Construction method and device of slope deformation and damage scale prediction model

Abstract

The invention relates to the technical field of slope engineering, and discloses a slope deformation damage scale prediction model construction method and device, and the method comprises the steps: defining m fuzzy factors affecting slope damage, and carrying out the quantification of the fuzzy factors; defining a fuzzy set of quantized fuzzy factors to obtain m fuzzy factor sets; performing rule base construction on the m fuzzy factor sets to obtain a rule base of fuzzy inference rules; and performing model construction based on the rule base by adopting a fuzzy neural network, and constructing a prediction model. According to the method, by introducing the fuzzy set theory and the fuzzy logic tool, the fuzzy factors in slope deformation and damage are quantified, the adaptability of the prediction model to the complex slope environment is improved, and the precision and reliability of the prediction result are improved; the theoretical model and monitoring data are effectively combined, the slope state can be dynamically evaluated in real time, a scientific decision basis is provided, precautionary measures can be taken in advance, and the slope instability risk can be reduced.

Description

Technical Field

[0001] The invention relates to the technical field of slope engineering, and in particular to a method and device for constructing a slope deformation and failure scale prediction model. Background Art

[0002] Safety assessment and deformation and damage scale prediction of slope engineering face major challenges, mainly because the factors affecting slope stability are numerous and complex, such as the heterogeneity of rock and soil, dynamic changes in groundwater levels, extreme climate events (especially heavy rainfall), etc. These factors have significant uncertainty and ambiguity, and are difficult to accurately quantify using traditional deterministic methods.

[0003] However, most existing technologies are based on simplified mechanical models or empirical formulas, ignoring the comprehensive influence of fuzzy factors, resulting in low accuracy of prediction results and making it difficult to effectively guide the design, construction and monitoring of slope projects. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention proposes a method and device for constructing a slope deformation and failure scale prediction model to solve the above technical problems.

[0005] In a first aspect, a method for constructing a slope deformation and failure scale prediction model is provided, comprising:

[0006] Define m fuzzy factors that affect slope failure and quantify them;

[0007] Define the fuzzy set of quantized fuzzy factors to obtain m fuzzy factor sets, each of which contains all the values ​​of the fuzzy factors and the corresponding membership functions;

[0008] Constructing a rule base for the m fuzzy factor sets to obtain a rule base of fuzzy reasoning rules;

[0009] A fuzzy neural network is used to construct a model based on the rule base to build a prediction model.

[0010] Furthermore, the m fuzzy factors affecting slope failure are defined and quantified, including:

[0011] directly using the fuzzy factors that can be directly quantified among the m fuzzy factors;

[0012] Using a linear interpolation method to quantify the fuzzy factors that cannot be directly quantified among the m fuzzy factors;

[0013] The m fuzzy factors affecting slope failure include geological conditions, rainfall intensity, vegetation development, rock and soil parameters, slope type, slope scale, geological structure, groundwater, construction excavation, and earthquake weathering.

[0014] Furthermore, the fuzzy set of quantized fuzzy factors is defined to obtain m fuzzy factor sets, including:

[0015] Define the values ​​of the corresponding set of fuzzy factors;

[0016] Define the membership functions of the fuzzy sets.

[0017] Furthermore, the rule base is constructed for the m fuzzy factor sets to obtain a rule base of fuzzy inference rules, including:

[0018] According to the historical data of slope deformation and failure and expert experience, k fuzzy reasoning rules are formulated and the matching degree of each rule is calculated;

[0019] Using the Borda ordinal value method to perform a preliminary sorting on the k fuzzy inference rules, and obtaining k fuzzy inference rules after the preliminary sorting;

[0020] The k fuzzy reasoning rules after the initial sorting are sorted again and weighted by using the hierarchical analysis method to obtain the weights corresponding to the k fuzzy reasoning rules;

[0021] The importance of all rules is calculated using the rule matching degree and corresponding weights to determine the final fuzzy reasoning rule base.

[0022] Furthermore, it also includes using a genetic algorithm to optimize the prediction model to obtain an optimized prediction model.

[0023] Furthermore, the method also includes using a cross-validation method to verify the optimized prediction model to verify the feasibility and reliability of the model.

[0024] In a second aspect, a device for constructing a slope deformation and failure scale prediction model is provided, comprising:

[0025] A definition module is used to define m fuzzy factors and fuzzy sets of quantitative factors that affect slope failure;

[0026] A first construction module is used to construct a rule base for the m fuzzy factor sets to obtain a rule base of fuzzy inference rules;

[0027] The second building module is used to construct a model based on the rule base using a fuzzy neural network to build a prediction model.

