Side slope rock-soil body parameter inversion method based on time sequence correlation analysis

Abstract

The invention discloses a slope rock-soil body parameter inversion method based on time sequence correlation analysis, and belongs to the technical field of civil engineering operation and maintenance. According to the method, the time sequence change and spatial correlation of the slope rock-soil material parameters are fully considered, and the accuracy of slope rock-soil body parameter inversion is improved. On-site monitoring data and geological exploration information are combined, a solution space of rock and soil parameters is constructed by adopting a positive algorithm, the solution space is efficiently optimized by utilizing a multi-objective optimization algorithm (SDMOM), the space-time consistency and rationality of material parameters are ensured, a parameter inversion problem is quickly solved by adopting a meta-heuristic algorithm, the calculation efficiency is remarkably improved, and the calculation cost is reduced. The calculation time is reduced, and the influence on displacement inversion caused by the sensing system error is further reduced.

Description

Technical Field

[0001] The invention relates to the technical field of civil engineering operation and maintenance, and in particular to a slope rock and soil parameter inversion method based on time series correlation analysis. Background Art

[0002] The physical and mechanical parameters of the rock and soil of the slope are the basis for slope stability analysis and structural analysis. At present, the physical and mechanical parameters of the rock and soil of the slope are often obtained through geological survey. However, in actual engineering, geological survey is usually carried out in the design stage, and the geological parameters obtained span the design, construction and operation stages. However, as a structure directly exposed to the natural environment, the slope is subjected to different physical and chemical effects during service, and the physical and mechanical parameters of the rock and soil have dynamic change characteristics. In particular, the effects of atmospheric precipitation and groundwater have been pointed out that water content will significantly change the physical and mechanical parameters of the soil. Therefore, in order to accurately analyze the stability state and service performance of the slope during the operation period, it is necessary to obtain the change law of the internal parameters of the slope with the service time. Traditional geological survey can only obtain the rock and soil parameters of some sparse measurement points in the slope area, and is limited by the high labor and economic costs. It is also destructive to the slope and is difficult to carry out for a long time in actual engineering.

[0003] Geotechnical parameter inversion has important theoretical and practical significance in geotechnical engineering. Its core lies in estimating the physical and mechanical parameters of the slope rock and soil, such as internal friction angle and cohesion, based on the monitoring data or test data of the on-site slope through a specific inversion method. Inversion technology updates the parameters by combining observation data with numerical simulation, which can effectively narrow the gap between the original geological survey parameters and the actual engineering conditions, thereby providing a reliable basis for slope safety operation and risk assessment. Although geotechnical parameter inversion has broad application prospects in slope stability analysis, the inversion problem is essentially an ill-posed problem, that is, there is a problem of non-uniqueness of the solution and sensitivity to the input data. Different observation data may lead to different inversion results, increasing the uncertainty of the model parameters.

[0004] In order to meet these challenges, researchers have begun to combine advanced optimization algorithms with machine learning methods in recent years to improve the accuracy and efficiency of the inversion process. For example, intelligent algorithms such as genetic algorithms and particle swarm optimization have been initially applied in geotechnical parameter inversion. These methods optimize the inversion results through global search and reduce the influence of local extreme values. Although the inversion method based on machine learning has improved the efficiency of inversion to a certain extent, the problem of determining the solution space for the inversion ill-posed problem is still a key issue restricting the high-precision inversion of soil parameters. Summary of the invention

[0005] In view of the above-mentioned deficiencies in the prior art, the present invention provides a slope rock and soil parameter inversion method based on time series correlation analysis.

[0006] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0007] A slope rock and soil parameter inversion method based on time series correlation analysis includes the following steps:

[0008] S1. Obtain monitoring / detection data of slope rock and soil parameters at different times, and perform data cleaning on the acquired monitoring / detection data;

[0009] S2. Use the positive algorithm to obtain the slope rock and soil parameter solution space that meets the requirements for the monitoring / detection data after data cleaning;

[0010] S3, based on the slope rock and soil parameter solution space obtained in S2, a multi-objective optimization algorithm model considering the spatial correlation of structured data is established;

[0011] S4. Use a multi-objective optimization algorithm to solve the spatial time series and obtain the inversion parameters of the slope rock and soil at different times.

