A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory

Abstract

A rapid inversion and identification method for static parameters of high earth-rock dams based on cloud theory, belonging to the technical field of hydraulic engineering. This rapid inversion method for static parameters of high earth-rock dams based on cloud theory starts from the fuzzy correlation between the dam materials of high earth-rock dams, establishes a non-linear mapping between the deformation of high earth-rock dams and the parameters of the constitutive model through cloud theory, and introduces the Jaya optimization algorithm to correct the internal control parameters of the constructed basic cloud proxy model. Finally, the parameters of the constitutive model are inverted using the deformation data of high earth-rock dams measured in the project. The constitutive parameters identified through inversion search have extremely high accuracy, and the running time of the entire inversion process is reduced by more than 98% compared with traditional machine learning methods and more than 99.9% compared with direct inversion using finite elements. This method effectively solves the problem of long time consumption of current inversion methods and can accurately identify the parameters of the constitutive model based on measured deformation while ensuring the transmission of uncertainty within the system.

Description

Technical Field

[0001] The invention belongs to the technical field of water conservancy engineering, and particularly relates to a method for quickly inversely identifying static parameters of high earth-rock dams based on cloud theory. Background Technique

[0002] During the construction and operation of high earth-rock dams, problems such as deformation that seriously threaten the safety of the dams will inevitably occur. Especially due to the increase in their construction height, the deformation will increase significantly, and the impact will be significantly enhanced. Under extreme conditions, accidents such as dam breaches may occur, and the severity of the losses is immeasurable. The deformation study of high earth-rock dams has become an important evaluation index to ensure the construction and operation safety of high earth-rock dams, and the selection of constitutive model parameters plays a very important role in accurately analyzing the deformation of high earth-rock dams. Therefore, the inversion method has become an effective method for determining constitutive model parameters due to its accurate parameter identification performance.

[0003] However, in the actual inversion analysis process of high earth-rock dams, although the inversion method using traditional machine learning methods as surrogate models to replace the direct finite element method has greatly improved the efficiency compared to the direct finite element method, it still takes a long time. The search time for a set of constitutive parameters in a large-scale search is as high as several hours. Moreover, if the number of search generations is small or the search space is insufficient, it is easy to lead to poor search results, making it difficult to combine the deformation analysis of high earth-rock dams with measured data for timely feedback and even more impossible to conduct effective engineering evaluations. For this reason, we propose a method for quickly inversely identifying static parameters of high earth-rock dams based on cloud theory. Summary of the Invention

[0004] The present invention is made to solve the above problems, and the purpose is to propose a new method for quickly inversely identifying static parameters of high earth-rock dams, which can improve the operation speed of the entire inversion analysis system under the condition of ensuring accurate parameter identification, so as to achieve the timely identification of constitutive parameters.

[0005] In order to achieve the above object, the present invention adopts the following scheme: a new method for quickly inversely identifying static parameters of high earth-rock dams, including the following steps:

[0006] (1) Establish a finite element model of a high earth-rock dam, determine the inversion reference points and select a constitutive model;

[0007] (2) Confirm the range of constitutive model parameters and generate a set of effective model parameters;

[0008] (3) Based on the set of constitutive model parameters, perform a forward analysis using finite elements;

[0009] (4) Establish a basic cloud proxy model through the set of constitutive model parameters and the deformation results of the finite element forward analysis;

[0010] (5) Adjust the internal control parameters of the cloud proxy model using the Jaya optimization algorithm to obtain the finally used cloud proxy model;

[0011] (6) Nest the corrected cloud proxy model into the Jaya optimization algorithm, and then invert the model parameters based on the measured deformation.

[0012] Further, the step (1) includes:

[0013] Step (1-1): Finite element modeling

[0014] According to the drawings and measured data of the high earth-rock dam, use professional finite element software (such as common finite element software like ABAQUS, ANSYS, GEODYNA, etc.) to establish a finite element model and reasonably divide the mesh. The mesh division can be carried out according to the calculation accuracy requirements, and there are no special requirements.

