A semi-submersible platform air gap extreme value prediction system and method based on numerical simulation

Abstract

The application provides a semi-submersible platform air gap extreme value prediction system and method based on numerical simulation, relates to the technical field of numerical simulation, and obtains an air gap response sequence of a semi-submersible platform in a target water area through numerical simulation; extracts a low-frequency response from the air gap response sequence, carries out nonlinear correction on the low-frequency response based on a data-driven model, and obtains a corrected low-frequency response sequence; reconstructs the corrected low-frequency response sequence and a wave frequency response to obtain a composite air gap response sequence; carries out non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain an air gap extreme value prediction result; and then obtains a test sequence of the air gap of the semi-submersible platform changing with time, and iteratively optimizes the data-driven model based on the difference between the test sequence and the air gap extreme value prediction result, so that the application can realize data-driven accurate correction of the low-frequency response in numerical simulation, and improve the air gap extreme value prediction accuracy of the semi-submersible platform.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of numerical simulation, more particularly, the present application relates to a semi-submersible platform air gap extreme value prediction system and method based on numerical simulation. BACKGROUND

[0002] Numerical simulation technology is the core means of current semi-submersible platform performance evaluation and safety prediction. By establishing a hydrodynamic model of the semi-submersible platform and solving the fluid dynamics control equation, the motion and load response of the platform in complex water areas are simulated to provide key data support for design optimization and safe operation. In the design and operation safety analysis of semi-submersible platforms, air gap extreme value prediction is crucial. A small air gap can cause severe impact of waves on the deck, seriously threatening the structural safety of the semi-submersible platform and the integrity of personnel and equipment.

[0003] Developing a high-precision air gap extreme value prediction is of great significance to ensuring the safety of the platform throughout its life cycle. Currently, numerical simulation based on the frequency domain or time domain is the mainstream method for air gap extreme value prediction. By calculating the relative relationship between the platform motion and wave elevation caused by waves, the statistical characteristics and extreme values of the air gap are predicted. However, the existing prediction methods still have obvious technical limitations. When dealing with low-frequency motion caused by second-order difference frequency forces, due to the simplification of complex physical phenomena such as viscous damping and vortex-induced effects, there is a systematic and inherent bias in the prediction of low-frequency response that cannot be completely eliminated by adjusting parameters. Moreover, there is a lack of understanding of the main sources of error. Errors in simulation are directly passed on and amplified in the final extreme value prediction results, making it difficult to effectively feedback test data to the numerical model to achieve automatic calibration of model parameters and continuous iterative optimization of prediction capability. Therefore, how to achieve data-driven precise correction of low-frequency response in numerical simulation to improve the accuracy of air gap extreme value prediction of semi-submersible platforms is a difficult problem faced by the industry. SUMMARY

[0004] The present application provides a semi-submersible platform air gap extreme value prediction system and method based on numerical simulation, which can achieve data-driven precise correction of low-frequency response in numerical simulation to improve the accuracy of air gap extreme value prediction of semi-submersible platforms.

[0005] In a first aspect, the present application provides a semi-submersible platform air gap extreme value prediction method based on numerical simulation, the prediction method comprising the following steps:

[0006] Obtaining an air gap response sequence of the semi-submersible platform in a target water area through numerical simulation, the air gap response sequence including wave frequency response and low-frequency response;

[0007] Extracting the low-frequency response from the air gap response sequence, and performing nonlinear correction on the low-frequency response based on a data-driven model to obtain a corrected low-frequency response sequence.

[0008] reconstructing the modified low-frequency response sequence and the wave-frequency response to obtain a composite air-gap response sequence;

[0009] performing non-Gaussian extreme value statistics based on the composite air-gap response sequence to obtain an air-gap extreme value prediction result of the target water area;

[0010] obtaining a test sequence of air gaps of the semi-submersible platform changing over time through a wave pool test, and iteratively optimizing the data-driven model based on a difference between the test sequence and the air-gap extreme value prediction result.

[0011] In this embodiment, obtaining the air-gap response sequence of the semi-submersible platform in the target water area through numerical simulation specifically includes:

[0012] constructing a transfer function of the semi-submersible platform motion and wave load based on structural parameters of the semi-submersible platform and wave parameters of the target water area;

[0013] determining the wave frequency motion response and the wave load through the transfer function;

[0014] performing coupling analysis on the wave frequency motion response and the wave load to obtain the air-gap response sequence.

[0015] In this embodiment, the low-frequency response is extracted from the air-gap response sequence based on a high-pass filter.

[0016] In this embodiment, the data-driven model is a machine learning model based on historical data, which is used to learn and correct inherent low-frequency response deviations in the numerical simulation through modal features of the low-frequency response.