[0012] Further, the following steps are included:

[0013] S11, sorting the monitoring / detection data of the slope rock and soil parameters at different times according to the time series, deleting the null values ​​in the sequence and normalizing them;

[0014] S12, establishing an additive model to decompose the ordered time series in S11 and extract the residuals of the monitoring / detection data of the slope rock and soil parameters at different times;

[0015] S13, the residuals at different times are combined into a residual sequence and the size of each data in the residual sequence is compared with the average of the sequence. If the preset conditions are met, it is regarded as abnormal data;

[0016] S14. Eliminate abnormal data to obtain a cleaned time series.

[0017] Furthermore, the additive model decomposition in S12 is expressed as:

[0018]

[0019] In the formula, Y(t) represents the actual observed value at time t, Represents the trend item, represents the periodic term, R(t) represents the residual term at time t, and a i 、b i 、c i ,d i and e iare the i-th learnable parameters, i is an integer greater than or equal to 0, n is the number of fitting items and satisfies:

[0020]

[0021] Where R(t) n Represents the residual at time t when it is set to n.

[0022] Furthermore, the S3 specifically includes the following steps:

[0023] S31, constructing inversion parameters according to the spatial position of the slope, with rows representing the horizontal direction in geological survey and columns representing the vertical direction in geological survey, forming a two-dimensional matrix g, and the position part that does not belong to the slope body is replaced by the number 0;

[0024] S32, taking the slope rock and soil parameter solution space obtained in S2 as the optimization domain, uniformly randomly sampling the sequence number in each subset of the optimization domain to obtain a set of solutions, and repeating multiple times to obtain the initial solution set;

[0025] S33, based on the correlation algorithm, respectively calculate the correlation between the soil material parameter matrix at time t and time t+1 in each set of solutions in the initial solution set;

[0026] S34, using L1 norm to represent the inverted monitoring error, and calculating the fitness function in combination with the correlation obtained in S33;

[0027] S35. A meta-heuristic algorithm is used, with the initial solution set as the initial population, and a multi-objective optimization algorithm is established that considers the spatial correlation of structured data in combination with the fitness function.

[0028] Furthermore, the specific calculation method of the correlation in S33 is:

[0029]

[0030] In the formula, C(t) is the correlation between geotechnical parameters at time t and t+1. The smaller the value, the better the correlation. t,i,j At time t, the matrix index is the value at position i and j, im is the maximum index value in the i direction, jm is the maximum index value in the j direction, and smooth is the minimum value set to ensure that the formula makes sense.

[0031] Furthermore, the specific calculation method of the fitness function in S34 is:

[0032]

[0033] In the formula, is the value of the monitoring point obtained by inversion parameter calculation, S tThe value obtained after the monitoring point is cleaned, smooth is the minimum value set to ensure the formula is meaningful, C(t) is the material correlation, P(G i ) is G i The fitness of the solution.

[0034] Furthermore, the meta-heuristic algorithm in S35 is specifically expressed as:

[0035]

[0036] In the formula, represents the flying speed of particle i in the dth dimension in the kth iteration, represents the position of particle i in the dth dimension in the kth iteration, c1 and c2 are acceleration constants, r1 and r2 are random numbers, and w is the inertia weight.