[0015] Step (1-2): Selection of inversion reference points

[0016] Based on the distribution of monitoring equipment on the dam cross-section during the construction of the high earth-rock dam, select the reference points for constitutive parameter inversion search from the points where the monitoring equipment is set on the dam cross-section and corresponding to the finite element mesh. Further, points with accurate monitoring results and corresponding to the finite element mesh division can be selected as the reference points for subsequent constitutive parameter inversion search (such as the dam apex, etc.). Only one reference point is selected for each inversion.

[0017] Step (1-3): Selection of constitutive model

[0018] Select a suitable constitutive model (such as common constitutive models like the Duncan-Chang E-B model, generalized plastic model, etc.) as the basis for subsequent finite element coupled forward analysis.

[0019] Further, the step (2) includes:

[0020] Step (2-1): Determine the constitutive model parameters and the fluctuation range

[0021] Based on the parameter values of the selected constitutive model in previous studies and the research on related dams in this field, the fluctuation center in this method is determined. For example, a set of parameter values of the generalized plasticity model determined in the article "Xu B, Zou D, Liu H. Three-dimensional simulation of the construction process of the Zipingpu concrete face rockfill dam based on a generalized plasticity model[J]. Computers&Geotechnics, 2012, 43(Jun.): 143-154." is used as the fluctuation center for the subsequent static inversion of the Zipingpu Dam. The fluctuation range of the parameter values is generally set to ±30%, and the fluctuation range can also be controlled according to the details of specific calculation examples.

[0022] Step (2-2): Generate a set of effective model parameters

[0023] The Latin hypercube sampling method is introduced to perform random sampling and combination within a reasonable fluctuation range (the fluctuation range determined in step 2-1), so as to obtain multiple sets of constitutive model parameter groups required for subsequent training.

[0024] Furthermore, the said step (3) includes:

[0025] All the constitutive model parameter groups obtained in the previous step are respectively brought into the finite element calculation process, so as to obtain multiple sets of deformation calculation results.

[0026] Furthermore, the said step (4) includes:

[0027] Step (4-1): Organize all the input and output data in step (3).

[0028] Step (4-2): "Execute eigenvalue calculation": Calculate the expected value of cloud droplets at the input and output ports respectively x i is the value of all data points in a certain type of dataset organized in step (4-1), and N is the number of data points. All the cloud droplets at the input and output ports are randomly grouped and cyclically combined according to a preset termination condition (the termination condition is defined as the number of loops, generally set to be slightly larger than the number of data points), and the "mean value of the sum of variances Ex" between each group of random groupings is defined as Y. Finally, there are q Ys that form a set: {Y1, Y2, Y3... Yq}. Finally, the remaining two eigenvalues are obtained: He 2 = EY - En 2 , where Ex is the expected value, En is the entropy, D is the variance, and He is the hyperentropy.

[0029] Step (4-3): Generate the features of all input and output clouds as needed according to step (4-2).

[0030] Step (4-4): "Deterministic determination". Here, the method for determining the degree of determination μ representing the position of each cloud droplet in a cloud cluster is as follows: Where Ex is the expectation generated by step 4-2, x is the value of a certain cloud droplet point, and En’ is the new entropy used in this step regenerated based on the entropy and hyperentropy generated in step 4-2. A normal random number is generated using the combination of (En, He ^2 )

[0031] Step (4-5): "Transfer and combination" process. A transfer channel is established between the input and output ports through the degree of determination μ. When there is only one input variable, a deterministic value of μ is transferred. When there are multiple input segments, the expression is used to determine each μ value, where ω is the weight of each component and m is the order of the input parameter, m = 1.....M, to obtain the final deterministic value.

[0032] Step (4-6): Input the final determined value μ transferred to the output end into the output cloud. The output cloud droplets are output according to the following rules, where B represents the output port, Ex represents the total expectation of this type of point, and En represents the entropy of this type of point:

[0033] If on the rising edge

[0034] If on the falling edge

[0035] Where En’ B is the new entropy of the output port generated following the process of step 4-2.

