A quick seismic inversion method for seismic data of mine rock roadway TBM

Abstract

The application discloses a kind of mine rock lane TBM quick seismic inversion method of excavation seismic data, it is related to mine seismic data processing technical field, first, according to the true velocity model, forward calculation is carried out, and seismic forward response is obtained;Second, the seismic full waveform inversion objective function is constructed;Then, the partial derivative is solved with respect to velocity, and the iterative formula of gradient update is obtained, and on the basis of conventional conjugate gradient parameters, by assigning different types of conjugate gradient parameter weight factor, the modified conjugate gradient parameter factor is constructed, and to avoid the occurrence of cycle at the unstable point of objective function, the maximum value of current conjugate gradient value and zero is taken as the next gradient value when the conjugate gradient parameter factor is updated;Finally, the iterative formula of velocity update is obtained, and the optimal solution of velocity model is obtained.The application guarantees the quick inversion requirement of TBM excavation seismic data, provides geological prediction results in time for the front of excavation working face, and guarantees the safety, high efficiency of TBM rapid excavation.

Description

Technical Field

[0001] This invention relates to the field of mine seismic data processing technology, and more specifically to a rapid seismic inversion method for seismic data generated during the excavation of a TBM (Tunnel Boring Machine) in a mine rock tunnel. Background Technology

[0002] With the increasing demands for safe and efficient tunneling in mines, a rapid tunneling technology using TBMs, distinct from the traditional drill-and-blast method, has been gradually applied to mine rock tunneling. However, during TBM construction, adverse geological conditions ahead of the working face, such as faults, water-rich areas, and zones of varying soft and hard rock, seriously threaten TBM construction safety. Furthermore, due to the high speed of TBM tunneling, conventional seismic advance detection technology can no longer meet its requirements for rapid and high-precision detection and imaging.

[0003] Therefore, how to utilize seismic data to achieve rapid and high-precision detection in order to meet the needs of TBM rapid tunneling for advanced geological prediction is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] In view of this, this invention provides a rapid seismic inversion method for TBM (Tunnel Boring Machine) seismic data during tunneling in mine rock tunnels. This method utilizes the conjugate gradient method to solve the nonlinear inversion objective function. It only uses first-order derivative information, avoiding the problem of calculating the second-order Hessian matrix in conventional inversion methods. Furthermore, to improve inversion speed, a modified conjugate gradient parameter factor is used, ensuring the inversion results are optimized, significantly improving both speed and accuracy. The inversion process converges overall and exhibits higher stability. This guarantees the need for rapid inversion of TBM seismic data during tunneling, providing timely geological prediction results for the tunneling face and ensuring the safety and efficiency of rapid TBM tunneling.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A rapid seismic inversion method for seismic data obtained during tunnel boring machine (TBM) excavation in mine rock tunnels includes the following steps:

[0007] Step 1: Define the initial velocity model, mesh the initial velocity model, and perform forward modeling based on the wave equation to obtain the seismic forward response f(m). The wave equation is expressed as:

[0008]

[0009] Where U represents displacement; x and z represent the ordinate and abscissa of any point after the initial velocity model is meshed, respectively; v(x,z) represents the velocity value at the position with abscissa x and ordinate z on the mesh; and t represents time.

[0010] Step 2: Based on the earthquake forward response f(m) and the actual observed earthquake data d obs Construct the inversion objective function;

[0011]

[0012] Where P(m) is the inversion objective function; m is the inversion parameter; f(m) is the seismic forward modeling response; d obs The data used are actual observed earthquake data, and λ is a weighting factor. For empirical models;

[0013] Step 3: Take the partial derivative of the inversion objective function with respect to the inversion parameter m to obtain the gradient, which is expressed as:

[0014]

[0015] In the formula, g(m) is the gradient; v i,j The velocity values ​​on the i-th row and j-th column of the mesh after the initial velocity model is meshed.