[0037] The present invention has the following beneficial effects:

[0038] The beneficial effects of the present invention are as follows: Different from the traditional slope rock and soil parameter inversion method, the present invention proposes a sequential inversion method based on meta-inspiration analysis, which fully considers the time series changes and spatial correlation of slope rock and soil material parameters, and improves the accuracy of slope rock and soil parameter inversion. By combining field monitoring data with geological exploration information, a forward algorithm is used to construct the solution space of rock and soil parameters, and a multi-objective optimization algorithm (SDMOM) is used to efficiently optimize the solution space to ensure the temporal and spatial consistency and rationality of material parameters. A meta-inspiration algorithm is used to quickly solve the parameter inversion problem, which significantly improves the calculation efficiency and reduces the calculation time. In addition, the present invention carries out inversion by monitoring displacement, and designs a monitoring data cleaning algorithm, which further reduces the impact of the perception system error on the displacement inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of the slope rock and soil mass parameter inversion method for time series correlation analysis of the present invention;

[0040] Figure 2 This is a comparison chart of time series data cleaning in an embodiment of the present invention;

[0041] Figure 3 This is a diagram of the timing inversion result of an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0043] like Figure 1 As shown in the figure, a sequential inversion method for slope rock and soil parameters based on meta-inspiration analysis includes the following steps:

[0044] S1. Analyze and organize the monitoring / detection data at different times, and clean the monitoring / detection data;

[0045] S2. Based on the slope monitoring / detection information and geological survey information, a forward algorithm is used to obtain the slope rock and soil parameter solution space that meets the analysis requirements, and the time series of the slope rock and soil solution space is obtained;

[0046] S3. Establish a multi-objective optimization algorithm SDMOM (Structured Data Multi-objective Optimization Method) that considers the spatial correlation of structured data;

[0047] S4, using SDMOM to analyze the time series of the solution space to obtain the inversion parameters of the slope rock and soil at different times;

[0048] Furthermore, in S1, it is required that the monitoring / detection information can include the displacement data of the slope, and data cleaning includes the following sub-steps:

[0049] S11, the time series is divided into S 0 ={s 0 1,s 0 2, …,s 0 t} to arrange, delete the data with null values ​​in the original sequence, ensure that the data at each moment is a real number, and normalize it. The normalization calculation formula used is:

[0050]

[0051] In the formula, represents the normalized sedimentation value, X i Indicates the i-th sedimentation value to be normalized, i = 1, 2, 3, ...; max_x is the maximum sedimentation value in the sample, and min_x is the minimum sedimentation value in the sample.

[0052] S12. Establish an additive model to decompose the original sequence, and its expression is:

[0053]

[0054] In the formula, Y(t) represents the actual observed value at time t, Represents the trend item, represents the periodic term, R(t) represents the residual term at time t, and ai , b i 、c i ,d i and e i are the i-th learnable parameters, i is an integer greater than or equal to 0, and n is the number of fitting items, which is used to control the accuracy of fitting. If the following formula is satisfied, it means that the setting of n is reasonable:

[0055]

[0056] Where R(t) n Represents the residual at time t when it is set to n.

[0057] When performing time series decomposition, commonly used solution methods include: quasi-Newton method, conjugate gradient method, L-BFGS (Limited-memory Broyden-Fletcher-Goldfarb-Shanno) algorithm, trust region method, etc. Among them, the L-BFGS algorithm is a variant of the BFGS algorithm. It uses limited memory space to store and calculate the approximate inverse of the Hessian matrix, thereby overcoming the limitations of the BFGS algorithm in memory usage. The L-BFGS algorithm is very popular in machine learning, especially for large-scale optimization problems.

[0058] S13, the residuals at different times are combined into a residual sequence R = {r1, r2, ..., r t}, compare the size of each data in the sequence with the average of the sequence, if it satisfies The data at time i is considered to be an outlier. is the average value of the R series.

[0059] S14, extract the data corresponding to the abnormal value at the time, and obtain the cleaned time series S = {s1,s 2, …,s t}.

[0060] Furthermore, in S2, the geological survey information is required to include the physical and mechanical parameter information to be inverted, and the forward algorithm adopted uses the finite element method as an analysis tool, wherein the judgment criterion for the analysis requirement is: the average error of the calculated displacement data is less than the monitoring error of the monitoring sensor.

