A multivariate linear regression in-situ stress inversion method based on single-hole partitioning
By using a single-hole partition multivariate linear regression method in tunnel engineering, a three-dimensional geological inversion model was established and loading simulation was carried out, which solved the problem of increased inversion error in the traditional method, and achieved higher inversion accuracy and error Reduce.
Patent Information
- Application Number
- CN202411116801.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In tunnel engineering, the traditional multivariate linear regression method fails to consider special landforms, resulting in an increase in the inversion error of the ground stress field, which cannot meet the engineering accuracy requirements.
Using a multivariate linear regression method based on single-hole partition, a three-dimensional geological inversion model is established by setting on-site ground stress measurement points, and a loading simulation is performed. The adjusted regression equation is established using a single-hole partition regression method to reduce errors and improve inversion accuracy. .
It effectively reduces the ground stress regression error, can better characterize the actual stress field, reduce inversion error, and meet engineering accuracy requirements.
Smart Images

Figure CN118643751B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of geotechnical engineering and numerical analysis, and in particular relates to a multivariate linear regression geostress inversion method based on single-hole partitioning. Background Art
[0002] Initial geostress refers to the rock stress field before the excavation of underground engineering, which is mainly composed of self-weight stress and tectonic stress. When building a tunnel in a high geostress area, the integrity of the rock mass is destroyed, and the initial geostress field is disturbed. Rock bursts may occur in hard rock areas, and large deformations of soft rocks may occur in soft rock areas. In order to ensure construction safety, the geostress of the project area should be tested before the construction of underground engineering to accurately grasp the distribution characteristics of the geostress in the project area. However, due to the high cost of field testing of geostress and the poor field testing conditions, the early engineering geostress test data are relatively limited. Therefore, it is very necessary to use the limited geostress test results to invert the geostress field of the entire project area.
[0003] Domestic and foreign researchers have proposed many methods for geostress inversion, which can be mainly divided into boundary load adjustment method, displacement back analysis method, neural network back analysis method and multivariate linear regression method. These methods have their own advantages and disadvantages. Among them, the multivariate linear regression method is widely used in geostress field inversion due to its convenience, speed and uniqueness of solution.
[0004] However, the traditional multivariate linear regression method regards the engineering area as a whole, and does not consider the influence of special landforms on the geostress field. Compared with coal mines and hydropower station projects, tunnel projects have the characteristic that the length of the axis is much longer than other directions. The geostress field often encounters structural influences such as fault fracture zones, river valley strata and intrusive rock strata, and the geostress value changes more dramatically. If all the borehole data are inverted using a regression equation in the traditional multivariate linear regression method, the regression error will inevitably increase under the influence of some data with large discreteness. Therefore, it is a key issue that needs to be solved urgently to develop a method suitable for tunnel projects that can explain special landforms and improve inversion accuracy. The traditional multivariate linear regression method regards the entire engineering area as a whole and uses one equation to invert the geostress field in the engineering area, which may have a "neutralizing" effect on these high geostress areas and increase the inversion error. Table 1 summarizes the inversion errors of typical domestic tunnel projects. It can be seen that with the increase in the number of boreholes (number of measuring points), the average inversion error increases.
[0005] Table 1 Inversion errors of typical domestic tunnel projects
[0006] Summary of the invention
[0007] In view of the above-mentioned deficiencies in the prior art, the present invention provides a multivariate linear regression geostress inversion method based on single-hole partitioning, which solves the problem that the average inversion error increases with the increase in the number of boreholes in the traditional multivariate linear regression method.
[0008] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a multivariate linear regression geostress inversion method based on single-hole partitioning, comprising the following steps:
[0009] S1. Set up on-site ground stress measuring points to obtain the measured stress values around the tunnel axis;
[0010] S2. Establish a three-dimensional geological inversion model based on engineering geological data and stratum lithology data;
[0011] S3, loading the three-dimensional geological inversion model to simulate different geological structural effects and obtain the calculated values of stress components at the on-site ground stress measuring points under different working conditions;
[0012] S4, perform stress field regression by using the overall regression method according to the calculated values of the stress components, establish a regression equation, perform statistical inspection of the regression results and regression accuracy analysis according to the regression equation, and judge whether the statistical inspection of the regression results is qualified and the regression accuracy meets the engineering requirements at the same time. If so, adopt the regression equation and enter S6, if not, enter S5;
[0013] S5, using the single-hole partition regression method to perform regression, establish an adjusted regression equation, perform statistical test on the regression results and regression accuracy analysis based on the adjusted regression equation, and determine whether the statistical test of the regression results is qualified and the regression accuracy meets the engineering requirements at the same time. If so, use the adjusted regression equation and enter S6. If not, adjust the three-dimensional geological inversion model and boundary conditions, and return to S2;
[0014] S6. Based on the adopted results, complete the geostress inversion and obtain the distribution characteristics of the geostress field at the tunnel axis.