[0017] In this embodiment, the data-driven model is a machine learning model based on historical data, which is used to learn and correct inherent low-frequency response deviations in the numerical simulation through modal features of the low-frequency response.

[0018] performing modal decomposition on the low-frequency response to obtain a plurality of proper orthogonal function components;

[0019] constructing a training set according to all the proper orthogonal function components;

[0020] inputting all the proper orthogonal function components into the data-driven model trained based on the training set for signal reconstruction to obtain the modified low-frequency response sequence.

[0021] In this embodiment, reconstructing the modified low-frequency response sequence and the wave-frequency response to obtain a composite air-gap response sequence specifically includes:

[0022] linearly superimposing the modified low-frequency response sequence and the wave-frequency response in the time domain to obtain an initial reconstruction sequence;

[0023] Performing phase consistency inspection and amplitude normalization processing on the initial reconstruction sequence to obtain a composite air gap response sequence.

[0024] In this embodiment, based on the composite air gap response sequence, non-Gaussian extreme value statistics are performed to obtain an air gap extreme value prediction result of the target water area, specifically including:

[0025] Performing high-order statistical analysis on the composite air gap response sequence to obtain skewness index and kurtosis index;

[0026] Based on the skewness index and the kurtosis index, a polynomial transformation model based on non-Gaussian characteristics is constructed;

[0027] Monte Carlo simulation is performed on the polynomial transformation model to generate a synthetic sequence, and then peak value threshold sampling is performed on the synthetic sequence to obtain extreme value samples;

[0028] The extreme value samples are fitted using a generalized Pareto distribution to obtain an extreme value distribution function, and then a characteristic value corresponding to a preset return period is extracted from the extreme value distribution function to obtain an air gap extreme value prediction result of the target water area.

[0029] In this embodiment, the test sequence of the air gap of the semi-submersible platform changing with time is obtained through a wave pool test, specifically including:

[0030] A physical model of the semi-submersible platform is constructed;

[0031] A wave field is generated according to the wave parameters of the target water area;

[0032] The physical model is placed in the wave field, and the change of the wave surface elevation around the semi-submersible platform is measured by a wave height sensor, and then the test sequence of the air gap of the semi-submersible platform changing with time is obtained.

[0033] In this embodiment, the data-driven model is iteratively optimized based on the difference between the test sequence and the air gap extreme value prediction result, specifically including:

[0034] A response residual sequence is determined through the test sequence and the air gap extreme value prediction result;

[0035] A loss function is constructed based on the response residual sequence, and the parameter gradient of the data-driven model is calculated through a gradient descent algorithm;

[0036] The internal weight parameters of the data-driven model are iteratively updated according to the parameter gradient until the loss function converges, and the iterative optimization of the data-driven model is completed.

[0037] In a second aspect, the application provides a numerical simulation-based semi-submersible platform air gap extreme value prediction system for performing a numerical simulation-based semi-submersible platform air gap extreme value prediction method, the prediction system comprising:

[0038] a numerical simulation module configured to obtain, through numerical simulation, an air gap response sequence of a semi-submersible platform in a target water area, the air gap response sequence comprising a wave frequency response and a low frequency response;

[0039] a low frequency correction module configured to extract the low frequency response from the air gap response sequence, perform nonlinear correction on the low frequency response based on a data-driven model, and obtain a corrected low frequency response sequence;

[0040] a response reconstruction module configured to reconstruct the corrected low frequency response sequence and the wave frequency response to obtain a composite air gap response sequence;

[0041] an extreme value prediction module configured to perform non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain an air gap extreme value prediction result of the target water area;

[0042] a model optimization module configured to obtain, through a wave pool test, a test sequence of air gap changes of the semi-submersible platform over time, and perform iterative optimization on the data-driven model based on a difference between the test sequence and the air gap extreme value prediction result.

[0043] The technical scheme provided by the embodiments disclosed in the application has the following beneficial effects:

[0044] The air gap response sequence of the semi-submersible platform in the target water area is obtained through numerical simulation, the air gap response sequence comprising a wave frequency response and a low frequency response; the low frequency response is extracted from the air gap response sequence, the low frequency response is corrected nonlinearly based on a data-driven model, and a corrected low frequency response sequence is obtained; the corrected low frequency response sequence and the wave frequency response are reconstructed to obtain a composite air gap response sequence; non-Gaussian extreme value statistics is performed based on the composite air gap response sequence to obtain an air gap extreme value prediction result of the target water area; a test sequence of air gap changes of the semi-submersible platform over time is obtained through a wave pool test, and iterative optimization is performed on the data-driven model based on a difference between the test sequence and the air gap extreme value prediction result.