[0061] In order to ensure consistency with the site, the variability and correlation of the soil are usually considered when performing finite element analysis, and random finite element analysis is carried out. When implementing random finite element analysis, firstly, based on the analysis of the correlation and variability parameters of the parameters to be inverted in the geological survey information, the random field is obtained according to the Cholesky decomposition. Combined with the standard normal distribution, the relevant random sample matrix is ​​obtained, and finally the standard normal distribution is transformed by equal probability to obtain the non-standard normal distribution random field sample. The calculation process of the distribution conversion is as follows:

[0062]

[0063] In the formula, ξ(x) is a random number sample under normal distribution, is the Cholesky decomposition upper triangular matrix of the cross-correlation coefficient, L2 is the Cholesky decomposition upper triangular matrix of the autocorrelation coefficient, Φ(·) is the cumulative distribution function of the standard normal distribution, is the inverse function of the target edge cumulative distribution; H i (x) is a discrete random field with correlation.

[0064] The random parameters are input into the finite element to carry out calculations, and the displacement data obtained by the finite element calculations at the monitoring points are extracted.

[0065] Furthermore, in S3, the calculation steps of the established multi-objective optimization algorithm SDMOM considering the spatial correlation of structured data are as follows:

[0066] For the problem to be optimized, P(G)→max(P(G)), where G is a set of parameter spaces input to the optimization model, including a set of solutions in the solution space obtained by inverting various physical and mechanical parameters at different times, P(G) is the adaptability function of the parameter group G, and the larger the P(G) function value is, the closer the required parameters are to the optimization goal, and max(P(G)) is the maximum function value obtained under various parameter G states.

[0067] S31, organizing the inversion parameters according to the spatial position of the slope, with rows representing the horizontal direction in geological survey and columns representing the vertical direction in geological survey, to form a two-dimensional matrix g, and replacing the position part of the matrix that does not belong to the slope body with the number 0;

[0068] S32, select the solution space obtained in S2 as the optimization domain H = {ζ1, ζ2, ..., ζ t}, where ζ t It represents the solution space obtained by the inversion parameters of the geotechnical field at time t. Each parameter combination is sorted according to the accuracy and assigned a serial number. In each subset of the optimization domain H, the serial number is uniformly randomly sampled to obtain a set of solutions G1 = {g1, g2, ..., g t}, repeat this process j times to obtain the initial solution set G = {G1, G2, ..., G j};

[0069] S33, according to the material correlation algorithm, respectively calculate the correlation between the soil material parameter matrix at time t and t+1 in each group of solutions in the solution set G, and the calculation formula is as follows:

[0070]

[0071] In the formula, C(t) is the correlation between geotechnical parameters at time t and t+1. The smaller the value, the better the correlation. t,i，j At time t, the matrix index is the value at position i and j, im is the maximum index value in the i direction, jm is the maximum index value in the j direction, and smooth is the minimum value set to ensure that the formula makes sense.

[0072] S34, using L1 norm to represent the inverted monitoring error, using C(t) to represent the similarity of material parameters, combining the similarity and monitoring error to calculate the fitness function value, the calculation formula is as follows:

[0073]

[0074] In the formula, is the value of the monitoring point obtained by inversion parameter calculation, S t The value obtained after the monitoring point is cleaned, smooth is the minimum value set to ensure the formula is meaningful, C(t) is the material correlation, P(G i ) is G i The fitness of the solution.

[0075] S35, using the meta-heuristic algorithm, with G = {G1, G2, ..., G j} is the initial population, with P(G i )A multi-objective optimization algorithm SDMOM is established for the fitness function, which considers the spatial correlation of structured data.