[0036] Furthermore, step (5) includes:

[0037] Step (5-1): Construct an objective function based on the termination requirement and use it to judge the selection of solutions: fitness = |(d gen,i -d real,i ) / d real,i |, where the fitness represents the judgment criterion of the objective function, d gen,i is the relative displacement predicted by the inversion method adopted in this study, and d real,i is the true measured displacement of the target point. The objective function is used to measure the closeness between the simulated value and the actual value. Inputting the position information of each particle into the objective function can obtain the fitness value for evaluating the quality of the position. The smaller the value, the better the simulation effect, and the deformation value corresponding to the parameter group obtained through the search is closer to the true value.

[0038] Step (5-2): Determine the number of particles, the number of iterations, and the initial position within the search space through basic settings, and calculate the change in the position of the particle during each iteration using the following formula:

[0039]

[0040] Where p, j, and k represent the iteration variable, the individual solution variable, and the solution in the population, respectively. Further, A(p,j,k) is the j-th variable of the k-th individual in the p-th iteration process, A(p + 1,j,k) represents the variable of the individual A(p,j,k) in the next iteration process, A(p,j,b) is the optimal solution in one iteration, A(p,j,w) is the worst solution in one iteration, and r represents a control index randomly selected within the range of [0,1].

[0041] Step (5-3): Continuously generate new solutions according to the position update formula in Step (5-2). In the next iteration process, compare the new and old excellence degrees of the solutions based on the objective function. If the new solution is better, replace the original solution and enter the next round of iteration; otherwise, keep the original solution position.

[0042] Step (5-4): Introduce the basic cloud proxy model obtained in Step (4) into the Jaya optimization algorithm proposed in Steps (5-1) - (5-4). By studying the required termination target, adjust the three eigenvalue (Ex, En, He) in the cloud proxy model through the optimization search correction coefficient, so as to obtain an improved cloud proxy model with a higher fitting degree of the simulation results.

[0043] Further, the said Step (6) includes:

[0044] Embed the optimized cloud proxy model into the Jaya optimization algorithm, set the basic parameters of the optimization algorithm that meet the research requirements (such as setting the control parameters in the Jaya optimization algorithm with reference to "Kang F, Liu X, Li J, et al. Multi-parameter inverse analysis of concrete dams using kernel extreme learning machines-based response surface model[J]. Engineering Structures, 2022, 256: 113999-.", and on this basis, if no appropriate value is found, appropriately increase the control parameters in the Jaya optimization algorithm), and input the measured values to be inverted as the target comparison value for inversion search. The position corresponding to the minimum fitness in the last generation of Jaya particle swarm is called the global optimal position, and its value is the global optimal solution. Outputting the value is the value of the constitutive model parameter group obtained through inversion.

[0045] Compared with the prior art, the beneficial effects brought by the technical solution provided by the present invention include: the present invention provides a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory. Starting from the fuzzy correlation between the dam materials of high earth-rock dams, a non-linear mapping between the deformation of high earth-rock dams and the parameters of the constitutive model is established through cloud theory. And the Jaya optimization algorithm is introduced to correct the internal control parameters of the constructed basic cloud proxy model. Finally, the parameters of the constitutive model are back-analyzed by using the deformation data of high earth-rock dams measured in engineering. The constitutive parameters identified through back-analysis search have extremely high accuracy, and the running time of the entire back-analysis process is reduced by more than 98% compared with traditional machine learning methods, and is reduced by more than 99.9% compared with direct back-analysis using finite elements. The idea of the present invention is clear and novel, and can effectively establish a non-linear correlation with short running time and accurate mapping, which can provide a more practical and effective algorithm for the timely condition evaluation and safety analysis of most water conservancy and civil engineering projects, making real-time feedback of deformation simulation possible. Therefore, it has broad application prospects. Description of the Drawings

[0046] Figure 1 It is a flowchart of a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory. Figure 2 It is a schematic diagram of the three-dimensional finite element mesh division of a high earth-rock dam for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory.

[0047] Figure 3 It is a schematic diagram of the setting and selection of reference points for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory.

[0048] Figure 4 It is a schematic diagram of the concept cloud for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory.