[0045] Therefore, the application can realize data-driven accurate correction of low-frequency response in numerical simulation. First, the air gap response sequence containing wave frequency response and low-frequency response is obtained through numerical simulation, a benchmark data model covering the dynamic characteristics of the full frequency band is established, and a complete analysis foundation is provided for subsequent accurate correction. Second, the low-frequency response is extracted from the air gap response sequence, and the data-driven model is used for nonlinear correction of the low-frequency response, which is beneficial to accurate intervention of numerical simulation error. Through the nonlinear mapping capability of the data-driven model, the inherent deviation of the potential flow theory in simulating low-frequency motion can be effectively compensated, which is beneficial to improving the prediction accuracy. Then, the corrected low-frequency response sequence and the wave frequency response with high reliability are reconstructed to obtain a composite air gap response sequence. The calibrated low-frequency component and the undisturbed high-frequency component can be organically integrated to generate a full-band time-domain response that is more accurate, which is beneficial to providing high-quality input data for extreme value statistics. Next, the non-Gaussian extreme value statistics is performed based on the high-precision composite air gap response sequence. Since the reliability of the input data is significantly improved, the extreme value statistics can more truly reflect the tail characteristics of the air gap response, thereby outputting an extreme value prediction result that is more consistent with the engineering practice. Finally, the verification sequence is obtained through the wave pool test, and the data-driven model is iteratively optimized based on the difference between the verification sequence and the prediction result. A complete evolutionary closed-loop mechanism can be constructed, and the data-driven model can be continuously optimized and corrected using the most reliable test data, so that the data-driven model has the generalization ability to adapt to new water areas, thereby ensuring the high accuracy of the air gap extreme value prediction.

[0046] In summary, the technical scheme adopted by the application can realize data-driven accurate correction of low-frequency response in numerical simulation to improve the air gap extreme value prediction accuracy of the semi-submersible platform. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0048] Figure 1 is a flowchart of a semi-submersible platform air gap extreme value prediction method based on numerical simulation provided by the application;

[0049] Figure 2 is an exemplary flowchart for determining the corrected low-frequency response sequence according to the application;

[0050] Figure 3 is an exemplary flowchart for determining the air gap extreme value prediction result of the target water area according to the application;

[0051] Figure 4 is a module structure diagram of a semi-submersible platform air gap extreme value prediction system based on numerical simulation according to the present application. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0053] The embodiments of the present application provide a semi-submersible platform air gap extreme value prediction system and method based on numerical simulation. The core is to obtain an air gap response sequence of a semi-submersible platform in a target water area through numerical simulation, wherein the air gap response sequence includes a wave frequency response and a low frequency response; extract the low frequency response from the air gap response sequence, perform nonlinear correction on the low frequency response based on a data-driven model to obtain a corrected low frequency response sequence; reconstruct the corrected low frequency response sequence and the wave frequency response to obtain a composite air gap response sequence; perform non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain an air gap extreme value prediction result of the target water area; obtain a test sequence of the air gap of the semi-submersible platform changing over time through a wave pool test, and iteratively optimize the data-driven model based on the difference between the test sequence and the air gap extreme value prediction result.

[0054] In order to better understand the above technical solutions, the above technical solutions will be described in detail below with reference to the drawings in the specification and specific embodiments. Referring to FIG. 1, the figure is a flowchart of a semi-submersible platform air gap extreme value prediction method based on numerical simulation according to the embodiments of the present application. The prediction method includes the following steps: Figure 1

[0055] In step S1, an air gap response sequence of a semi-submersible platform in a target water area is obtained through numerical simulation, wherein the air gap response sequence includes a wave frequency response and a low frequency response.

[0056] In the embodiments, the air gap response sequence of the semi-submersible platform in the target water area obtained through numerical simulation can be obtained in the following manner, that is:

[0057] Construct a transfer function of semi-submersible platform motion and wave load based on the structural parameters of the semi-submersible platform and the wave parameters of the target water area;

[0058] Determine the wave frequency motion response and wave load through the transfer function;

[0059] ​The wave frequency motion response and the wave load are coupled to obtain a gap response sequence.

