Earth and rockfill dam deformation prediction method based on PZOA-VMD-TCN

Abstract

The invention provides an earth and rockfill dam deformation prediction method based on PZOA-VMD-TCN, and the method comprises the steps: collecting dam deformation data and environmental quantity data, integrating the data into a dam deformation measurement value sequence based on a time sequence, dividing a training set and a test set according to a set proportion, and carrying out the normalization processing of all data; then optimizing a parameter modal number K and a penalty factor alpha in the VMD decomposition parameters by using a PZOA optimization algorithm; a value obtained through optimization is applied to a VMD decomposition algorithm to decompose the deformation data to obtain a plurality of IMF sub-sequence values; based on the deformation influence factors, screening the deformation influence factors by utilizing a PLS (Partial Least Squares) method, and finding out an optimal input deformation influence factor set and the number of principal components corresponding to each IMF component; and finally, constructing a deformation prediction input data set according to the obtained optimal input deformation factor set, and predicting an earth and rockfill dam deformation value by using a PZOA-TCN model.

Description

Technical Field

[0001] The invention belongs to the technical field of dam deformation prediction, and in particular relates to a deformation prediction method for earth-rock dam based on PZOA-VMD-TCN. Background Art

[0002] As a type of dam widely used and rapidly developed in water conservancy construction, earth-rock dams have gradually become more and more vulnerable to safety hazards caused by the aging of the dam body over time. In dam health monitoring, deformation data is a key indicator for evaluating its status. Therefore, establishing a scientific and efficient prediction model to grasp the deformation trend of the dam is extremely important for ensuring the structural safety and functional stability of the dam. However, the deformation data of earth-rock dams show strong nonlinear characteristics, and the noise interference in the monitoring data makes the construction of the prediction model challenging, thus affecting the accuracy of deformation prediction. As one of the core projects of earth-rock dam safety monitoring, how to establish an efficient, accurate and effective dam deformation prediction model for the needs of earth-rock dam safety monitoring is an urgent problem to be solved. Summary of the invention

[0003] In order to overcome the problem of low prediction accuracy caused by strong nonlinearity and noise in the data in the earth-rock dam deformation prediction method, the present invention provides an earth-rock dam deformation prediction method based on PZOA-VMD-TCN.

[0004] The implementation process includes the following steps: 1) collecting dam deformation data and environmental quantity data, and integrating them into a dam deformation measurement sequence based on time series, dividing the training set and the test set according to the set ratio, and normalizing all the data; 2) constructing the PZOA-VMD model to realize the optimal calculation of two important parameters (K, alpha); 3) decomposing the dam deformation measurement sequence <IMF n >, use the MIC algorithm to calculate the correlation coefficient between each IMF and the actual dam deformation measurement, identify the noisy components, and use the wavelet-Gaussian integrated filtering method to perform complementary double denoising on the noisy components; 4) Based on the deformation influencing factors, a PLS-based technology for screening earth-rock dam deformation influencing factors is proposed to find the optimal number of input deformation influencing factor sets corresponding to each IMF component; 5) According to the optimal deformation influencing factor set, a deformation prediction input data set is constructed, and the PZOA-TCN model is used to predict the deformation value of the earth-rock dam. The present invention obtains the original characteristics of the dam deformation measurement sequence according to the signal decomposition method, and obtains the optimal input dam deformation factor set corresponding to each decomposed subsequence through the PLS algorithm; the model prediction input factor set is constructed in this way, and then loaded into the model PZOA-TCN to predict the deformation value of the earth-rock dam. Selecting the optimal number of input factor sets can effectively improve the prediction accuracy of the model.