[0076] Metaheuristic algorithm is an optimization algorithm, including simulated annealing algorithm, genetic algorithm, particle swarm optimization algorithm, ant colony algorithm and gray wolf optimization algorithm. The idea of ​​particle swarm optimization algorithm comes from the study of bird predation behavior. By simulating the behavior of bird flocks flying and foraging, the group can achieve the optimal goal. It does not have the crossover and mutation operations in genetic algorithms, but finds the global optimal solution by following the currently known optimal value. Compared with other optimization algorithms, particle swarm optimization algorithm requires few parameters to be adjusted, converges quickly, and is simple and easy to implement. We take the solution space H = {ζ1, ζ2, ..., ζ t The inversion parameter matrix number of each solution is taken as the item to be optimized, and P(G i ) is used as the fitness function, and the particle velocity and position update formula used are as follows:

[0077]

[0078] In the formula, represents the flying speed of particle i in the dth dimension in the kth iteration, represents the position of particle i in the dth dimension in the kth iteration, c1 and c2 are acceleration constants used to adjust the learning step size, r1 and r2 are random numbers between [0, 1], used to increase the randomness of the search, and w is the inertia weight, a non-negative number, used to adjust the search range of the solution space.

[0079] Furthermore, for S4, the method established in S3 is used to solve the problem, and the solution with the highest fitness function of S34 is selected as the inversion parameter of the slope rock and soil mass at different times.

[0080] The following describes the invention in conjunction with specific embodiments.

[0081] In S1, displacement monitoring of the slope is carried out while geological survey is carried out to obtain the soil parameter status of the slope at the beginning of monitoring, analyze and organize the monitoring / detection data at different times, and perform data cleaning on the monitoring / detection data;

[0082] The monitored time series is converted into S 0 ={s 0 1,s 0 2, ..., s 0 t} to arrange, delete the data with null values ​​in the original sequence, ensure that the data at each moment is a real number, and normalize it. The normalization calculation formula used is:

[0083]

[0084] In the formula, represents the normalized sedimentation value, X i Indicates the i-th sedimentation value to be normalized, i = 1, 2, 3, ...; max_x is the maximum sedimentation value in the sample, and min_x is the minimum sedimentation value in the sample.

[0085] Based on the L-BFGS algorithm, an additive model is established to decompose the original sequence, and its expression is:

[0086]

[0087] In the formula, Y(t) represents the actual observed value at time t, Represents the trend item, represents the periodic term, R(t) represents the residual term at time t, which is the part that cannot be fitted by the trend term and the periodic term, where a i 、b i 、c i ,d i and e iare i learnable parameters, i is an integer greater than or equal to 0, and n is the number of fitting items, which is used to control the accuracy of fitting. If the following formula is satisfied, it means that the setting of n is reasonable:

[0088]

[0089] Where R(t) n Represents the residual at time t when it is set to n.

[0090] The residuals at different times are combined into a residual sequence R = {r1, r2, ..., r t}, compare the size of each data in the sequence with the average of the sequence, if it satisfies The data at time i is considered to be an outlier. is the average value of the R series.

[0091] Extract the data corresponding to the outliers and obtain the cleaned time series S = {s1,s 2, …,s t},like Figure 2 shown.

[0092] In S2, based on the slope monitoring / detection information and geological survey information, a positive algorithm is used to obtain the slope rock and soil parameter solution space that meets the analysis requirements, and the time series of the slope rock and soil solution space is obtained; in order to ensure consistency with the site, the variability and correlation of the soil are usually considered during finite element analysis, and random finite element analysis is carried out. When implementing random finite element analysis, firstly, based on the correlation and variability parameters of the parameters to be inverted in the geological survey information, the random field is obtained according to the Cholesky decomposition. Combined with the standard normal distribution, the relevant random sample matrix is ​​obtained, and finally the standard normal distribution is transformed by equal probability to obtain the non-standard normal distribution random field sample. The calculation process of the distribution conversion is as follows:

[0093]

[0094] In the formula, ξ(x) is a random number sample under normal distribution, is the Cholesky decomposition upper triangular matrix of the cross-correlation coefficient, L2 is the Cholesky decomposition upper triangular matrix of the autocorrelation coefficient, Φ(·) is the cumulative distribution function of the standard normal distribution, is the inverse function of the target edge cumulative distribution; H i (x) is a discrete random field with correlation.