[0060] In the implementation, first, the structure parameters include platform main dimensions, column spacing, draft depth and mass distribution, which can be obtained from platform design drawings; the wave parameters include wave spectrum type, significant wave height and spectrum peak period, which can be obtained from long-term observation statistics of target water area; a three-dimensional potential flow theory software is used to establish a platform wet surface model, a boundary element method is used to solve the velocity potential under the action of unit wave amplitude regular wave, and then the platform six-degree-of-freedom motion (i.e. sway, surge, heave, roll, pitch and yaw) and the wave load (wave force and moment in each direction) relative to wave frequency and direction are calculated to obtain complex response amplitudes, and a matrix formed by the complex response amplitudes is used as a transfer function of the semi-submersible platform motion and the wave load; then, the wave parameters can be discretized into multiple regular wave components by using the JONSWAP spectrum, for each regular wave component, the wave amplitude of the regular wave component is multiplied by the complex response amplitude in the transfer function to obtain the complex form of the semi-submersible platform motion response and the wave load under the regular wave component, and then the complex form of the semi-submersible platform motion response and the load under all regular wave components are obtained, and then the motion response obtained after superposition is used as the wave frequency motion response, and the load obtained after superposition is used as the wave load; finally, according to the linear wave theory, the incident wave surface can be generated from the JONSWAP spectrum, and the difference between the instantaneous wave surface elevation (the diffraction and radiation effects of the wave need to be considered) and the vertical motion (synthesized from heave, pitch and roll in the wave frequency motion response) of the multiple gap monitoring points under the deck of the semi-submersible platform is obtained by using the sensor monitoring technology, and the result of arranging all the difference values in time sequence is used as the gap response sequence.

[0061] It should be noted that in the embodiment, the transfer function describes the linear relationship between the motion and the wave load of the semi-submersible platform and the incident wave in the frequency domain, and the essence of the transfer function is a complex matrix, wherein the modulus of the complex matrix represents the response amplitude, and the amplitude angle of the complex matrix represents the phase lag; the wave frequency motion response specifically refers to the motion of the semi-submersible platform caused by the wave in the wave frequency (usually 0.05-2.0 rad / s) range, but does not include the low-frequency motion caused by the second-order difference frequency force; the wave load refers to the total wave force and moment obtained by integrating the first-order pressure and the second-order pressure induced by the wave on the platform wet surface; the gap response sequence refers to a numerical sequence of the clearance height above the static water surface changing with time, wherein the negative value in the numerical sequence indicates that the wave may impact the bottom of the deck.

[0062] In step S2, the low-frequency response is extracted from the air gap response sequence, and the low-frequency response is nonlinearly corrected based on a data-driven model to obtain a corrected low-frequency response sequence.

[0063] In a specific implementation, in this embodiment, the low-frequency response is extracted from the air gap response sequence based on a high-pass filter. In actual implementation, first, the technical parameters of the high-pass filter are determined, and the cutoff frequency of the filter can be set to 0.04 Hz. The cutoff frequency is a key threshold for distinguishing the wave frequency response from the low-frequency response. The wave component below the cutoff frequency is usually excited by the second-order difference frequency force. Then, the phase information interference in the air gap response sequence is reduced by using a Butterworth filter, and the air gap response sequence is filtered by the high-pass filter to retain the sequence with a frequency below the cutoff frequency, and the sequence is taken as the low-frequency response.

[0064] It should be noted that, in this embodiment, the high-pass filter is an electronic or digital filter device that allows high-frequency signals to pass through while suppressing or attenuating low-frequency signals. In this application, the sequence with a frequency below the cutoff frequency in the air gap response sequence can be separated by the high-pass filter. The low-frequency response specifically refers to the air gap variation component excited by the wave second-order difference frequency force, and the frequency of the low-frequency response is usually below the cutoff frequency. The low-frequency response is the main source of system error in the prediction of air gap extreme values.

[0065] Preferably, in this embodiment, the low-frequency response is nonlinearly corrected based on a data-driven model. Figure 2 As shown in the figure, the figure is an exemplary flowchart for determining the corrected low-frequency response sequence according to the present application. In this embodiment, the low-frequency response is nonlinearly corrected based on a data-driven model to obtain a corrected low-frequency response sequence. The following steps can be used to achieve this:

[0066] First, in step S21, the low-frequency response is modal decomposed to obtain a plurality of proper orthogonal function components.

[0067] Then, in step S22, a training set is constructed according to all the proper orthogonal function components.

[0068] Finally, in step S23, all the proper orthogonal function components are input into the data-driven model trained by the training set for signal reconstruction to obtain the corrected low-frequency response sequence.

[0069] It should be noted that in the embodiment, the data-driven model is a machine learning model based on historical data, which is used to learn and correct the inherent low-frequency response deviation in the numerical simulation through the modal characteristics of the low-frequency response. Preferably, the machine learning model can use a long short-term memory neural network model, which can be used to establish a mapping relationship from the erroneous modal characteristics to the corrected values closer to the true state; wherein the historical data refers to a pre-accumulated data pair set containing multiple sets of "numerical simulation obtained low-frequency response" and "corresponding low-frequency response obtained by physical model test or high-precision sensor measurement", which can constitute a knowledge base for training the machine learning model and is the basis for learning and capturing the simulation deviation rule.

