Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind plant based on CPT

Abstract

The invention discloses a Bayesian hierarchical probability prediction method for a small strain shear modulus Gmax of an offshore wind plant soil body based on CPT, and the method comprises the steps: obtaining Gmax data at different CPT positions of an offshore wind plant, and constructing a training sample; constructing a Bayesian regression model, including constructing a complete hybrid model and a partial hybrid model; the constructed Bayesian regression model is subjected to Markov chain Monte Carlo MCMC simulation, posterior distribution of model parameters is obtained, the fitting effect of the model is evaluated, and a target model meeting the prediction effect requirement is obtained; and according to the target model, dividing a plurality of scene prediction Gmax profiles, and further judging whether a CPT position and a soil body unit are in a calibration data set or not to complete probability prediction. According to the embodiment of the invention, the CPT data can be utilized more effectively, a more accurate soil parameter representative profile can be generated, the accuracy and reliability of designing the fan foundation are improved, and the method can be widely applied to the technical field of offshore wind power.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore wind power, and in particular to a Bayesian stratified probability prediction method for small strain shear modulus of soil in an offshore wind farm based on CPT. Background Art

[0002] As the scale of offshore wind farms continues to expand, it becomes crucial to accurately determine the representative or characteristic profiles of soil parameters used in wind turbine foundation design. Modern geotechnical limit state design (LSD) standards require that representative values ​​of soil parameters be specified at a certain confidence level or exceedance probability to reflect the uncertainty in soil parameter characterization. In the design of offshore wind farms, a combination of cone penetration test (CPT) and borehole sampling indoor testing is usually used to determine the spatial distribution of soil parameters. CPT can provide continuous, rapid and indirect in-situ soil parameter measurements, while borehole sampling is used to obtain direct measurements of soil parameters for calibrating empirical models based on CPT, thereby obtaining probabilistic characterization profiles of soil parameters based on their CPT measurement curves for different wind turbine locations, such as the five representative ABCDE profiles reflecting different probability levels commonly used in the field of marine geotechnical exploration. Among them, profiles A and E represent G max The low and high estimates (boundaries of the 90% prediction interval) of G are shown in Figure 2. max Best estimate, low estimate, and high estimate of the mean (bounds of the 90% confidence interval). Small strain shear modulus G max It is a key parameter in the design of offshore wind turbine foundations and is used to estimate the structural characteristic frequency and foundation stiffness (under fatigue and service limit states). In the site investigation of offshore wind farms, seismic static penetration test (SCPT) is widely used to obtain G max This method uses in-situ shear wave velocity measurements to derive G max , avoiding the disturbance of soil layer sampling in laboratory testing and being able to obtain a relatively large amount of data.

[0003] Although the traditional drilling sampling method can directly obtain soil samples for laboratory testing, it has many limitations. Drilling sampling in marine environments is costly and time-consuming. For large-scale offshore wind farm projects, soil samples can only be obtained at specific depths at limited wind turbine locations for indoor testing. The data is relatively sparse and cannot meet the requirements for accurate characterization of the spatial distribution of soil parameters at all wind turbine locations. Moreover, soil samples are easily disturbed during the seabed drilling sampling and transportation process, which can cause a certain deviation between the test results and the actual characteristics of the in-situ soil. CPT, as an in-situ testing technology, has been widely used in site surveys of offshore wind farms. It can continuously and quickly obtain the cone tip resistance (q t), lateral friction (fs) and pore water pressure (u), thereby indirectly reflecting the engineering properties of the soil. CPT testing has the advantages of being relatively non-invasive, low cost and high efficiency. It can provide a large amount of test data in a short time, making it possible to fully understand the soil parameters of the site.

[0004] Currently, based on CPT data, G max The conventional approach is the fully mixed approach. This approach assumes that the empirical relationship based on the CPT is constant throughout the wind farm, ignoring the significant differences that may exist between different CPT locations and between different soil elements. The fully mixed model uses all the data from the entire site to calibrate a single CPT-G max The empirical model does not take into account the hierarchical structure and spatial variability of the data. This leads to misleading probability predictions and unreasonable G max Representative profiles affect the accuracy and safety of wind turbine foundation design. For example, when selecting representative profiles of soil property parameters for design (such as profile B with a cautious low estimate of its mean), the fully mixed model may be too optimistic because it ignores the variation between data groups, resulting in a large deviation between the selected estimate and the actual situation, which in turn affects the design reliability of the wind turbine foundation. Summary of the invention

[0005] The main purpose of the embodiment of the present invention is to propose a Bayesian layered probability prediction method for the small strain shear modulus of an offshore wind farm based on CPT, which can more effectively utilize CPT data, generate more accurate representative profiles of soil layers, and improve the accuracy and reliability of designing wind turbine foundations.

[0006] To achieve the above object, one aspect of an embodiment of the present invention proposes a Bayesian hierarchical probability prediction method for small strain shear modulus of an offshore wind farm based on CPT, comprising the following steps:

[0007] Obtain soil small strain shear modulus data at different CPT positions in offshore wind farms and construct training samples;

[0008] Constructing a Bayesian regression model based on the training samples, including constructing a complete mixed model and a partial mixed model;

[0009] Perform Markov Chain Monte Carlo (MCMC) simulation on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, and then evaluate the fitting effect of the model to obtain the target model that meets the prediction effect requirements;

[0010] According to the target model, multiple scenarios are divided into prediction G max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, completing the probabilistic prediction.

[0011] In some embodiments, the step of obtaining soil small strain shear modulus data at different CPT positions of an offshore wind farm and constructing a training sample comprises the following steps:

[0012] Collect geological, geophysical and geotechnical data for the offshore wind farm from designated data sources, including cone tip resistance and pore water pressure data obtained from static cone penetration tests at multiple CPT locations, and in situ G measured by seismic static cone penetration tests (SCPTs) or other methods. max value;

[0013] The effective vertical stress is calculated based on the geological conditions of the site and the measurement depth. Specifically, the effective vertical stress corresponding to each data point is calculated by multiplying the depth below the seabed by the average effective stress gradient;

[0014] G max The values ​​are converted from the original measurement units into MPa, and are classified according to the CPT positions and soil units, a cross-classification data structure is constructed, the number of samples in each classification group is counted, a data summary is formed, and the training samples are constructed.

