A seismic input method for long-span tunnel design in soft soil areas

Abstract

The present invention relates to a seismic motion input method suitable for designing long-span tunnels in soft soil areas, comprising the following steps: generating a non-uniform seismic acceleration time history curve on the surface based on random vibration theory; inverting the non-uniform seismic acceleration time history curve on the surface to bedrock, vertically corresponding to the non-uniform seismic acceleration generation curves generated at multiple positions on the surface at the bedrock; filtering and baseline adjustment of the non-uniform seismic wave acceleration time history curve at the bedrock; and calculating the seismic response of the tunnel in the soft soil area under the action of a non-uniform earthquake using a finite element method.

Description

Technical Field

[0001] The invention relates to the field of geotechnical seismic research, and in particular to a seismic motion input method suitable for designing long-span tunnels in soft soil areas. Background Art

[0002] The prevention and control of tunnel earthquake disasters is a major issue related to the safety of life and property. It is an important part of improving public security and disaster prevention, mitigation, resistance and relief capabilities. Large-scale subway tunnel construction is booming in soft soil areas such as the Bohai Rim, the Yangtze River Delta, and the Pearl River Delta in my country. However, the engineering geological environment in soft soil areas is complex (thick soft soil layer, poor engineering properties, high groundwater level) and the seismic fortification intensity is high (≥7 degrees). Once an earthquake occurs, tunnels, as the main life engineering (traffic lifeline), will cause huge losses of life and property. A large number of earthquake geological disasters show that tunnel damage will experience a continuous collapse process such as cracking of pipe segment joints → water and sand leakage → tunnel cracking → foundation collapse. This is closely related to the inconsistency of earthquake excitation (the tunnel length exceeds 1 / 4 of the earthquake wavelength, and the difference in the propagation time of seismic waves causes different parts of the tunnel to bear inconsistent earthquake effects). Inconsistent earthquake excitation will increase the soil-structure interaction force and the internal force of the tunnel structure.

[0003] At present, the seismic design methods of tunnels mainly include cross-sectional design methods (free field deformation method, flexibility coefficient method, response displacement method) and tunnel longitudinal design methods (analytical solution method and response displacement method). The design methods are all consistent seismic excitation inputs, that is, the ground motion is applied as a whole at the bottom of the numerical or experimental model. However, a large number of studies have shown that due to the large length of long-span tunnels, non-uniform earthquake characteristics are obvious. For example, across different soil layers, due to the traveling wave effect of seismic waves, there is an obvious time difference effect in the seismic response of the tunnel. In addition, after refraction or reflection, the seismic waves show certain coherence effects and energy attenuation effects.

[0004] At present, the dynamic response analysis of long-span bridges, stadium structures and dams under non-uniform seismic excitation has been widely studied, but the relevant research on underground tunnels is still extremely rare. There are two main ways to generate non-uniform ground motions: first, using multi-vibration table experiments to simulate non-uniform seismic excitation (considering the traveling wave effect); second, using random vibration theory to generate random ground motions. For example, the patent "A multi-point earthquake synthesis method and system" (CN112069569A) has disclosed a method for generating non-uniform ground motions using random theory. However, this method only describes the input of surface ground motions, which is suitable for the design of above-ground structures and cannot be used for the design of underground structures. This method has not yet clarified the input mode of non-uniform ground motions, such as the acceleration time history curves, velocity time history curves and displacement time history curves of longitudinal non-uniform and transverse non-uniform seismic excitations.

[0005] Based on the above reasons, a seismic motion input method suitable for the design of long-span tunnels was developed, which provides an important reference for the seismic design of long-span tunnels in geotechnical engineering and has important scientific research value and academic significance for minimizing or avoiding safety accidents. Summary of the invention

[0006] The object of the present invention is to provide an analysis method with high accuracy that can be applied to the seismic design of long-span tunnels in soft soil areas. The technical solution of the present invention is as follows:

[0007] A seismic input method suitable for designing long-span tunnels in soft soil areas comprises the following steps:

[0008] Step 1: Generate the surface non-uniform earthquake acceleration time history curve using random vibration theory;

[0009] Step 2: Invert the surface non-uniform earthquake acceleration time history curve to the bedrock. The method is as follows:

