JP4480155B2 - Prediction method of geological composition of natural ground - Google Patents

Description

  The present invention relates to a geological composition prediction method for identifying a natural ground constant in the process of excavating the natural ground.

  For example, in the construction of rockfill dams, the rock material used is taken from the rough rock mountain near the site, but the rock classification of the rock material differs for each dam site, so the rock material required efficiently from the rough rock mountain. Are required to be collected. For this reason, drilling surveys of rough rocks are conducted in advance, and the rocks of the required quality are collected efficiently after a rough understanding of the natural rock composition.

  However, it is difficult to accurately understand the geological composition of the natural ground only by the drilling survey conducted in advance, and by collecting rocks of the rock classification that is not expected during the excavation, In some cases, rocks were collected excessively, and necessary rock-grade rocks were excavated and insufficient, resulting in a delay in the process.

  On the other hand, in recent years, various geophysical exploration methods have been proposed and put into practical use as an alternative to boring surveys. Specifically, methods using surface waves (Rayleigh waves), electromagnetic waves, and reflected waves have been proposed.

  As a method of using the surface wave, for example, in Patent Document 1 below, a Rayleigh wave is generated by exciting a face with a vibrator attached at substantially the center of a tunnel face, and a predetermined distance in the radial direction from the vibrator. A method is disclosed in which a Rayleigh wave propagating on a face is detected by two detectors installed at a distance from each other, and a ground state such as a fault fracture zone is investigated by analyzing a propagation time difference of the Rayleigh wave.

  As a method of using electromagnetic waves, for example, in Patent Document 2 below, the velocity of electromagnetic waves is determined from reflection data representing the current geological situation, the relative permittivity of the current ground is calculated, and the amplitude generated by electromagnetic wave propagation Estimate the distance attenuation with respect to the propagation distance from the attenuation, and calculate the original reflection coefficient from the reflection phase by correcting this, and estimate the relative dielectric constant of the geology that appears after the reflection surface. A method for predicting the geology of the soil is disclosed.

As a method using a reflected wave, for example, in Patent Documents 3 and 4 below, an exciter and a receiving device are installed in an already excavated mine, and the reflected wave reflected at the velocity boundary in the rock is observed. A method for estimating the position of a fault, a fractured zone, or the like is disclosed.
Japanese Patent Laid-Open No. 5-113097 JP 2002-106291 A JP 2001-249186 A Japanese Patent Laid-Open No. 2002-156459

  However, in the case of the geophysical exploration method using the surface wave (Rayleigh wave), electromagnetic wave, reflected wave, etc., since these geophysical exploration measuring instruments are brought to the excavation site, they are specialized in handling the instruments. In addition to the need for an engineer, the drilling was temporarily interrupted, which caused the process to be delayed. In addition, since a lot of setup and labor is required, there is a problem that it cannot be performed as part of daily construction management.

  Therefore, the main problem of the present invention is that the geological configuration of the natural ground is accurately and easily performed as a part of daily construction management, without requiring special geophysical measurement equipment and in parallel with excavation work. It is to provide a method that can be grasped.

In order to solve the above-mentioned problem, as the present invention according to claim 1 , an arbitrary number of observation points are set in a peripheral region of an excavation part, and the time history displacement speed, time history displacement and time history acceleration of the observation point at the time of blasting In the three-dimensional analysis model, by performing a reverse analysis using any one of the measurement data of the time history displacement speed, the time history displacement and the time history acceleration as a given condition, A method for predicting geological composition of natural ground characterized by identifying ground constants is provided. Since this method can know the geological configuration over a wide range, it is preferably performed prior to or in the middle of the geological configuration prediction method of the natural ground according to any one of claims 2 to 5 below . Is desirable.

As the present invention according to claim 2 , an arbitrary number of observation points are set in the surrounding area of the light excavation part of the natural ground, and displacement measurement data is obtained from the result of measuring the displacement of the observation points before and after the blast excavation, In a three-dimensional analysis model reflecting the progress of excavation, blocks are divided based on the geological distribution in advance, and the displacement measurement data is reversed under the condition that the ground constant is constant in each block area. The geological configuration prediction method according to claim 1 is performed prior to or in the middle of the geological configuration prediction method for identifying the ground constant of the natural ground for each block region by performing analysis. A featured geological composition prediction method is provided.