[0070] In specific implementation, first, a variational modal decomposition algorithm can be used for processing, for example: setting the intrinsic modal function component number K value to 6 and the penalty parameter α value to 2000, searching the center frequency and signal distribution that minimizes the sum of each modal estimated bandwidth through iteration, decomposing the low-frequency response into 6 modal components arranged from low frequency to high frequency, and taking the 6 modal components obtained by decomposition as intrinsic modal function components; then, all intrinsic modal function components extracted from historical numerical simulation cases can be taken as input features, and the intrinsic modal function components obtained by variational modal decomposition of the low-frequency response under the same parameters can be taken as target output, and the combination of the training subset and the test subset can be randomly divided according to a 7:3 ratio to form a training set; finally, the long short-term memory neural network model is trained using the training set containing the training subset and the test subset, all intrinsic modal function components are input into the trained long short-term memory neural network input network model in order from low to high frequency, all corrected modal components are output through the long short-term memory neural network input network model, and linear superposition is performed on all corrected modal components in the time domain, and then the superimposed sequence is taken as the corrected low-frequency response sequence.

[0071] It should be noted that in the embodiment, the intrinsic modal function component refers to a quasi-orthogonal sub-signal with different center frequencies, each intrinsic modal function component represents the oscillation mode of a specific frequency band in the original signal, and can decompose a complex mixed signal into multiple relatively simple sub-signals; the training set in the application is a labeled data set for training the data-driven model, which can provide learning samples to enable the data-driven model to master the mapping rule from the erroneous simulation modal to the real modal; signal reconstruction refers to the process of recombining each intrinsic modal function component corrected by the data-driven model into a complete time domain signal, and signal reconstruction is the inverse operation of modal decomposition, which can fuse the corrected frequency band signal components into a complete and corrected low-frequency response sequence.

[0072] In step S3, the modified low-frequency response sequence and the wave-frequency response are reconstructed to obtain a composite air-gap response sequence.

[0073] In this embodiment, the modified low-frequency response sequence and the wave-frequency response are reconstructed to obtain a composite air-gap response sequence in the following manner, namely:

[0074] The modified low-frequency response sequence and the wave-frequency response are linearly superimposed in the time domain to obtain an initial reconstruction sequence.

[0075] The initial reconstruction sequence is subjected to phase consistency inspection and amplitude normalization processing to obtain a composite air-gap response sequence.

[0076] In specific implementation, first, the modified low-frequency response sequence and the wave-frequency response are aligned according to the same time step, the amplitudes at each corresponding time point are added through array addition operation, and the sequence obtained after addition is taken as the initial reconstruction sequence. Then, the cross-correlation coefficient of the initial reconstruction sequence and the wave-frequency response on the main frequency component is calculated, and the phase consistency is evaluated by finding the time shift amount corresponding to the maximum cross-correlation coefficient, for example, when the time shift amount exceeds 0.1 second, the initial reconstruction sequence is subjected to corresponding time shift correction, and the sequence after time shift correction is taken as the phase alignment sequence. The phase alignment sequence is subjected to amplitude normalization processing to obtain the composite air-gap response sequence, wherein the arithmetic mean and the standard deviation of the phase alignment sequence can be calculated by Z-score standardization, each data point in the sequence is subtracted from the mean and then divided by the standard deviation, and the sequence after the above standardization processing is taken as the composite air-gap response sequence.

[0077] It should be noted that the initial reconstruction sequence in this application refers to the preliminary composite signal obtained by simple linear superposition and not subjected to phase and amplitude optimization, including the modified low-frequency and wave frequency; the phase consistency inspection refers to the quality control process of verifying and ensuring that the reconstruction sequence has correct time sequence relationship between the frequency components through signal processing technology, which can eliminate the phase deviation possibly introduced by the signal processing link, thereby ensuring the accuracy of the time corresponding relationship of each component in the reconstruction sequence; the amplitude normalization processing refers to the standardization operation of adjusting the signal to a unified dimension range, which is used to eliminate the magnitude difference possibly existing in different signal sources; and the composite air-gap response sequence refers to the final time domain sequence fused with the modified low-frequency response and the wave-frequency response after phase alignment and amplitude standardization processing, which can provide high-quality and reliable input data basis for subsequent non-Gaussian extreme value statistics.

[0078] In step S4, non-Gaussian extreme value statistics is performed based on the composite air-gap response sequence to obtain the air-gap extreme value prediction result of the target water area.

