A tunnel excavation deformation prediction method based on numerical simulation and parameter inversion

Abstract

The application relates to the technical field of tunnel construction, and particularly discloses a tunnel excavation deformation prediction method based on numerical simulation and parameter inversion, which comprises the following steps: (1) collecting deformation modulus and Poisson's ratio as parameters to be studied; (2) establishing a two-dimensional finite element model through abapus; (3) setting multiple test groups for the parameters to be studied, obtaining a functional relationship between vault displacement and deformation modulus and Poisson's ratio and a functional relationship between peripheral displacement and deformation modulus and Poisson's ratio under a first excavation condition; (4) substituting monitored values of vault and peripheral displacement under the first excavation condition into the functional relationship to obtain inversion values E1 and v1 of the deformation modulus and Poisson's ratio under the first excavation condition, and substituting E1 and v1 into the two-dimensional finite element model to obtain predicted values S of the vault and peripheral displacement under the first excavation condition. 拱顶2 and S 周边2 The tunnel excavation deformation prediction method based on numerical simulation and parameter inversion can improve the accuracy of tunnel excavation deformation prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel construction, in particular to a tunnel excavation deformation prediction method based on numerical simulation and parameter inversion. BACKGROUND

[0002] In the numerical simulation of rock tunnel excavation, due to the inaccuracy of rock constitutive model parameters, the prediction error of the rock tunnel model for subsequent excavation conditions is large, which is not conducive to guiding construction; at present, due to the development of computer theory, more scholars use intelligent algorithms to invert rock constitutive parameters, such inversion is based on training data and has a strong mathematical relationship, but the mechanical properties of the rock itself are ignored in the calculation, so that the tunnel excavation deformation prediction is still not accurate enough and the applicability is poor. SUMMARY

[0003] The present application aims to at least solve one of the above-mentioned technical problems, and provides a tunnel excavation deformation prediction method based on numerical simulation and parameter inversion, which uses mathematical relationship to invert soil layer parameter values, and re-substitutes the two-dimensional finite element model for calculation, if the error criterion is not met, the parameter gradient is set to re-calculate, which can consider the soil layer properties and improve the accuracy of tunnel excavation deformation prediction.

[0004] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: a tunnel excavation deformation prediction method based on numerical simulation and parameter inversion, comprising the following steps:

[0005] (1) collecting initial parameters, taking the deformation modulus and Poisson's ratio in the initial parameters as the parameters to be studied;

[0006] (2) according to the section under the actual working condition of the tunnel, a two-dimensional finite element model is established by abapus;

[0007] (3) taking the collected parameters to be studied as the benchmark, setting multiple test groups for the parameters to be studied, and substituting them into the two-dimensional finite element model for calculation, and using the improved sparrow algorithm to obtain the functional relationship between the crown displacement and the deformation modulus and Poisson's ratio under the first excavation condition, and the functional relationship between the peripheral displacement and the deformation modulus and Poisson's ratio;

[0008] (4) substituting the monitoring values of the crown and peripheral displacement under the first excavation condition into the above functional relationship to obtain the inversion values E1 and v1 of the deformation modulus and Poisson's ratio under the first excavation condition, and substituting E1 and v1 into the two-dimensional finite element model to obtain the predicted values S 拱顶2 and S 周边2If the predicted value meets the error criterion compared with the monitored value of the crown and peripheral displacement under the first excavation condition, the crown and peripheral displacement under the second excavation condition is predicted by using E1 and v1; if the error criterion is not met, the E1 and V1 are set with multiple sets of parameter gradients and substituted into the two-dimensional finite element model for calculation until E2 and V2 meeting the error criterion are found, the crown and peripheral displacement under the second excavation condition is predicted by using E2 and V2, and the above operation is repeated until all excavation conditions are predicted.

[0009] Preferably, the initial parameters of the model include specific gravity, deformation modulus, Poisson's ratio, shear strength and tensile strength.

[0010] Preferably, the error criterion is: |U 拱顶 -U 拱顶监测 | / U 拱顶监测 ≤10%; and |U 周边 -U 周边监测 | / U 周边监测 ≤10%.

[0011] Preferably, the parameter gradients of E1 and V1 are set to 90% E1, 95% E1, 105% E1, 110% E1, 90% V1, 95% V1, 105% V1 and 110% V1.

[0012] Preferably, the total length of the tunnel excavation is evenly divided into multiple excavation sections of the same length, each excavation section is taken as an excavation condition, each excavation condition is unified into a section form, and a two-dimensional finite element model is established.