[0095] The random parameters are input into the finite element to carry out calculations, the displacement data obtained by the finite element calculation of the monitoring points are extracted, and at each monitoring moment, the parameter field in which the error between the calculated displacement data of the monitoring points and the cleaned monitoring data is less than the error measured by the system is selected as the solution space at that moment.

[0096] In S3, a multi-objective optimization algorithm SDMOM (Structured Data Multi-objective Optimization Method) considering the spatial correlation of structured data is established. For the problem to be optimized, P(G)→max(P(G)), where G is a set of parameter spaces input to the optimization model, including a set of solutions in the solution space obtained by inverting various physical and mechanical parameters at different times, P(G) is the adaptability function of the parameter group G, and the larger the P(G) function value, the closer the required parameters are to the optimization goal, and max(P(G)) is the maximum function value obtained under various different parameter G states.

[0097] The inversion parameters are organized according to the spatial position of the slope, with rows representing the horizontal direction in geological exploration and columns representing the vertical direction in geological exploration, forming a two-dimensional matrix g. The positions that do not belong to the slope body in the matrix are replaced by the number 0; the solution space obtained in S2 is selected as the optimization domain H = {ζ1, ξ2, ..., ζ t}, where ζ t It represents the solution space obtained by the inversion parameters of the geotechnical field at time t. Each parameter combination is sorted according to the accuracy and assigned a serial number. In each subset of the optimization domain H, the serial number is uniformly randomly sampled to obtain a set of solutions G1 = {g1, g2, ..., g t}, repeat this process j times to obtain the initial solution set G = {G1, G2, ..., G j};

[0098] According to the material correlation algorithm, the correlation of the soil material parameter matrix at time t and t+1 in each solution in the solution set G is calculated respectively. The calculation formula is as follows:

[0099]

[0100] In the formula, C(t) is the correlation between geotechnical parameters at time t and t+1. The smaller the value, the better the correlation. t，i，j At time t, the matrix index is the value at position i and j, im is the maximum index value in the i direction, jm is the maximum index value in the j direction, and smooth is the minimum value set to ensure that the formula makes sense.

[0101] The L1 norm is used to represent the inverted monitoring error, and C(t) is used to represent the similarity of material parameters. The fitness function value is calculated by combining the similarity and the monitoring error. The calculation formula is as follows:

[0102]

[0103] In the formula, is the value of the monitoring point obtained by inversion parameter calculation, St The value obtained after the monitoring point is cleaned, smooth is the minimum value set to ensure the formula is meaningful, C(t) is the material correlation, P(G i ) is G i The fitness of the solution.

[0104] Using the meta-heuristic algorithm, G = {G1, G2, ..., G j} is the initial population, with P(G i )A multi-objective optimization algorithm SDMOM is established for the fitness function, which considers the spatial correlation of structured data.

[0105] In this embodiment, the particle swarm optimization algorithm is used for optimization. We take the solution space H = {ζ1, ζ2, ..., ζ t The inversion parameter matrix number of each solution is taken as the item to be optimized, and P(G i ) is used as the fitness function, and the particle velocity and position update formula used are as follows:

[0106]

[0107] In the formula, represents the flying speed of particle i in the dth dimension in the kth iteration, represents the position of particle i in the dth dimension in the kth iteration, c1 and c2 are acceleration constants used to adjust the learning step size, r1 and r2 are two random numbers between [0, 1], used to increase the randomness of the search, and w is the inertia weight, a non-negative number, used to adjust the search range of the solution space.

[0108] In S4, SDMOM is used to analyze the time series of the solution space to obtain the inversion parameters of the slope rock and soil at different times; the solution with the highest fitness function is selected as the inversion parameters of the slope rock and soil at different times, such as Figure 3 shown.

[0109] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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 produce 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.

[0110] 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.