[0005] The technical solution specifically adopted by the present invention to solve the technical problem is:

[0006] A deformation prediction method for earth-rock dam based on PZOA-VMD-TCN: firstly, dam deformation data and environmental quantity data are collected and integrated into a dam deformation measurement sequence based on time series, and the training set and the test set are divided according to the set ratio, and all the data are normalized; then the PZOA optimization algorithm is used to optimize the parameter mode number K and the penalty factor alpha in the VMD decomposition parameters; and the optimized values ​​are applied to the VMD decomposition algorithm to decompose the deformation data to obtain a number of IMF subsequence values; based on the deformation influencing factors, the PLS method is used to screen the deformation influencing factors, and the optimal input deformation influencing factor set and the number of principal components corresponding to each IMF component are found; finally, the deformation prediction input data set is constructed according to the obtained optimal input deformation factor set, and the deformation value of the earth-rock dam is predicted using the PZOA-TCN model.

[0007] Furthermore, the training set and the test set are selected according to the following ratios:

[0008] disp_set[:]=DISP_SET[:]*proportion

[0009] proportion is the selected proportionality factor;

[0010] The specific steps for normalizing the data are as follows:

[0011] max=disp_set[:] max

[0012] min = disp_set[:] min

[0013]

[0014] Among them, disp_set[:], They are the actual value of the dam deformation measurement and the value obtained after normalization;

[0015] min, max are the minimum and maximum values ​​in the sample data of dam deformation measurement respectively.

[0016] Furthermore, the use of the PZOA optimization algorithm to optimize the parameter mode number K and the penalty factor alpha in the VMD decomposition parameters is specifically as follows: in the initialization stage of PZOA optimization of VMD, each dimension of the particle corresponds to a parameter, so each particle will have an initial value of alpha and K. When they are input into the fitness function for decomposition, the corresponding fitness value will be obtained.

[0017] Initialize the position X of the zebra individuals in the population i and the velocity vector V i , the setting parameters include: maximum number of iterations, population size, inertia weight w, learning factors c1 and c2, and exploration and development weight factors b1, b2, b3;

[0018] Then the following iteration process is performed until the maximum number of iterations is reached:

[0019] The position vector and velocity vector of the individuals in the population are updated as shown in the following formula: r1, r2, r3, r4, r5 are the generated random numbers, G(t) is the global optimal position, P i is the optimal position of an individual;

[0020] V i (t+1)=w×V i (t)+c1×r4×(P i (t)-X i (t)+c2×r5×(G(t)-X i (t))

[0021] X i (t+1)=X i (t)+b1×r1×(G(t)-X i (t))+b2×r2×(P i (t)-X i (t)+b3×r3+V i (t+1)

[0022] Confirm that the individual position is within the allowed search space and update the fitness value of each individual;

[0023] Update the individual optimal position and the global optimal position, corresponding to the global optimal fitness value, and assign a value;

[0024] Pi=Xi

[0025] G(t)=Xi

[0026] Finally, the output optimal value is verified: the fitness function value is calculated within the obtained optimal value range, and if the calculated result is consistent with the optimization result, it is output.

[0027] Furthermore, in the process of applying the optimal value to the VMD decomposition algorithm to decompose the deformation data to obtain a number of IMF subsequence values, the dam deformation measurement value decomposition sequence is calculated based on the result of the optimal calculation. <IMF n>The MIC algorithm is used to calculate the correlation coefficient between each IMF and the measured deformation of the dam, identify the components containing noise, and use the wavelet-Gaussian integrated filtering method to perform complementary double denoising on the components containing noise;

[0028] Specifically:

[0029] Substitute the best K and alpha values ​​obtained by optimization into the VMD decomposition algorithm, load the data of the dam deformation measurement sequence, and obtain the decomposition subsequence set Num-IMF based on the number of K;

[0030] IMF 1 =[v 1 ,v 2 ,v 3 ,v 4 ,v 5 ,....v n ]

[0031] IMF 2 =[v 1 ,v 2 ,v 3 ,v 4 ,v 5 ,....v n ]

[0032] …

[0033] IMF K =[v 1 ,v 2 ,v 3 ,v 4 ,v 5 ,....v n The elements in the group are different and not consistent according to different groups. The number of decomposition groups depends on K. Then, according to the maximum information coefficient MIC algorithm, the correlation coefficient between each decomposition sequence IMF and the original measurement value of the dam is calculated to identify the noise sequence in the decomposition sequence.