[0015] Further: S1 comprises the following sub-steps:
[0016] S11, collecting measured stress values through on-site geostress measuring points, performing stress component conversion on the measured stress values, and obtaining six geostress components in the geodetic coordinate system;
[0017] S12. The six geostress components in the geodetic coordinate system are converted into six stress components in the calculation coordinate system by using a coordinate transformation formula, thereby obtaining the measured value components.
[0018] Further: In S11, the expression for stress component conversion is specifically the following formula:
[0019]
[0020] In the formula, , and is the measured stress value, , , , , and is the geostress component, T is the transposed sign, is the stress component conversion matrix, and its specific expression is:
[0021]
[0022] In the formula, is the measured value of the i-th stress Direction cosines for the x-axis, i=1,2,3, is the measured value of the i-th stress The direction cosines about the y-axis, is the measured value of the i-th stress Direction cosines about the z-axis.
[0023] Further: In said S12, the method for obtaining the stress component under the calculation coordinate is specifically:
[0024] The ground stress components are transformed into stress components under calculation coordinates through the coordinate transformation formula, where the coordinate transformation formula is specifically as follows:
[0025]
[0026] In the formula, , , , , and To calculate the stress components in coordinates, is the coordinate transformation matrix, and its specific expression is:
[0027] .
[0028] Further: in S2, the three-dimensional geological inversion model includes distribution information of fault fracture zones, river valley strata, intrusive rock strata and different rock strata, wherein the distribution information includes: strike, dip, inclination and fault width.
[0029] Further: in said S3, the loading method of the three-dimensional geological inversion model includes: self-weight stress field simulation, x-direction compression tectonic movement, y-direction compression tectonic movement and shear tectonic movement in the horizontal plane;
[0030] The specific method of simulating the loading of the self-weight stress field is as follows: using the measured weight of the rock mass, generating the self-weight stress field under the action of the self-weight stress, and constraining the displacement of the bottom and side surfaces of the model during calculation;
[0031] The x-direction extrusion structural motion loading method is specifically as follows: simulating by applying horizontal displacement in the x-direction of the model, and applying displacement constraints to the non-loaded interface and the bottom surface;
[0032] The method of loading the y-direction extrusion structural movement is specifically as follows: simulating by applying a horizontal displacement in the y-direction of the model, and applying a displacement constraint to the non-loaded interface and the bottom surface;
[0033] The loading method of the shear tectonic movement in the horizontal plane is specifically: to achieve it by applying a horizontal displacement inversely proportional to the boundary length, and to impose a displacement constraint on the bottom surface.
[0034] Further: S4 includes the following sub-steps:
[0035] S41. Establish the residual sum of squares formula of the least squares method based on the measured stress value component and the calculated stress value component ;
[0036]
[0037] In the formula, For the Measuring point The measured stress component, For the Influence factors of geostress Measuring point Calculated values of stress components, , m is the total number of on-site geostress measuring points set up, is the regression coefficient of the corresponding independent variable, i=1,2,3,…,n, n is the total number of regression coefficients of the corresponding independent variable;
[0038] S42. According to the principle of least square method, the residual sum of squares is minimized and the regression coefficient is calculated;
[0039]
[0040] S43, establishing a regression equation according to the regression coefficient, calculating the stress regression calculation value of any point in the region, and obtaining a regression result;
[0041]
[0042] In the formula, is a single-column matrix of calculated values, is the stress regression calculated value, is the regression coefficient of the corresponding independent variable;
[0043] S44, perform statistical test on the regression result and regression accuracy analysis. If the statistical test on the regression result is qualified and the regression accuracy meets the engineering requirements, the regression equation obtained in S43 is adopted and the process proceeds to S6. If the statistical test on the regression result is unqualified or the regression accuracy does not meet the engineering requirements, the process proceeds to S5.
[0044] Further: In S44, the method of statistical test of regression results is specifically:
[0045] According to the regression results, the regression sum of squares and the residual sum of squares are calculated, and then the total deviation, the multiple correlation coefficient and the test parameter F are calculated. If the multiple correlation coefficient is within the qualified interval and the test parameter F is greater than the test threshold, the statistical test of the regression result is qualified;
[0046] Among them, the expression of regression square sum U is:
[0047]
[0048] In the formula, For the Measuring point Stress regression calculation value;
[0049] The expression of the residual sum of squares Q is:
[0050]
[0051] In the formula, For the Influence factors of geostress Measuring point Stress regression calculation value;
[0052] Total deviation S yy :
[0053]
[0054] The expression of the multiple correlation coefficient R is:
[0055]
[0056] The expression of the test parameter F is:
[0057] .
