Advanced prediction method and system of tunnel adverse geology based on data fusion theory

Abstract

The present invention provides a method and system for advanced prediction of bad geology in tunnels based on data fusion theory, and relates to the technical field of data processing systems or methods for prediction purposes. The method determines tunnel mileage sections requiring advanced prediction of bad geology, bad geological conditions requiring advanced prediction, bad geological problems requiring advanced prediction, and corresponding multiple advanced prediction methods. The same mileage sections corresponding to multiple advanced prediction methods are used as basic units for data fusion, and a data fusion index system corresponding to each basic unit of data fusion is determined. Each index in the data fusion index system is assigned a value according to an assignment standard to obtain an assignment result of each index in the data fusion index system. Multi-source data fusion calculation and analysis are performed based on the Kalman filter principle to obtain the probability of the bad geological problem requiring prediction, thereby solving the problem that the probability of the bad geological problem cannot be calculated in the prior art. The present invention is suitable for probability calculation of bad geological problems.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing systems or methods for prediction purposes, and in particular to a method and system for advanced prediction of poor tunnel geology based on data fusion theory. Background Art

[0002] Tunnel construction is largely restricted by geological conditions. During the tunnel construction process, the complexity, particularity and suddenness of various adverse geological problems often become the most important factors restricting tunnel construction. Therefore, advanced geological forecasting work is of great significance to the safe construction of tunnels.

[0003] At present, the advanced geological forecast in railway tunnel and other projects mainly adopts the simple combination of geological exploration information, drilling information and various geological detection information, and empirical judgment. It fails to effectively combine multi-source advanced geological forecast data, cannot calculate the probability of adverse geological problems, and cannot conduct risk assessment. Summary of the invention

[0004] The technical problem solved by the present invention is to provide a method and system for advanced prediction of tunnel adverse geology based on data fusion theory, so as to solve the problem that the probability of adverse geological problems cannot be calculated in the prior art.

[0005] The present invention solves the above technical problems by adopting a technical solution: a method for advanced prediction of poor tunnel geology based on data fusion theory, comprising the following steps:

[0006] S1. Summarize the adverse geological conditions and adverse geological problems of the tunnel, analyze and obtain the advanced prediction methods corresponding to each adverse geological condition and each adverse geological problem, and establish a data fusion index system library for adverse geological conditions and adverse geological problems;

[0007] S2. Divide the weight level of each indicator in each data fusion indicator system;

[0008] S3, determine the value assignment standard of each indicator in each data fusion indicator system;

[0009] S4. Determine the tunnel mileage sections that require advanced prediction of adverse geology, adverse geological conditions that require advanced prediction, adverse geological problems that require advanced prediction, and corresponding multiple advanced prediction methods, use the same mileage sections corresponding to multiple advanced prediction methods as basic units of data fusion, determine the data fusion index system corresponding to each basic unit of data fusion, and assign values ​​to each index in the data fusion index system according to the assignment standard to obtain the assignment results of each index in the data fusion index system;

[0010] S5. Obtain a first error covariance matrix corresponding to the adverse geological conditions that need to be predicted through the advanced prediction cases of adverse geological conditions, obtain a second error covariance matrix corresponding to the adverse geological problems that need to be predicted through the advanced prediction cases of adverse geological problems, and set a third error covariance matrix corresponding to each advanced prediction method according to the weight level based on the adopted advanced prediction method;

[0011] S6. Based on the assignment results of each indicator in the data fusion indicator system, the first error covariance matrix, the second error covariance matrix and the third error covariance matrix, multi-source data fusion calculation and analysis are performed based on the Kalman filter principle to obtain the probability of adverse geological problems that need to be predicted.

[0012] Furthermore, the adverse geological conditions include high ground stress, faults, broken zones, weak rock zones, karst, rocks containing toxic and harmful gases, rocks containing radioactive sources, high rock temperature and high water temperature; the adverse geological problems include high ground stress hard rock rock bursts and large deformation of soft rocks, fault collapse, broken zone collapse, water gushing, karst caves, sudden water and mud, deformation and collapse of weak rock zones, high geothermal heat hazards, toxic and harmful gas hazards and radioactive hazards.

[0013] Furthermore, the weight level includes four levels.

[0014] Furthermore, the value assignment standard is that the value assigned to each indicator is between 0 and 1.