[0079] Preferably, in this embodiment, the referenceFigure 3 As shown in the figure, the figure is an example flow chart for determining the air gap extreme value prediction result of the target water area according to the present application. In the embodiment, the air gap extreme value prediction result of the target water area is obtained by performing non-Gaussian extreme value statistics based on the composite air gap response sequence, and the following steps can be used to realize the specific implementation:

[0080] Firstly, in step S41, the high-order statistical analysis is performed on the composite air gap response sequence to obtain skewness index and kurtosis index;

[0081] Secondly, in step S42, a polynomial transformation model based on non-Gaussian characteristics is constructed based on the skewness index and the kurtosis index;

[0082] Then, in step S43, the Monte Carlo simulation is used on the polynomial transformation model to generate a synthetic sequence, and then the peak value threshold sampling is performed on the synthetic sequence to obtain extreme value samples;

[0083] Finally, in step S44, the generalized Pareto distribution is used to fit the extreme value samples to obtain an extreme value distribution function, and then the characteristic value corresponding to a preset return period is extracted from the extreme value distribution function to obtain the air gap extreme value prediction result of the target water area.

[0084] In a specific implementation, first, a skewness function can be used to calculate a ratio of a third central moment of a sequence to a cube of a standard deviation, and the ratio can be taken as a skewness index; a kurtosis function can be used to calculate a ratio of a fourth central moment of the sequence to a fourth power of the standard deviation, and the ratio can be taken as a kurtosis index; second, a third Hermite polynomial expansion method can be used to take a standard Gaussian random variable as a basic variable, a cubic term coefficient can be determined by the skewness index, a quartic term coefficient can be determined by the kurtosis index, and a third Hermite polynomial containing the basic variable, the cubic term coefficient, and the quartic term coefficient can be taken as a polynomial transformation model based on non-Gaussian characteristics; then, Monte Carlo simulation can be used on the polynomial transformation model to generate a synthetic sequence, for example, 10,000 standard normal distribution random numbers can be generated as input, and the input can be converted into a non-Gaussian sequence by the polynomial transformation model, and the sequence containing 10,000 data points obtained after the conversion can be taken as the synthetic sequence, a threshold can be set as a 95% quantile of the synthetic sequence according to actual air gap prediction experience, all peak values exceeding the threshold can be extracted, and then all the extracted peak values can be taken as extreme value samples; finally, a generalized Pareto distribution can be used to fit the extreme value samples to obtain an extreme value distribution function, and then a characteristic value corresponding to a preset return period can be extracted from the extreme value distribution function to obtain an air gap extreme value prediction result of the target water area, that is, a maximum likelihood estimation method can be used to determine a scale parameter and a shape parameter of the generalized Pareto distribution, a distribution function containing the scale parameter and the shape parameter can be taken as the extreme value distribution function, a quantile of the extreme value distribution function can be calculated through an inverse function, and then the quantile value can be taken as the air gap extreme value prediction result of the target water area.

[0085] It should be noted that in this embodiment, the skewness index is a third-order statistical quantity representing the asymmetry of a probability distribution, which is used to quantify the degree of deviation of a composite air gap response sequence from a symmetric distribution; the kurtosis index is a fourth-order statistical quantity representing the steepness of a probability distribution, which is used to measure the thickness of the tail of a distribution curve of a composite air gap response sequence; the polynomial transformation model is a mathematical model for converting a Gaussian random process into a non-Gaussian random process through polynomial expansion, which is used to describe the non-Gaussian characteristics of a composite air gap response sequence; and the extreme value distribution function is a probability distribution function describing the statistical characteristics of a random process extreme value, which can provide a quantitative basis for air gap extreme value prediction under different return periods.

[0086] In step S5, an experimental sequence of the air gap of the semi-submersible platform changing over time is obtained through a wave pool test, and the data-driven model is iteratively optimized based on a difference between the experimental sequence and the air gap extreme value prediction result.

[0087] In this embodiment, the experimental sequence of the air gap of the semi-submersible platform changing over time obtained through the wave pool test can be obtained in the following manner, that is:

[0088] constructing a physical model of the semi-submersible platform;

[0089] generating a wave field according to wave parameters of a target water area;

[0090] placing the physical model in the wave field, and measuring changes in wave surface elevation around the semi-submersible platform through a wave height sensor, thereby obtaining a test sequence of changes in air gap of the semi-submersible platform over time.

[0091] In implementation, first, a physical model of the semi-submersible platform is constructed, that is, a platform model can be made of glass fiber reinforced plastic material according to a geometric scale ratio of 1:50, to ensure that the mass distribution, center of gravity position and moment of inertia of the model meet the Froude similarity criterion, and the platform entity made according to the similarity criterion is taken as the physical model of the semi-submersible platform; then, a wave field is generated according to wave parameters of a target water area, that is, in a wave pool with a length of 50 meters, a width of 30 meters and a depth of 5 meters, a hydraulic servo wave generator system is used to input scale ratio parameters corresponding to a significant wave height of 2.5 meters and a spectral peak period of 8.5 seconds of the target water area, to generate irregular waves, and the water surface movement generated in the pool that simulates the wave characteristics of the actual water area is taken as the wave field; finally, the physical model is placed in the wave field, and changes in wave surface elevation around the semi-submersible platform are measured through a wave height sensor, thereby obtaining a test sequence of changes in air gap of the semi-submersible platform over time, for example, the platform model can be moored in the center of the pool by an elastic cable, 8 capacitive wave height sensors are arranged around the four columns of the platform, the time history of the wave surface can be recorded synchronously at a sampling frequency of 50 Hz, the vertical distance between the wave surface and the bottom surface of the platform deck can be calculated through data processing, and then all the vertical distance measurement values arranged in time sequence are taken as the test sequence of changes in air gap of the semi-submersible platform over time.