[0058] Further: In S44, the method of regression accuracy analysis is specifically:
[0059] S441. The relative error between the inversion results and the measured values of each measuring point of the project is represented by the following formula:
[0060]
[0061]
[0062] In the formula, For the The relative error between the inversion result of the geostress at each measuring point and the measured result, i=1,2,3,…,n; For the The inversion results of the ground stress at each measuring point are: Indicates The measured results of the ground stress at each measuring point; Represents the average value of the absolute value of the inversion relative error;
[0063] S442, according to the average value of the absolute value of the inversion relative error and the actual engineering accuracy requirements The specific method for judging whether the engineering inversion accuracy requirements are met is as follows:
[0064] when When , the calculation formula of stress regression calculation value meets the accuracy requirement;
[0065] when When , the calculation formula of the stress regression calculation value does not meet the accuracy requirements.
[0066] Further: S5 comprises the following sub-steps:
[0067] S51, based on the p geostress measuring holes set up on site, the measured values of the stress components and the calculated values of the stress components are grouped into p groups according to different measuring holes, and each group of data is used in turn to invert and calculate the regression coefficient to obtain the regression equation of the p geostress measuring holes, and use the regression equation as the adjusted regression equation;
[0068] S52, extracting stress components at the geostress measuring holes, and substituting the stress components into the regression equation of each geostress measuring hole to obtain regression values of p geostress measuring holes;
[0069] S53, perform statistical test on the regression result and regression accuracy analysis based on the regression values of p geostress measuring holes. If the statistical test on the regression result is qualified and the regression accuracy meets the engineering requirements, use the adjusted regression equation and enter S6. If the statistical test on the regression result is unqualified or the regression accuracy does not meet the engineering requirements, return to S2, adjust the three-dimensional geological inversion model and boundary conditions, and re-simulate and regress.
[0070] The beneficial effects of the present invention are:
[0071] (1) The present invention provides a multivariate linear regression geostress inversion method based on single-hole partitioning. As the measured data increases in the traditional regression method, the single equation cannot explain the stress mutation in some areas caused by special landforms, which leads to an increase in the geostress regression error and cannot meet the engineering accuracy requirements or the regression effect test fails. The present invention uses a single-hole partitioning regression method to reduce the error, better characterize the actual stress field, and reduce the inversion error.
[0072] (2) Generally, fault fracture zones, river valley strata, intrusive rock strata, etc. are irregularly distributed, so artificial zoning is difficult to achieve. The present invention establishes a three-dimensional geological inversion model through engineering geological data and stratum lithology data, establishes a regression equation through a single-hole zoning method, substitutes the stress components between two adjacent measuring points into the regression equation of the two measuring points, and finds the minimum difference, thereby achieving the minimum stress mutation value. BRIEF DESCRIPTION OF THE DRAWINGS
[0073] Figure 1 The present invention is a flow chart of a multivariate linear regression geostress inversion method based on single-hole partitioning.
[0074] Figure 2 This is a three-dimensional geological inversion model diagram.
[0075] Figure 3 This is the principal stress distribution diagram of the tunnel axis. DETAILED DESCRIPTION
[0076] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.
[0077] Embodiment 1:
[0078] like Figure 1 As shown, in one embodiment of the present invention, a multivariate linear regression geostress inversion method based on single-hole partitioning includes the following steps:
[0079] S1. Set up on-site ground stress measuring points to obtain the measured stress values around the tunnel axis;
[0080] S2. Establish a three-dimensional geological inversion model based on engineering geological data and stratum lithology data;
[0081] S3, loading the three-dimensional geological inversion model to simulate different geological structural effects and obtain the calculated values of stress components at the on-site ground stress measuring points under different working conditions;
[0082] S4, perform stress field regression by using the overall regression method according to the calculated values of the stress components, establish a regression equation, perform statistical inspection of the regression results and regression accuracy analysis according to the regression equation, and judge whether the statistical inspection of the regression results is qualified and the regression accuracy meets the engineering requirements at the same time. If so, adopt the regression equation and enter S6, if not, enter S5;
[0083] S5, using the single-hole partition regression method to perform regression, establish an adjusted regression equation, perform statistical test on the regression results and regression accuracy analysis based on the adjusted regression equation, and determine whether the statistical test of the regression results is qualified and the regression accuracy meets the engineering requirements at the same time. If so, use the adjusted regression equation and enter S6. If not, adjust the three-dimensional geological inversion model and boundary conditions, and return to S2;
[0084] S6. Based on the adopted results, complete the geostress inversion and obtain the distribution characteristics of the geostress field at the tunnel axis.
[0085] In this embodiment, on-site geostress measuring points are set according to the hydrogeological data of the project area to obtain the measured stress values around the tunnel axis. The on-site geostress measuring points should be arranged in combination with different topographic and geomorphic areas such as regional fault structure distribution, river valley strata, and intrusive rock strata to ensure that the inversion of special geomorphic areas has on-site measured data. The measured geostress values obtained at each measuring point include three principal stress values: , and The three principal stresses obtained are in the geodetic coordinate system, which has the E direction as the x-axis, the N direction as the y-axis, and the vertical direction as the z-axis, which is different from the coordinate system of the three-dimensional geological inversion model, so a coordinate system conversion is required.