[0013] Preferably, the function relationship establishment in step 2 includes the following steps: taking the crown and peripheral displacement under the first excavation condition as input variables, taking the deformation modulus and Poisson's ratio as output variables, establishing an ELM neural network; introducing a Logistic function in the sparrow algorithm to update the position of the joiner in the sparrow population, searching the optimal weight and threshold value of the network based on the ELM network, and optimizing the ELM neural network.

[0014] The beneficial effect is: compared with the prior art, the tunnel excavation deformation prediction method based on numerical simulation and parameter inversion of the present application uses the function relationship between the crown displacement under the first excavation condition and the deformation modulus and Poisson's ratio, and the function relationship between the peripheral displacement and the deformation modulus and Poisson's ratio, when the predicted value does not meet the error criterion, the parameter to be studied is set with gradients, and then the parameter to be studied is substituted into the finite element model for cyclic calculation until the error criterion is met, in the deformation prediction of each excavation condition, the mechanical properties of the rock itself are considered, the accuracy of the tunnel excavation deformation prediction is improved, and the applicability is stronger. BRIEF DESCRIPTION OF DRAWINGS

[0015] The specific embodiments of the present application will be further described in conjunction with the accompanying drawings, in which:

[0016] Figure 1 is a schematic diagram of a two-dimensional finite element model with the parameters to be studied being brought in after calculation.

[0017] Figure 2 is a schematic diagram of a two-dimensional finite element model with the parameters to be studied being brought in after calculation. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0019] It should be noted that when a component is referred to as being "fixed" to another component, it can be directly on the other component or there can be intervening components. When a component is referred to as being "connected" to another component, it can be directly connected to the other component or there can be intervening components. When a component is referred to as being "disposed on" another component, it can be directly on the other component or there can be intervening components. When a component is referred to as being "disposed in the middle", it is not only disposed in the middle position, but also falls within the range defined by the middle. The terms "vertical", "horizontal", "left", "right", and similar terms used herein are for illustrative purposes only.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0021] A tunnel excavation deformation prediction method based on numerical simulation and parameter inversion is disclosed in the present application, comprising the following steps:

[0022] (1) Collect initial parameters, and take the deformation modulus and Poisson's ratio in the initial parameters as parameters to be studied;

[0023] (2) According to the section under the actual working condition of the tunnel, a two-dimensional finite element model is established through abapus;

[0024] (3) Based on the collected parameters to be studied, set up multiple test groups for the parameters to be studied and substitute them into the two-dimensional finite element model for calculation. Use the improved sparrow algorithm to obtain the functional relationship between the crown displacement and deformation modulus and Poisson's ratio under the first excavation condition, as well as the functional relationship between the perimeter displacement and deformation modulus and Poisson's ratio.

[0025] (4) Substitute the monitored values ​​of the crown and perimeter displacements under the first excavation condition into the above functional relationship to obtain the inverse values ​​E1 and v1 of the deformation modulus and Poisson's ratio under the first excavation condition. Substitute E1 and v1 into the two-dimensional finite element model to obtain the predicted values ​​S of the crown and perimeter displacements under the first excavation condition. 拱顶2 and S 周边2 If the predicted value meets the error criterion compared with the monitored values ​​of the crown and surrounding displacement under the first excavation condition, then continue to use E1 and v1 to predict the crown and surrounding displacement under the second excavation condition; if the error criterion is not met, set multiple sets of parameter gradients for E1 and V1 and substitute them into the two-dimensional finite element model for calculation until E2 and V2 that meet the error criterion are found. Use E2 and V2 to predict the crown and surrounding displacement under the second excavation condition. Repeat the above operation until all excavation conditions are predicted.

[0026] In one specific implementation, the error criterion can be designed as: |U 拱顶 -U 拱顶监测 | / U 拱顶监测 ≤10%; and |U 周边 -U 周边监测 | / U 周边监测 ≤10%. Meanwhile, the parameter gradients of E1 and V1 can be set to 90%E1, 95%E1, 105%E1, 110%E1, 90%V1, 95%V1, 105%V1, 110%V1, forming 16 sets of parameters which are substituted into the two-dimensional finite element model for calculation. The output is obtained when the error criterion is met.

[0027] When designing excavation conditions, the total length of the tunnel excavation can be divided into multiple excavation segments of equal length, and each excavation segment can be considered as an excavation condition. For example, if the total length of the tunnel is 120m, then 3m can be considered as an excavation condition. Therefore, the tunnel excavation can be designed with 40 excavation conditions, and typical cross-sections of the tunnel under each condition can be selected and a two-dimensional finite element model can be established.