[0028] Furthermore, it comprises a processor and a memory storing program instructions, wherein the processor is used to execute a slope stability probability assessment method as described in any one of the above when running the program instructions.

[0029] In a third aspect, an electronic device is provided, comprising a device for constructing a slope deformation and damage scale prediction model as described above.

[0030] The invention adopting the above technical solution has the following advantages:

[0031] The present invention introduces mathematical tools such as fuzzy set theory and fuzzy logic to quantify the fuzzy factors in slope deformation and failure, significantly improving the adaptability of the prediction model to complex slope environments, thereby improving the accuracy and reliability of the prediction results; the theoretical model and monitoring data are effectively combined to dynamically evaluate the slope status in real time, provide a basis for scientific decision-making, and help to take preventive measures in advance and reduce the risk of slope instability. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the specific implementation of the present invention, the following will briefly introduce the drawings required for use in the specific implementation. In all the drawings, each element or part is not necessarily drawn according to the actual scale.

[0033] Figure 1 A flow chart of a method for constructing a slope deformation and failure scale prediction model according to the present invention;

[0034] Figure 2 A flowchart of matrix consistency test in a method for constructing a slope deformation and failure scale prediction model according to the present invention;

[0035] Figure 3 The present invention is a flow chart of a device for constructing a slope deformation and failure scale prediction model. DETAILED DESCRIPTION

[0036] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention.

[0037] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should be the common meanings understood by technicians in the field to which the present invention belongs. The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so as to implement the embodiments of the present disclosure described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. Unless otherwise specified, the term "multiple" means two or more. In the embodiments of the present disclosure, the character " / " indicates that the objects before and after are in an "or" relationship. For example, A / B means: A or B. The term "and / or" is a description of the association relationship of objects, indicating that there can be three relationships. For example, A and / or B means: A or B, or, A and B. The term "corresponding" can refer to an association relationship or a binding relationship, and A and B correspondingly refer to an association relationship or a binding relationship between A and B.

[0038] like Figure 1 to Figure 3 As shown, a method for constructing a slope deformation and failure scale prediction model of the present invention comprises:

[0039] Step S01, defining m fuzzy factors affecting slope failure, and quantifying the fuzzy factors;

[0040] Step S02: define a fuzzy set of quantized fuzzy factors to obtain m fuzzy factor sets, each of which contains all values ​​of the fuzzy factors and corresponding membership functions;

[0041] Step S03, constructing a rule base for the m fuzzy factor sets to obtain a rule base of fuzzy reasoning rules;

[0042] Step S04: Use fuzzy neural network to build a model based on the rule base to build a prediction model.

[0043] Specifically, by introducing mathematical tools such as fuzzy set theory and fuzzy logic, the fuzzy factors in slope deformation and failure are quantitatively processed, which significantly improves the adaptability of the prediction model to complex slope environments, thereby improving the accuracy and reliability of the prediction results; the effective combination of theoretical models and monitoring data can dynamically evaluate the slope status in real time, provide a basis for scientific decision-making, and help take preventive measures in advance and reduce the risk of slope instability.

[0044] In some embodiments, m fuzzy factors affecting slope failure are defined and quantified, including:

[0045] Directly use the fuzzy factors that can be directly quantified among the m fuzzy factors;

[0046] The linear interpolation method is used to quantify the fuzzy factors that cannot be directly quantified among the m fuzzy factors;

[0047] The m fuzzy factors that affect slope failure include geological conditions, rainfall intensity, vegetation development, rock and soil parameters, slope type, slope scale, geological structure, groundwater, construction excavation, and earthquake weathering.

[0048] In some embodiments, a fuzzy set of quantized fuzzy factors is defined to obtain m fuzzy factor sets, including:

[0049] Define the values ​​of the corresponding set of fuzzy factors;

[0050] Define the membership functions of the fuzzy sets.

[0051] Specifically, (1) define fuzzy factors and establish fuzzy sets

[0052] 1) Fuzzy factor quantification: define the fuzzy factors that affect slope failure, such as geological conditions, rainfall intensity, vegetation development, rock and soil parameters, slope type, slope scale, geological structure, groundwater, construction excavation, earthquake weathering, etc. The data that can be directly quantified are used directly, and the data that cannot be directly quantified are quantified by linear interpolation, as shown in Formula 1:

[0053]

[0054] Wherein: RS is the probability value of slope failure caused by fuzzy factors, RS∈[RS1,RS2]; RS1 is the lower limit of the probability value; RS2 is the upper limit of the probability value; RP is the possibility value of the fuzzy factor, RP∈[RP1,RP2]; RP1 is the lower limit of the possibility value of the fuzzy factor; RP2 is the upper limit of the possibility value of the fuzzy factor; RL is the quantified value of the fuzzy factor, RL∈[RL1,RL2]; RL1 is the lower limit of the quantified value; RL2 is the upper limit of the quantified value.