In the present invention as set forth in claim 2 , the ground constant of the natural ground (by a reverse analysis using the displacement measurement data as a given condition by a three-dimensional analysis model based on the displacement measurement data observed in the peripheral area of the light excavation part ( Mainly Young's modulus) was identified. Therefore, it can be carried out as part of daily construction management without requiring special physical exploration measuring equipment and in parallel with excavation work. The calculation of the inverse analysis is a method of optimizing a natural ground physical property value (Young's modulus) that minimizes the residual sum of squares between the calculated displacement of the observation point caused by excavation in a predetermined section and the actual displacement of the observation point. Is what you want.

As the present invention according to claim 3 , the observation point is in a 1.5D region centering on the upper end of the excavation surface and a 1.5D region centering on the lower end of the excavation surface when the height of the excavation surface is D. The geological composition prediction method of the natural ground of Claim 2 installed in the natural ground surface of the inside is provided.

In the present invention described in claim 3 , in the 1.5D region centering on the upper end of the excavation surface and excavation when the height of the excavation surface (slope or wall surface) is D as the observation point of the displacement measurement data It is installed on the natural ground surface in the 1.5D area centering on the lower end of the surface. This region is also a region where the results of the three-dimensional analysis can be evaluated with reliability.

According to a fourth aspect of the present invention, there is provided the geological configuration prediction method for a natural ground according to any one of the second and third aspects, wherein the displacement measurement data of the observation point is one or a combination of 1-3 axial direction displacement components. Provided.

As the present invention according to claim 5, with the progress of the drilling, based on observation and / or the three-dimensional analysis of excavating surface, modifying the block allocation sequentially based on the geological distribution according to any one of claims 2-4 A geological prediction method for natural ground is provided.

In the present invention according to the fifth aspect, the block allocation based on the geological distribution is sequentially corrected with the progress of excavation and / or based on the three-dimensional analysis result. The geological distribution assumed by prior drilling surveys, etc. is assumed to contain a certain amount of error due to restrictions on the number of drilling points, etc., so the excavation surface is observed as the excavation progresses and / or 3D The prediction accuracy can be improved by correcting the sequential block allocation based on the analysis result.

  As described above in detail, according to the present invention, the geological configuration of the natural ground is simple and highly accurate as a part of daily construction management without requiring a special geophysical measurement instrument and while performing excavation work. Can be grasped.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a flowchart of a geological configuration prediction method for natural ground according to the present invention.

  First, as shown in FIG. 1, based on a geological survey such as boring in advance, the natural ground is divided into blocks and a rock classification is set for each block. The rock classification can be classified as shown in Table 1 below based on, for example, “prediction and actual condition of rock mass behavior” (Soil Engineering Society Library 33, Soil Engineering Society, 1988, p29). In addition, a three-dimensional analysis model reflecting the natural ground shape is set.

  In natural ground excavation (light excavation), for example, a lift is divided in the height direction, blast excavation is performed for each lift, and a required portion is excavated according to a rock class of a required rock material.

  At that time, an arbitrary number of observation points are set in the surrounding area of the excavation part, displacement measurement data is obtained from the result of measuring the displacement of the observation point before and after the blast excavation, and the ground constant is constant in each block area. Under the condition that it is a value, the ground constant of the ground is identified for each block by performing inverse analysis using the displacement measurement data as a given condition.

  In addition, the geological configuration prediction method is intended to improve the quarrying accuracy by gradually correcting the initially established stratum configuration, so the excavation surface is observed after excavation, and the geological configuration can be changed if necessary. At the same time, the block allocation of the three-dimensional analysis model is changed. If necessary, the geological configuration is changed and the block allocation of the three-dimensional analysis model is changed based on the three-dimensional analysis result. Finally, the next drilling plan will be drafted.

