CN-120449404-B - Hydropower high slope response updating method and device based on constraint Bayesian reasoning

CN120449404BCN 120449404 BCN120449404 BCN 120449404BCN-120449404-B

Abstract

The application relates to the technical field of slope engineering response updating, in particular to a hydropower high slope response updating method and device based on constraint Bayesian reasoning, comprising the steps of determining prior distribution of key material parameters in an actual slope of a hydropower station, and constructing a monitoring data likelihood function by utilizing slope monitoring data of the actual slope; and combining the prior distribution, the monitoring data likelihood function and the constraint likelihood function by using a Bayesian theory to determine posterior distribution of the key material parameters, thereby determining equivalent samples of the key material parameters by using a slope response proxy model and a target sampling algorithm to update the slope response of the hydropower station and obtain a hydropower high slope response updating result based on constraint Bayesian reasoning. Therefore, the problem that uncertainty of rock-soil material parameters cannot be effectively reduced when the types and the quantity of monitoring data are limited by traditional Bayesian reasoning in the related technology is solved.

Inventors

TANG XIAOSONG
HU BO
LI DIANQING

Assignees

武汉大学

Dates

Publication Date: 20260508
Application Date: 20250327

Claims (8)

1. A hydropower high slope response updating method based on constraint Bayesian reasoning is characterized by comprising the following steps: Determining prior distribution of target key material parameters in an actual slope of a target hydropower station, and constructing a monitoring data likelihood function by using slope monitoring data of the actual slope; Determining a target constraint condition of the target key material parameter based on the target physical meaning of the rock-soil body parameter in the actual slope, so as to construct a target constraint likelihood function according to the target constraint condition, wherein the target constraint condition is divided into an equality constraint and an inequality constraint according to constraint properties, and is divided into a soft constraint and a hard constraint according to constraint degrees, the target constraint likelihood function comprises an equality soft constraint likelihood function, an inequality soft constraint likelihood function and an inequality hard constraint likelihood function, and the equality soft constraint likelihood function is characterized by: Wherein, the ; (L=1, 2,., N c ) is the first constraint on θ, N c is the total constraint number; Is that Standard deviation of (2); The inequality soft constraint likelihood function is characterized by: Wherein, the Is a normalization constant, ensure Integration to 1 over the parameters possible; the inequality hard constraint likelihood function is characterized by: wherein mu g(θ) and sigma g(θ) are the mean value and standard deviation of g (theta) respectively, a and b are the left and right boundaries of g (theta) respectively, and phi is the standard normal cumulative distribution function; Combining the prior distribution, the monitoring data likelihood function and the target constraint likelihood function by using a Bayesian theory to determine posterior distribution of the target key material parameter, determining an equivalent sample of the target key material parameter by using a pre-constructed slope response proxy model and a target sampling algorithm based on the posterior distribution, and updating the slope response of the target hydropower station by using the equivalent sample to obtain a hydropower high slope response updating result based on constraint Bayesian inference, wherein the posterior distribution is expressed as: where P (θ) is an a priori distribution, P (Z|θ) is a monitor data likelihood function, P (G (θ) |θ) is a constraint likelihood function, Is a normalization constant that ensures that P (theta|z, G (theta)) is integrated to 1 over the range of parameters possible.
2. The method of claim 1, wherein constructing a monitoring data likelihood function using the slope monitoring data of the actual slope comprises: acquiring slope monitoring data of the actual slope, and determining a measurement error of the slope monitoring data; and constructing the monitoring data likelihood function based on the slope monitoring data and the measurement error.
3. The method of claim 1, wherein said determining an equivalent sample of said critical material parameter using a pre-constructed slope response proxy model and a target sampling algorithm comprises: Determining a target posterior sample by using the slope response agent model and a target Markov chain Monte Carlo simulation (MCMC) algorithm based on the posterior distribution; an equivalent sample of the critical material parameter is determined based on the target posterior sample.
4. The method of claim 1, wherein the hydropower high slope response update result is expressed as: where Θ represents an equivalent sample, and N represents the number of equivalent samples.
5. The utility model provides a water and electricity high slope response updating device based on constraint Bayesian inference which characterized in that includes: the determining module is used for determining prior distribution of target key material parameters in an actual side slope of a target hydropower station and constructing a monitoring data likelihood function by utilizing side slope monitoring data of the actual side slope; The construction module is used for determining target constraint conditions of the target key material parameters based on target physical meanings of the rock-soil body parameters in the actual side slope, so as to construct target constraint likelihood functions according to the target constraint conditions, wherein the target constraint conditions are divided into equality constraint and inequality constraint according to constraint properties, the equality constraint likelihood functions are divided into soft constraint and hard constraint according to constraint degrees, the target constraint likelihood functions comprise equality soft constraint likelihood functions, inequality soft constraint likelihood functions and inequality hard constraint likelihood functions, and the equality soft constraint likelihood functions are characterized by: Wherein, the ; (L=1, 2,., N c ) is the first constraint on θ, N c is the total constraint number; Is that Standard deviation of (2); The inequality soft constraint likelihood function is characterized by: Wherein, the Is a normalization constant, ensure Integration to 1 over the parameters possible; the inequality hard constraint likelihood function is characterized by: wherein mu g(θ) and sigma g(θ) are the mean value and standard deviation of g (theta) respectively, a and b are the left and right boundaries of g (theta) respectively, and phi is the standard normal cumulative distribution function; The updating module is used for combining the prior distribution, the monitoring data likelihood function and the target constraint likelihood function to determine posterior distribution of the target key material parameter, determining an equivalent sample of the target key material parameter by utilizing a pre-constructed slope response agency model and a target sampling algorithm based on the posterior distribution, and updating the slope response of the target hydropower station by utilizing the equivalent sample to obtain a hydropower high slope response updating result based on constraint Bayesian inference, wherein the posterior distribution is expressed as: where P (θ) is an a priori distribution, P (Z|θ) is a monitor data likelihood function, P (G (θ) |θ) is a constraint likelihood function, Is a normalization constant that ensures that P (theta|z, G (theta)) is integrated to 1 over the range of parameters possible.
6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the constrained bayesian-based hydropower highside response updating method according to any of claims 1-4.
7. A computer readable storage medium having stored thereon a computer program, the program being executable by a processor for implementing a hydropower highside response updating method based on constrained bayesian reasoning according to any of claims 1-4.
8. A computer program product comprising a computer program, wherein the computer program is executed by a processor for implementing a hydropower highslope response updating method based on constrained bayesian reasoning as claimed in any one of claims 1-4.