[0015] Furthermore, S6 specifically includes the following steps:

[0016] S601, taking the result of assigning a certain indicator in the data fusion indicator system as an initial estimated value;

[0017] S602, determining a state transfer matrix, and calculating a priori estimate values ​​according to the initial estimate values ​​and the state transfer matrix;

[0018] S603, calculating a priori state mean square error matrix according to the state transfer matrix, the first error covariance matrix and the second error covariance matrix;

[0019] S604, determining an input relationship matrix, and calculating Kalman filter coefficients according to the input relationship matrix, the priori state mean square error matrix and the third error covariance matrix;

[0020] S605, calculating the posterior estimation value according to the Kalman filter coefficient, the prior estimation value, the assignment result of the prediction result of the current advance prediction method and the input relationship matrix;

[0021] S606, updating the posterior state mean square error matrix according to the Kalman filter coefficient, the input relationship matrix and the prior state mean square error matrix;

[0022] S607, repeat S602 to S606 until the indicator fusion iteration corresponding to each advanced prediction method is completed, and the probability of the adverse geological problem that needs to be predicted is obtained.

[0023] Furthermore, an indicator in the data fusion indicator system is one of advanced exploration drilling information, preliminary geological survey data, excavation geological cataloging and mapping, advanced geophysical prospecting, excavation construction information, monitoring and detection information and experimental testing information.

[0024] Furthermore, the formula for calculating the prior estimate is: ,in, represents the prior estimate, represents the current estimated value, is the number of fusion iterations, when hour, represents the initial estimate, represents the state transition matrix.

[0025] Furthermore, the formula for calculating the prior state mean square error matrix is: ,in, represents the prior state mean square error matrix, express The transposed matrix of represents the first error covariance matrix, represents the second error covariance matrix.

[0026] Furthermore, the formula for calculating the Kalman filter coefficient is: ,in, represents the Kalman filter coefficient, represents the third error covariance matrix, represents the input relationship matrix, express The transposed matrix of ; the formula for calculating the posterior estimate is: ,in, represents the posterior estimate, It represents the assignment result of the forecast result of the current advance forecast method; the update formula of the posterior state mean square error matrix is: ,in, represents the posterior state mean square error matrix, Represents the identity matrix.

[0027] The present invention also provides a tunnel adverse geology advance prediction system based on data fusion theory for implementing the above method, realizing the above method for tunnel adverse geology advance prediction based on data fusion theory, the system includes an indicator system storage module, a parameter and matrix analysis module and a data fusion calculation module; the indicator system storage module is used to store the data fusion indicator system of adverse geological conditions and adverse geological problems, as well as the weight and assignment of each indicator; the parameter and matrix analysis module is used to analyze and store the parameters and matrices required by the data fusion calculation module, the data fusion calculation module uses the Kalman filtering algorithm to perform multi-source data fusion calculation and analysis on all indicators in the indicator system through iterative calculation to obtain the probability of the adverse geological problems that need to be predicted.

[0028] Beneficial effects of the invention: The invention provides a method and system for advanced prediction of bad geological conditions in tunnels based on data fusion theory, which summarizes bad geological conditions and bad geological problems in tunnels, and analyzes and obtains the advanced prediction methods corresponding to each bad geological condition and the advanced prediction methods corresponding to each bad geological problem, establishes a data fusion index system library for bad geological conditions and bad geological problems, determines the tunnel mileage sections that need advanced prediction of bad geological conditions, bad geological conditions that need advanced prediction, bad geological problems that need advanced prediction, and corresponding multiple advanced prediction methods, takes the same mileage sections corresponding to multiple advanced prediction methods as the basic units of data fusion, determines the data fusion index system corresponding to each basic unit of data fusion, and assigns the corresponding data fusion index system to the basic units of data fusion. The value standard is used to assign values ​​to each indicator in the data fusion indicator system to obtain the assignment results of each indicator in the data fusion indicator system. Through the advance prediction cases of adverse geological conditions, the first error covariance matrix corresponding to the adverse geological conditions that need to be predicted is obtained. Through the advance prediction cases of adverse geological problems, the second error covariance matrix corresponding to the adverse geological problems that need to be predicted is obtained. Based on the adopted advance prediction method, the third error covariance matrix corresponding to each advance prediction method is set according to the weight level. Based on the Kalman filter principle, multi-source data fusion calculation and analysis are performed to obtain the probability of the adverse geological problems that need to be predicted, which solves the problem that the probability of adverse geological problems cannot be calculated in the existing technology, and is helpful for the risk assessment of tunnel engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flow chart of the method for advanced prediction of adverse geology in tunnels based on data fusion theory provided by the present invention. DETAILED DESCRIPTION