[0086] The S1 comprises the following sub-steps:
[0087] S11. Collect the measured stress values through the on-site ground stress measuring points. , and Perform stress component transformation to obtain 6 geostress components in the geodetic coordinate system;
[0088] S12. The six geostress components in the geodetic coordinate system are converted into six stress components in the calculation coordinate system by using a coordinate transformation formula, thereby obtaining the measured value components.
[0089] In S11, the expression for stress component conversion is specifically as follows:
[0090]
[0091] In the formula, , and is the measured stress value, , , , , and is the geostress component, T is the transposed sign, is the stress component conversion matrix, and its specific expression is:
[0092]
[0093] In the formula, is the measured value of the i-th stress Direction cosines for the x-axis, i=1,2,3, is the measured value of the i-th stress The direction cosines about the y-axis, is the measured value of the i-th stress Direction cosines about the z-axis.
[0094] In S12, the method for obtaining the stress component under the calculation coordinate is specifically:
[0095] The ground stress components are transformed into stress components under calculation coordinates through the coordinate transformation formula, where the coordinate transformation formula is specifically as follows:
[0096]
[0097] In the formula, , , , , and To calculate the stress components in coordinates, is the coordinate transformation matrix, and its specific expression is:
[0098] .
[0099] In this embodiment, the established three-dimensional geological inversion model needs to be able to reflect the actual terrain undulations, including the distribution information of fault fracture zones, river valley strata, intrusive rock strata and different rock strata, wherein the distribution information includes: strike, dip, inclination and fault width, and the physical and mechanical parameters of the strata should include: density, elastic modulus and Poisson's ratio.
[0100] In said S3, the loading method of the three-dimensional geological inversion model includes: self-weight stress field simulation, x-direction compression tectonic movement, y-direction compression tectonic movement and shear tectonic movement in the horizontal plane;
[0101] The specific method of simulating the loading of the self-weight stress field is as follows: using the measured weight of the rock mass, generating the self-weight stress field under the action of the self-weight stress, and constraining the displacement of the bottom and side surfaces of the model during calculation;
[0102] The x-direction extrusion structural motion loading method is specifically as follows: simulating by applying horizontal displacement in the x-direction of the model, and applying displacement constraints to the non-loaded interface and the bottom surface;
[0103] The method of loading the y-direction extrusion structural movement is specifically as follows: simulating by applying a horizontal displacement in the y-direction of the model, and applying a displacement constraint to the non-loaded interface and the bottom surface;
[0104] The loading method of the shear tectonic movement in the horizontal plane is specifically: to achieve it by applying a horizontal displacement inversely proportional to the boundary length, and to impose a displacement constraint on the bottom surface.
[0105] The S4 comprises the following sub-steps:
[0106] S41. Establish the residual sum of squares formula of the least squares method based on the measured stress value component and the calculated stress value component ;
[0107]
[0108] In the formula, For the Measuring point The measured stress component, For the Influence factors of geostress Measuring point Calculated values of stress components, , m is the total number of on-site geostress measuring points set up, is the regression coefficient of the corresponding independent variable, i=1,2,3,…,n, n is the total number of regression coefficients of the corresponding independent variable;
[0109] S42. According to the principle of least square method, the residual sum of squares is minimized and the regression coefficient is calculated;
[0110]
[0111] From the above formula, we can get the regression coefficients of n corresponding independent variables: By substituting the regression coefficient into the formula, the regression calculation value of the geostress at any point in the calculation area can be obtained.
[0112] S43, establishing a regression equation according to the regression coefficient, calculating the stress regression calculation value of any point in the region, and obtaining a regression result;
[0113]
[0114] In the formula, is a single-column matrix of calculated values, is the stress regression calculated value, is the regression coefficient of the corresponding independent variable;
[0115] S44, perform statistical test on the regression result and regression accuracy analysis. If the statistical test on the regression result is qualified and the regression accuracy meets the engineering requirements, the regression equation obtained in S43 is adopted and the process proceeds to S6. If the statistical test on the regression result is unqualified or the regression accuracy does not meet the engineering requirements, the process proceeds to S5.
[0116] The specific method of statistical test of regression results is as follows:
[0117] The regression sum of squares and the residual sum of squares are calculated based on the regression results, and then the total deviation, the multiple correlation coefficient and the test parameter F are calculated. If the multiple correlation coefficient is within the qualified interval and the test parameter F is greater than the test threshold, the statistical test of the regression result is qualified.
[0118] Regression sum of squares U:
[0119]
[0120] In the formula, For the Measuring point Stress regression calculation value;
[0121] Residual sum of squares Q:
[0122]
[0123] In the formula, For the Influence factors of geostress Measuring point Stress regression calculation value;
[0124] Total deviation S yy :
[0125]
[0126] Multiple correlation coefficient R:
[0127]
[0128] In this embodiment, The qualified interval range is set to [0,1]. Generally, The closer the value is to 1, the better the regression effect. hour, Should be greater than 5~10 times of the There will always be In this case, test.