[0055] Define m fuzzy factors, denoted as F i (F1, F2, F3...F m ), let the prediction of slope deformation and failure scale be to find a mapping function f, which transforms the fuzzy factor set F = {F1, F2, ..., F m} is mapped to the slope deformation failure scale S, that is:

[0056] S=f(F1,F2,…,F m ) (Formula 2)

[0057] 2) Establish fuzzy sets: For each fuzzy factor F i Define its fuzzy set Among them, m is the number of fuzzy factors, that is, the domain. For example, the fuzzy set of rainfall intensity is defined as 0 means no rainfall, 1 means extreme rainfall, and other factors are referenced to establish the set. And give its membership function The specific x i The value is mapped to the interval [0, 1], indicating that x i belong The degree of each fuzzy factor in the membership function. The membership function can be Gaussian or Sigmoid. The Gaussian membership function is shown in formula 3:

[0058]

[0059] Among them, c im is the center point of the fuzzy factor, indicating the typical value or the most likely value of the fuzzy factor; σ im is the range of the fuzzy factor, which determines the degree of fuzziness. The larger the range, the stronger the fuzziness. The initial values ​​of both can be obtained through historical data and field surveys, and the value update is carried out using the maximum likelihood estimation method, as shown in Equation 4 and Equation 5:

[0060]

[0061] Where K is the total number of data points used for estimation and k is a data point.

[0062] The Sigmoid membership function is shown in Formula 4:

[0063]

[0064] Among them, α im is the slope parameter of the Sigmoid membership function, which controls the slope of the function and affects the transition speed of the function from 0 to 1; β im is the center point parameter of the function, indicating the midpoint position of the function curve, and determines where the Sigmoid function begins to change significantly. The initial values ​​of both can also be obtained through historical data and field surveys, and the value update is performed using the gradient descent method, as shown in Equation 7 and Equation 8:

[0065]

[0066] Where L is the loss function MSE and η is the learning rate.

[0067]

[0068] n is the number of samples; y is i is the actual value of the i-th sample; is the predicted value for the i-th sample.

[0069] In some embodiments, a rule base is constructed for the m fuzzy factor sets to obtain a rule base of fuzzy inference rules, including:

[0070] According to the historical data of slope deformation and failure and expert experience, k fuzzy reasoning rules are formulated and the matching degree of each rule is calculated;

[0071] Borda ordinal value method is used to perform initial sorting on k fuzzy inference rules, and k fuzzy inference rules after initial sorting are obtained;

[0072] The hierarchical analysis method is used to perform secondary sorting and weight allocation on the k fuzzy reasoning rules after the initial sorting, and the corresponding weights of the k fuzzy reasoning rules are obtained;

[0073] The importance of all rules is calculated using the rule matching degree and corresponding weights to determine the final fuzzy reasoning rule base.

[0074] Specifically, (2) establish a fuzzy reasoning rule base

[0075] For a given input (F1 = x1, F2 = x2, ..., F n =x n ), each rule R under the two membership functions k The matching degree α k The calculation method is the same, Calculated by the corresponding membership function. Each rule R k Assign a weight w k , to reflect the importance of different rules in sorting and prediction. The importance of different rules is sorted according to the Borda ordinal value method for the first time, as shown in formula 10:

[0076]

[0077] Where: b i is the Borda number corresponding to rule i; i represents the i-th rule; k is the risk criterion for sorting the rules; n is the number of criteria; N is the total number of rules; R ik is the number of rules whose importance is higher than that of rule i under the k risk criterion.

[0078] By Borda number b i It can be further sorted to obtain the Borda ordinal number B i , B i Indicates the number of influencing factors with a greater Borda number than influencing factor i. That is, the smaller the Borda number of the influencing factor, the higher the corresponding correlation ranking. After the initial ranking of the importance of the rules, the hierarchical analysis method is used to perform a secondary ranking of the importance and the corresponding weight calculation. According to the process Figure 2The judgment matrix constructed based on all rules is subjected to consistency check. In the figure, m represents the total number of indicators, and RI is the average random consistency index (obtained by table lookup). Then, the eigenvector corresponding to the maximum eigenvalue of the judgment matrix that passes the consistency check is used as the influence weight of each influencing factor, that is, w k .