[0010] ① Assuming the number of ground soil layers is N, the acceleration time history curve of the non-uniform earthquake on the surface is converted into the sum of finite harmonics in the frequency domain by Fourier transformation method, and the undetermined cosine coefficient A in the Fourier expansion formula is obtained: k and sine coefficient B k , as shown in formula (1):

[0011]

[0012] In the formula, f j (t) is the discrete acceleration time history signal of the non-uniform earthquake at the jth node on the surface, A0 is f j (t); ND is the total number of data points of the acceleration time series with equal step length, ND is an integer power of 2; k is a positive integer ranging from 1 to ND / 2-1; ω k is the discrete circular frequency, determined by formula (2):

[0013]

[0014] Where Dt is the time step;

[0015] ② The undetermined cosine coefficient A obtained by Fourier transform k With the sine coefficient B k , combined with formula (1) and (2), determine the amplitude A(ω) in the acceleration time history curve in the first layer of soil k ) and phase angle φ(ω k ) value;

[0016] ③According to the amplitude A(ω) of the acceleration time history curve in the first layer of soilk ) and phase angle φ(ω k ) and combined with formula (3) to determine the transfer parameters E1 and F1 in the first layer of soil:

[0017]

[0018] In the formula, i is an imaginary number and e is a natural constant;

[0019] ④Use m to represent the number of soil layers and calculate the effective self-weight stress σ' of each soil layer m [m], using indoor small strain test to obtain the empirical coefficient k 2,max , combined with formula (4) to determine the maximum dynamic shear modulus G of each soil layer max [m]:

[0020]

[0021] In the formula, p a is the atmospheric pressure;

[0022] ⑤According to the engineering characteristics of each layer of soil, obtain the maximum damping ratio β of each layer of soil max [m]; Combined with soil mechanics theory, determine the maximum shear strength τ of each layer of soil max [m], and the reference shear strain γ of each layer of soil is further calculated according to formula (5): r [m]:

[0023]

[0024] ⑥ The maximum dynamic shear modulus G of each layer max [m] and the maximum damping ratio β of each soil layer max [m] is used as the initial parameter, and the soil parameter K[m] of each layer is obtained by substitution formula (6):

[0025]

[0026] In the formula, ρ[m] is the density of each layer of soil, and η is the viscosity coefficient, which is determined by the damping ratio;

[0027] ⑦ After determining the transfer parameters E1, F1 and the first layer soil parameter K[1] in the first layer of soil, the partial derivative of the displacement equation (7) with respect to depth z is obtained to obtain the peak shear strain γ in the first layer of soil. p1 The expression (8) is:

[0028] U(z)1=E1e iK[1]z +F1e -iK[1]z , (7)

[0029] γ p1 =iK[1](E1e iK[1]z -F1e-iK[1]z ), (8)

[0030] ⑧For each soil layer m, calculate the equivalent dynamic shear modulus G in the soil layer according to equations (9) and (10): eq [m] and equivalent damping ratio β eq [m]; and the maximum dynamic shear modulus G max [m] and maximum damping ratio β max [m], if the error exceeds 1%, the equivalent dynamic shear modulus G eq [m] and equivalent damping ratio β eq [m] replaces G max [m] and β max Return to step ⑥ and recalculate the value of [m] until the accuracy is less than 1%:

[0031]

[0032]

[0033] In the formula, p(γ p ) is the peak probability density function of shear strain, which is determined according to random vibration theory; G(γ p )[m] is the dynamic shear modulus function of the shear strain of the mth layer of soil; η(γ p )[m] is the viscosity coefficient function of the shear strain of the mth layer of soil, which is related to the damping ratio;

[0034] ⑨ Combine formula (11) and formula (12), and infer the transfer parameters in the second layer of soil from the transfer parameters in the first layer of soil. Repeat steps ⑥ to ⑨, and infer the transfer parameters in the third layer of soil from the transfer parameters in the second layer of soil. And so on, until the transfer is inferred to the bedrock:

[0035]

[0036]

[0037] In the formula, h m-1 is the thickness of the m-1th soil layer; α m is the complex impedance ratio, E m and F m Transfer parameters to the mth layer of soil;