[0129]
[0130] Take the significance level , if F is greater than the test threshold , indicating that this The variable has a significant effect. Can be viewed by Obtain the test critical value table.
[0131] Among them, the method of regression accuracy analysis is specifically as follows:
[0132] The relative error between the inversion results of each measuring point in the project and the measured value is represented by the relative error calculation process as shown in the following formula.
[0133]
[0134]
[0135] In the formula, For the The relative error between the inversion result of the geostress at each measuring point and the measured result, i=1,2,3,…,n; For the The inversion results of the ground stress at each measuring point are: Indicates The measured results of the ground stress at each measuring point; Represents the average value of the absolute value of the inversion relative error;
[0136] According to the average value of the absolute value of the inversion relative error and the actual engineering accuracy requirements The specific method for judging whether the engineering inversion accuracy requirements are met is as follows:
[0137] when When , the calculation formula of stress regression calculation value meets the accuracy requirement;
[0138] when When , the calculation formula of the stress regression calculation value does not meet the accuracy requirements.
[0139] The S5 comprises the following sub-steps:
[0140] S51, based on the p geostress measuring holes set up on site, the measured values of the stress components and the calculated values of the stress components are grouped into p groups according to different measuring holes, and each group of data is used in turn to invert and calculate the regression coefficient to obtain the regression equation of the p geostress measuring holes, and use the regression equation as the adjusted regression equation;
[0141] S52, extracting stress components at the geostress measuring holes, and substituting the stress components into the regression equation of each geostress measuring hole to obtain regression values of p geostress measuring holes;
[0142] In this embodiment, the single-hole partition regression method also requires partition interface identification. In order to minimize the sudden change of adjacent stress fields, the stress component at the tunnel axis is extracted, and the stress components between two adjacent measuring holes are respectively substituted into two adjacent regression equations, and the minimum value of the principal stress difference is obtained. The position of the minimum difference point is used as the interface of the tunnel axis geostress field, and p-1 interfaces and p partitions are obtained.
[0143] S53, perform statistical test on the regression result and regression accuracy analysis based on the regression values of p geostress measuring holes. If the statistical test on the regression result is qualified and the regression accuracy meets the engineering requirements, use the adjusted regression equation and enter S6. If the statistical test on the regression result is unqualified or the regression accuracy does not meet the engineering requirements, return to S2, adjust the three-dimensional geological inversion model and boundary conditions, and re-simulate and regress.
[0144] In this embodiment, the geostress measuring hole can be set according to the actual situation, and the method for calculating the regression equation is the same as the method of S41 to S43;
[0145] Embodiment 2:
[0146] This embodiment provides a specific experimental data for embodiment 1.
[0147] Step 1: Conduct geostress measurements based on the hydrogeological data of the project area to obtain the measured value components near the tunnel axis:
[0148] Step 11: Arrange 7 geostress measurement holes along the tunnel axis at the construction site, with a total of 29 measurement points, and use hydraulic fracturing to conduct in-situ geostress testing. The measured stress values are shown in Table 2:
[0149] Table 2 Measured stress values
[0150]
[0151] Step 12, convert the measured stress value into 6 stress components in the geodetic coordinate system according to the stress component conversion formula, and convert it into 6 stress components in the calculation coordinate system according to the coordinate conversion formula, as shown in Table 3:
[0152] Table 3 Stress components in the calculation coordinate system
[0153]
[0154] In the hydraulic fracturing test, it is assumed that the vertical principal stress is equal to the deadweight of the overlying rock mass, and the other two principal stresses are in the horizontal direction. and Both are 0.
[0155] Step 2: Establish a three-dimensional geological inversion model that can reflect the actual terrain and stratum lithology based on engineering geological data and stratum lithology data:
[0156] Step 21, establish a three-dimensional geological inversion model based on the geological survey report and engineering contour map, and divide the model into grids. The three-dimensional geological inversion model is required to reflect the actual terrain undulations, including the actual strata, the direction, dip, inclination and fault width of the fault fracture zone, such as Figure 2 shown.
[0157] Step 22, establish a three-dimensional geological inversion model. To reduce the impact of boundary effects, the calculation range should be appropriately expanded according to the size of the study area. Taking into account the topography and structural factors of the Zhaotong Tunnel, it is determined that 15,600 meters along the tunnel axis is taken as the long side of the model, 4,000 meters along the tunnel cross section is taken as the short side of the model, and the actual terrain altitude is used as the model height to establish a three-dimensional numerical inversion model.
[0158] Step 23, the three-dimensional geological inversion model has a total of 167,222 nodes and 896,223 units, and is calculated using an elastic constitutive model.
[0159] Step 24, select the physical and mechanical parameters of the formation according to the recommended values in the Railway Engineering Design Technical Manual - Tunnels and Tunnel Engineering Geological Description: Density , elastic modulus E, Poisson's ratio , as shown in Table 4.