[0111] 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. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0112] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0113] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A slope rock and soil parameter inversion method based on time series correlation analysis, characterized in that: The steps include: S1. Obtain monitoring / detection data of slope rock and soil parameters at different times, and perform data cleaning on the acquired monitoring / detection data; S2. Use the positive algorithm to obtain the slope rock and soil parameter solution space that meets the requirements for the monitoring / detection data after data cleaning; S3, based on the slope rock and soil parameter solution space obtained in S2, a multi-objective optimization algorithm model considering the spatial correlation of structured data is established; S4. Use a multi-objective optimization algorithm to solve the spatial time series and obtain the inversion parameters of the slope rock and soil at different times.

2. The slope rock and soil mass parameter inversion method based on time series correlation analysis according to claim 1 is characterized in that: The steps include: S11, sorting the monitoring / detection data of the slope rock and soil parameters at different times according to the time series, deleting the null values ​​in the sequence and normalizing them; S12, establishing an additive model to decompose the ordered time series in S11 and extract the residuals of the monitoring / detection data of the slope rock and soil parameters at different times; S13, the residuals at different times are combined into a residual sequence and the size of each data in the residual sequence is compared with the average of the sequence. If the preset conditions are met, it is regarded as abnormal data; S14. Eliminate abnormal data to obtain a cleaned time series.

3. The slope rock and soil mass parameter inversion method based on time series correlation analysis according to claim 2 is characterized in that: The additive model decomposition in S12 is expressed as: In the formula, Y(t) represents the actual observed value at time t, Represents the trend item, represents the periodic term, R(t) represents the residual term at time t, and a i 、b i 、c i ,d i and e i are the i-th learnable parameters, i is an integer greater than or equal to 0, n is the number of fitting items and satisfies: Where R(t) n Represents the residual at time t when it is set to n.

4. The slope rock and soil mass parameter inversion method based on time series correlation analysis according to claim 2 is characterized in that: The S3 specifically includes the following steps: S31, constructing inversion parameters according to the spatial position of the slope, with rows representing the horizontal direction in geological survey and columns representing the vertical direction in geological survey, forming a two-dimensional matrix g, and the position part that does not belong to the slope body is replaced by the number 0; S32, taking the slope rock and soil parameter solution space obtained in S2 as the optimization domain, uniformly randomly sampling the sequence number in each subset of the optimization domain to obtain a set of solutions, and repeating multiple times to obtain the initial solution set; S33, based on the correlation algorithm, respectively calculate the correlation between the soil material parameter matrix at time t and time t+1 in each set of solutions in the initial solution set; S34, using L1 norm to represent the inverted monitoring error, and calculating the fitness function in combination with the correlation obtained in S33; S35. A meta-heuristic algorithm is used, with the initial solution set as the initial population, and a multi-objective optimization algorithm is established that considers the spatial correlation of structured data in combination with the fitness function.

5. The slope rock and soil mass parameter inversion method based on time series correlation analysis according to claim 4 is characterized in that: The specific calculation method of the correlation in S33 is: In the formula, C(t) is the correlation between geotechnical parameters at time t and t+1. The smaller the value, the better the correlation. t,i,j At time t, the matrix index is the value at position i and j, im is the maximum index value in the i direction, jm is the maximum index value in the j direction, and smooth is the minimum value set to ensure that the formula makes sense.

6. The slope rock and soil mass parameter inversion method based on time series correlation analysis according to claim 4 is characterized in that: The specific calculation method of the fitness function in S34 is: In the formula, is the value of the monitoring point obtained by inversion parameter calculation, S t The value obtained after the monitoring point is cleaned, smooth is the minimum value set to ensure the formula is meaningful, C(t) is the material correlation, P(G i ) is G i The fitness of the solution.

7. The slope rock and soil mass parameter inversion method based on time series correlation analysis according to claim 4 is characterized in that: The meta-heuristic algorithm in S35 is specifically expressed as: In the formula, represents the flying speed of particle i in the dth dimension in the kth iteration, represents the position of particle i in the dth dimension in the kth iteration, c1 and c2 are acceleration constants, r1 and r2 are random numbers, and w is the inertia weight.

Images