[0038] Assuming that the bedrock surface is located at the M-layer soil, the displacement time history curve equation (13) at the bedrock is obtained, and the second partial derivative with respect to time is obtained to obtain the acceleration time history curve equation (14) of the non-uniform earthquake at the bedrock:

[0039]

[0040]

[0041] After Fourier inverse transformation, the acceleration curve corresponding to the frequency is converted into the acceleration time history curve;

[0042] For the non-uniform seismic acceleration generation curves generated at multiple locations on the surface, it is necessary to follow steps ① to ⑨ to vertically correspond to the non-uniform seismic acceleration generation curves at the bedrock;

[0043] Step 3: Filtering and baseline adjustment of non-uniform seismic wave acceleration time history curves at bedrock

[0044] ① Read the non-uniform seismic acceleration time history curve at the bedrock, integrate the inverted non-uniform seismic acceleration time history curve at the bedrock to obtain the non-uniform seismic velocity time history curve at the bedrock;

[0045] ② Fitting the quadratic polynomial c1t of the non-uniform seismic velocity time history at the bedrock 2 +c2t+c3, where c1, c2 and c3 are unknown parameters that can be determined by the least squares method;

[0046] ③ Subtract the derivative 2c1t+c2 of the quadratic polynomial of the non-uniform seismic velocity time history at the bedrock obtained by step ② from the non-uniform seismic acceleration time history curve at the bedrock obtained by step ①, and re-obtain the corrected non-uniform seismic acceleration time history curve at the bedrock;

[0047] ④ Using a fourth-order Butterworth filter to filter the corrected non-uniform earthquake acceleration time history curve at the bedrock obtained in step ③;

[0048] ⑤ Integrate and quadratically integrate the non-uniform seismic acceleration time history curve at the bedrock obtained after filtering in step ④, and obtain the non-uniform seismic velocity time history at the bedrock and the non-uniform seismic displacement time history at the bedrock respectively, and use the fourth-order Butterworth filter to filter the obtained non-uniform seismic velocity time history at the bedrock and the non-uniform seismic displacement time history at the bedrock respectively, and obtain the corrected non-uniform seismic velocity time history and non-uniform seismic displacement time history;

[0049] Step 4: Use the finite element method to calculate the seismic response of the tunnel in the soft soil area under the action of non-uniform earthquake.

[0050] Step 4 is as follows:

[0051] ① Combine the length of the long-span tunnel and the engineering geological survey report to establish a numerical model of the tunnel and soil

[0052] ② Boundary condition setting: The filtered non-uniform earthquake acceleration time history curve at the bedrock, the corrected non-uniform earthquake velocity time history at the bedrock and the non-uniform earthquake displacement time history at the bedrock are obtained by step 3 as the numerical simulation input signal; in the numerical analysis, the bottom of the tunnel is divided into n regional units, which is the same as the number of acceleration time history curves of the surface non-uniform earthquake generated in step 1, and the inversion depth is consistent with the formation depth established by the numerical model; the numerical model is set as a viscoelastic boundary around to prevent the refraction and reflection of seismic waves; the upper boundary of the model is a free boundary;

[0053] ③Constitutive selection: set appropriate soil layer parameters and select the equivalent nonlinear Hardin-Drenvich constitutive model;

[0054] ④ Calculate the seismic response of tunnels in soft soil areas under non-uniform earthquake action. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 Schematic diagram of the steps of the present invention

[0056] Figure 2 Flow chart for inversion of non-uniform earthquake acceleration time history at bedrock

[0057] Figure 3 The inversion result of the non-uniform earthquake acceleration time history curve at the bedrock

[0058] Figure 4 Flowchart for filtering and baseline of non-uniform seismic wave acceleration time history curves at bedrock

[0059] Figure 5 Schematic diagram of seismic input for underground tunnel in soft soil Specific embodiments

[0060] The present invention is described in detail below in conjunction with specific embodiments and accompanying drawings. In the following description, for the purpose of explanation and not limitation, specific details are set forth to help fully understand the present invention. However, it is apparent to those skilled in the art that the present invention may also be practiced in other embodiments that depart from these specific details.