[0049] Figure 5 It is a schematic diagram of the modified cloud for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory.

[0050] Figure 6 It is a schematic diagram of the multi-condition single-rule process for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory.

[0051] Figure 7 It is a schematic diagram of the Jaya optimization algorithm search for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory.

[0052] Figure 8 It is a comparison chart of the back-analysis running time of Jaya-ICSM and Jaya-BP for a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory. Detailed Embodiment

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Subsequently, a comparative example is described by introducing the traditional method based on the embodiments to verify the innovation, effectiveness of the present invention and its beneficial help in solving existing problems.

[0054] Embodiment

[0055] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Refer to Figure 1 As shown, a rapid back-analysis method for static parameters of high earth-rock dams based on cloud theory provided by an embodiment of the present invention may include the following steps:

[0056] Step (1):

[0057] Step (1-1): Finite element modeling

[0058] According to the drawings of the dam and the measured data of the vertical deformation of the main section (specifically, refer to the paper "Song Yangang, Deng Liangsheng, Cai Dewen, etc. Settlement monitoring and analysis during the construction period of Zipingpu concrete face rockfill dam [J]. Sichuan Water Power, 2006, 25(1): 7."), use the professional finite element software GEODYNA to establish a finite element model, as shown in Figure 2 shown.

[0059] Step (1-2): Selection of back-analysis reference points

[0060] Based on the distribution of monitoring equipment on the dam cross-section during the construction of the high earth-rock dam, select the points with accurate monitoring results and corresponding to the finite element mesh division as the reference points for subsequent constitutive parameter back-analysis search, as shown in Figure 3 shown.

[0061] Step (1-3): Selection of constitutive model

[0062] Select a suitable constitutive model as the basis for subsequent finite element coupled forward analysis (in this embodiment, the Duncan-Chang E-B model is selected as the model basis for static deformation calculation of high earth-rock dams).

[0063] Table 1 Cloud characteristics

[0064]

[0065] Step (2):

[0066] Step (2-1): Determine the constitutive model parameters and the fluctuation range

[0067] Based on the selected parameter combinations in the previous research (the research on the sensitivity analysis of Duncan-Chang model parameters is included in the references "Yang He. Inversion and Analysis of Material Parameters of Shuibuya Concrete Face Rockfill Dam Based on Improved Response Surface Method [D]. Wuhan University, 2017." and "Fang Xing. Inversion Analysis of Parameters of High Core Rockfill Dam Based on Neural Network and Artificial Bee Colony Algorithm [D]. Wuhan University, 2016."), the constitutive model parameter values shown in Table 2 are selected. Through the previous sensitivity analysis research, K, K b , r f , are selected as the parameters to be inverted. The parameter values in Table 2 are used as the fluctuation center, and the fluctuation range is set to ±30%.

[0068] Table 2 Duncan-Chang E-B Constitutive Parameters of Dam Rockfill Area

[0069]

[0070] Step (2-2): Generate valid model parameter groups

[0071] Introduce the Latin hypercube sampling method, and conduct random sampling and combination within the fluctuation range of step (2-2) to obtain the training group and test group samples of the constitutive model parameters.

[0072] Step (3):

[0073] Bring all the constitutive model parameter groups obtained in step (2-2) into the finite element calculation process respectively to obtain the calculation results of the vertical deformations of all reference points in the training group and test group. Part of the calculation results are given in Table 3.

[0074] Table 3 Partial Finite Element Calculation Results of Vertical Deformations of Reference Points of High Earth-Rock Dam

[0075]

[0076]

[0077] Step (4):

[0078] Step (4-1): Organize all the input and output data in step (3), as shown in Table 3.

[0079] Step (4-2): "Execute eigenvalue calculation": Calculate the expected value of cloud droplets at the input and output ports respectively Randomly group and cycle-combine all the x i cloud droplets at the input and output ports according to the preset termination conditions (that is, the number of loops is set to 500 times, and the operation ends when 500 loops are completed) to obtain the variance between groups and the mean value of Ex {Y1, Y2, Y3... Y q}. Finally, the remaining two features are obtained: He 2 = EY - En 2 , where E is the expected value and D is the deviation.