Hereinafter, an example will be specifically described in detail.
A geological survey such as horizontal boring is performed on the natural ground shown in FIG. 2, and the natural ground is divided into a plurality of regions, and in the example shown in FIG. In each block division area, rock classification is set based on the soil test. In the illustrated example, blocks 1, 3, and 7 are set in the rock classification section CL, and blocks 2, 4, 5, and 6 are set in the rock classification section CM.

  FIG. 3 shows a state in the middle of excavation. With respect to the excavation zone (inclined surface), the peripheral region, specifically, the excavation slope upper end when the height of the excavation slope is D is the center. An observation point is set on the ground surface in the 1.5D region and the 1.5D region centered on the lower end of the excavation slope. It is desirable to provide a plurality of observation points, preferably about 5 to 10 points, in order to improve analysis accuracy.

  The reason why the observation points are provided in the above range is that, in the three-dimensional analysis model shown in FIG. 7, when the ground constant is excavated for two meshes (corresponding to 2 m) under certain conditions, In the result of vector analysis of the vertical displacement component after excavation, when the height of the excavation slope is D, in the 1.5D area centered on the upper end of the excavation slope and in the 1.5D area centered on the lower end of the excavation slope This is because it has been found that the vertical displacement component is superior to other regions. This region is set as an observation point installation zone, but is also a region where the results of the three-dimensional analysis can be evaluated with reliability.

  If the observation point is installed in the vicinity of the excavation zone, use a surveying instrument such as a total station (three-dimensional displacement measuring instrument) or measure the three-dimensional coordinates using GPS or the like before the blast excavation. Then, excavation is performed, and the three-dimensional coordinate measurement of the observation point is performed again after excavation, and displacement measurement data is acquired from the coordinate measurement before and after excavation.

  Then, in the three-dimensional analysis model reflecting the progress of excavation, the block constant is obtained by performing reverse analysis using the displacement measurement data as a given condition under the condition that the ground constant is constant in each block region. The ground constant of the ground is identified every time.

  As a result, as shown in FIG. 4, the result of identifying the Young's modulus of each block by the three-dimensional analysis is as shown in FIG. Then, the boundary line between the blocks 3 and 4 is corrected by observing the excavation cross section. Then, in the next excavation, in the modified three-dimensional analysis model, the inverse analysis is performed using the displacement measurement data before and after the blast excavation as a given condition, and the Young's modulus for each block is identified. As a result, as shown in FIG. 6, if the blocks 3 and 4 are clearly divided, the next excavation plan is made according to this block allocation.

Hereinafter, identification analysis of Young's modulus by a three-dimensional analysis model will be described in detail.
《Prediction of ground constant of natural ground》
(Analysis using a three-dimensional finite element method analysis model)
Displacement measurement data measured by a surveying instrument such as the total station is input to a computer installed in a field office, and in a three-dimensional analysis model, a reverse analysis is performed using the displacement measurement data as a given condition. The ground constant of ground is identified. That is, the natural ground property value (Young's modulus) is subject to inverse analysis, and the natural ground constant that minimizes the residual sum of squares between the calculated displacement of the observation point caused by excavation in the predetermined section and the actual displacement of the observation point ( The ground constant (Young's modulus) is identified for each block by obtaining the Young's modulus by an optimization method. As the ground constant, the deformation coefficient is an unknown parameter in the present embodiment, but the Poisson's ratio, the adhesive force, the lateral pressure coefficient, and the like can also be targeted.

  In the inverse analysis for identifying the Young's modulus, when performing an analysis assuming that all elements can take an arbitrary Young's modulus, it takes a lot of displacement measurement data and requires a long time for calculation. Inverse analysis is performed under the condition that the ground constant is constant in each block area.

Hereinafter, the inverse analysis for Young's modulus will be described in detail according to the procedure.
<< Formulation of inverse analysis >>
1. Forward analysis The finite element equation of the three-dimensional linear elastic body that is the basis of the identification calculation is expressed by the following equation (1).

  Here, K is a stiffness matrix expressed by Young's modulus and Poisson's ratio, u is a displacement vector, and Γ is a load vector.

  In forward analysis, as shown in the following equation (2), the displacement is calculated by giving the analysis region and boundary conditions, Young's modulus and Poisson's ratio, and load.