[0034] In the MIC algorithm, we first define the decomposition subsequence set Num-IMF, MI[v:X] is denoted as v k The mutual information value between the original displacement measurement sequence X(t) and p(t); v X[,] represents the joint probability distribution function; p v , pX represents the marginal probability distribution function, as shown below:

[0035]

[0036] Then calculate the mutual information value between each subsequence and the original displacement measurement sequence:

[0037]

[0038] …

[0039] Until the mutual information values ​​of all subsequences are obtained, take:

[0040] I[v:X]=maxMI[v:X]

[0041] That is, the maximum mutual information value is obtained, and then the maximum mutual information value is normalized:

[0042]

[0043] Then add the MIC values ​​of all subsequences to get the average value:

[0044]

[0045] After standardization, the MIC value range is [0,1]; if MIC k >MIC m , it is considered that the Kth subsequence is strongly correlated with the original dam deformation measurement value and there is no noise; otherwise, there is noise and the subsequence needs to be denoised and reconstructed.

[0046] Furthermore, for the subsequences with noise, Gaussian filtering and wavelet filtering are used to perform overlapping denoising on the original data.

[0047] Furthermore, based on the deformation influencing factors, the deformation influencing factors are screened using the PLS method to find out the optimal input deformation influencing factor set and the number of principal components corresponding to each IMF component:

[0048] The dam deformation factor set and the dam deformation measurement value subsequence set Num-IMF are standardized to obtain the standardized data, and then the training set and the test set are divided according to the proportional factor; the dam deformation factor set I(t) is used as the independent variable, and the dam deformation measurement value sequence subsequence set Num-IMF is used as the dependent variable:

[0049] Num_IMF = [IMF 1 ,IMF 2 ,IMF 3 ,....,IMF K ] n×K

[0050] I(t)=[I 1 ,I 2 ,I 3 ,....,I q ] n×q

[0051] Where q is the total number of factor sets;

[0052] Solve the principal components u1 and v1 of the independent variable and the dependent variable; then establish the regression variance of the independent variable I, the dependent variable Num-IMF denoted as F, and the first principal component u1 and v1, and calculate the residual matrix I1 and F1; then use I1 and F1 to replace I and F to form new independent variables and dependent variables, and solve the first principal component u2 and t2 of the new independent variable and dependent variable, which are the second principal component of the original independent variable and dependent variable; then establish the residual matrix composed of the new independent variable I1 and dependent variable F1 and the regression equation of the second principal component u2 and t2, and calculate the residual matrix I2 and F2;

[0053] Repeat the above steps until all principal components are found or the conditions are met; then perform a cross-test to determine the number of principal components P that meet the conditions; each IMF has a corresponding P value, and according to the dam deformation factor set and P value, find the optimal deformation factor set corresponding to each IMF:

[0054] D 1 =I×P 1 ×(P 1 T ×P 1 ) -1

[0055] D 2 =I×P 2 ×(P 2 T ×P 2 ) -1

[0056] …

[0057] D K =I×P K ×(P K T ×P K ) -1

[0058] D is the final optimal input deformation factor set.

[0059] Furthermore, the deformation prediction input data set is constructed according to the obtained optimal input deformation factor set, and the deformation value of the earth-rock dam is predicted by using the PZOA-TCN model as follows: first, an initialization calculation is performed, and data feature analysis is performed according to the deformation sequence characteristics of the earth-rock dam to generate a learning rate learning_rate, an output neuron output_dim, an iteration number num_epochs, a value range lval_range[:], oval_range[:], and eval_range[:], and an initial optimal objective function value best_fit is set to satisfy best_fit>0; then the objective function value obj_fun is calculated, and if it satisfies obj_ fun>optimal objective function value best_fit, then assign best_fit=obj_fun; then use the PZOA optimization algorithm to optimize and update the population position corresponding to the learning rate learning_rate, output neuron output_dim, and number of iterations num_epochs, and calculate obj_fun and best_fit; finally, determine whether the number of iterations num_epochs is less than the maximum number of iterations max_iter. If not, stop the calculation, otherwise output the optimal parameters of the learning rate, output neuron, and number of iterations and the earth-rock dam deformation measurement value sequence prediction result Prediction_y[:] calculated by the TCN model.