[0016] Further simplification of the gradient yields the gradient search direction, expressed as:

[0017] d k+1 =g k +d k β k+1 (4)

[0018] In the formula, d k+1 Let β be the gradient search direction for the (k+1)th iteration. k+1 The modified conjugate gradient parameter factor for the (k+1)th iteration; g k d is the gradient of the k-th iteration; k This indicates the gradient search direction in the k-th iteration;

[0019] Step 4: Select the modified conjugate gradient parameter factor;

[0020] Fast hybrid conjugate gradient parameter factor β inverted by the hybrid fast conjugate gradient method HFCG As shown in (5):

[0021]

[0022] HS-type conjugate gradient parameter factor As in formula (6)

[0023]

[0024] DY-type conjugate gradient parameter factor As shown in formula (7)

[0025]

[0026] By combining (5), (6), and (7) and introducing a weighting factor and a gradient restart mechanism, we obtain the modified conjugate gradient parameter factor. As in equation (8):

[0027]

[0028] Where α represents the weighting factor; α n α n+1 These represent the weight factors at the nth and (n+1)th iterations, respectively; d k-1 This indicates the gradient search direction in the (k-1)th iteration; Let g represent the gradient of the k-th iteration. k transpose; d represents the gradient search direction in the (k-1)th iteration. k-1 transpose of; y k-1 y represents the difference in gradient between two consecutive iterations. k-1 =g k -g k-1 ;

[0029] Step 5: Substitute formulas (1), (3), (4), and (8) into formula (2) to further simplify and obtain the velocity model update formula as shown in formula (9):

[0030] m k+1 =m k +γ k d k+1 (9)

[0031] In the formula: γ k Let m be the step size for the k-th iteration. k Let m be the velocity model value for the k-th iteration. k+1 For the (k+1)th iteration velocity model, d k+1 This represents the direction of the (k+1)th gradient search.

[0032] Step 6: Determine if the iteration termination condition is met. If it is, output the optimal solution m. k+1 Otherwise, return to step 1 until the iteration termination condition is met.

[0033] Preferably, the iteration termination condition is that the error between two consecutive updated velocity models is less than or equal to 2%, i.e., m k+1 -m k ≤2%. The updated velocity module that meets the iteration termination condition is the optimal velocity model, and high-precision detection imaging is obtained based on the model results.

[0034] Preferably, the gradient restart mechanism is as follows: during the iterative solution of the objective function, when the conjugate gradient parameter factor is calculated to be negative, zero is introduced to recalculate the gradient.

[0035] As can be seen from the above technical solution, compared with the prior art, this invention discloses a rapid seismic inversion method for TBM-driven seismic data in mine tunnels. To meet the rapid inversion requirements of TBM rapid tunneling for advanced geological prediction of seismic data during tunneling, the method first performs forward modeling calculations based on the actual velocity model to obtain the seismic forward response; secondly, it constructs a seismic full-waveform inversion objective function; then, it calculates the partial derivative with respect to velocity to obtain the iterative formula for gradient update, and constructs a modified conjugate gradient parameter. This is achieved by assigning different types of weighting factors to the conventional conjugate gradient parameters to obtain modified conjugate gradient parameter factors. To avoid loops at unstable points in the objective function, the maximum value between the current conjugate gradient value and zero is taken as the next gradient value during conjugate gradient parameter factor update; finally, the iterative formula for velocity update is obtained, leading to the optimal solution of the velocity model. This invention significantly improves both computational speed and convergence, not only with faster inversion calculations but also with a gradually converging overall inversion process, meeting the requirements for rapid inversion processing of TBM-driven seismic data. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0037] Figure 1 The attached figure is a flowchart of a rapid inversion method for seismic data during tunnel boring machine (TBM) excavation in mine rock tunnels provided by the present invention;

[0038] Figure 2 The attached figure is a schematic diagram of the actual velocity model provided by the present invention;

[0039] Figure 3 The attached figure is a schematic diagram of the initial velocity model provided by the present invention;

[0040] Figure 4 The attached figure is a schematic diagram of the conventional conjugate gradient full waveform inversion result provided by the present invention;

[0041] Figure 5 The attached figure is a schematic diagram of the fast full waveform inversion result based on the modified conjugate gradient provided by the present invention;