2. Parameter identification analysis Among the ground constants, identification analysis when Young's modulus is an unknown parameter is described.

  In the Young's modulus identification analysis, the analysis area, boundary conditions, Poisson's ratio, and load are given, and the Young's modulus is calculated based on a finite number of displacements (displacements that are already known by measurement, etc.).

  There are various identification calculation methods, but here they are treated as minimization problems. Therefore, the evaluation function shown in the following formula (3) is defined.

  Here, u: displacement calculated by the finite element method, u *: observation data.

  The evaluation function J is expressed as the sum of squares of the difference between the displacement obtained by calculation and the observed data. The evaluation function J is a positive scalar. If the Young's modulus is correctly determined, the evaluation function J approaches zero. That is, if a Young's modulus that minimizes the evaluation function J is found, the identification calculation is completed. The Young's modulus identification calculation is a problem of minimizing the evaluation function J.

  Of the methods for solving the minimization problem for obtaining the Young's modulus that minimizes the evaluation function J, the conjugate gradient method is used here. In this example, the Fletcher-Reeves Method, which is a kind of conjugate gradient method, is employed. This method is said to have a simple calculation algorithm and good convergence of repeated calculations (calculation is stable).

Next, the formulation for calculation is described below.
(1) Calculation of the gradient of the evaluation function The gradient d of the evaluation function related to the Young's modulus E, which is an unknown parameter of the evaluation function, is calculated. The evaluation function J in the above equation (3) is a displacement functional. As shown in the finite element equation (1), the displacement uses the Young's modulus as an unknown parameter. Therefore, the gradient d is expressed by the following formula (4).

The sensitivity matrix represents the degree of influence of the parameter (Young's modulus) on the displacement.
(2) Calculation of sensitivity matrix As described above, the finite element equation of the three-dimensional linear elastic body is expressed by equation (1).
When the above equation (1) is partially differentiated with respect to the mth unknown parameter Em, the following equation (5) is obtained.

  Further, when the above equation (5) is expanded, the following equation (6) is obtained.

  When this is calculated by the number of unknown parameters and expressed in a matrix, the following equation (7) is obtained.

  At this time, since the amount of displacement does not change at the boundary where the displacement is applied, the following equation (8) is established.

Therefore, the sensitivity matrix regarding the unknown parameter (Young's modulus) E using the above equation (8) can be obtained by calculation for each parameter.
3. Calculation of Step Width α For the convergence calculation, the convergence step width α is calculated. The step width is obtained by Taylor expansion of the evaluation function J of the above formula (3) to the first order term and partial differentiation with α.

  The Taylor expansion up to the first term of the evaluation function J is expressed by the following equation (9).

  When the above equation (9) is partially differentiated by α, the following equation (10) is obtained.

  To obtain α from the above equation (10), the following equation (11) is calculated.

  By determining this α, all parameters are updated as shown in the following equation (12).

  Next, the gradient after the second time can be calculated by the following equation (13) by the Fletcher-Reeves Method (Fletcher-Reeves Method).

  Here, β is calculated by the following equation (14).

4). Identification calculation algorithm for unknown parameter (Young's modulus E)
[step 0]: An initial value of the parameter E (0) and a convergence determination value εJ, which is a value that is considered to have converged when the evaluation function becomes equal to or less than this value, are determined.
[step 1]: Set the number of repetitions i = 0.
[step 2]: {u} ⁽ⁱ ) in the equation (3) is calculated as shown in the following equation (15).

[step 3]: The evaluation function J (i) times (i) times is calculated as shown in the following expression (16) by the expression (3).

[step 4]: The sensitivity matrix {∂u / ∂E _m } ⁽ⁱ⁾ is calculated as shown in the following equation (17) using equation (7).

[step 5]: The initial gradient {d} ⁽ⁱ⁾ of the evaluation function is calculated by the equation (4) as shown in the following equation (18).

[step 6]: The step width α is calculated by the equation (11) as shown in the following equation (19).

[step 7]: The parameter (Young's modulus) is corrected with the gradient and step width as shown in the following equation (20).