[0030] Aiming at the problem that the probability of adverse geological problems cannot be calculated, the present invention provides a method for advanced prediction of adverse geological conditions in tunnels based on data fusion theory. Figure 1 As shown, the following steps are included:

[0031] S1. Summarize the adverse geological conditions and adverse geological problems of the tunnel, analyze and obtain the advanced prediction methods corresponding to each adverse geological condition and each adverse geological problem, and establish a data fusion index system library for adverse geological conditions and adverse geological problems;

[0032] Specifically, the adverse geological conditions include high ground stress, faults, broken zones, weak rock zones, karst, rocks containing toxic and harmful gases, rocks containing radioactive sources, high rock temperature and high water temperature; the adverse geological problems include high ground stress hard rock rock bursts and large deformation of soft rocks, fault collapse, broken zone collapse, water gushing, karst caves, sudden water and mud, deformation and collapse of weak rock zones, high geothermal heat hazards, toxic and harmful gas hazards and radioactive hazards.

[0033] S2. Divide the weight level of each indicator in each data fusion indicator system;

[0034] Specifically, for certain adverse geological conditions or adverse geological problems, different forecasting methods have different accuracy, so the weights of different indicators in the data fusion indicator system are also different. After completing the construction of the indicator system library, the weight level of each indicator should be divided based on factors such as method principles and engineering practice. For example, the indicator weights can be divided into four levels from high to low: A, B, C, and D, which respectively represent important, relatively important, less important, and unimportant.

[0035] S3, determine the value assignment standard of each indicator in each data fusion indicator system;

[0036] Specifically, the data information obtained by different forecasting methods includes qualitative and quantitative categories, and there are differences in dimensions, orders of magnitude, etc. Therefore, the assignment rules of data fusion indicators covered by various types of adverse geological conditions and adverse geological problems are established, so that the assignment of each indicator is between 0 and 1, in order to clarify the data fusion indicator assignment standards and realize the dimensionless normalization of indicators; for quantitative indicators, their dimensionless normalized assignments should be determined based on their original values ​​and their value ranges; for qualitative indicators, it is necessary to assign indicators based on their response to specific adverse geological problems and their corresponding results.

[0037] S4. Determine the tunnel mileage sections that require advanced prediction of adverse geology, adverse geological conditions that require advanced prediction, adverse geological problems that require advanced prediction, and corresponding multiple advanced prediction methods, use the same mileage sections corresponding to multiple advanced prediction methods as basic units of data fusion, determine the data fusion index system corresponding to each basic unit of data fusion, and assign values ​​to each index in the data fusion index system according to the assignment standard to obtain the assignment results of each index in the data fusion index system;

[0038] Specifically, since different prediction methods have different effective distances, when conducting data fusion analysis on tunnel mileage sections that are predicted by multiple prediction methods, the mileage sections covering the same prediction methods should be used as the basic unit sections for data fusion analysis, and the specific data fusion indicator system for various adverse geological conditions and problems of each basic unit of data fusion should be determined to ensure the stability of the indicator system during a single data fusion analysis; as the tunnel is excavated, the advance prediction method with a shorter detection distance will be implemented again, resulting in the mileage sections that were not originally covered by the advance prediction method with a shorter detection distance needing to be re-included in the data information predicted by the advance prediction method with a shorter detection distance.

[0039] S5. Obtain a first error covariance matrix corresponding to the adverse geological conditions that need to be predicted through the advanced prediction cases of adverse geological conditions, obtain a second error covariance matrix corresponding to the adverse geological problems that need to be predicted through the advanced prediction cases of adverse geological problems, and set a third error covariance matrix corresponding to each advanced prediction method according to the weight level based on the adopted advanced prediction method;

[0040] Specifically, the first error covariance matrix of the adverse geological conditions that need to be predicted has nothing to do with the specific prediction method; the second error covariance matrix of the adverse geological problems that need to be predicted is also independent of the specific prediction method; the third error covariance matrix corresponding to each prediction method represents the error analysis result of the specific prediction method, and when determining it, it should be corrected in combination with the weight level of the indicator. For indicators with higher weight levels, the value of the covariance matrix of the corresponding prediction method can be appropriately reduced. On the contrary, for indicators with lower weight levels, the value of the covariance matrix of the corresponding prediction method can be appropriately increased.