[0061] It should be noted that, in order to avoid obscuring the present invention due to unnecessary details, only the device structure and / or processing steps closely related to the solution according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0062] The embodiment of the present invention provides a method and system for inputting earthquake motion suitable for designing a long-span tunnel in a soft soil area, as shown in FIG1 , comprising the following steps:

[0063] Step 1: Generation of surface non-uniform earthquake acceleration time history curve

[0064] By using random vibration theory and referring to the book Virtual Excitation Method of Random Vibration by Lin Jiahao and Zhang Yahui, the sixth chapter "Application of Virtual Excitation Method in Earthquake Engineering" is used to generate the acceleration time history curve of non-uniform earthquake on the surface.

[0065] Step 2: Inversion of non-uniform surface earthquake acceleration time history curves to bedrock

[0066] The implementation of this step includes nine sub-steps, and its flow chart is as follows Figure 2 shown.

[0067] ① Assuming the number of ground soil layers is N, the acceleration time history curve of the non-uniform earthquake on the surface is converted into the sum of finite harmonics in the frequency domain by Fourier transformation method, and the undetermined cosine coefficient A in the Fourier expansion formula is obtained: k and sine coefficient B k , as shown in formula (1):

[0068]

[0069] In the formula, f j (t) is the discrete acceleration time history signal of the non-uniform earthquake at the jth node on the surface, A0 is f j ND is the total number of data points of the acceleration time series with equal step length. ND is generally equal to an integer power of 2. k is a positive integer ranging from 1 to ND / 2-1. k is the discrete circular frequency, determined by formula (2):

[0070]

[0071] Where Dt is the time step.

[0072] ② The undetermined cosine coefficient A obtained by Fourier transform k With the sine coefficient B k , combined with formula (1) and (2), determine the amplitude A(ω) in the acceleration time history curve in the first layer of soil k ) and phase angle φ(ω k ) value.

[0073] ③According to the amplitude A(ω) of the acceleration time history curve in the first layer of soil k ) and phase angle φ(ω k ) and combined with formula (3) to determine the transfer parameters E1 and F1 in the first soil layer.

[0074]

[0075] In the formula, i is an imaginary number and e is a natural constant, which is 2.72.

[0076] ④ Calculate the effective self-weight stress σ' of each soil layer m [m], using indoor small strain test to obtain the empirical coefficient k 2,max , combined with formula (4) to determine the maximum dynamic shear modulus G of each soil layer max [m] (m represents the number of soil layers, for example, m = 1 is the first soil layer). The determination of the various coefficients or parameters involved in formula (4) can be referred to the literature "Chi Shichun, Guo Xiaoxia, Yang Jun, Lin Gao. Study on small strain characteristics and threshold strain of dynamic Hardin-Drnevich model of soil [J]. Chinese Journal of Geotechnical Engineering, 2008 (02): 243-249".

[0077]

[0078] In the formula, p a is atmospheric pressure.

[0079] ⑤According to the engineering characteristics of each layer of soil, obtain the maximum damping ratio β of each layer of soil max Combined with soil mechanics theory, the maximum shear strength τ of each layer of soil is determined max [m], and the reference shear strain γ of each layer of soil is further calculated according to formula (5): r [m].

[0080]

[0081] ⑥ The maximum dynamic shear modulus G of each layer max [m] and the maximum damping ratio β of each soil layer max [m] is used as the initial parameter, and the soil parameter K[m] of each layer can be obtained by substitution formula (6).

[0082]

[0083] Where ρ[m] is the density of each layer of soil, and η is the viscosity coefficient, which can be determined by the damping ratio.

[0084] ⑦ After the transfer parameters E1, F1 and the first layer soil parameter K[1] in the first layer soil are determined, the partial derivative of the displacement equation (7) with respect to the depth z is obtained to obtain the shear strain peak value γ in the first layer soil p1 Expression (8).

[0085] U(z)1=E1e iK[1]z +F1e -iK[1]z , (7)

[0086] γ p1 =iK[1](E1e iK[1]z -F1e -iK[1]z ), (8)

[0087] ⑧ According to equations (9) and (10), calculate the equivalent dynamic shear modulus G in each layer of soil: eq [m] and equivalent damping ratio β eq [m] and maximum dynamic shear modulus G max [m] and maximum damping ratio β max [m], if the error exceeds 1%, the equivalent dynamic shear modulus G eq [m] and equivalent damping ratio β eq [m] replaces G max [m] and β max Return to step ⑥ and recalculate until the accuracy is less than 1%.