[0060] And, a deformation prediction system for earth-rock dam based on PZOA-VMD-TCN, characterized by comprising: a preprocessing module for collecting dam deformation data and environmental quantity data, and integrating them into a dam deformation measurement value sequence based on a time series, dividing the training set and the test set according to a set ratio, and normalizing all the data;

[0061] An optimization module is used to optimize the parameter mode number K and penalty factor alpha in the VMD decomposition parameters using the PZOA optimization algorithm;

[0062] The VMD decomposition module is used to apply the optimal value to the VMD decomposition algorithm to decompose the deformed data to obtain a number of IMF subsequence values;

[0063] A screening module is used to screen the deformation influencing factors based on the deformation influencing factors using the PLS method to find out the optimal input deformation influencing factor set and the number of principal components corresponding to each IMF component;

[0064] And a prediction module is used to construct a deformation prediction input data set according to the obtained optimal input deformation factor set, and use the PZOA-TCN model to predict the deformation value of the earth-rock dam.

[0065] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a deformation prediction method for earth-rock dam based on PZOA-VMD-TCN as described above are implemented.

[0066] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a deformation prediction method for earth-rock dams based on PZOA-VMD-TCN as described above.

[0067] Compared with the prior art, the present invention and its preferred solution obtain the original features of the dam deformation measurement sequence according to the signal decomposition method, and obtain the optimal input dam deformation factor set corresponding to each decomposed subsequence through the PLS algorithm; thereby constructing a model prediction input factor set, and then loading it into the model PZOA-TCN to predict the deformation value of the earth-rock dam. Selecting the optimal number of input factor sets can effectively improve the prediction accuracy of the model. In this way, the problem of low prediction accuracy caused by strong nonlinearity and noise in the data in the earth-rock dam deformation prediction method is overcome. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0069] Figure 1 This is a flow chart of a method according to an embodiment of the present invention;

[0070] Figure 2 Schematic diagram of historical measurement values ​​of deformation monitoring points of an earth-rock dam according to an embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail as follows:

[0072] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0073] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0074] The purpose of the present invention is to overcome the problem of low prediction accuracy caused by strong nonlinearity and noise in the data in the deformation prediction method of earth-rock dam, and propose an earth-rock dam deformation prediction method based on PZOA-VMD-TCN. First, collect dam deformation data and environmental quantity data, and integrate them into a dam deformation measurement sequence based on time series, divide the training set and the test set according to the set ratio, and normalize all data at the same time; secondly, use the PZOA optimization algorithm to optimize the VMD decomposition parameters, and apply the optimized values ​​to the VMD decomposition algorithm to decompose the deformation data to obtain several IMF subsequence values; then, based on the deformation influencing factors, use the PLS method to screen the deformation influencing factors to find the optimal input deformation influencing factor set corresponding to each IMF component; finally, construct a deformation prediction input data set based on the obtained optimal input deformation factor set, and apply the PZOA-TCN model to predict the deformation value of the earth-rock dam. The specific implementation flow chart is as follows: Figure 1 As shown. It mainly includes the following 5 steps:

[0075] Step 1: Collect dam deformation data and environmental quantity data, and integrate them into a dam deformation measurement sequence based on time series, divide the training set and test set according to the set ratio, and normalize all data;

[0076] Step 2: Construct the PZOA-VMD model to realize the optimization calculation of two important parameters, the modal number and the penalty factor (K, alpha), and verify the accuracy of the optimization results;