[0160] Table 4 Physical and mechanical parameters of rock formations
[0161]
[0162] Step 3: Load the three-dimensional geological inversion model in different ways to simulate different geological tectonic actions and obtain stress component values at measuring points under different working conditions.
[0163] Step 31, the different loading modes include: self-weight stress field simulation, x-direction compression tectonic movement, y-direction compression tectonic movement and shear tectonic movement in the horizontal plane.
[0164] (1) The self-weight stress field is generated by using the measured weight of the rock mass. The displacement of the bottom and side surfaces of the model is constrained during calculation.
[0165] (2) The extrusion tectonic movement in the x-direction is simulated by applying horizontal displacement in the x-direction of the model and applying displacement constraints to the non-loaded interface and bottom surface;
[0166] (3) The extrusive tectonic movement in the y direction is simulated by applying horizontal displacement in the y direction of the model and applying displacement constraints to the non-loaded interface and bottom surface;
[0167] (4) Shear tectonic movement in the horizontal plane is achieved by applying a horizontal displacement inversely proportional to the boundary length, while imposing a displacement constraint on the bottom surface.
[0168] Step 32, through numerical simulation, respectively extract the calculated values of stress components at the locations of the measured ground stress measuring points.
[0169] Step 4: Use the overall regression method to regress the stress field. If it meets the engineering requirements, it is adopted.
[0170] Step 41, extract all calculated values of stress components as independent variables, take the transformed measured values of stress components as dependent variables, and obtain the regression equation through the principle of least square method:
[0171]
[0172] In the formula, is the regression calculation value of the geostress field, is the stress field caused by the deadweight of the rock mass, is the stress field caused by the horizontal compression tectonic movement along the x direction, is the stress field caused by the horizontal extrusion motion along the y direction, It is the stress field caused by shear tectonic movement in the horizontal plane.
[0173] Step 42, perform a statistical test on the regression results, as shown in Table 5, and after calculation, obtain the complex correlation coefficient , , indicating that the regression effect is significant, the independent variable has a significant explanatory effect on the dependent variable, and the regression result statistical test is qualified.
[0174] Table 5 Comparison between calculated values and measured values using the overall regression method
[0175]
[0176] Step 43, performing regression accuracy analysis, after calculation, the maximum horizontal principal stress relative error is 21.07%, the minimum horizontal principal stress average relative error is 17.89%, the vertical principal stress relative error is 19.54%, and the average relative error is 19.50%.
[0177] Step 44, according to engineering experience, the error of multivariate linear regression inversion of geostress is generally less than 15%, indicating that the inversion result is relatively reliable. The average relative error obtained by the overall regression method in this project is 19.50%, and the accuracy does not meet the engineering requirements.
[0178] Step 5: If the regression result of the stress regression equation obtained in step 4 is qualified by statistical test, but the regression accuracy does not meet the engineering requirements, the single-hole partition regression method is used for regression to obtain the distribution characteristics of the geostress field in the engineering area.
[0179] Step 51: Group the measured stress components and calculated stress components according to different measuring holes, and divide them into 7 groups of data in total.
[0180] Step 52: Use each set of data to perform regression according to step 4, and obtain 7 regression equations, which are:
[0181]
[0182]
[0183]
[0184]
[0185]
[0186]
[0187]
[0188] In the formula, Representative The regression calculation value of the inversion of the borehole data.
[0189] Step 53, extract the stress components at the geostress measuring holes and bring them into 7 regression equations respectively to obtain the regression values of the 7 geostress measuring holes.
[0190] Step 54, in order to minimize the sudden change of adjacent stress fields, the stress component at the tunnel axis is extracted, the stress component between two adjacent measuring holes is respectively substituted into two adjacent regression equations, and the minimum value of the principal stress difference is obtained. The position where the point with the minimum difference is located is taken as the interface of the geostress field of the tunnel axis. At this time, the stress field obtained by inversion of the two regression equations has the minimum stress sudden change at the interface. A total of 6 interfaces are obtained along the tunnel axis, and the mileages are: K376+700, K378+100, K379+200, K380+900, K383+200, K383+800.
[0191] Step 55, perform a statistical test on the regression results. As shown in Table 6, after calculation, the seven multiple correlation coefficients are 0.987, 0.992, 0.979, 0.993, 0.988, 0.977, and 0.960, and the seven F values are 173.55, 290.04, 109.70, 353.48, 199.67, 97.94, and 73.49, all greater than , indicating that the regression effect is significant, the independent variable has a significant explanatory effect on the dependent variable, and the regression result statistical test is qualified.
[0192] Step 56, performing regression accuracy analysis, after calculation, the maximum horizontal principal stress relative error is 9.94%, the minimum horizontal principal stress average relative error is 10.37%, the vertical principal stress relative error is 1.77%, and the average relative error is 7.36%.