[0092] It should be noted that in this embodiment, the physical model refers to a scale platform entity made according to the similarity theory, which can accurately reflect the main physical characteristics of the actual ship, and can reproduce the motion and response characteristics of the real semi-submersible platform in waves in a controlled test environment; the wave field refers to a controllable water surface movement environment generated in the pool by a wave making device, which can simulate the actual wave conditions of the target water area, and is the physical experiment basis for verifying the numerical simulation results and obtaining the real air gap response data; the wave height sensor is a contact or non-contact measuring device for measuring the wave height of the water surface, which is used to capture the instantaneous changes of the wave surface near the semi-submersible platform and provide raw data for air gap calculation; the test sequence refers to a discrete data sequence of changes in air gap value at a specific position of the platform over time directly measured through the physical model test, which is the benchmark true value for verifying and correcting the numerical simulation results.

[0093] In this embodiment, the data-driven model is iteratively optimized based on the difference between the test sequence and the air gap extreme value prediction result, which can be implemented in the following manner, that is,

[0094] determining a response residual sequence based on the test sequence and the air gap extreme value prediction result;

[0095] constructing a loss function based on the response residual sequence, and calculating a parameter gradient of the data-driven model by a gradient descent algorithm;

[0096] iteratively updating internal weight parameters of the data-driven model according to the parameter gradient until the loss function converges, and completing the iterative optimization of the data-driven model.

[0097] In a specific implementation, first, the test sequence can be taken as a benchmark true value, the air gap extreme value prediction result can be expanded into a time sequence with the same length as the test sequence, and then the difference between the two sequences can be calculated point by point through array subtraction operation, and then the array of all the difference values arranged in time sequence can be taken as the response residual sequence; then, the square difference of all data points in the response residual sequence can be summed and averaged by using the mean square error calculation formula, and the result can be taken as the loss function, and then the partial derivative of the loss function with respect to all weight parameters in the data-driven model can be calculated by using the random gradient descent algorithm with a learning rate of 0.001, and then the partial derivative value can be taken as the parameter gradient; finally, the internal weight parameters of the data-driven model can be iteratively updated according to the parameter gradient, that is, the current weight parameters can be subtracted by the product of the learning rate and the parameter gradient to obtain new weight parameter values, and the process can be repeated, and when the curve of the change amount of the loss function value converges to smoothness in continuous 10 iterations, the above adjustment process can be taken as the iterative optimization of the data-driven model.

[0098] It should be noted that in the embodiment, the response residual sequence refers to the difference sequence between the test measurement value and the model prediction value, which is used to quantify the deviation degree between the current prediction ability of the data-driven model and the true physical response; the loss function is a mathematical function used to evaluate the prediction effect of the model, and in the present application, it specifically refers to the mean square error function constructed based on the response residual, and the numerical value of the loss function represents the prediction accuracy of the data-driven model; the parameter gradient refers to the change rate of the loss function with respect to each weight parameter of the model, which can provide an accurate adjustment direction and amplitude for the parameter update of the data-driven model; the internal weight parameter refers to the connection strength value between neurons in the data-driven model, and iterative optimization of these parameters can make the model gradually approach the true physical law; convergence refers to the state in which the loss function value tends to be stable and does not decrease significantly in the optimization process, indicating that the data-driven model has reached the optimal performance.

[0099] In summary, the technical scheme adopted in the present application can realize data-driven precise correction for low-frequency response in numerical simulation, so as to improve the air gap extreme value prediction accuracy of the semi-submersible platform.