[0079] In getting w k Then, the matching degree between the rule and the prediction model is calculated by formula 11.

[0080] α′ k =α k ·w k (Formula 11)

[0081] Among them, α k is the matching degree of the kth rule. It indicates the matching degree between the input data and the premise of the rule.

[0082] α′ k It is the weighted rule matching degree, which is used to consider the importance of the rule in the aggregation process.

[0083] In some embodiments, a fuzzy neural network is used in combination with a rule base to construct a prediction model.

[0084] Specifically, (3) prediction model construction

[0085] Fuzzy neural network is used to construct the prediction model, including input layer, fuzzy rule layer, fuzzy reasoning layer and output layer.

[0086] 1) Input layer: receives the original input data x = [x1, x2, ..., x n ] T

[0087] 2) Fuzzy rule layer: Fuzzify the input data and generate multiple fuzzy set memberships for each input variable. i Corresponding to m i fuzzy sets, the output of the fuzzification layer is the membership matrix μ im ,in Represents x i Belongs to the mth fuzzy set The degree of , is calculated using Formula 3 or Formula 6.

[0088] 3) Fuzzy reasoning layer: Based on the output of the fuzzy rule layer and the predefined fuzzy rules, the activation of each rule (also called rule strength or rule matching) is calculated. The input membership is combined using T-norm or algebraic product to obtain the activation of the rule α k .

[0089] Suppose there are R fuzzy rules, each of which is in the form of:

[0090]

[0091] α k First, T-norm or algebraic product is used for calculation. The T-norm calculation is shown in formula 12:

[0092]

[0093] The algebraic product calculation is shown in formula 13:

[0094]

[0095] 4) Output layer: Based on the output of the inference layer, the final output y is calculated through aggregation operations (such as weighted average). The output y is calculated by weighted average or other aggregation methods, as shown in Formula 14:

[0096]

[0097] Among them, y k is the output value corresponding to the kth rule, and y is the scale of slope failure.

[0098] In some embodiments, it also includes using a genetic algorithm to optimize the prediction model to obtain an optimized prediction model.

[0099] Specifically, (4) model optimization

[0100] Genetic algorithm is used for model optimization. The steps are as follows:

[0101] 1) Initialization: Randomly generate an initial population, each individual represents a possible hyperparameter combination (such as learning rate η, membership threshold τ, rule weight w k wait).

[0102] 2) Fitness evaluation: Calculate the fitness of each individual and use the inverse of the cross validation score (CV Score) as the fitness function, as shown in Formula 15:

[0103]

[0104] s i is the individual fitness.

[0105] 3) Selection: Individuals are selected for reproduction based on their fitness, and individuals with high fitness have a higher probability of being selected.

[0106] 4) Crossover: The selected individuals are paired up, and some genes are exchanged through crossover operations (such as single-point crossover and multi-point crossover) to generate new individuals.

[0107] 5) Mutation: Randomly change certain genes in individuals with a certain probability to increase the diversity of the population.

[0108] 6) Iteration: Repeat steps 2)-5) until the stopping condition is met (such as reaching the maximum number of iterations, fitness no longer significantly improving, etc.).

[0109] Finally, after obtaining the hyperparameter combination, the prediction model is input to output the y value to optimize the model.

[0110] In some embodiments, a cross-validation method is also used to validate the optimized prediction model to verify the feasibility and reliability of the model.

[0111] Specifically, (5) model validation

[0112] Cross-validation is used to verify the prediction model. The steps are as follows:

[0113] 1) Dataset partitioning: Divide the dataset D into K mutually exclusive subsets of similar size, i.e., D = D1∪D2∪D3∪.....∪D k ,and (for all i≠j), each subset D i Obtained from D by stratified sampling.

[0114] 2) Repeat training and testing: For each k=1, 2, ..., K, set D k As the test set, the rest D to D k As the training set, use training set D to D k The prediction model constructed in step (3) is trained using the test set D k To evaluate the performance of the model, the mean square error (MSE) of Equation 16 was used to calculate the performance index.

[0115]

[0116] Among them, MSE k The mean square error of the k-fold cross validation; |D k | is the test set D k The size of (x i ,y i ) is a sample point in the test set; is the model for x i The predicted value of .

[0117] In formula 15, we have:

[0118]

[0119] 3) Average performance index calculation: Calculate the average performance index of all K tests as the final estimate of model performance.

[0120] (6) Index evaluation

[0121] Evaluate the fuzzy factor F i When considering the importance of the impact on the scale of slope failure, for the test set (x i ,y i ):

[0122] 1) Replace x i Medium F i The value of , get the new sample x i ′(where F i The values ​​of are randomly permuted or set to a certain baseline value).