[0015] In some embodiments, the step of constructing a complete mixed model in the step of constructing a Bayesian regression model according to the training samples comprises the following steps:

[0016] Based on empirical formula Construct a complete mixed regression model, where q t represents the total cone tip resistance of the static penetration test (CPT) after correction of pore pressure; p a represents atmospheric pressure; σ′ v represents the effective vertical stress; α0, α1, α2 represent the three unknown parameters of the model;

[0017] Setting prior distributions for model parameters, using weak information priors to reflect the lack of specific prior knowledge;

[0018] A preliminary estimate of the possible range of values ​​for a parameter is represented by a prior distribution.

[0019] In some embodiments, the step of constructing a partial mixed model in the step of constructing a Bayesian regression model according to the training samples includes the following steps:

[0020] Considering the variation between CPT locations and soil units, a cross-classification hierarchical model was constructed based on the set exchangeability intercept.

[0021] In some embodiments, the Markov chain Monte Carlo (MCMC) simulation is performed on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, comprising the following steps:

[0022] The constructed Bayesian regression model was fitted using the probabilistic programming language Stan combined with the Markov Chain Monte Carlo simulation method;

[0023] Run multiple Markov chains in parallel, and each chain performs a corresponding number of simulated samplings to obtain the posterior distribution of model parameters;

[0024] Among them, during the simulation process, the posterior distribution of the parameters is judged whether it converges by checking the statistical indicators within and between chains. If the posterior distribution does not converge, the simulation parameters are adjusted and the simulation is re-run until a stable posterior distribution is obtained; among them, the simulation parameters include increasing the number of iterations and adjusting the prior distribution.

[0025] In some embodiments, the fitting effect of the evaluation model is obtained to obtain a target model that meets the prediction effect requirements, including the following steps:

[0026] Analyze the posterior distribution of model parameters, calculate the statistical indicators of the posterior mean, standard deviation, and 95% confidence interval, and evaluate the model's fit to the data;

[0027] Calculate the residual standard deviation τ ε , by comparing the τ of the fully mixed model and the cross-classification hierarchical model ε The value determines the goodness of fit of the model;

[0028] Calculate the intraclass correlation coefficient, the formula is The degree of similarity between data groups is measured by the intraclass correlation coefficient; wherein ICC represents the intraclass correlation coefficient; represents the variance of the local group (such as position j) (representing G max variance among different CPT positions); represents the variance of the parent group (such as unit k) (G max variance between different soil elements); Represents G max The variance of the residual;

[0029] The leave-one-out cross-validation method was used to compare the expected log-prediction density of different models and evaluate the prediction accuracy of the models;

[0030] Based on the values ​​of the predicted densities and their differences and standard errors, the target model with the best prediction performance is selected.

[0031] In some embodiments, the target model is divided into multiple scene predictions G max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, and the probability prediction is completed, including the following steps:

[0032] According to the CPT position and whether the soil unit is in the calibration data set, G is predicted in the following scenarios. max profile:

[0033] Existing locations and existing units with observed data: Use the estimated Calculate the group-specific β0 and substitute it into the model formula (3a) to predict G max profile; where β 0(jk) represents the intercept between CPT location j and soil unit k; represents the population mean of the intercept β0; represents the local deviation of position j; represents the local deviation of unit k; β0 represents the overall basic intercept;

[0034] Existing locations and existing units but no observations: Use β estimated from calibration data 0(jk) Calculate β0 for prediction;

[0035] Existing locations and new units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates;

[0036] New Locations and Existing Units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates;

[0037] New location and new unit: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates;

[0038] A leave-one-group (LOGO) analysis was performed, where the data of one soil unit at one CPT position were kept as the validation set each time, the remaining data were used to fit the model and make a priori predictions, and then the validation data were added to make a posteriori predictions; the G of the a priori and posteriori predictions were compared. max By comparing the changes in these profiles, we can analyze the degree of improvement of the target model after obtaining new data.

[0039] Another aspect of the embodiment of the present invention further provides a Bayesian hierarchical probability prediction system for small strain shear modulus of an offshore wind farm based on CPT, comprising:

[0040] The first module is used to obtain soil small strain shear modulus data at different CPT positions in offshore wind farms and construct training samples;

[0041] The second module is used to construct a Bayesian regression model according to the training samples, including constructing a complete mixed model and a partial mixed model;

[0042] The third module is used to perform Markov Chain Monte Carlo (MCMC) simulation on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, and then evaluate the fitting effect of the model to obtain the target model that meets the prediction effect requirements;

[0043] The fourth module is used to divide multiple scenes into prediction G according to the target model max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, completing the probabilistic prediction.

[0044] To achieve the above objective, another aspect of an embodiment of the present invention provides an electronic device, the electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the above-mentioned method when executing the computer program.

[0045] To achieve the above objective, another aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method is implemented.

[0046] The embodiment of the present invention also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above method.