[0080] Step (4-3): For each type of data set, use Step (4-2) to process all the data point values included in this type, generating three features of the cloud proxy model for the input end (as shown in Table 4) and the output end (as shown in Table 5). For the output-end cloud model, the entropy and hyper-entropy in the cloud features are output separately for the left and right sides. The cloud model formed by aggregating all cloud droplets is as Figure 4 shown.

[0081] Table 4 Input-end features of the cloud proxy model

[0082]

[0083] Table 5 Output-end features of the cloud proxy model

[0084]

[0085] Step (4-4): "Determination of certainty". Here, the method for determining the degree of certainty μ representing the position of each cloud droplet in a cloud cluster is as follows: where A represents the input port of the cloud droplet, x Aii is the value of a data point for which the degree of certainty is to be determined in a major category of data sets, Ex Aii is the expectation of this major category of data sets, and En' is the new entropy of this major category of data sets regenerated based on the entropy and hyper-entropy generated in Step 4-2.

[0086] Step (4-5): The "transmission and combination" process. Establish a transmission channel between the input and output ports through the degree of certainty μ calculated in Step (4-4).

[0087] Step (4-6): Input the final determined value μ transmitted to the output end into the output cloud. The output cloud droplets are output according to the following rules:

[0088] If on the rising edge

[0089] If on the falling edge

[0090] Step (5):

[0091] Step (5-1): Construct an objective function based on the termination requirement and use this to judge the acceptance or rejection of the solution: fitness = |(d gen,i - d real,i ) / d real,i |, where the fitness represents the judgment criterion of the objective function. d gen,iThe relative displacement predicted by the inversion method adopted in this study (a set can be obtained each time the Jaya optimization algorithm is used for optimization), d real,i is the true measured displacement of the target point (in this example, a measured vertical deformation value of -0.956 m is selected and input into the fitness formula). The objective function is used to measure the degree of closeness between the simulated value and the actual value. By inputting the position information of each particle into the objective function, the fitness value for evaluating the quality of the position can be obtained. The smaller the value, the better the simulation effect, and the deformation value corresponding to the parameter group obtained through the search is closer to the true value.

[0092] Step (5-2): Determine that the number of particles is 1000, the number of iterations is 500, and the initial position in the search space is a random value within a 30% fluctuation range above and below the fluctuation base value through basic settings. Calculate the change in the position of the particle during each iteration through the following formula (the search path is as Figure 6 shown):

[0093]

[0094] where p, j, and k represent the iteration variable, the individual solution variable, and the solution in the population, respectively. Further, A(p,j,k) is the j-th variable of the k-th individual in the p-th iteration process, A(p + 1,j,k) represents the variable of the individual A(p,j,k) in the next iteration process, A(p,j,b) is the optimal solution in one iteration, A(p,j,w) is the worst solution in one iteration, and r represents a control index randomly selected within the range of [0,1].

[0095] Step (5-3): Continuously generate new solutions according to the position update formula in step (5-2). In the next iteration process, compare the quality of the new and old solutions based on the objective function. If the new solution is better, replace the original solution and enter the next round of iteration; otherwise, still retain the original solution position.

[0096] Step (5-4): Introduce the basic cloud proxy model obtained in step (4) into the Jaya optimization algorithm proposed in steps (5-1) - (5-4). Set the termination target (i.e., the number of iterations is 500) according to the research requirements, and adjust the three characteristic values (Ex, En, He) in the cloud proxy model through the optimization search correction coefficient, so as to obtain an improved cloud proxy model with a higher fitting degree of the simulation result (the deformation cloud form is as Figure 5 shown).

[0097] Further, step (6) includes:

[0098] Embed the optimized cloud proxy model into the Jaya optimization algorithm, and according to the cloud information transmission rule (such as Figure 7Search for input and output of information as shown, set the basic parameters of the Jaya optimization algorithm the same as in step (5-2), and set the termination criterion to stop when the maximum number of iterations is reached. Input the measured vertical deformation value (-0.956 m) of the reference point to be inverted as the target comparison value for inversion search. The position corresponding to the minimum fitness in the last generation of Jaya particle swarm is called the global optimal position, and outputting its value gives the required constitutive model parameter group values obtained through inversion, as shown in Table 6.