[0077] Step 3: Perform VMD decomposition on the dam deformation measurement sequence and obtain several decomposition sequences <IMF n >, using the MIC algorithm to calculate the correlation coefficient between each IMF and the dam deformation measurement, identify the noisy components, and use the wavelet-Gaussian integrated filtering method to perform complementary double denoising on the noisy components;

[0078] Step 4: Based on the deformation influencing factors, a PLS-based screening technology for deformation influencing factors of earth-rock dams is proposed to find the optimal number of input deformation influencing factor sets corresponding to each IMF component;

[0079] Step 5: Construct the deformation prediction input data set based on the optimal deformation influencing factor set, and apply the PZOA-TCN model to predict the deformation value of the earth-rock dam.

[0080] This embodiment takes a deformation monitoring point of an earth-rock dam as an example. The historical monitoring value of this monitoring point is as follows: Figure 2 As shown, according to the method flow chart, Figure 1 As shown, a detailed description is given:

[0081] Step 1: Collect dam deformation data and environmental quantity data, and integrate them into a dam deformation measurement sequence based on time series. Divide the training set and test set according to the set ratio, and normalize all data at the same time. Specifically:

[0082] 1) According to the deformation monitoring point data of the earth-rock dam, it is sorted into a dam deformation measurement value sequence based on the time series, and divided into a training set and a test set according to the proportional factor proportion = 0.8:

[0083] disp_set[:]=DISP_SET[:]*proportion (1)

[0084] 2) Normalize the dam measurement series data:

[0085] max=disp_set[:] max (2)

[0086] min = disp_set[:] min

[0087]

[0088] Step 2: Construct the PZOA-VMD model to realize the optimization calculation of two important parameters, the modal number and the penalty factor (K, alpha), and verify the accuracy of the optimization results;

[0089] Specifically:

[0090] 1) Initialize the position X of the zebra individuals in the population i and the velocity vector V i , set the algorithm parameters: maximum number of iterations, population size, inertia weight w, learning factors c1 and c2, and exploration and development weight factors b1, b2, b3;

[0091] 2) Update the position vector and velocity vector of the individuals in the population, as shown in equations (4) and (3): r1, r2, r3, r4, r5 are the generated random numbers, G(t) is the global optimal position, P i is the optimal position of an individual;

[0092] V i (t+1)=w×V i (t)+c1×r4×(P i (t)-X i (t)+c2×r5×(G(t)-X i (t)) (3)

[0093] X i (t+1)=X i(t)+b1×r1×(G(t)-X i (t))+b2×r2×(P i (t)-X i (t)+b3×r3+V i (t+1)(4)

[0094] 3) Confirm that the individual position is within the allowed search space and update the fitness value of each individual;

[0095] 4) Update the individual optimal position and the global optimal position and assign values;

[0096] Pi=Xi (5)

[0097] G(t)=Xi (6)

[0098] 5) Load the training set data into the PZOA algorithm, find the optimal values ​​of the number of modes and penalty factors (K, alpha), and verify them

[0099] 6) Verify the optimal value of the final output.

[0100] Step 3: Decompose the dam deformation measurement sequence using the optimal value of the VMD decomposition parameter obtained in step 2; and use the MIC algorithm to calculate the correlation coefficient between each IMF and the dam deformation measurement, identify the components containing noise, and use the wavelet-Gaussian integrated filtering method to perform complementary double denoising on the components containing noise, specifically:

[0101] 1) Based on the VMD parameters obtained by optimization, the dam deformation measurement value sequence is decomposed to obtain the decomposed dam deformation measurement value subsequence Num-IMF:

[0102]

[0103] 2) The MIC algorithm is used to calculate the correlation coefficient between each IMF component and the dam deformation sequence, and to identify the IMF components containing noise:

[0104]

[0105] 3) The obtained MIC values ​​are normalized and the average value is calculated as the criterion for judging whether the subsequence contains noise:

[0106]

[0107]

[0108] 4) Use Gaussian filtering and wavelet filtering algorithms to perform overlapping denoising on the identified noisy subsequences.