[0047] The embodiments of the present invention include at least the following beneficial effects: the present invention provides a Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT, the scheme constructs training samples by acquiring soil small strain shear modulus data at different CPT positions of offshore wind farms; constructs a Bayesian regression model based on the training samples, including constructing a complete mixed model and a partial mixed model; performs Markov chain Monte Carlo MCMC simulation on the constructed Bayesian regression model to obtain the posterior distribution of model parameters, and then evaluates the fitting effect of the model to obtain a target model that meets the prediction effect requirements; divides multiple scenarios into prediction G according to the target model maxThe CPT data can be used to generate a more accurate representative soil profile, thereby improving the accuracy and reliability of the design of the wind turbine foundation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of an implementation environment provided by an embodiment of the present invention;

[0049] Figure 2 is a flow chart of overall steps provided by an embodiment of the present invention;

[0050] Figure 3 is an overview of a TNW wind farm with CPT drilling locations provided by an embodiment of the present invention;

[0051] Figure 4 is a directed acyclic graph of the Bayesian regression model provided by an embodiment of the present invention;

[0052] Figure 5 is a posterior estimation diagram of CPT position and soil unit effect provided by an embodiment of the present invention;

[0053] Figure 6 is a comparison diagram of β0 estimation values ​​provided by an embodiment of the present invention;

[0054] Figure 7 is a posterior normalized residual map provided by an embodiment of the present invention;

[0055] Figure 8 is a LOGO analysis result diagram provided by an embodiment of the present invention;

[0056] Fig. 9 This is a final prediction example of the entire wind farm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present invention. They are only examples of devices and methods consistent with some aspects of the embodiments of the present invention as detailed in the attached claims.

[0058] It is understood that the terms "first", "second", etc. used in the present invention may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present invention, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0059] The terms "at least one", "multiple", "each", "any", etc. used in the present invention, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein are only for the purpose of describing the embodiments of the present invention and are not intended to limit the present invention.

[0061] The Bayesian hierarchical probability prediction method for the small strain shear modulus of an offshore wind farm based on CPT provided in an embodiment of the present invention relates to the field of offshore wind power technology. The Bayesian hierarchical probability prediction method for the small strain shear modulus of an offshore wind farm based on CPT provided in an embodiment of the present invention can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application of the Bayesian hierarchical probability prediction method for the small strain shear modulus of an offshore wind farm based on CPT, etc., but is not limited to the above forms.

[0062] The present invention can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present invention can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0063] like Figure 1 FIG. 1 is a schematic diagram of an implementation environment provided by an embodiment of the present invention. Figure 1 The implementation environment includes at least one terminal 102 and a server 101. The terminal 102 and the server 101 can be connected to a network wirelessly or wired to complete data transmission and exchange.

[0064] Server 101 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as big data and artificial intelligence platforms.

[0065] In addition, the server 101 can also be a node server in the blockchain network. Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithm.

[0066] The terminal 102 may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc. The terminal 102 may also be a vehicle-mounted terminal of various device types described above, but is not limited thereto. The terminal 102 and the server 101 may be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment of the present invention.

[0067] Based on the example Figure 1In the implementation environment shown, an embodiment of the present invention provides a Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT. The following is explained by taking the application of the Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT in server 101 as an example. It can be understood that the method can also be applied to terminal 102.

[0068] Reference Figure 2 , Figure 2 A flowchart of a Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT applied to a server provided in an embodiment of the present invention. The execution subject of the method can be any of the aforementioned computer devices (including servers or terminals). Figure 2 , the method may include the following steps:

[0069] Obtain soil small strain shear modulus data at different CPT positions in offshore wind farms and construct training samples;

[0070] Constructing a Bayesian regression model based on the training samples, including constructing a complete mixed model and a partial mixed model;

[0071] Perform Markov Chain Monte Carlo (MCMC) simulation on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, and then evaluate the fitting effect of the model to obtain the target model that meets the prediction effect requirements;

[0072] According to the target model, multiple scenarios are divided into prediction G max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, completing the probabilistic prediction.

[0073] In some embodiments, the step of obtaining soil small strain shear modulus data at different CPT positions of an offshore wind farm and constructing a training sample comprises the following steps:

[0074] Collect geological, geophysical and geotechnical data for the offshore wind farm from designated data sources, including cone tip resistance and pore water pressure data obtained from static cone penetration tests at multiple CPT locations, and in situ G measured by seismic static cone penetration tests (SCPTs) or other methods. max value;

[0075] The effective vertical stress is calculated based on the geological conditions of the site and the measurement depth. Specifically, the effective vertical stress corresponding to each data point is calculated by multiplying the depth below the seabed by the average effective stress gradient;

[0076] G maxThe values ​​are converted from the original measurement units into MPa, and are classified according to the CPT positions and soil units, a cross-classification data structure is constructed, the number of samples in each classification group is counted, a data summary is formed, and the training samples are constructed.

[0077] In some embodiments, the step of constructing a complete mixed model in the step of constructing a Bayesian regression model according to the training samples comprises the following steps:

[0078] Based on empirical formula Construct a complete mixed regression model, where q t represents the total cone tip resistance of the static penetration test (CPT) after correction of pore pressure; p a represents atmospheric pressure; σ′ v represents the effective vertical stress; α0, α1, α2 represent the three unknown parameters of the model;

[0079] Setting prior distributions for model parameters, using weak information priors to reflect the lack of specific prior knowledge;

[0080] A preliminary estimate of the possible range of values ​​for a parameter is represented by a prior distribution.

[0081] In some embodiments, the step of constructing a partial mixed model in the step of constructing a Bayesian regression model according to the training samples includes the following steps:

[0082] Considering the variation between CPT locations and soil units, a cross-classification hierarchical model was constructed based on the set exchangeability intercept.

[0083] In some embodiments, the Markov chain Monte Carlo (MCMC) simulation is performed on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, comprising the following steps:

[0084] The constructed Bayesian regression model was fitted using the probabilistic programming language Stan combined with the Markov Chain Monte Carlo simulation method;

[0085] Run multiple Markov chains in parallel, and each chain performs a corresponding number of simulated samplings to obtain the posterior distribution of model parameters;

[0086] Among them, during the simulation process, the posterior distribution of the parameters is judged whether it converges by checking the statistical indicators within and between chains. If the posterior distribution does not converge, the simulation parameters are adjusted and the simulation is re-run until a stable posterior distribution is obtained; among them, the simulation parameters include increasing the number of iterations and adjusting the prior distribution.

