to 8.