[step 8]: The displacement is calculated by the following equation (21) based on the corrected parameter {E} ^{(i + 1)} by the equation (2).

[step 9]: An evaluation function is calculated as shown in the following equation (22) from the corrected displacement u (i + 1) obtained by the equation (3).

[step 10]: A new sensitivity matrix {∂u / ∂E _m } ^{(i + 1)} is calculated by the following equation (23).

[step 11]: β is calculated as shown in the following equation (24).

[step 12]: The gradient of the evaluation function is corrected as shown in the following formula (25).

[step 13 (convergence judgment)]:

If the above equation (26) is satisfied, {E} ^{(i + 1)} is a parameter (Young's modulus) to be obtained.
[Step 14 (convergence judgment)]:

If the above equation (27), i = i + 1 and return to [Step 6] ”.
[Second analysis method]
This second analysis method is performed prior to or during the geological composition prediction method of the natural ground, and in the light excavation of the natural ground or tunnel tunnel excavation, an arbitrary number of observation points are set, and at the time of blasting Any one of the time history displacement speed, the time history displacement and the time history acceleration of the observation point is measured, and in the three-dimensional analysis model, any one of the time history displacement speed, the time history displacement and the time history acceleration is measured. The ground constant (mainly Young's modulus) of the natural ground is identified by repeatedly performing inverse analysis given measurement data as given conditions. If the above-described analysis method is a static analysis method, this second analysis method is a method for identifying ground constants of ground by dynamic analysis.

  In the second analysis method, the observation point can be set in a wide range as long as the blasting impact is propagated. Moreover, the measurement can use a speed sensor, a displacement meter, or an accelerometer directly.

  For example, as shown in FIG. 15, the time history behavior is observed at observation points B1 and B2 installed inside the mine and / or observation points A1 and A2 installed outside the mine during tunnel excavation and underground cavity excavation. Identification analysis such as Young's modulus is performed by three-dimensional vibration analysis, and the geological situation ahead of the face is predicted from Young's modulus. And if it is confirmed that a weak ground layer exists in front of the face, the construction can be efficiently carried out by preparing a support work in advance.

(Other embodiment examples)
(1) In the above example, the three-dimensional analysis model is a finite element method analysis model. However, in addition to the finite element method, the analysis method includes a method using an analytical solution using a mapping function, a finite difference method, a boundary element. You may use the analysis method which consists of a method (Green function method), an individual element method, a finite volume method, RBSM, etc., and these combination.
(2) In the above configuration example, natural rock excavation for rock material sampling was taken as an example, but in the foundation construction work of the dam dam body, it is necessary to confirm that the hard rock was excavated, but this method is It can also be applied to support layer identification for root excavation.

(1) Example 1
In Example 1, a simple analysis model was set, and the validity of this analysis method was verified. In the verification analysis, the forward analysis was first performed to calculate the displacement of the observation point, and this was assumed as the observation data. Then, in the identification analysis by reverse analysis, the observation data was given as given conditions to identify the ground constant, and it was verified whether the result would be the ground constant set in the forward analysis.
(1) Displacement calculation by forward analysis Specifically, the analysis model shown in FIG. 8 was assumed. The ground constant used for forward analysis was set as shown in Table 2 below.

  Then, from the three-dimensional analysis model shown in Fig. 9 (A) (excavation up to X = 9), the observation point at the time when the state shown in Fig. 9 (B) was excavated for one mesh scale. The displacement of was calculated.

(2) Identification calculation by reverse analysis As shown in FIG. 10, the reverse analysis is an analysis model in which the number of blocks is 4, and the displacement amount by the above-mentioned forward analysis is given to the observation point as displacement measurement data.

In the analysis, the initial Young's modulus of all blocks was set to 10000 (tf / m ² ) as an initial value, and the Young's modulus of each block was obtained by repeated calculation. FIG. 11 shows a situation in which the calculated values converge by each repeated calculation. From Fig. 11, blocks 1, 3 and 4 have Young's modulus converged to 45000-55000 (tf / m ² ), and block 2 has a Young's modulus converged to 5000 (tf / m ² ). The validity of the method was verified.