[0087] In some embodiments, the fitting effect of the evaluation model is obtained to obtain a target model that meets the prediction effect requirements, including the following steps:

[0088] Analyze the posterior distribution of model parameters, calculate the statistical indicators of the posterior mean, standard deviation, and 95% confidence interval, and evaluate the model's fit to the data;

[0089] Calculate the residual standard deviation τ ε , by comparing the τ of the fully mixed model and the cross-classification hierarchical model ε The value determines the goodness of fit of the model;

[0090] Calculate the intraclass correlation coefficient, the formula is The degree of similarity between data groups is measured by the intraclass correlation coefficient; wherein ICC represents the intraclass correlation coefficient; represents the variance of the local group (such as position j) (representing G max variance among different CPT positions); represents the variance of the parent group (such as unit k) (G max variance between different soil elements); Represents G max The variance of the residual;

[0091] The leave-one-out cross-validation method was used to compare the expected log-prediction density of different models and evaluate the prediction accuracy of the models;

[0092] Based on the values ​​of the predicted densities and their differences and standard errors, the target model with the best prediction performance is selected.

[0093] In some embodiments, the target model is divided into multiple scene predictions G max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, and the probability prediction is completed, including the following steps:

[0094] According to the CPT position and whether the soil unit is in the calibration data set, G is predicted in the following scenarios. max profile:

[0095] Existing locations and existing units with observed data: Use the estimated Calculate the group-specific β0 and substitute it into the model formula (3a) to predict G max profile; where β 0(jk) represents the intercept between CPT location j and soil unit k; represents the population mean of the intercept β0; represents the local deviation of position j; represents the local deviation of unit k; β0 represents the overall basic intercept;

[0096] Existing locations and existing units but no observations: Use β estimated from calibration data 0(jk) Calculate β0 for prediction;

[0097] Existing locations and new units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates;

[0098] New Locations and Existing Units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates;

[0099] New location and new unit: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates;

[0100] A leave-one-group LOGO analysis was performed, where the data of one soil unit at one CPT position were kept as the validation set each time, the remaining data were used to fit the model and make a priori predictions, and then the validation data were added to make a posteriori predictions; the G of the a priori and posteriori predictions were compared. max By comparing the changes in these profiles, we can analyze the degree of improvement of the target model after obtaining new data.

[0101] The following describes the specific implementation process of the embodiment of the present invention in detail by taking a specific application scenario as an example:

[0102] In view of the problems existing in the prior art, the present invention relates to a soil profile prediction method for offshore wind farm foundation design, in particular to a small strain shear modulus G based on static penetration test (CPT) data. max As the scale of offshore wind farms expands, CPT and its empirical statistical models are often relied upon to derive soil profiles. However, traditional statistical modeling methods fail to fully utilize the widespread increase in G max The amount of data is too large, and the differences between CPT positions and soil units are ignored, which may lead to misleading probability predictions. This paper proposes a Bayesian cross-classification hierarchical model that takes the differences between CPT positions and soil units into consideration, thereby achieving G max "High-resolution" prediction of profiles. The actual case application verifies the significant improvement of the model in terms of probability prediction and profile rationality, and confirms the effective posterior update and uncertainty reduction after new data observation through leave-one-group analysis.

[0103] Specifically, a method for calculating the small strain shear modulus G of soil based on CPT data is proposed. max The method for constructing a Bayesian cross-classification hierarchical model for high-resolution prediction includes the following steps:

[0104] 1. Obtain soil small strain shear modulus data at different CPT positions in offshore wind farms:

[0105] Collect geological, geophysical and geotechnical data from designated data sources for offshore wind farms (e.g. Ten noorden van de Waddeneilanden (TNW) wind farm in the Dutch North Sea), including cone tip resistance q obtained from static penetration tests at multiple CPT locations (at least 162 CPT locations) t , pore water pressure data, and in-situ G measured by seismic static penetration test (SCPT) or other related methods. max value.

[0106] Calculate the effective vertical stress σ′ based on the site geological conditions and measurement depth v Specifically, the effective vertical stress corresponding to each data point is calculated by multiplying the depth below the seabed by the average effective stress gradient (for example, the average effective stress gradient in the North Sea is 10 kPa / m).

[0107] G max The value is converted from the original measurement unit to MPa using the conversion formula G max =ρV s 2 (where ρ is soil density, determined according to soil type and actual measurement, V s is shear wave velocity), and is classified according to CPT position and soil unit, a cross-classification data structure is constructed, and the number of samples in each classification group (combination of CPT position and soil unit) is counted to form a data summary to ensure the integrity and accuracy of the data for subsequent analysis.

[0108] 2. Bayesian regression model construction:

[0109] 1) Completely mixed model construction

[0110] Based on empirical formula The fully mixed regression model is constructed and expressed in natural logarithmic scale as

[0111]

[0112] Where i=1,…,N represents the data point number. is the regression coefficient to be estimated, ε i is the random error term, p aAtmospheric pressure (100 kPa) is used for normalization to avoid unit inconsistency issues.

[0113] E represents ln(G max(i) )’s expected value; q t(i) represents the cone tip resistance of the ith data point; β0 represents the intercept of the regression model; represent The regression coefficient of represent The regression coefficient of .

[0114] Weak information priors are used to set the prior distribution of model parameters to reflect the lack of specific prior knowledge, specifically: β0~Normal(0,5 2 )(2a)

[0115]

[0116] τ ε ~Half-normal(0,1 2 )(2d)

[0117] Among them, τ ε represents the standard deviation of the error term in the model; these prior distributions reflect a preliminary estimate of the possible range of parameter values ​​in the absence of specific prior knowledge, while avoiding the problems that may arise from traditional fuzzy / flat priors (such as biased posterior results when the data is complex or the model is complex).

[0118] 2) Partially mixed (cross-classification and stratification) model construction:

[0119] A cross-classification hierarchical model was constructed to consider the variation between CPT locations and soil units, assuming an exchangeability intercept β0 and other parameters. The model formula is the same between different CPT positions and soil units.

[0120]

[0121] Among them, G max(ijk) represents the i-th (i=0,…,N k-th) group in the cross-classification of location j and unit jk ) observation value; β 0(jk) is the intercept corresponding to site j and unit k, and its population distribution is specified as a common normal prior distribution reflecting the exchangeability assumption.