[0123] 2) Use the model to calculate x i ' Make a prediction and get the predicted value

[0124] 3) Calculate the value due to the permutation F i The resulting increase in prediction error is:

[0125]

[0126] 4) Average all test samples to get the fuzzy factor F i The global importance of , that is, the influence of a single fuzzy factor on the scale of slope failure:

[0127]

[0128] The influence of a single fuzzy factor on the scale of slope damage can provide an important decision-making tool for timely discovering potential slope instability risks, and can also provide important technical support for slope reinforcement and management.

[0129] In some other embodiments, a device for constructing a slope deformation and failure scale prediction model is provided, comprising:

[0130] A definition module is used to define m fuzzy factors and fuzzy sets of quantitative factors that affect slope failure;

[0131] The first construction module is used to construct a rule base for the m fuzzy factor sets to obtain a rule base of fuzzy reasoning rules;

[0132] The second building module is used to construct a model based on a rule base using a fuzzy neural network to build a prediction model.

[0133] In some embodiments, a processor and a memory storing program instructions are included, and the processor is used to execute a slope stability probability assessment method as described in any of the above items when running the program instructions.

[0134] In other embodiments, an electronic device (eg, a mobile phone) is provided, comprising a device for constructing a slope deformation and damage scale prediction model as described above.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has 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 by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A method for constructing a slope deformation and failure scale prediction model, characterized in that: include: Define m fuzzy factors that affect slope failure and quantify them; Define the fuzzy set of quantized fuzzy factors to obtain m fuzzy factor sets, each of which contains all the values ​​of the fuzzy factors and the corresponding membership functions; Constructing a rule base for the m fuzzy factor sets to obtain a rule base of fuzzy reasoning rules; A fuzzy neural network is used to construct a model based on the rule base to build a prediction model.

2. The method for constructing a slope deformation and failure scale prediction model according to claim 1, characterized in that: The m fuzzy factors affecting slope failure are defined and quantified, including: directly using the fuzzy factors that can be directly quantified among the m fuzzy factors; Using a linear interpolation method to quantify the fuzzy factors that cannot be directly quantified among the m fuzzy factors; The m fuzzy factors affecting slope failure include geological conditions, rainfall intensity, vegetation development, rock and soil parameters, slope type, slope scale, geological structure, groundwater, construction excavation, and earthquake weathering.

3. The method for constructing a slope deformation and failure scale prediction model according to claim 1, characterized in that: The fuzzy set of quantized fuzzy factors is defined to obtain m fuzzy factor sets, including: Define the values ​​of the corresponding set of fuzzy factors; Define the membership functions of the fuzzy sets.

4. The method for constructing a slope deformation and failure scale prediction model according to claim 3, characterized in that: The rule base is constructed for the m fuzzy factor sets to obtain a rule base of fuzzy inference rules, including: According to the historical data of slope deformation and failure and expert experience, k fuzzy reasoning rules are formulated and the matching degree of each rule is calculated; Using the Borda ordinal value method to perform a preliminary sorting on the k fuzzy inference rules, and obtaining k fuzzy inference rules after the preliminary sorting; The k fuzzy reasoning rules after the initial sorting are sorted again and weighted by using the hierarchical analysis method to obtain the weights corresponding to the k fuzzy reasoning rules; The importance of all rules is calculated using the rule matching degree and corresponding weights to determine the final fuzzy reasoning rule base.

5. The method for constructing a slope deformation and failure scale prediction model according to claim 4, characterized in that: Also includes: The prediction model is optimized by using a genetic algorithm to obtain an optimized prediction model.

6. The method for constructing a slope deformation and failure scale prediction model according to claim 5, characterized in that: Also includes: The optimized prediction model was validated by cross-validation method to verify the feasibility and reliability of the model.

7. A device for constructing a slope deformation and failure scale prediction model, characterized in that: include: A definition module is used to define m fuzzy factors and fuzzy sets of quantitative factors that affect slope failure; A first construction module is used to construct a rule base for the m fuzzy factor sets to obtain a rule base of fuzzy inference rules; The second building module is used to build a prediction model by using a fuzzy neural network in combination with the rule base.

8. A device for constructing a slope deformation and failure scale prediction model, comprising a processor and a memory storing program instructions, characterized in that: The processor is used to execute a slope stability probabilistic assessment method as described in any one of claims 1 to 6 when running the program instructions.

9. An electronic device, characterized in that: It comprises a device for constructing a slope deformation and failure scale prediction model as described in claim 7 or 8.

Images