[0088]

[0089]

[0090] In the formula, p(γ p ) is the peak probability density function of shear strain, which can be determined by random vibration theory. p )[m] is the dynamic shear modulus function of the shear strain of the mth layer of soil. p )[m] is the viscosity coefficient function of the shear strain of the mth layer of soil and is related to the damping ratio.

[0091] ⑨ Combine formula (11) and formula (12), and infer the transfer parameters in the second layer of soil from the transfer parameters in the first layer of soil. Repeat steps ⑥ to ⑨, and infer the transfer parameters in the third layer of soil from the transfer parameters in the second layer of soil, and so on, until the transfer is inferred to the bedrock.

[0092]

[0093]

[0094] In the formula, h m-1 is the thickness of the m-1th soil layer. m is the complex impedance ratio, E m and F m Transfer parameters to the mth layer of soil.

[0095] Assuming that the bedrock surface is located at the M-layer soil, the displacement time history curve equation at the bedrock can be obtained (13), and the acceleration time history curve equation of the non-uniform earthquake at the bedrock can be obtained by taking the second partial derivative with respect to time (14):

[0096]

[0097]

[0098] Finally, the acceleration curve corresponding to the frequency is converted into an acceleration time-history curve through inverse Fourier transformation.

[0099] The non-uniform earthquake acceleration curves generated at multiple locations on the surface need to be generated vertically corresponding to the non-uniform earthquake acceleration curves at the bedrock according to steps ① to ⑨. Combining the above theory, three non-uniform earthquake acceleration time history curves at the bedrock position were randomly selected, such as Figure 3 shown.

[0100] Step 3: Filtering and baseline adjustment of non-uniform seismic wave acceleration time history curves at bedrock

[0101] The implementation of this step includes five sub-steps, and its flow chart is as follows Figure 4 shown.

[0102] ① Read in the non-uniform seismic acceleration time history curve at the bedrock, and integrate the inverted non-uniform seismic acceleration time history curve at the bedrock to obtain the non-uniform seismic velocity time history curve at the bedrock.

[0103] ② Fitting the quadratic polynomial c1t of the non-uniform seismic velocity time history at the bedrock 2 +c2t+c3, where c1, c2 and c3 are unknown parameters that can be determined by the least squares method.

[0104] ③ Subtract the derivative 2c1t+c2 of the quadratic polynomial of the non-uniform seismic velocity time history at the bedrock obtained by fitting in step ② from the non-uniform seismic acceleration time history curve at the bedrock obtained in step ①, and re-obtain the corrected non-uniform seismic acceleration time history curve at the bedrock.

[0105] ④ A fourth-order Butterworth filter is used to filter the corrected non-uniform earthquake acceleration time history curve at the bedrock obtained in step ③.

[0106] ⑤ Integrate and quadratically integrate the non-uniform seismic acceleration time history curve at the bedrock obtained after filtering in step ④, and obtain the non-uniform seismic velocity time history at the bedrock and the non-uniform seismic displacement time history at the bedrock respectively. Use a fourth-order Butterworth filter to filter the obtained non-uniform seismic velocity time history at the bedrock and the non-uniform seismic displacement time history at the bedrock respectively, and obtain the corrected non-uniform seismic velocity time history and non-uniform seismic displacement time history.

[0107] Step 4: Combine the finite element calculation method to calculate the seismic response of the tunnel in the soft soil area under the action of non-uniform earthquake. The implementation of this step includes four sub-steps, and its flow chart is as follows: Figure 5 shown.

[0108] ① Combine the length of the long-span tunnel and the engineering geological survey report to establish a numerical model of the tunnel and soil

[0109] ② Boundary condition setting: Use step three to obtain the filtered non-uniform earthquake acceleration time history curve at the bedrock, the corrected non-uniform earthquake velocity time history at the bedrock, and the non-uniform earthquake displacement time history at the bedrock as the numerical simulation input signal. In the numerical analysis, the bottom of the tunnel is divided into n regional units, which is the same number as the acceleration time history curves of the surface non-uniform earthquake generated in step one, and the inversion depth is consistent with the stratigraphic depth established by the numerical model. The numerical model is set as a viscoelastic boundary on all sides to prevent refraction and reflection of seismic waves. The upper boundary of the model is a free boundary, such as Figure 5 shown.