Description

Hydropower high slope response updating method and device based on constraint Bayesian reasoning Technical Field The application relates to the technical field of slope engineering response updating, in particular to a hydropower high slope response updating method and device based on constraint Bayesian reasoning. Background Along with the development of the monitoring technology and the birth of an automatic monitoring instrument, the slope engineering practice has monitoring data such as ground deformation, pile top displacement, anchoring stress and the like besides limited actual measurement data. The monitoring data are utilized to make up for the defect of rareness of measured data, thereby reducing the uncertainty of rock-soil mass parameters and improving the accuracy of slope response calculation, and in view of the fact, various parameter inversion methods are developed to reduce the uncertainty of parameters, such as neural networks, ensemble Kalman filtering, bayesian reasoning and the like. However, in the related art, the conventional bayesian reasoning is mainly to invert the geotechnical material parameters by fusing the monitoring data, the variety and the quantity of the monitoring data for most geotechnical engineering are still very limited, the collection is time-consuming and labor-consuming, the uncertainty of the geotechnical material parameters cannot be effectively reduced, and accurate slope response cannot be obtained for the conventional bayesian reasoning, so that the problem is to be solved. Disclosure of Invention The present application is based on the inventors' knowledge and knowledge of the following problems: The river in southwest China has rich water energy resources due to the large land level difference. In recent years, hydropower development in southwest areas is vigorous, and a large number of high dams are built successively. In the construction and operation process, besides the stability of the dam, the stability of the slope of the reservoir area is also a key factor for controlling the safety of the whole engineering. The instability of the reservoir side slope can cause landslide-surge-dam break-flood chain disasters, and the life and property safety of people is seriously threatened. Therefore, accurate response calculation of the pool side slope is important. However, due to the complex geological conditions in southwest China, the geological activities are frequent, and accurate slope response calculation is a very challenging task. And the rock-soil body material is used as a natural material, so that the material has great uncertainty due to stress history, physical and chemical actions and the like, and in-situ experiments and laboratory experiments are complex and time-consuming, and the actual measurement data of the rock-soil body parameters of a specific field are usually relatively rare. Obviously, the rock-soil body parameter values estimated from a small amount of measured data cannot represent the true values thereof, which results in significant errors in calculating the slope response using numerical simulation. In order to improve the slope response calculation precision, the uncertainty of the rock-soil body parameters is greatly reduced. Along with the development of the monitoring technology and the birth of an automatic monitoring instrument, the slope engineering practice has monitoring data such as ground deformation, pile top displacement, anchoring stress and the like besides limited actual measurement data. The defect of rareness of measured data is made up by using the monitoring data, thereby reducing uncertainty of rock-soil body parameters and improving accuracy of slope response calculation. In view of this, various parametric inversion methods have been developed to reduce uncertainty of parameters, such as neural networks, ensemble kalman filtering, bayesian reasoning, and the like. The Bayesian inference has at least two advantages of (1) integrating prior information and monitoring data into posterior distribution to provide reliable rock-soil mass parameter probability distribution for slope response update, and (2) not only can definitely model uncertainty of rock-soil mass parameters, but also can utilize the monitoring data to reduce the uncertainty of the rock-soil mass parameters, so that the Bayesian inference is widely used in geotechnical engineering parameter probability inversion. However, the complexity and specificity of geotechnical engineering have the disadvantages of (1) conventional Bayesian reasoning, which is to invert geotechnical material parameters mainly by fusing monitoring data. Despite the great development of monitoring technology in recent years, in most geotechnical engineering practices, the variety and quantity of monitoring data are still generally very limited, which makes Bayesian reasoning not very effective in reducing uncertainty of geotechnical material parameters, and (2) most of geo