[0042] Figure 6The attached figure is a schematic diagram of the convergence result of the fast full waveform inversion based on the modified conjugate gradient provided by the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] This invention discloses a rapid seismic inversion method for seismic data obtained during the excavation of a TBM (Tunnel Boring Machine) in a mine rock tunnel. The process is as follows: Figure 1 As shown, the specific steps are as follows:

[0045] S1: Initial velocity model as follows Figure 3 As shown, the initial velocity model is gridded, and forward modeling is performed based on the wave equation, i.e., formula (1), to obtain the earthquake forward modeling response f(m);

[0046]

[0047] Where U represents displacement; x and z represent the ordinate and abscissa of any point after the initial velocity model is meshed, respectively; v(x,z) represents the velocity value at the position with abscissa x and ordinate z on the mesh; and t represents time.

[0048] S2: Construct the inversion objective function based on the measured seismic records and seismic forward modeling data as shown in formula (2);

[0049]

[0050] Where P(m) is the inversion objective function; m is the inversion parameter; f(m) is the seismic forward modeling response; d obs The data used are actual observed earthquake data, and λ is a weighting factor. For empirical models;

[0051] S3: Take the partial derivative of the inversion objective function, i.e., formula (2), with respect to the parameter m to obtain the gradient expression.

[0052]

[0053] In the formula, g(m) is the gradient; v i,j The velocity values ​​on the i-th row and j-th column of the mesh after the initial velocity model is meshed.

[0054] Further simplification of equation (3) yields the gradient search direction as shown in equation (4):

[0055] dk+1 =g k +d k β k+1 (4)

[0056] In the formula, d k+1 Let β be the gradient search direction for the (k+1)th iteration. k+1 g is the conjugate gradient parameter factor for the (k+1)th iteration; k d is the gradient of the k-th iteration; k This indicates the gradient search direction in the k-th iteration;

[0057] S4: To improve the inversion speed, the conjugate gradient parameter factor β... k+1 An improvement was made, and a modified conjugate gradient parameter factor was proposed. The conventional conjugate gradient full-wave inversion results are as follows: Figure 4 The fast full-waveform inversion results based on the modified conjugate gradient are as follows: Figure 5 As shown;

[0058] Fast hybrid conjugate gradient parameter factor β inverted by the hybrid fast conjugate gradient method HFCG As shown in formula (5):

[0059]

[0060] HS-type conjugate gradient parameter factor As shown in formula (6):

[0061]

[0062] DY-type conjugate gradient parameter factor As shown in formula (7)

[0063]

[0064] Combining equations (5), (6), and (7), and introducing a weighting factor and a gradient restart mechanism, we obtain the modified conjugate gradient parameter factor as shown in equation (8).

[0065]

[0066] Where α represents the weighting factor; α n α n+1 These represent the weight factors at the nth and (n+1)th iterations, respectively; d k-1 This indicates the gradient search direction in the (k-1)th iteration; Let g represent the gradient of the k-th iteration. k transpose; d represents the gradient search direction in the (k-1)th iteration. k-1 transpose of; y k-1y represents the difference in gradient between two consecutive iterations. k-1 =g k -g k-1 ;

[0067] S5: Substituting formulas (1), (3), (4), and (8) into formula (2) and further simplifying, we obtain the velocity model update formula as shown in formula (9):

[0068] m k+1 =m k +γ k d k+1 (9)

[0069] In the formula: γ k Let m be the step size for the k-th iteration. k Let m be the velocity model value for the k-th iteration. k+1 For the (k+1)th iteration velocity model, d k+1 This represents the gradient search direction value for the (k+1)th iteration.

[0070] S6: Determine if the iteration termination condition is met, such as the residual between two adjacent updated velocity models being less than or equal to 2%, m k+1 -m k If the value is ≤2%, then output the optimal solution m. k+1 Conversely, if the condition is not met, the iteration returns to S1 until the termination condition is satisfied. The convergence result of the fast full-waveform inversion based on the modified conjugate gradient is as follows: Figure 6 As shown, the optimal solution m k+1 That is, the real speed model, such as Figure 2 As shown, high-precision detection imaging is obtained based on the model results.