[0193] Table 6 Comparison between calculated values and measured values by single-hole partition regression method
[0194]
[0195] Step 57: The regression result passes the statistical test, and the single-hole partition regression accuracy meets the engineering requirements, so it is adopted.
[0196] Step 6: Divide the tunnel axis into 7 parts according to the interface, and substitute the stress components at the tunnel axis into the 7 regression equations respectively, so as to obtain Figure 3 The stress distribution characteristics of the entire engineering area are shown in Figure 1. Figure 3 In the figure, the left ordinate is stress in MPa, the right ordinate is altitude in meters, and the abscissa is tunnel mileage in meters.
[0197] The beneficial effects of the present invention are as follows: the present invention provides a multivariate linear regression geostress inversion method based on single-hole partitioning. As the measured data increases in the traditional regression method, a single equation cannot explain the stress mutation in some areas caused by special landforms, which leads to an increase in the geostress regression error and cannot meet the engineering accuracy requirements or the regression effect test fails. The present invention uses a single-hole partitioning regression method to reduce the error, better characterize the actual stress field, and reduce the inversion error.
[0198] Generally, fault fracture zones, river valley strata, intrusive rock strata, etc. are irregularly distributed, so artificial zoning is difficult to achieve. The present invention establishes a three-dimensional geological inversion model through engineering geological data and stratum lithology data, establishes a regression equation through a single-hole zoning method, substitutes the stress components between two adjacent measuring points into the regression equation of the two measuring points, and finds the minimum difference, thereby achieving the minimum stress mutation.
[0199] In the description of the present invention, it is necessary to understand that the orientation or positional relationship indicated by the terms "center", "thickness", "upper", "lower", "horizontal", "top", "bottom", "inner", "outer", "radial", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", and "third" are used only for descriptive purposes, and cannot be understood as indicating or implying the relative importance or the number of implicitly specified technical features. Therefore, the features defined by "first", "second", and "third" may explicitly or implicitly include one or more of the features.
Claims
1. A multivariate linear regression in-situ stress inversion method based on single-hole partitioning, characterized in that: The following steps are involved: S1. Set up on-site ground stress measuring points to obtain the measured stress values around the tunnel axis; S2. Establish a three-dimensional geological inversion model based on engineering geological data and stratum lithology data; S3, loading the three-dimensional geological inversion model to simulate different geological structural effects and obtain the calculated values of stress components at the on-site ground stress measuring points under different working conditions; S4, perform stress field regression by using the overall regression method according to the calculated values of the stress components, establish a regression equation, perform statistical inspection of the regression results and regression accuracy analysis according to the regression equation, and judge whether the statistical inspection of the regression results is qualified and the regression accuracy meets the engineering requirements at the same time. If so, adopt the regression equation and enter S6, if not, enter S5; S5, using the single-hole partition regression method to perform regression, establish an adjusted regression equation, perform statistical test on the regression results and regression accuracy analysis based on the adjusted regression equation, and determine whether the statistical test of the regression results is qualified and the regression accuracy meets the engineering requirements at the same time. If so, use the adjusted regression equation and enter S6. If not, adjust the three-dimensional geological inversion model and boundary conditions, and return to S2; S6. Based on the adopted results, complete the inversion of geostress and obtain the distribution characteristics of the geostress field at the tunnel axis; In said S3, the loading method of the three-dimensional geological inversion model includes: self-weight stress field simulation, x-direction compression tectonic movement, y-direction compression tectonic movement and shear tectonic movement in the horizontal plane; The specific method of simulating the loading of the self-weight stress field is as follows: using the measured weight of the rock mass, generating the self-weight stress field under the action of the self-weight stress, and constraining the displacement of the bottom and side surfaces of the model during calculation; The x-direction extrusion structural motion loading method is specifically as follows: simulating by applying horizontal displacement in the x-direction of the model, and applying displacement constraints to the non-loaded interface and the bottom surface; The method of loading the y-direction extrusion structural movement is specifically as follows: simulating by applying a horizontal displacement in the y-direction of the model, and applying a displacement constraint to the non-loaded interface and the bottom surface; The loading method of the shear tectonic movement in the horizontal plane is specifically: to achieve the loading by applying a horizontal displacement inversely proportional to the boundary length, and to apply a displacement constraint to the bottom surface; The S4 comprises the following sub-steps: S41. Establish the residual sum of squares formula of the least squares method based on the measured stress value component and the calculated stress value component ; In the formula, For the Measuring point The measured stress component, For the Influence factors of geostress Measuring point Calculated values of stress components, , m is the total number of on-site geostress measuring points set up, is the regression coefficient of the corresponding independent variable, i=1,2,3,…,n, n is the total number of regression coefficients of the corresponding independent variable; S42. According to the principle of least square method, the residual sum of squares is minimized and the regression coefficient is calculated; S43, establishing a regression equation according to the regression coefficient, calculating the stress regression calculation value of any point in the region, and obtaining a regression result; In the formula, is