[0110] ③Constitutive selection: Set appropriate soil layer parameters and select the equivalent nonlinear Hardin-Drenvich constitutive model.

[0111] ④Numerical analysis and calculation.

[0112] According to the above method, the non-uniform earthquake acceleration time history curve, non-uniform earthquake velocity time history curve and non-uniform earthquake displacement time history curve of the long-span tunnel in the soft soil area can be obtained, which can be combined with the finite element calculation method to calculate the seismic response of the tunnel in the soft soil area under the action of non-uniform earthquake.

[0113] This patent is a new algorithm for quickly and accurately calculating the seismic response of long-span tunnels in soft soil areas, which can provide a design reference for the seismic design of long-span tunnels.

Claims

1. A seismic input method suitable for designing long-span tunnels in soft soil areas, comprising the following steps: Step 1: Generate the surface non-uniform earthquake acceleration time history curve based on random vibration theory; Step 2: Invert the surface non-uniform earthquake acceleration time history curve to the bedrock; the method is as follows: ① Assuming the number of ground soil layers is N, the acceleration time history curve of the non-uniform earthquake on the surface is converted into the sum of finite harmonics in the frequency domain by Fourier transformation method, and the undetermined cosine coefficient A in the Fourier expansion formula is obtained: k and sine coefficient B k , as shown in formula (1): In the formula, f j (t) is the discrete acceleration time history signal of the non-uniform earthquake at the jth node on the surface, A0 is f j (t); ND is the total number of data points of the acceleration time series with equal step length, and ND is an integer power of 2; k is a positive integer ranging from 1 to ND / 2-1; ω k is the discrete circular frequency, determined by formula (2): in, Dt is the time step; ② The undetermined cosine coefficient A obtained by Fourier transform k With the sine coefficient B k , combined with formula (1) and (2), determine the amplitude A(ω) in the acceleration time history curve in the first layer of soil k ) and phase angle φ(ω k ) value; ③According to the amplitude A(ω) of the acceleration time history curve in the first layer of soil k ) and phase angle φ(ω k ) and combined with formula (3) to determine the transfer parameters E1 and F1 in the first layer of soil: In the formula, i is an imaginary number and e is a natural constant; ④Use m to represent the number of soil layers and calculate the effective self-weight stress σ' of each soil layer m [m], using indoor small strain test to obtain the empirical coefficient k 2,max , combined with formula (4) to determine the maximum dynamic shear modulus G of each soil layer max [m]: In the formula, p a is the atmospheric pressure; ⑤According to the engineering characteristics of each layer of soil, obtain the maximum damping ratio β of each layer of soil max [m]; Combined with soil mechanics theory, determine the maximum shear strength τ of each layer of soil max [m], and the reference shear strain γ of each layer of soil is further calculated according to formula (5): r [m]: ⑥ The maximum dynamic shear modulus G of each layer max [m] and the maximum damping ratio β of each soil layer max [m] is used as the initial parameter, and the soil parameter K[m] of each layer is obtained by substitution formula (6): In the formula, ρ[m] is the density of each layer of soil, and η is the viscosity coefficient, which is determined by the damping ratio; ⑦ After determining the transfer parameters E1, F1 and the first layer soil parameter K[1] in the first layer of soil, the partial derivative of the displacement equation (7) with respect to depth z is obtained to obtain the peak shear strain γ in the first layer of soil. p1 The expression (8) is: U(z)1=E1e iK[1]z +F1e -iK[1]z , (7) γ p1 =iK[1](E1e iK[1]z -F1e -iK[1]z ), (8) ⑧For each soil layer m, calculate the equivalent dynamic shear modulus G in the soil layer according to equations (9) and (10): eq [m] and equivalent damping ratio β eq [m]; and the maximum dynamic shear modulus G max [m] and maximum damping ratio β max [m], if the error exceeds 1%, the equivalent dynamic shear modulus G eq [m] and equivalent damping ratio β eq [m] replaces G max [m] and β max Return to step ⑥ and recalculate the value of [m] until the accuracy is less than 1%: In the formula, p(γ p ) is the peak probability density function of shear strain, which is