[0109] Step 4: Use the PLS-based screening technology for deformation influencing factors of earth-rock dams to find the optimal deformation influence subset corresponding to each IMF, and use it as the final input deformation factor set, specifically:

[0110] 1) Standardize the dam deformation factor set and the dam deformation measurement value subsequence Num-IMF to obtain standardized data, and then divide the training set and the test set according to the proportional factor;

[0111]

[0112] 2) Solve the principal components of the independent variables and dependent variables, and determine the number of principal components P that meet the conditions;

[0113] 3) According to the P value corresponding to the IMF, the optimal deformation factor set corresponding to each IMF can be obtained:

[0114]

[0115] Step 5: According to the optimal input factor set (D 1 , D 2 D K ), forming an input factor set for each IMF component, and applying the PZOA-TCN prediction model to calculate the deformation prediction value of each IMF subsequence, and finally superimposing the prediction values ​​of all subsequences to obtain the final dam deformation prediction value. Specifically:

[0116] 1) Initialize the calculation, perform data feature analysis according to the deformation sequence characteristics of the earth-rock dam, generate the learning rate learning_rate, output neuron output_dim, number of iterations num_epochs, value range lval_range[0.001-0.01], oval_range[1-100], eval_range[0-200], and set the initial optimal objective function value best_fit=1, satisfying best_fit>0;

[0117] 2) Calculate the objective function value obj_fun again. If it satisfies obj_fun>optimal objective function value best_fit, assign best_fit=obj_fun;

[0118] 3) Then use the PZOA optimization algorithm to optimize and update the population position corresponding to the learning rate learning_rate, output neuron output_dim, and number of iterations num_epochs, and calculate obj_fun and best_fit;

[0119] 4) Determine whether the number of iterations num_epochs is less than the maximum number of iterations max_iter. If not, stop the calculation and output the optimal parameters of the learning rate, output neurons, number of iterations and the prediction result Prediction_y[:] of the earth-rock dam deformation measurement sequence calculated by the TCN model. If it is satisfied, repeat the above steps 2) and 3).

[0120] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0121] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0122] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0123] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0124] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0125] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

[0126] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive other forms of a deformation prediction method for earth-rock dam based on PZOA-VMD-TCN under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.

Claims

1. A deformation prediction method for earth-rock dam based on PZOA-VMD-TCN, characterized by: Firstly, the dam deformation data and environmental quantity data are collected and integrated into a dam deformation measurement sequence based on time series. The training set and the test set are divided according to the set ratio, and all the data are normalized. Then, the PZOA optimization algorithm is used to optimize the parameter mode number K and the penalty factor alpha in the VMD decomposition parameters. The optimized values ​​are applied to the VMD decomposition algorithm to decompose the deformation data to obtain several IMF subsequence values. Based on the deformation influencing factors, the PLS method is used to screen the deformation influencing factors to find the optimal input deformation influencing factor set and the number of principal components corresponding to each IMF component. Finally, the deformation prediction input data set is constructed according to the obtained optimal input deformation factor set, and the PZOA-TCN model is used to predict the deformation value of the earth-rock dam.

2. The deformation prediction method of earth-rock dam based on PZOA-VMD-TCN according to claim 1 is characterized by: The training set and test set are selected according to the following ratios: disp_set[:]=DISP_SET[:]*proportion proportion is the selected proportionality factor; The specific steps for normalizing the data are as follows: max=disp_set[:] max min=disp_set[:] min in, They are the actual value of the dam deformation measurement and the value obtained after normalization; min, max are the minimum and maximum values ​​in the sample data of dam deformation measurement respectively.