[0122]

[0123] Among them, ε ijkrepresents the random error term for the ith data point (position j, soil unit k); j = 1,…, J represents the CPT position, and k = 1,…, K represents the soil unit.

[0124] Set prior distribution for hyperparameters:

[0125]

[0126] in, The prior distribution of is the same as that of the complete mixture model.

[0127] 3. Model fitting and analysis:

[0128] The constructed model is fitted using the probabilistic programming language Stan combined with the Markov chain Monte Carlo simulation method No-U-TurnSampler, and multiple Markov chains (at least three) are run in parallel. Each chain performs a large number of simulation sampling (such as 5000 times) to obtain the posterior distribution of the model parameters. During the simulation process, the convergence of the chain is closely monitored, and the posterior distribution of the parameters is judged by checking the statistical indicators within and between chains (such as Rhat values). If the Rhat value is close to 1 (usually within 1.05), the model is considered to be converged; otherwise, the simulation parameters need to be adjusted (such as increasing the number of iterations, adjusting the prior distribution, etc.), and the simulation is re-run until a stable posterior distribution is obtained.

[0129] Analyze the posterior distribution of model parameters, calculate the posterior mean, standard deviation, 95% confidence interval (CIs) and other statistical indicators, and evaluate the model's fit to the data. Calculate the residual standard deviation τ ε , by comparing the τ of the fully mixed model and the cross-classification hierarchical model ε The value of τ can be used to judge the goodness of fit of the model. ε A value of usually indicates that the model fits the data better. At the same time, the intraclass correlation coefficient (ICC) is calculated, and the formula is To measure the degree of similarity between data groups and further evaluate the rationality of the model assumptions, the closer the ICC value is to 0, the more similar the groups are, and the closer it is to 1, the greater the difference between the groups.

[0130] The leave-one-out cross-validation (LOO-CV) method was used to compare the expected logarithmic prediction density (elpd) of different models and evaluate the prediction accuracy of the models. The larger the elpd value, the stronger the prediction ability of the model outside the sample. The model with the best prediction performance was selected based on the elpd value and its difference (Δelpd) and standard error (SE). When the absolute value of Δelpd was greater than several times (such as more than 3 times) of its standard error, it was considered that there was a significant difference between the models; otherwise, the difference was not significant.

[0131] 4. G maxProfile prediction and update:

[0132] According to the CPT position and whether the soil unit is in the calibration data set, G is predicted in the following scenarios. max profile:

[0133] Existing locations and existing units with observed data: Directly use the estimated Calculate the group-specific β0 and substitute it into the model formula (3a) to predict G max profile.

[0134] Existing locations and existing units but no observations: Also use the β estimated from the calibration data 0(jk) β0 is calculated for prediction since the position and unit effects have been estimated in the calibration data.

[0135] Existing locations and new units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates.

[0136] New Locations and Existing Units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates.

[0137] New location and new unit: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates.

[0138] A leave-one-group (LOGO) analysis was performed, where the data of one soil unit at one CPT position were kept as the validation set each time, the remaining data were used to fit the model and make a priori predictions, and then the validation data were added to make a posteriori predictions. max The changes in the representative profile (i.e., AE profile) are analyzed to update the model after observing new data to evaluate the model's fusion performance of the new data and the improvement effect of the prediction of the representative profile.

[0139] The Bayesian cross-classification hierarchical model employed leave-one-group cross-validation analysis to calibrate model parameters to provide reliable a priori predictions when site data were insufficient.

[0140] The Bayesian model is used to estimate the G of future unobserved soil units and CPT locations.max The values ​​are predicted probabilistically and representative profiles are generated.

[0141] Through the a posteriori update of the model, G is further optimized after obtaining new CPT or soil layer data. max prediction accuracy.

[0142] In summary, the present invention can more effectively utilize CPT data by using the Bayesian cross-classification hierarchical model and consider the potential differences of different CPT locations and soil layer units to provide the small strain shear modulus G of the soil layer in the offshore wind farm. max This method generates more accurate soil profile representations, improving accuracy and reliability when designing wind turbine foundations.

[0143] The following takes the TNW offshore wind farm geotechnical survey public dataset as an example to describe the implementation process of the present invention in detail. It should be noted that the present invention introduces a general method and does not rely on the specific wind farm TNW dataset in the example.

[0144] 1. Data preparation stage:

[0145] 1) Data collection:

[0146] refer to Figure 3 , Figure 3 Overview of the TNW wind farm with CPT borehole locations: The geographical location of the wind farm, the CPT borehole locations and the SCPT borehole annotations are shown, illustrating the spatial distribution of data collection. Geological, geophysical and geotechnical data for the Tennoor denvande Waddeneilanden (TNW) offshore wind farm in the Dutch North Sea were obtained from public data sources. The data included detailed measurement records at 162 CPT locations, where 14 sand units and 7 clay units were identified, forming a total of 896 cross-classification groups. At the same time, 407 pairs of SCPT and in situ G max The values ​​were used as a calibration dataset for the CPT-based empirical model, which covers 125 cross-classification groups involving 13 sand units and 6 clay units from 31 CPT locations.

[0147] 2) Data preprocessing:

[0148] Calculate the effective vertical stress σ′ v , according to the formula σ′ v = depth below the seafloor × 10 kPa / m (based on the average effective stress gradient in the North Sea) to calculate the effective vertical stress corresponding to each data point. max The value is obtained from the in-situ shear wave velocity V s To convert, use the formula (where ρ is the soil density, which is determined according to the soil type and actual measurement), and G max The values ​​were converted to MPa units to ensure the consistency and accuracy of the data in subsequent calculations.

[0149] The data were sorted according to CPT positions and soil units to construct a cross-classification data structure. The number of samples for each CPT position, soil unit, and cross-classification group was counted to form a sample quantity summary table 1 in order to clearly understand the distribution of the data and provide a basis for subsequent model construction and analysis.