[0028] In one embodiment, the function relationship in step 2 is established by the following steps: establishing an ELM neural network with the vault and peripheral displacement displacement in the first excavation condition as input variables and the deformation modulus and Poisson's ratio as output variables; introducing a Logistic function into the sparrow algorithm to update the position of the joiner in the sparrow population, searching for the optimal weight and threshold 0 of the network based on the ELM network, and optimizing the ELM neural network. Specifically, the training group input variable can be set as inputn; the training group output variable as outputn; the prediction group input variable as inputtest; and the prediction group output variable as outputtest.

[0029] The input and output variables are normalized, and the corresponding matlab code is as follows:

[0030] [inputn,inputps]=mapminmax(input_train,-1,1);

[0031] [outputn,outputps]=mapminmax(output_train,-1,1);

[0032] An ELM neural network is created, with the number of input layers being 6, the number of output layers being 2, and the number of hidden layers being 13:

[0033] [IW,B,LW,TF,TYPE]=elmtrain1(inputn,outputn,5,'sig',0);

[0034] The initial parameters of the improved sparrow algorithm are determined: the population size is 20, the producer ratio is 20%, the maximum iteration number is 100, and the joiner ratio is 20%. The sparrow algorithm will continuously update the position according to the population foraging rule until the maximum iteration number is reached.

[0035] First, calculate the dimension dim of the solution

[0036] d=inputnum+inputnum×hiddennum+hiddennum×outputnum+outputnum

[0037] Where: d is the sum of the number of nodes in the input layer, hidden layer and output layer, inputnum, hiddennum and outputnum are the number of nodes in the input layer, hidden layer and output layer, i.e. all the weights and thresholds of the neural network;

[0038] Then generate the initial solution:

[0039] Xij=lb+(ub-lb)×rand(1,dim)

[0040] where Xijis the position information of sparrow in space, [lb, ub] represents the solution space, and rand(0, dim) means random value in the range of 0-1.

[0041] The position update of the finder during the iterative search is as follows:

[0042]

[0043] where X i,j is the position information of the ith sparrow in the jth dimension in the solution space; t is the current iteration number; M is the maximum iteration number; r1 is a random number, 0≤r1<1; r2 is a random number, 0≤r1<1, representing the signal value emitted when the sparrow realizes the presence of the predator; Y is the early warning threshold; G is a random value in [-1, 1]; and L is a matrix with elements of 1 in the dimension l*d.

[0044] A part of the joiners compete with the finder for food, and a part of the joiners go to other places to forage, and the position update is as follows:

[0045]

[0046] where X f is the position where the fitness value of the finder is optimal; X worse is the position where the fitness value is worst; and A is a matrix with elements of 1 or -1 in the dimension l*d.

[0047] The position of the joiner is disturbed by using the Logistic function, so that the solution is searched faster. The Logistic function is expressed as:

[0048] X i+1 = 4×X n ×(1-X n )

[0049] The sparrow individuals will gather when they encounter danger. The position update of the sparrow individuals in this process is as follows:

[0050]

[0051] where X best is the optimal position of the sparrow individual; a is a random value in [-1, 1]; ω is a random value in [-1, 1]; ε is a very small number to avoid the denominator being 0, and the value is 1e-50; f i is the fitness value of the ith sparrow individual in the current iteration; f best is the optimal value of the fitness value; and f worse is the worst value of the fitness value.

[0052] The above process is repeated, when the maximum number of iterations is reached, the search is ended, the value of fitness searched by the algorithm is taken as the weight and threshold of the ELM network structure, and the LSSA-ELM neural network is established, so that the functional relationship between the vault displacement and the deformation modulus and Poisson's ratio under the first excavation condition and the functional relationship between the peripheral displacement and the deformation modulus and Poisson's ratio are obtained. The prediction method of the application must obtain the monitoring values of the vault and peripheral displacement under the first excavation condition when performing prediction, and can only be used for prediction of adjacent excavation conditions.