[0099] Table 6 Comparison of inversion parameter values with measured values and running time of the inversion process

[0100]

[0101] Comparison example

[0102] This comparison verifies the effect and unique superiority of the inversion of constitutive parameters in the example by using the proposed new rapid inversion method of static parameters for high rock and earth dams, Jaya-ICSM, and compares it with the existing traditional machine learning method (BP) as a surrogate model method.

[0103] In the experimental test, the ICSM algorithm proposed in the present invention and the optimized BP algorithm in the traditional machine learning method are combined with the Jaya optimization algorithm at the same time, and comparative analysis is carried out based on the same training data and test database. Before the comparison, the basic control parameters of each algorithm are debugged.

[0104] Table 7 Built-in control parameter settings of the optimized BP algorithm

[0105] Project Name Training Frequency Learning Rate Minimum Error of Training Target Value 10000 0.05 0.0001

[0106] Table 8 Built-in control parameter settings of the ICSM algorithm

[0107]

[0108] Table 9 Inversion results of constitutive model parameters

[0109]

[0110]

[0111] Table 9 lists the running results of the two combined algorithms, Figure 8 indicating the running time consumption of the two combined algorithms in the whole process of inversion. It can be seen that in the process of inversion and identification of constitutive parameters, the ICSM proposed in the present invention can achieve extremely high search accuracy, and compared with the combined algorithm using the optimized BP algorithm as a surrogate model, it can significantly save the inversion search time.

[0112] The above embodiments are merely illustrative examples of the technical solutions of the present invention. The constitutive model parameter inversion method and the corresponding program involved in the present invention are not limited to the content described in the above embodiments, but are subject to the scope defined by the claims. Any modification, supplement, or equivalent replacement made by those skilled in the art to which the present invention pertains based on this embodiment is within the scope protected by the claims of the present invention.

Claims

1. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory, characterized in that The method includes the following steps: 1) Establish a finite element model of a high earth-rock dam, determine the inversion reference points, and select a constitutive model; 2) Confirm the range of constitutive model parameters and generate a set of effective model parameters; 3) Based on the set of constitutive model parameters, perform a forward analysis using finite element method; 4) Establish a basic cloud proxy model through the set of constitutive model parameters and the deformation results of the finite element forward analysis; 5) Use the Jaya optimization algorithm to adjust the internal control parameters of the cloud proxy model to obtain the finally used cloud proxy model; 6) Nest the corrected cloud proxy model into the Jaya optimization algorithm, and then invert the model parameters according to the measured vertical deformation values.

2. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory according to claim 1, characterized in that, The method for establishing a finite element model of a high earth-rock dam and confirming the constitutive model is as follows: According to the drawings and measured data of the high earth-rock dam, use finite element software to establish a finite element model and divide the grid. Set monitoring devices on the dam cross-section and select the reference points for constitutive parameter inversion search from the points corresponding to the finite element grid. Select a suitable constitutive model according to research needs as the basis for subsequent finite element coupled forward analysis.

3. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory according to claim 1, characterized in that The method for confirming the range of constitutive model parameters and generating a set of effective model parameters is as follows: Select a parameter combination according to the constitutive model determined in step 1, determine the fluctuation center, and set the fluctuation range within ±10% - ±30% according to the calculation limit of the constitutive model; Introduce the Latin hypercube sampling method, perform random sampling and combination within the set fluctuation range to obtain the set of constitutive model parameters required for training.

4. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory according to claim 1, characterized in that The method for performing a forward analysis using finite element method based on the set of constitutive model parameters: Substitute the set of constitutive model parameters determined in step 2 into the finite element model for calculation to obtain multiple sets of deformation calculation results.

5. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory according to claim 1, characterized in that The method for establishing a basic cloud proxy model through the model parameter set and the forward analysis deformation calculation results is as follows: 4-1) Organize all the input and output data in step 3; 4-2) Perform eigenvalue calculation: Use the following formula to calculate the cloud droplet expectation Ex of the input and output ports respectively: where x i is the value of the i-th data point in a certain large class of data sets in an input port or an output port, and N is the number of data points; Randomly group and cycle-combine all the cloud droplets of the input and output ports according to the preset termination conditions respectively, obtain the variance between the groups of random groupings and the mean value Y of the cloud droplet expectations, and use the following formula to calculate the eigenvalues En and He: He 2 = EY - En 2 , where En is entropy, He is hyperentropy, E is the expected value, and D is the variance; 4-3) Generate the eigenvalues of all the input and output cloud sets according to step 4-2; 4-4) Deterministic determination: The determination degree μ representing the position of each cloud droplet in a cloud cluster is calculated using the following formula: Among them, Ex is the expectation generated in step 4-2, x is the value of the cloud droplet point to be solved, En’ is the new entropy used in this step regenerated based on the entropy and hyperentropy generated in step 4-2, and a normal random number is generated by using the combination of (En, He ^2 ); 4-5) Transfer and combination process: Establish a transfer channel between the input and output ports through the determination degree μ. When there is only one input variable, transfer a deterministic value of μ. When there are multiple input segments, use the following formula to determine each μ value: where ω is the weight of each component and m is the order of the input parameter, obtaining the final deterministic value; 4-6) Input the final determined value μ transferred to the output end into the output cloud, and the cloud droplets of the output port B are output according to the following rules: If on the rising edge, If on the falling edge, Among them, En’ B is the new entropy of the output port generated by following the process of step 4-2.

6. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory according to claim 1, characterized in that, The method for using the Jaya optimization algorithm to adjust the internal control parameters of the cloud proxy model to obtain the finally used cloud proxy model is as follows: 5-1) Construct the objective function based on the termination requirement and use it to judge the selection of solutions: where the fitness represents the judgment criterion of the objective function, n is the number of observation values, and d gen,i is the relative displacement predicted by the inversion method adopted in this study, and d real,i is the true measured displacement of the target point; 5-2): Determine the number of particles, the number of iterations, and the initial position in the search space through basic settings, and calculate the change in the position of the particles in each iteration through the following formula: A(p + 1,j,k) = A(p,j,k) + r(p,j,1)(A(p,j,b) - |A(p,j,k)|) - r(p,j,2)(A(p,j,b) - |A(p,j,w)| Among them, p is the iteration variable, j is the individual solution variable, k is the solution in the population, A(p,j,k) is the j-th variable of the k-th individual in the p-th iteration process, A(p + 1,j,k) represents the variable of the individual A(p,j,k) in the next iteration process, A(p,j,b) is the optimal solution in one iteration, A(p,j,w) is the worst solution in one iteration, and r represents the control index randomly selected within the range of [0,1]; 5 - 3): Continuously generate new solutions according to the position update formula in step 5 - 2, compare the new and old excellence degrees of the solutions based on the objective function in the next iteration process. If the new solution is better, replace the original solution and enter the next round of iteration; otherwise, still keep the position of the original solution; 5 - 4): Introduce the basic cloud proxy model obtained in step 4 into the Jaya optimization algorithm proposed in steps 5 - 1 to 5 - 3. By studying the termination target to be set, adjust the three eigenvalues Ex, En, and He in the cloud proxy model through the optimization search correction coefficient to obtain a cloud proxy model with a higher fitting degree.

7. A rapid inverse identification method for static parameters of high earth-rock dams based on cloud theory according to claim 1, characterized in that, Embed the cloud proxy model into the Jaya optimization algorithm and invert the model parameters based on the measured deformation values. The method is as follows: Embed the cloud proxy model obtained in step 5 into the Jaya optimization algorithm, set the basic parameters of the optimization algorithm to meet the research requirements, input the measured values to be inverted as the target comparison values for inversion search. The position corresponding to the minimum fitness in the last generation of Jaya particle swarm is called the global optimal position, and the value output is the value of the constitutive model parameter group obtained through inversion.

Images