[0150] Table 1. Sample sizes for different combinations of CPT locations and soil units in the TNW dataset

[0151]

[0152] 2. Model building and fitting stage

[0153] refer to Figure 4 , Figure 4 is a directed acyclic graph of Bayesian regression models, including cross-classification hierarchical models (such as Figure 4 (a)), the position classification partial hybrid model (as shown in Figure 4 (b)) and the diagram of the mixed model of unit classification (as shown in Figure 4 (c) shows the relationship between the model structure and parameters.

[0154] 1) Complete hybrid model construction and programming implementation

[0155] Based on the assumption of the complete mixed model, a regression model is constructed based on the empirical formula. In a probabilistic programming environment (such as Stan), the model formula is converted into executable code.

[0156] 2) Construction and programming of cross-classification hierarchical model

[0157] Construct a cross-classification hierarchical model, consider the variation between groups, assume an exchangeable intercept β0, and other parameters The code was written in Stan based on the partial mixed model formulation which is the same across different CPT locations and soil elements.

[0158] 3) Model fitting and convergence evaluation

[0159] Run MCMC simulations and use the No-U-TurnSampler in Stan for parameter estimation. Run three Markov chains in parallel, each containing 5000 simulated posterior samples. During the simulation, closely monitor the convergence of the chains and check the statistical indicators (such as Rhat values) within and between chains to determine whether the posterior distribution of the parameters has converged. If the Rhat value is close to 1 (usually within 1.05, it can be considered that the convergence is good), it means that the model has converged; otherwise, it is necessary to adjust the simulation parameters (such as increasing the number of iterations, adjusting the prior distribution, etc.) and rerun the simulation until a stable posterior distribution is obtained.

[0160] like Figure 5 As shown, Figure 5 The posterior estimation diagrams of CPT position and soil unit effect are shown as follows: The CPT position and soil unit effect of TNW sand data and clay data are shown as follows: and ) are used to analyze the location and unit-to-unit variations and their uncertainties, where Figure 5 (a) shows the data of sandy soil. Figure 5 (b) shows clay.

[0161] Figure 6 is a comparison plot of β0 estimates: β0 estimates from the complete mixing model and the partial mixing model are compared for specific sand groups (such as the groups associated with 083-BH-CPT and 072-BH-CPT), highlighting the inadequacy of the complete mixing model in accounting for variation.

[0162] Figure 7 The posterior standardized residual plot shows the distribution of the posterior standardized residuals of the complete and partial mixture models on the TNW dataset, which is used to preliminarily check the model fit and determine the degree of deviation between the data points and the model predictions.

[0163] Figure 8 LOGO analysis result graph: presents the measurement q of multiple selected examples (such as different positions and unit combinations) t Representative G of profiles, prior and posterior predictions max The profiles, as well as the relative changes of profiles A and B, intuitively demonstrate the prediction and updating capabilities of the model. Figure 8 (a) is the new clay unit GGM31B at the new position 081-SCPT; Figure 8 Middle (b) is the existing clay unit GGM52 at the new location 016-BH-CPT; Figure 8 Middle (c) is the existing sand unit GGM42 at the existing location 072-BH-CPT; Figure 8 Middle (d) is the existing sand unit GGM31A at the existing location 034-SCPT.

[0164] Fig. 9 This is an example of the final prediction of the entire wind farm: Taking the existing location 034-SCPT and the new location 006-PCPT as examples, the predicted G of all soil elements based on CPT measurement data are shown. max Cross-section, covering different prediction scenarios, illustrating the application effect of the model on the entire wind farm.

[0165] 3. Model evaluation and comparison stage:

[0166] 1) Posterior distribution analysis:

[0167] Calculate the posterior mean, 95% confidence intervals (CIs), and residual standard deviation τ of the model parameters ε , intraclass correlation coefficient (ICC) and other indicators. The posterior mean can be used as the best estimate of the parameter, and the 95% confidence interval reflects the uncertainty range of the parameter estimate. By analyzing these indicators, the model's fitting effect and explanatory power for the data are evaluated. For example, a smaller residual standard deviation τ ε It usually indicates that the model fits the data well; the ICC value can be used to measure the degree of similarity between data groups and help determine whether the model assumptions (such as a complete mixed model or a partial mixed model) are reasonable.

[0168] 2) Model comparison and selection

[0169] The leave-one-out cross validation (LOO-CV) method was used to calculate the expected logarithmic prediction density (elpd) of the complete mixed model, the cross-classification hierarchical model, and other alternative models (such as the position classification and unit classification partial mixed model). The larger the elpd value, the higher the prediction accuracy of the model. The difference (Δelpd) and its standard error (SE) between the elpd value of each model and the elpd value of the best model (the cross-classification hierarchical model in the present invention) were calculated, and the advantages and disadvantages of the models were judged according to the size of the difference and the standard error. If the absolute value of Δelpd is greater than several times its standard error (such as more than 3 times), it can be considered that there are significant differences between the models; otherwise, the difference is not significant. According to the comparison results, the model with the highest prediction accuracy is selected as the final model for subsequent G max Profile prediction.

[0170] 4.G max Profile prediction and update phase

[0171] 1) β0 calculation under different prediction scenarios

[0172] Different prediction scenarios are determined depending on whether the CPT location and soil unit are in the calibration data set. For existing locations and existing units with observed data (prediction scenario 1), the estimated Calculate group-specific β0. For existing locations and existing units but no observed data (forecast scenario 2), β0 is calculated using the same formula as above, since the location and unit effects have been estimated in the calibration data. For existing locations and new units (forecast scenario 3), β0 is calculated by For new locations and existing units (prediction scenario 4), β0 is obtained by sampling from For new locations and new units (prediction scenario 5), β0 is sampled from Medium sampling.