[0041] S6, according to the assignment results of each indicator in the data fusion indicator system, the first error covariance matrix, the second error covariance matrix and the third error covariance matrix, based on the Kalman filter principle, multi-source data fusion calculation and analysis are performed to obtain the probability of adverse geological problems that need to be predicted, specifically including the following steps:

[0042] S601, taking the result of assigning a certain indicator in the data fusion indicator system as an initial estimated value;

[0043] Specifically, an indicator in the data fusion indicator system is one of the following: advance exploration drilling information, preliminary geological survey data, geological cataloging and mapping of excavation, advance geophysical prospecting, excavation construction information, monitoring and detection information, and experimental test information. In particular, if there is advance exploration drilling information, the assignment result corresponding to the advance exploration drilling information is preferentially selected as the initial estimate.

[0044] S602, determining a state transfer matrix, and calculating a priori estimate values ​​according to the initial estimate values ​​and the state transfer matrix;

[0045] Specifically, the formula for calculating the prior estimate is: ,in, represents the prior estimate, represents the current estimated value, ,when hour, represents the initial estimate, Represents the state transfer matrix. Usually, the state transfer matrix is ​​the identity matrix.

[0046] S603, calculating a priori state mean square error matrix according to the state transfer matrix, the first error covariance matrix and the second error covariance matrix;

[0047] Specifically, the formula for calculating the prior state mean square error matrix is: ,in, represents the prior state mean square error matrix, express The transposed matrix of represents the first error covariance matrix, represents the second error covariance matrix.

[0048] S604, determining an input relationship matrix, and calculating Kalman filter coefficients according to the input relationship matrix, the priori state mean square error matrix and the third error covariance matrix;

[0049] Specifically, the formula for calculating the Kalman filter coefficient is: ,in, represents the Kalman filter coefficient, represents the prior state mean square error matrix, represents the third error covariance matrix, represents the input relationship matrix, express The transposed matrix of

[0050] S605, calculating the posterior estimation value according to the Kalman filter coefficient, the prior estimation value, the assignment result of the prediction result of the current advance prediction method and the input relationship matrix;

[0051] Specifically, the formula for calculating the posterior estimate is: ,in, represents the posterior estimate, Indicates the assignment result of the forecast result of the current advance forecast method, represents the prior estimate, represents the Kalman filter coefficient, Represents the input relationship matrix.

[0052] S606, updating the posterior state mean square error matrix according to the Kalman filter coefficient, the input relationship matrix and the prior state mean square error matrix;

[0053] Specifically, the update formula of the posterior state mean square error matrix is: ,in, represents the posterior state mean square error matrix, represents the identity matrix, represents the Kalman filter coefficient, represents the input relationship matrix, represents the prior state mean square error matrix.

[0054] S607, repeat S602 to S606 until the fusion iteration of the indicators corresponding to each advanced prediction method is completed, and the final posterior estimation value is obtained, that is, the probability of the adverse geological problem that needs to be predicted is obtained.

[0055] The present invention also provides a tunnel adverse geology advance prediction system based on data fusion theory, which realizes the above-mentioned tunnel adverse geology advance prediction method based on data fusion theory. The system includes an indicator system storage module, a parameter and matrix analysis module and a data fusion calculation module; the indicator system storage module is used to store the data fusion indicator system of adverse geological conditions and adverse geological problems, as well as the weight and assignment of each indicator; the parameter and matrix analysis module is used to analyze and store the parameters and matrices required by the data fusion calculation module, and the data fusion calculation module uses the Kalman filtering algorithm to perform multi-source data fusion calculation and analysis on all indicators in the indicator system through iterative calculation to obtain the probability of the adverse geological problems that need to be predicted.

[0056] Specifically, the parameter and matrix analysis module is also used to retrieve the parameters or matrices used in any iterative calculation at any time, which is convenient for debugging the calculation program and recalculating when the indicator value is updated.