3. The deformation prediction method of earth-rock dam based on PZOA-VMD-TCN according to claim 1 is characterized by: The specific method of optimizing the parameter mode number K and penalty factor alpha in the VMD decomposition parameters by using the PZOA optimization algorithm is as follows: in the initialization stage of PZOA optimization of VMD, each dimension of the particle corresponds to a parameter, and each particle corresponds to an initial value of alpha and K, which are input into the fitness function for decomposition to obtain the corresponding fitness value: Initialize the position X of the zebra individuals in the population i and the velocity vector V i , the setting parameters include: maximum number of iterations, population size, inertia weight w, learning factors c1 and c2, and exploration and development weight factors b1, b2, b3; Then the following iteration process is performed until the maximum number of iterations is reached: The position vector and velocity vector of the individuals in the population are updated as shown in the following formula: r1, r2, r3, r4, r5 are the generated random numbers, G(t) is the global optimal position, P i is the optimal position of an individual; V i (t+1)=w×V i (t)+c1×r4×(P i (t)-X i (t)+c2×r5×(G(t)-X i (t)) X i (t+1)=X i (t)+b1×r1×(G(t)-X i (t))+b2×r2×(P i (t)-X i (t)+b3×r3+V i (t+1) Confirm that the individual position is within the allowed search space and update the fitness value of each individual; Update the individual optimal position and the global optimal position, corresponding to the global optimal fitness value, and assign a value; Pi=Xi G(t)=Xi Finally, the output optimal value is verified: the fitness function value is calculated within the obtained optimal value range, and if the calculated result is consistent with the optimization result, it is output.

4. The deformation prediction method of earth-rock dam based on PZOA-VMD-TCN according to claim 1 is characterized by: In the process of applying the optimal value to the VMD decomposition algorithm to decompose the deformation data to obtain a number of IMF subsequence values, the dam deformation measurement value decomposition sequence is calculated based on the result of the optimal calculation. <IMF n >The MIC algorithm is used to calculate the correlation coefficient between each IMF and the measured deformation of the dam, identify the components containing noise, and use the wavelet-Gaussian integrated filtering method to perform complementary double denoising on the components containing noise; Specifically: Substitute the best K and alpha values ​​obtained by optimization into the VMD decomposition algorithm, load the data of the dam deformation measurement sequence, and obtain the decomposition subsequence set Num-IMF based on the number of K; Then, according to the maximum information coefficient MIC algorithm, the correlation coefficients between each decomposition sequence IMF and the original measured value of the dam are calculated to identify the noise sequence in the decomposition sequence; In the MIC algorithm, we first define the decomposition subsequence set Num-IMF, and MI[v:X] is recorded as the element v in IMF. k The mutual information value between the original displacement measurement sequence X(t) and p(t); v X[,] represents the joint probability distribution function; p v , pX represents the marginal probability distribution function, as shown below: Then calculate the mutual information value between each subsequence and the original displacement measurement sequence: … Until the mutual information values ​​of all subsequences are obtained, take: I[v:X]=maxMI[v:X] That is, the maximum mutual information value is obtained, and then the maximum mutual information value is normalized: Then add the MIC values ​​of all subsequences to get the average value: After standardization, the MIC value range is [0,1]; if MIC k >MIC m , it is considered that the Kth subsequence is strongly correlated with the original dam deformation measurement value and there is no noise; otherwise, there is noise and the subsequence needs to be denoised and reconstructed.

5. The deformation prediction method of earth-rock dam based on PZOA-VMD-TCN according to claim 4 is characterized by: For the subsequences with noise, Gaussian filtering and wavelet filtering are used to perform overlapping denoising on the original data.