[0173] 2) G max Profile prediction and uncertainty analysis

[0174] Using the calculated β0 value, the G of each CPT position and soil unit is predicted according to the constructed model formula. max For each predicted profile, the statistical indicators such as the mean, 95% confidence interval and 90% prediction interval are calculated to reflect the uncertainty of the prediction results. max The profile variation trend is used to evaluate the impact of the spatial variability of soil properties on the design of wind turbine foundations. For example, a larger prediction interval indicates that the G max There is a large uncertainty in the value, and more conservative design parameters need to be considered when designing the wind turbine foundation.

[0175] 3) Leave one group (LOGO) for analysis

[0176] Perform LOGO analysis, select a soil unit at a CPT position from the original data set each time as a validation set, and keep it to observe the prediction ability of the model before and after the new data. Use the remaining data to fit the model, and make a priori prediction on the data reserved for verification to obtain the G of the prior prediction. max Then add the validation data to the model, make a posteriori prediction, and calculate the G of the posteriori prediction again. max Profiles and their related statistical indicators. Compare the changes in the prior prediction and the posterior prediction profiles, and analyze the updates of the model after observing new data, such as the relative changes of profiles A and B (δ A and δ B ) to reflect the uncertainty reduction effect of the model and its adaptability to new data. Through multiple LOGO analysis examples with different positions and unit combinations, the model's prediction and update capabilities in different situations are verified to ensure that the model has good generalization and reliability.

[0177] 5. Practical application stage

[0178] 1) Full wind farm G max Profile prediction

[0179] The final selected cross-classification hierarchical model was applied to the entire TNW wind farm, and G was performed on all CPT locations and soil elements that were not included in the calibration dataset. max Profile prediction. According to different prediction scenarios, the β0 value of each group is calculated according to the above calculation method, and the corresponding G max Generate G of the entire wind farm max Profile prediction diagram, intuitively showing the G at different locations and soil units max The distribution of wind turbines provides comprehensive data support for the design of offshore wind turbine foundations.

[0180] 2) Design optimization and decision support

[0181] According to the predicted G max The prediction results can also be used to evaluate the bearing capacity and deformation characteristics of wind turbine foundations at different locations, providing a basis for decision-making in the layout planning and operation management of wind farms. For example, for G max In areas with lower values ​​or greater variability, additional reinforcement measures can be taken or the layout of wind turbines can be adjusted to reduce risks and improve the overall performance of the wind farm.

[0182] 3) Model update and continuous improvement

[0183] During the construction and operation of the wind farm, as new data is continuously acquired (such as new CPT test data or long-term monitoring data), the model can be updated regularly. Incorporating new data into the model calibration process, re-evaluating model parameters and prediction results, and further optimizing G max At the same time, based on actual engineering experience and feedback, the model's assumptions and structure are continuously improved to improve the model's adaptability to complex marine geological conditions, ensuring that the model can always provide reliable technical support for the design and operation of offshore wind farms.

[0184] The final selected model was applied to the entire TNW wind farm, and G was performed for all CPT locations and soil elements not included in the calibration dataset. max Profile prediction provides a basis for offshore wind turbine foundation design, ensuring that the design takes into account the spatial variability of soil parameters and improves the reliability and safety of the design. At the same time, the model and prediction results can be further optimized based on actual project feedback and the acquisition of new data.

[0185] Compared with the prior art, the present invention has the following advantages:

[0186] 1. The Bayesian cross-classification hierarchical model of the present invention can effectively capture the changes between CPT positions and soil units. Compared with the traditional complete mixing model, it can provide more reasonable parameter uncertainty estimation and better fit the data.

[0187] 2. Through LOO-CV analysis, the model is verified in G max The out-of-sample prediction accuracy of the wind turbine foundation of offshore wind farms has been significantly improved, which can provide more reliable G max The representative profile prediction will help to design wind turbine foundations more reasonably and improve the safety and economy of offshore wind farms.

[0188] Another aspect of the embodiment of the present invention further provides a Bayesian hierarchical probability prediction system for small strain shear modulus of an offshore wind farm based on CPT, comprising:

[0189] The first module is used to obtain soil small strain shear modulus data at different CPT positions in offshore wind farms and construct training samples;

[0190] The second module is used to construct a Bayesian regression model according to the training samples, including constructing a complete mixed model and a partial mixed model;

[0191] The third module is used to perform Markov Chain Monte Carlo (MCMC) simulation on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, and then evaluate the fitting effect of the model to obtain the target model that meets the prediction effect requirements;

[0192] The fourth module is used to divide multiple scenes into prediction G according to the target model max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, completing the probabilistic prediction.

[0193] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0194] The embodiment of the present invention further provides an electronic device, the electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the above-mentioned Bayesian hierarchical probability prediction method of small strain shear modulus of offshore wind farm based on CPT when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0195] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0196] An embodiment of the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT.

[0197] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0198] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0199] It should be noted that in various specific embodiments of the present invention, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present invention needs to obtain the user's sensitive personal information, it will obtain the user's separate permission or consent through a pop-up window or jump to a confirmation page, and after clearly obtaining the user's separate permission or consent, it will obtain the necessary user-related data for the normal operation of the embodiment of the present invention.

[0200] The embodiments described in the embodiments of the present invention are intended to more clearly illustrate the technical solutions of the embodiments of the present invention, and do not constitute a limitation on the technical solutions provided by the embodiments of the present invention. Those skilled in the art can appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present invention are also applicable to similar technical problems.

[0201] Those skilled in the art will appreciate that the technical solutions shown in the figures do not limit the embodiments of the present invention and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0202] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0203] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0204] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0205] It should be understood that in the present invention, "at least one (item)" refers to one or more, and "plurality" refers to two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can represent: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0206] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0207] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0208] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0209] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including multiple instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0210] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the embodiments of the present invention is not limited thereby. Any modification, equivalent substitution and improvement made by those skilled in the art without departing from the scope and essence of the embodiments of the present invention shall be within the scope of the rights of the embodiments of the present invention.