[0071] In this embodiment, the advantage of the hybrid conjugate gradient method over the traditional conjugate gradient method lies in the introduction of not only a hybrid conjugate gradient factor but also a gradient restart mechanism. This measure greatly improves the calculation speed of the objective function iteration while ensuring the convergence of the iteration process and the high accuracy of the inversion results. Therefore, its model iteration solution process is faster than the traditional inversion method, resulting in shorter processing time for TBM-driven seismic data and facilitating faster TBM data processing and imaging.

[0072] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0073] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A rapid seismic inversion method for seismic data obtained during tunnel boring machine (TBM) excavation in mine rock tunnels, characterized in that, Includes the following steps: Step 1: Set up an initial velocity model, mesh the initial velocity model, perform forward modeling based on the wave equation, and obtain the earthquake forward modeling response; Step 2: Collect actual observed seismic data and construct an inversion objective function based on the seismic forward modeling response; Step 3: Obtain the gradient by taking the partial derivative of the inverted objective function, and simplify to obtain the conjugate gradient direction; Based on the conjugate gradient direction and gradient, the fast hybrid conjugate gradient method is used to calculate the fast hybrid conjugate gradient parameter factor. HS-type conjugate gradient parameter factor and DY-type conjugate gradient parameter factor , respectively represented as: ； ； ； Solve for the modified conjugate gradient parameter factor by simultaneously solving for the three types of conjugate gradient parameter factors, based on the weighting factor and gradient restart mechanism. , is represented as: ； in, Indicates the weighting factor; , These represent the weight factors at the nth and (n+1)th iterations, respectively; , These represent the gradient search directions for the k-th and (k-1)-th iterations, respectively. Describes the gradient of the k-th iteration. transpose; This indicates the gradient search direction in the (k-1)th iteration. transpose; y represents the difference in gradient between two consecutive iterations. k-1 = g k -g k-1 ; Step 4: Using the hybrid fast conjugate gradient method, calculate the modified conjugate gradient parameter factor based on the conjugate gradient direction and gradient; Step 5: Substitute the wave equation, gradient, conjugate gradient direction, and modified conjugate gradient parameter factor into the inversion objective function to obtain the update velocity model; Step 6: If the iteration termination condition is met, output the updated velocity model; otherwise, set the updated velocity model to the initial velocity model and return to step 1.

2. The rapid seismic inversion method for TBM-driven seismic data in mine rock tunnels according to claim 1, characterized in that, The wave equation is expressed as: ； in, represents displacement; x and z represent the ordinate and abscissa of any point after the initial velocity model is meshed, respectively; v(x,z) represents the velocity value at the position with abscissa x and ordinate z on the mesh; t represents time.

3. The rapid seismic inversion method for TBM-driven seismic data in mine rock tunnels according to claim 1, characterized in that, The inversion objective function is expressed as: ； in, The objective function is the inversion function; These are the inversion parameters; Forward modeling of earthquake response; To obtain actual earthquake data, As a weighting factor, It is an empirical model.

4. A rapid seismic inversion method for seismic data obtained during TBM excavation in mine rock tunnels according to claim 3, characterized in that, In step 3, the inversion objective function is analyzed with respect to the inversion parameters. The partial derivative is expressed as: ； in, For gradient; The velocity value in the i-th row and j-th column after the initial velocity model is meshed. The gradient search direction is obtained by simplifying the gradient, and is represented as: ； in, This represents the gradient search direction for the (k+1)th iteration. For the modified conjugate gradient parameter factor in the (k+1)th iteration; d is the gradient of the k-th iteration; k This indicates the gradient search direction in the k-th iteration.

5. A rapid seismic inversion method for seismic data from TBM excavation in mine rock tunnels according to claim 4, characterized in that, The updated velocity model in step 6 is represented as follows: ； in, Let k be the step size of the kth iteration. This represents the velocity model value for the k-th iteration. For the (k+1)th iteration velocity model, This represents the direction of the gradient search in the (k+1)th iteration.

6. A rapid seismic inversion method for seismic data obtained during tunnel boring machine (TBM) excavation in mine rock tunnels according to claim 1, characterized in that, The iteration terminates when the error between two consecutive updates of the rate model is less than or equal to 2%.

Images