Claims

1. The method for advanced prediction of tunnel adverse geology based on data fusion theory is characterized by: The following steps are involved: S1. Summarize the adverse geological conditions and adverse geological problems of the tunnel, analyze and obtain the advanced prediction methods corresponding to each adverse geological condition and each adverse geological problem, and establish a data fusion index system library for adverse geological conditions and adverse geological problems; S2. Divide the weight level of each indicator in each data fusion indicator system; S3, determine the value assignment standard of each indicator in each data fusion indicator system; S4. Determine the tunnel mileage sections that require advanced prediction of adverse geology, adverse geological conditions that require advanced prediction, adverse geological problems that require advanced prediction, and corresponding multiple advanced prediction methods, use the same mileage sections corresponding to multiple advanced prediction methods as basic units of data fusion, determine the data fusion index system corresponding to each basic unit of data fusion, and assign values ​​to each index in the data fusion index system according to the assignment standard to obtain the assignment results of each index in the data fusion index system; S5. Obtain a first error covariance matrix corresponding to the adverse geological conditions that need to be predicted through the advanced prediction cases of adverse geological conditions, obtain a second error covariance matrix corresponding to the adverse geological problems that need to be predicted through the advanced prediction cases of adverse geological problems, and set a third error covariance matrix corresponding to each advanced prediction method according to the weight level based on the adopted advanced prediction method; S6. Based on the assignment results of each indicator in the data fusion indicator system, the first error covariance matrix, the second error covariance matrix and the third error covariance matrix, multi-source data fusion calculation and analysis are performed based on the Kalman filter principle to obtain the probability of adverse geological problems that need to be predicted.

2. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 1 is characterized in that: The adverse geological conditions include high ground stress, faults, broken zones, weak rock zones, karst, rocks containing toxic and harmful gases, rocks containing radioactive sources, high rock temperature and high water temperature; the adverse geological problems include high ground stress hard rock rock bursts and large deformation of soft rocks, fault collapse, broken zone collapse, water gushing, karst caves, sudden water and mud, deformation and collapse of weak rock zones, high geothermal heat hazards, toxic and harmful gas hazards and radioactive hazards.

3. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 1 is characterized in that: The weight levels include four levels.

4. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 1 is characterized in that: The value assignment standard is that the value assigned to each indicator is between 0 and 1.

5. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 1 is characterized in that: S6 specifically includes the following steps: S601, taking the result of assigning a certain indicator in the data fusion indicator system as an initial estimated value; S602, determining a state transfer matrix, and calculating a priori estimate values ​​according to the initial estimate values ​​and the state transfer matrix; S603, calculating a priori state mean square error matrix according to the state transfer matrix, the first error covariance matrix and the second error covariance matrix; S604, determining an input relationship matrix, and calculating Kalman filter coefficients according to the input relationship matrix, the priori state mean square error matrix and the third error covariance matrix; S605, calculating the posterior estimation value according to the Kalman filter coefficient, the prior estimation value, the assignment result of the prediction result of the current advance prediction method and the input relationship matrix; S606, updating the posterior state mean square error matrix according to the Kalman filter coefficient, the input relationship matrix and the prior state mean square error matrix; S607, repeat S602 to S606 until the indicator fusion iteration corresponding to each advanced prediction method is completed, and the probability of the adverse geological problem that needs to be predicted is obtained.

6. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 5 is characterized in that: An indicator in the data fusion indicator system is one of advanced exploration drilling information, preliminary geological survey data, excavation geological cataloging and mapping, advanced geophysical prospecting, excavation construction information, monitoring and detection information and experimental test information.

7. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 5 is characterized in that: The formula for calculating the prior estimate is: ,in, represents the prior estimate, represents the current estimated value, is the number of fusion iterations, when hour, represents the initial estimate, represents the state transition matrix.

8. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 7 is characterized in that: The formula for calculating the prior state mean square error matrix is: ,in, represents the prior state mean square error matrix, express The transposed matrix of represents the first error covariance matrix, represents the second error covariance matrix.

9. The method for advanced prediction of tunnel adverse geology based on data fusion theory according to claim 8 is characterized in that: The formula for calculating the Kalman filter coefficient is: ,in, represents the Kalman filter coefficient, represents the third error covariance matrix, represents the input relationship matrix, express The transposed matrix of ; the formula for calculating the posterior estimate is: ,in, represents the posterior estimate, It represents the assignment result of the forecast result of the current advance forecast method; the update formula of the posterior state mean square error matrix is: ,in, represents the posterior state mean square error matrix, Represents the identity matrix.

10. The tunnel adverse geological advance prediction system based on data fusion theory is characterized by: Implementing the method for advanced prediction of poor tunnel geology based on data fusion theory as described in any one of claims 1 to 9, the system includes an index system storage module, a parameter and matrix analysis module and a data fusion calculation module; The indicator system storage module is used to store the data fusion indicator system of adverse geological conditions and adverse geological problems, as well as the weight and assignment of each indicator; the parameter and matrix analysis module is used to analyze and store the parameters and matrices required by the data fusion calculation module. The data fusion calculation module uses the Kalman filter algorithm to perform multi-source data fusion calculation and analysis on all indicators in the indicator system through iterative calculation to obtain the probability of adverse geological problems that need to be predicted.

Images