Claims

1. A Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT, characterized in that: The following steps are involved: Obtain soil small strain shear modulus data at different CPT positions in offshore wind farms and construct training samples; Constructing a Bayesian regression model based on the training samples, including constructing a complete mixed model and a partial mixed model; Perform Markov Chain Monte Carlo (MCMC) simulation on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, and then evaluate the fitting effect of the model to obtain the target model that meets the prediction effect requirements; According to the target model, multiple scenarios are divided into prediction G max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, completing the probabilistic prediction.

2. According to claim 1, a Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farm based on CPT is characterized in that: The method of obtaining small strain shear modulus data of each soil unit at different CPT positions of the offshore wind farm and constructing a training sample comprises the following steps: Collect geological, geophysical and geotechnical data for the offshore wind farm from designated data sources, including cone tip resistance and pore water pressure data obtained from static cone penetration tests at multiple CPT locations, and in situ G measured by seismic static cone penetration tests (SCPTs) or other methods. max value; The effective vertical stress is calculated based on the geological conditions of the site and the measurement depth. Specifically, the effective vertical stress corresponding to each data point is calculated by multiplying the depth below the seabed by the average effective stress gradient; G max The values ​​are converted from the original measurement units into MPa, and are classified according to the CPT positions and soil units, a cross-classification data structure is constructed, the number of samples in each classification group is counted, a data summary is formed, and the training samples are constructed.

3. The Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT according to claim 1 is characterized in that: The step of constructing a complete mixed model in the step of constructing a Bayesian regression model according to the training samples comprises the following steps: Based on empirical formula Construct a complete mixed regression model, where q t represents the total cone tip resistance of the static penetration test (CPT) after correction of pore pressure; p a represents atmospheric pressure; σ v ′ represents the effective vertical stress; α0, α1, α2 represent the three unknown parameters of the model; Setting prior distributions for model parameters, using weak information priors to reflect the lack of specific prior knowledge; A preliminary estimate of the possible range of values ​​for a parameter is represented by a prior distribution.

4. The Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT according to claim 1 is characterized in that: The step of constructing a partial mixed model in the step of constructing a Bayesian regression model according to the training samples comprises the following steps: Considering the variation between CPT locations and soil units, a cross-classification hierarchical model was constructed based on the set exchangeability intercept.

5. The Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT according to claim 1 is characterized in that: The Markov Chain Monte Carlo (MCMC) simulation is performed on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, including the following steps: The constructed Bayesian regression model was fitted using the probabilistic programming language Stan combined with the Markov Chain Monte Carlo simulation method; Run multiple Markov chains in parallel, and each chain performs a corresponding number of simulated samplings to obtain the posterior distribution of model parameters; Among them, during the simulation process, the posterior distribution of the parameters is judged whether it converges by checking the statistical indicators within and between chains. If the posterior distribution does not converge, the simulation parameters are adjusted and the simulation is re-run until a stable posterior distribution is obtained; among them, the simulation parameters include increasing the number of iterations and adjusting the prior distribution.

6. The Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT according to claim 5 is characterized in that: The fitting effect of the evaluation model is evaluated to obtain a target model that meets the prediction effect requirements, including the following steps: Analyze the posterior distribution of model parameters, calculate the statistical indicators of the posterior mean, standard deviation, and 95% confidence interval, and evaluate the model's fit to the data; Calculate the residual standard deviation τ ε , by comparing the τ of the fully mixed model and the cross-classification hierarchical model ε The value determines the goodness of fit of the model; Calculate the intraclass correlation coefficient, the formula is The degree of similarity between data groups is measured by the intraclass correlation coefficient; wherein ICC represents the intraclass correlation coefficient; Represents the variance of the local group, also represents G max variance among different CPT positions; Represents the variance of the parent group, and also represents G max Variance between different soil elements; Represents G max The variance of the residual; The leave-one-out cross-validation method was used to compare the expected log-prediction density of different models and evaluate the prediction accuracy of the models; Based on the values ​​of the predicted densities and their differences and standard errors, the target model with the best prediction performance is selected.

7. The Bayesian hierarchical probability prediction method for small strain shear modulus of offshore wind farms based on CPT according to claim 1 is characterized in that: According to the target model, multiple scene predictions G are divided max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, and the probability prediction is completed, including the following steps: According to the CPT position and whether the soil unit is in the calibration data set, G is predicted in the following scenarios. max profile: Existing locations and existing units with observed data: Use the estimated Calculate group-specific β0 and predict G max profile; where β 0(jk) represents the intercept between CPT location j and soil unit k; represents the population mean of the intercept β0; represents the local deviation of position j; represents the local deviation of unit k; β0 represents the overall basic intercept; Existing locations and existing units but no observations: Use β estimated from calibration data 0(jk) Calculate β0 for prediction; Existing locations and new units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates; New Locations and Existing Units: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates; New location and new unit: From Sampling β0 is obtained, and then substituting it into the model formula to predict G max Section, where and are the corresponding hyperparameters and parameter estimates; A leave-one-group LOGO analysis was performed. Each time, the data in a soil unit at a certain CPT position was reserved as a validation set. The remaining data was used to fit the model and perform a priori prediction. Then, the validation data was added to perform a posteriori prediction. The G of the a priori and posteriori predictions were compared. max By comparing the changes in these profiles, we can analyze the degree of improvement of the target model after obtaining new data.

8. A Bayesian hierarchical probability prediction system for small strain shear modulus of offshore wind farms based on CPT, characterized in that: include: The first module is used to obtain soil small strain shear modulus data at different CPT positions in offshore wind farms and construct training samples; The second module is used to construct a Bayesian regression model according to the training samples, including constructing a complete mixed model and a partial mixed model; The third module is used to perform Markov Chain Monte Carlo (MCMC) simulation on the constructed Bayesian regression model to obtain the posterior distribution of the model parameters, and then evaluate the fitting effect of the model to obtain the target model that meets the prediction effect requirements; The fourth module is used to divide multiple scenes into prediction G according to the target model max The profile is then used to determine whether the CPT position and soil unit are in the calibration data set, completing the probabilistic prediction.

9. An electronic device, characterized in that: including a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a program, and the program is executed by a processor to implement the method according to any one of claims 1 to 7.

Images