(2) Example 2
In Example 2, a simple analysis model was set and the validity of the second analysis method was verified. In the verification analysis, first, forward analysis is performed to calculate the time-history displacement velocity of the observation point, and this is given as observation data in the identification calculation by inverse analysis to identify the ground constant, and the result is set by forward analysis. The ground constant was verified.

(1) Time history displacement speed calculation by forward analysis Specifically, the analysis model shown in FIG. 12 is set, and a uniform distribution load of 1000 (kN / m ² ) is applied to the entire left side of the analysis region at time t = 0.01. The time history displacement velocity at the observation points (A to C) when only (s) was applied was obtained by dynamic forward analysis. In addition, the ground constant used for forward analysis was made into Young's modulus E = 10000 (tf / m < ² >) of the whole area | region, ground density (rho) = 2.0 (kN / m < ³ >), and Poisson's ratio micro = 0.3. The time history displacement speeds at the observation points A to C by the above-described forward analysis are as shown in FIG.

(2) Identification calculation by inverse analysis Based on the above results, the time history displacement speed shown in FIG. 13 is given as a given condition by inverse analysis, and the Young's modulus initial value is 5000 (tf / m ² ). Young's modulus was obtained by repeated calculation. FIG. 14 shows the identification calculation result. From FIG. 14, a correct Young's modulus of 10000 (tf / m ² ) was obtained after 40 iterations, and the validity of the second analysis method could be verified.

It is a flowchart of the geological structure prediction method of the natural ground which concerns on this invention. It is a figure which shows the block allocation example of a natural mountain. It is a longitudinal cross-sectional view which shows the block division and excavation condition of a natural ground. It is a figure which shows the identification calculation result of the natural ground ground constant. It is a longitudinal cross-sectional view which shows the block division of the natural ground at the time of the next step, and the excavation condition. It is a figure which shows the identification calculation result of the natural ground ground constant. It is a vertical displacement vector diagram by excavation. 2 is an analysis model in Example 1. FIG. It is a state change figure by excavation in an analysis model. It is an analysis model figure. It is a figure which shows the identification calculation result of a Young's modulus. 6 is an analysis model diagram in Example 2. FIG. It is a time history displacement velocity diagram of observation point AC acquired by forward analysis. It is a figure which shows the identification calculation result by a reverse analysis. It is a schematic diagram of the example of an identification analysis by the 2nd analysis method.

Claims (5)

  Set an arbitrary number of observation points in the peripheral area of the excavation part, measure any one of the time history displacement speed, time history displacement and time history acceleration of the observation point at the time of blasting, in a three-dimensional analysis model, Geological composition prediction of natural ground characterized by identifying ground constant of ground by performing inverse analysis using measurement data of time history displacement speed, time history displacement and time history acceleration as given conditions Method. Set up an arbitrary number of observation points in the surrounding area of the light excavation part of the natural ground, obtain displacement measurement data from the result of measuring the displacement of the observation point before and after blast excavation, 3D analysis reflecting the progress of excavation In the model, blocks are divided based on the geological distribution in advance, and under the condition that the ground constant is a constant value in each block area, a reverse analysis is performed using the displacement measurement data as a given condition. The geological composition prediction of the natural ground according to claim 1, wherein the geological composition prediction method of the natural ground according to claim 1 is performed prior to or in the middle of the geological composition prediction method for identifying the ground constant of the natural ground. Method. The observation point is installed on the ground surface in the 1.5D region centered on the upper end of the excavation surface and in the 1.5D region centered on the lower end of the excavation surface, where D is the height of the excavation surface. The geological composition prediction method of a natural ground according to claim 2 . The geological configuration prediction method for a natural ground according to any one of claims 2 and 3, wherein the displacement measurement data of the observation point is any one or combination of 1-3 axial direction displacement components. The geological prediction method of the natural ground according to any one of claims 2 to 4 , wherein the block division based on the geological distribution is sequentially corrected based on observation of the excavation surface and / or the three-dimensional analysis result as the excavation progresses.

Images