a single-column matrix of calculated values, is the stress regression calculated value, is the regression coefficient of the corresponding independent variable; S44, perform statistical test on the regression result and regression accuracy analysis. If the statistical test on the regression result is qualified and the regression accuracy meets the engineering requirements, the regression equation obtained in S43 is adopted and the process proceeds to S6. If the statistical test on the regression result is unqualified or the regression accuracy does not meet the engineering requirements, the process proceeds to S5. In S44, the method for statistical testing of regression results is specifically: According to the regression results, the regression sum of squares and the residual sum of squares are calculated, and then the total deviation, the multiple correlation coefficient and the test parameter F are calculated. If the multiple correlation coefficient is within the qualified interval and the test parameter F is greater than the test threshold, the statistical test of the regression result is qualified; Among them, the expression of regression square sum U is: In the formula, For the Measuring point Stress regression calculation value; The expression of the residual sum of squares Q is: In the formula, For the Influence factors of geostress Measuring point Stress regression calculation value; Total deviation S yy : The expression of the multiple correlation coefficient R is: The expression of the test parameter F is: In S44, the method of regression accuracy analysis is specifically as follows: S441. The relative error between the inversion results and the measured values of each measuring point of the project is represented by the following formula: In the formula, For the The relative error between the inversion result of the geostress at each measuring point and the measured result, i=1,2,3,…,n; For the The inversion results of the ground stress at each measuring point are: Indicates The measured results of the ground stress at each measuring point; Represents the average value of the absolute value of the inversion relative error; S442, according to the average value of the absolute value of the inversion relative error and the actual engineering accuracy requirements The specific method for judging whether the engineering inversion accuracy requirements are met is as follows: when When , the calculation formula of stress regression calculation value meets the accuracy requirement; when When , the calculation formula of stress regression calculation value does not meet the accuracy requirement; The S5 comprises the following sub-steps: S51, based on the p geostress measuring holes set up on site, the measured values of the stress components and the calculated values of the stress components are grouped into p groups according to different measuring holes, and each group of data is used in turn to invert and calculate the regression coefficient to obtain the regression equation of the p geostress measuring holes, and use the regression equation as the adjusted regression equation; S52, extracting stress components at the geostress measuring holes, and substituting the stress components into the regression equation of each geostress measuring hole to obtain regression values of p geostress measuring holes; S53, perform statistical test on the regression result and regression accuracy analysis based on the regression values of p geostress measuring holes. If the statistical test on the regression result is qualified and the regression accuracy meets the engineering requirements, use the adjusted regression equation and enter S6. If the statistical test on the regression result is unqualified or the regression accuracy does not meet the engineering requirements, return to S2, adjust the three-dimensional geological inversion model and boundary conditions, and re-simulate and regress.
2. The multivariate linear regression geostress inversion method based on single-hole partitioning according to claim 1 is characterized in that: The S1 comprises the following sub-steps: S11, collecting measured stress values through on-site geostress measuring points, performing stress component conversion on the measured stress values, and obtaining six geostress components in the geodetic coordinate system; S12. The six geostress components in the geodetic coordinate system are converted into six stress components in the calculation coordinate system by using a coordinate transformation formula, thereby obtaining the measured value components.
3. The multivariate linear regression geostress inversion method based on single-hole partitioning according to claim 2 is characterized in that: In S11, the expression for stress component conversion is specifically as follows: In the formula, , and is the measured stress value, , , , , and is the geostress component, T is the transposed sign, is the stress component conversion matrix, and its specific expression is: In the formula, is the measured value of the i-th stress Direction cosines for the x-axis, i=1,2,3, is the measured value of the i-th stress The direction cosines about the y-axis, is the measured value of the i-th stress Direction cosines about the z-axis.
4. The multivariate linear regression geostress inversion method based on single-hole partitioning according to claim 3 is characterized in that: In S12, the method for obtaining the stress component under the calculation coordinate is specifically: The ground stress components are transformed into stress components under calculation coordinates through the coordinate transformation formula, where the coordinate transformation formula is specifically as follows: In the formula, , , , , and To calculate the stress components in coordinates, is the coordinate transformation matrix, and its specific expression is: 。 5. The multivariate linear regression geostress inversion method based on single-hole partitioning according to claim 1 is characterized in that: In S2, the three-dimensional geological inversion model includes distribution information of fault fracture zones, river valley strata, intrusive rock strata and different rock strata, wherein the distribution information includes: strike, dip, inclination and fault width.
Citation Information
Patent Citations
Geostress field distributed data management (DDM) retrieval method based on actual measurement results and linear programming method
CN103091727A
Stability monitoring method for cold high-altitude steep slope
CN107067333A
Ground stress inversion analysis method and system for tunnel soft rock deformation grade evaluation
CN115688237A
Method and system for online monitoring and optimization of mining and mineral processing operations
IN201821040541A
Multi-functional true triaxial rock drilling test system and method
WO2018006585A1