[0053] In a specific application example, the specific data of the tunnel excavation construction are: length x height 200 x 100, tunnel buried depth 25 m, diameter 12 m, rock V-class surrounding rock, deformation modulus 2.43 Gpa, Poisson's ratio 0.405, and the two-dimensional finite element model of the tunnel section is as shown in Figure 1 The calculation result after the research parameters are brought into the two-dimensional finite element model is as shown in Figure 2 The vault displacement prediction value of the application is compared with the vault displacement prediction value of the traditional method, and the peripheral displacement prediction value of the application is compared with the peripheral displacement prediction value of the traditional method, respectively, for five adjacent excavation conditions, and the following comparison data are obtained:

[0054] Vault displacement

[0055]

[0056] Peripheral displacement

[0057]

[0058] It can be seen that, compared with the prior art, the tunnel excavation deformation prediction method based on numerical simulation and parameter inversion of the application utilizes the functional relationship between the vault displacement and the deformation modulus and Poisson's ratio under the first excavation condition and the functional relationship between the peripheral displacement and the deformation modulus and Poisson's ratio, when the prediction value does not satisfy the error criterion, the gradient of the research parameter is set by using the corresponding functional relationship, and then the research parameter is substituted into the model for cyclic calculation until the error criterion is satisfied, the mechanical properties of the rock itself are considered in the deformation prediction of each excavation condition, the accuracy of the tunnel excavation deformation prediction is improved, and the applicability is stronger.

[0059] The above examples are only used to illustrate the technical solutions of the application and not to limit them, any modification or equivalent replacement within the spirit and scope of the application should be covered in the scope of the technical solutions of the application.

Claims

1. A tunnel excavation deformation prediction method based on numerical simulation and parameter inversion, characterized in that, Includes the following steps: (1) Collect initial parameters, and take the deformation modulus and Poisson's ratio in the initial parameters as the parameters to be studied; (2) Based on the cross-section of the tunnel under actual working conditions, a two-dimensional finite element model was established using Abapus; (3) Based on the collected parameters to be studied, multiple test groups are set up for the parameters to be studied, and they are substituted into the two-dimensional finite element model for calculation. The improved sparrow algorithm is used to obtain the functional relationship between the crown displacement and the deformation modulus and Poisson's ratio under the first excavation condition, as well as the functional relationship between the peripheral displacement and the deformation modulus and Poisson's ratio. The establishment of the functional relationship includes the following steps: using the crown displacement and peripheral displacement under the first excavation condition as input variables, and the deformation modulus and Poisson's ratio as output variables, an ELM neural network is established; the Logistic function is introduced into the sparrow algorithm to update the position of the sparrow population. The Logistic function is expressed as: ； The individual sparrow update location is: ； where, is the optimal position of the sparrow individual; is a random value between [-1, 1]; is a random value between [-1, 1]; is a very small number to avoid division by zero, with a value of 1e-50; is the th individual in the current iteration; is the fitness value of the th individual in the current iteration; is the optimal fitness value; is the worst fitness value; Then, based on the ELM network, the optimal weights and thresholds of the network are searched to optimize the ELM neural network; (4) Substitute the monitored values ​​of the crown and surrounding displacements under the first excavation condition into the above functional relationship to obtain the inverse values ​​E1 and v1 of the deformation modulus and Poisson's ratio under the first excavation condition. Substitute E1 and v1 into the two-dimensional finite element model to obtain the predicted values ​​of the crown and surrounding displacements under the first excavation condition. and If the predicted value meets the error criterion compared with the monitored values ​​of the crown and surrounding displacement under the first excavation condition, then continue to use E1 and v1 to predict the crown and surrounding displacement under the second excavation condition; if the error criterion is not met, set multiple sets of parameter gradients for E1 and V1 and substitute them into the two-dimensional finite element model for calculation until E2 and V2 that meet the error criterion are found. Use E2 and V2 to predict the crown and surrounding displacement under the second excavation condition. Repeat the above operation until all excavation conditions are predicted.

2. The method for predicting tunnel excavation deformation based on numerical simulation and parameter inversion according to claim 1, characterized in that, The error criterion is: | - | / ≤10%; and | - | / ≤10%.

3. The method for predicting tunnel excavation deformation based on numerical simulation and parameter inversion according to claim 1, characterized in that, The parameter gradients of E1 and V1 are set to 90%E1, 95%E1, 105%E1, 110%E1, 90%V1, 95%V1, 105%V1, 110%V1.

4. The method for predicting tunnel excavation deformation based on numerical simulation and parameter inversion according to claim 1, characterized in that, The total length of the tunnel excavation is divided into multiple excavation segments of equal length, and each excavation segment is treated as an excavation condition. Each excavation condition is given a unified cross-sectional form, and a two-dimensional finite element model is established.

Images