determined according to random vibration theory; G(γ p )[m] is the dynamic shear modulus function of the shear strain of the mth layer of soil; η(γ p )[m] is the viscosity coefficient function of the shear strain of the mth layer of soil, which is related to the damping ratio; ⑨ Combine formula (11) and formula (12), and infer the transfer parameters in the second layer of soil from the transfer parameters in the first layer of soil. Repeat steps ⑥ to ⑨, and infer the transfer parameters in the third layer of soil from the transfer parameters in the second layer of soil. And so on, until the transfer is inferred to the bedrock: In the formula, h m-1 is the thickness of the m-1th soil layer; α m is the complex impedance ratio, E m and F m Transfer parameters to the mth layer of soil; Assuming that the bedrock surface is located at the M-layer soil, the displacement time history curve equation (13) at the bedrock is obtained, and the second partial derivative with respect to time is obtained to obtain the acceleration time history curve equation (14) of the non-uniform earthquake at the bedrock: After Fourier inverse transformation, the acceleration curve corresponding to the frequency is converted into the acceleration time history curve; For the non-uniform seismic acceleration generation curves generated at multiple locations on the surface, it is necessary to follow steps ① to ⑨ to vertically correspond to the non-uniform seismic acceleration generation curves at the bedrock; Step 3: Filtering and baseline adjustment of non-uniform seismic wave acceleration time history curves at bedrock ① Read the non-uniform seismic acceleration time history curve at the bedrock, integrate the inverted non-uniform seismic acceleration time history curve at the bedrock to obtain the non-uniform seismic velocity time history curve at the bedrock; ② Fitting the quadratic polynomial c1t of the non-uniform seismic velocity time history at the bedrock 2 +c2t+c3, where c1, c2 and c3 are unknown parameters that can be determined by the least squares method; ③ Subtract the derivative 2c1t+c2 of the quadratic polynomial of the non-uniform seismic velocity time history at the bedrock obtained by step ② from the non-uniform seismic acceleration time history curve at the bedrock obtained by step ①, and re-obtain the corrected non-uniform seismic acceleration time history curve at the bedrock; ④ Using a fourth-order Butterworth filter to filter the corrected non-uniform earthquake acceleration time history curve at the bedrock obtained in step ③; ⑤ Integrate and quadratically integrate the non-uniform seismic acceleration time history curve at the bedrock obtained after filtering in step ④, and obtain the non-uniform seismic velocity time history at the bedrock and the non-uniform seismic displacement time history at the bedrock respectively, and use the fourth-order Butterworth filter to filter the obtained non-uniform seismic velocity time history at the bedrock and the non-uniform seismic displacement time history at the bedrock respectively, and obtain the corrected non-uniform seismic velocity time history and non-uniform seismic displacement time history; Step 4: Use the finite element method to calculate the seismic response of the tunnel in the soft soil area under the action of non-uniform earthquake.

2. The input method according to claim 1, characterized in that: The method for step 4 is as follows: ① Combine the length of the long-span tunnel and the engineering geological survey report to establish a numerical model of the tunnel and soil ② Boundary condition setting: Use step 3 to obtain the filtered non-uniform earthquake acceleration time history curve at the bedrock, the corrected non-uniform earthquake velocity time history at the bedrock, and the non-uniform earthquake displacement time history at the bedrock as the numerical simulation input signal; In the numerical analysis, the tunnel bottom is divided into n regional units, which is the same number as the acceleration time history curves of the surface non-uniform earthquake generated in step 1, and the inversion depth is consistent with the stratum depth established by the numerical model; the numerical model is set as a viscoelastic boundary around to prevent the refraction and reflection of seismic waves; the upper boundary of the model is a free boundary; ③Constitutive selection: set appropriate soil layer parameters and select the equivalent nonlinear Hardin-Drenvich constitutive model; ④ Calculate the seismic response of tunnels in soft soil areas under non-uniform earthquake action.

Images