6. The deformation prediction method of earth-rock dam based on PZOA-VMD-TCN according to claim 1 is characterized by: Based on the deformation influencing factors, the PLS method is used to screen the deformation influencing factors, and the optimal input deformation influencing factor set and the number of principal components corresponding to each IMF component are found as follows: The dam deformation factor set and the dam deformation measurement value subsequence set Num-IMF are standardized to obtain the standardized data, and then the training set and the test set are divided according to the proportional factor; the dam deformation factor set I(t) is used as the independent variable, and the dam deformation measurement value sequence subsequence set Num-IMF is used as the dependent variable: Num_IMF=[IMF1,IMF2,IMF3,....,IMF K ] n×K I(t)=[I1,I2,I3,....,I q ] n×q Where q is the total number of factor sets; Solve the principal components u1 and v1 of the independent and dependent variables; then establish the independent variable I and the dependent variable Num denoted as F - The regression variance of IMF and the first principal component u1, v1, and the residual matrix I1, F1 are calculated; I1, F1 are used to replace I, F to form new independent variables and dependent variables, and the first principal component u2, t2 of the new independent variable and dependent variable are solved, which are the second principal component of the original independent variable and dependent variable; the regression equation of the residual matrix composed of the new independent variable I1 and dependent variable F1 and the second principal component u2, t2 is established, and the residual matrix I2, F2 is calculated; Repeat the above steps until all principal components are found or the conditions are met; then perform a cross-test to determine the number of principal components P that meet the conditions; each IMF has a corresponding P value, and according to the dam deformation factor set and P value, find the optimal deformation factor set corresponding to each IMF: D1=I×P1×(P1 T ×P1) -1 <h2 style=";text-align:left;direction:ltr">D2 = I × P2 × (P2<h2 style=";text-align:left;direction:ltr"> T <h2 style=";text-align:left;direction:ltr"> ×P2)<h2 style=";text-align:left;direction:ltr"> -1 … D K =I×P K ×(P K T ×P K ) -1 D is the final optimal input deformation factor set.

7. The deformation prediction method of earth-rock dam based on PZOA-VMD-TCN according to claim 1 is characterized by: The deformation prediction input data set is constructed according to the obtained optimal input deformation factor set, and the deformation value of the earth-rock dam is predicted by using the PZOA-TCN model. Specifically, an initialization calculation is first performed, and data feature analysis is performed according to the deformation sequence characteristics of the earth-rock dam to generate a learning rate learning_rate, an output neuron output_dim, an iteration number num_epochs, a value range lval_range[:], oval_range[:], and eval_range[:], and an initial optimal objective function value best_fit is set to satisfy best_fit>0; then the objective function value obj_fun is calculated, and if it satisfies obj_fu n>optimal objective function value best_fit, then assign best_fit=obj_fun; then use the PZOA optimization algorithm to optimize and update the population position corresponding to the learning rate learning_rate, output neuron output_dim, and number of iterations num_epochs, and calculate obj_fun and best_fit; finally, determine whether the number of iterations num_epochs is less than the maximum number of iterations max_iter. If not, stop the calculation, otherwise output the optimal parameters of the learning rate, output neuron, and number of iterations and the earth-rock dam deformation measurement value sequence prediction result Prediction_y[:] calculated by the TCN model.

8. A deformation prediction system for earth-rock dam based on PZOA-VMD-TCN, characterized in that: include: The preprocessing module is used to collect dam deformation data and environmental quantity data, and integrate them into a dam deformation measurement sequence based on a time series, divide the training set and the test set according to the set ratio, and normalize all the data; The optimization module is used to optimize the parameter mode number K and penalty factor alpha in the VMD decomposition parameters using the PZOA optimization algorithm; The VMD decomposition module is used to apply the optimal value to the VMD decomposition algorithm to decompose the deformed data to obtain a number of IMF subsequence values; A screening module is used to screen the deformation influencing factors based on the deformation influencing factors using the PLS method to find out the optimal input deformation influencing factor set and the number of principal components corresponding to each IMF component; And a prediction module is used to construct a deformation prediction input data set according to the obtained optimal input deformation factor set, and use the PZOA-TCN model to predict the deformation value of the earth-rock dam.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the earth-rock dam deformation prediction method based on PZOA-VMD-TCN as described in any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a deformation prediction method for earth-rock dam based on PZOA-VMD-TCN as described in any one of claims 1 to 7 are implemented.

Images