[0100] In the second embodiment, the application provides a numerical simulation based air gap extreme value prediction system for a semi-submersible platform, referring to Figure 4 Fig. 1 is a module structure diagram of a numerical simulation based air gap extreme value prediction system for a semi-submersible platform according to the application, the prediction system comprising:

[0101] a numerical simulation module 100 configured to obtain an air gap response sequence of the semi-submersible platform in a target water area through numerical simulation, the air gap response sequence comprising a wave frequency response and a low frequency response;

[0102] a low frequency correction module 200 configured to extract the low frequency response from the air gap response sequence, and perform nonlinear correction on the low frequency response based on a data driven model to obtain a corrected low frequency response sequence;

[0103] a response reconstruction module 300 configured to reconstruct the corrected low frequency response sequence and the wave frequency response to obtain a composite air gap response sequence;

[0104] an extreme value prediction module 400 configured to perform non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain an air gap extreme value prediction result of the target water area;

[0105] a model optimization module 500 configured to obtain a test sequence of air gap changes of the semi-submersible platform over time through a wave pool test, and perform iterative optimization on the data driven model based on a difference between the test sequence and the air gap extreme value prediction result.

[0106] The application is described with reference to flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The functions specified in one flow or multiple flows and / or blocks

[0107] Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiments can be completed by instructing the relevant hardware by means of a program, and the program can be stored in a computer readable storage medium, including Read-Only Memory (ROM), Random Access Memory (RAM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), One-time Programmable Read-Only Memory (OTPROM), Electrically-Erasable Programmable Read-Only Memory (EEPROM), Compact Disc Read-Only Memory (CD-ROM) or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other medium that can be used to carry or store data in a computer readable manner.

[0108] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the statement "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or apparatus that includes the element.

Claims

1. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform, the method comprising: The prediction method comprises the following steps: obtaining an air gap response sequence of the semi-submersible platform in the target water area through numerical simulation, the air gap response sequence comprising a wave frequency response and a low frequency response; extracting the low frequency response from the air gap response sequence, and performing nonlinear correction on the low frequency response based on a data-driven model to obtain a corrected low frequency response sequence; reconstructing the corrected low frequency response sequence and the wave frequency response to obtain a composite air gap response sequence; performing non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain an air gap extreme value prediction result of the target water area; iteratively optimizing the data-driven model based on differences between a test sequence of air gap changes of the semi-submersible platform over time obtained through a wave pool test and the air gap extreme value prediction result; wherein the data-driven model is a machine learning model based on historical data, used to learn and correct inherent low frequency response deviations in the numerical simulation through modal characteristics of the low frequency response; wherein iteratively optimizing the data-driven model based on differences between the test sequence and the air gap extreme value prediction result specifically comprises: determining a response residual sequence from the test sequence and the air gap extreme value prediction result; constructing a loss function based on the response residual sequence, and calculating a parameter gradient of the data-driven model through a gradient descent algorithm; iteratively updating internal weight parameters of the data-driven model according to the parameter gradient until the loss function converges, completing the iterative optimization of the data-driven model.

2. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform as claimed in claim 1, wherein, obtaining an air gap response sequence of the semi-submersible platform in the target water area through numerical simulation specifically comprises: constructing a transfer function of semi-submersible platform motion and wave load based on structural parameters of the semi-submersible platform and wave parameters of the target water area; determining wave frequency motion response and wave load through the transfer function; performing coupling analysis on the wave frequency motion response and the wave load to obtain an air gap response sequence.

3. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform as claimed in claim 1, wherein, extracting the low frequency response from the air gap response sequence based on a high-pass filter.

4. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform as recited in claim 1, wherein, performing nonlinear correction on the low frequency response based on a data-driven model to obtain a corrected low frequency response sequence specifically comprises: performing modal decomposition on the low frequency response to obtain a plurality of proper orthogonal function components; constructing a training set according to all the proper orthogonal function components; inputting all the proper orthogonal function components into the data-driven model trained based on the training set to reconstruct a signal and obtain a corrected low frequency response sequence.

5. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform as recited in claim 1, wherein, reconstructing the corrected low frequency response sequence and the wave frequency response to obtain a composite air gap response sequence specifically comprises: linearly superimposing the corrected low frequency response sequence and the wave frequency response in the time domain to obtain an initial reconstruction sequence; performing phase consistency inspection and amplitude normalization processing on the initial reconstruction sequence to obtain a composite air gap response sequence.

6. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform as recited in claim 1, wherein, performing non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain an air gap extreme value prediction result of the target water area specifically comprises: performing high-order statistical analysis on the composite air gap response sequence to obtain skewness and kurtosis indexes; constructing a polynomial transformation model based on non-Gaussian characteristics based on the skewness index and the kurtosis index; performing Monte Carlo simulation on the polynomial transformation model to generate a synthetic sequence, and then performing peak over threshold sampling on the synthetic sequence to obtain an extreme value sample; fitting the extreme value sample using a generalized Pareto distribution to obtain an extreme value distribution function, and then extracting a characteristic value corresponding to a preset return period from the extreme value distribution function to obtain the air gap extreme value prediction result of the target water area.

7. A numerical simulation based method for predicting air gap extreme values for a semi-submersible platform as recited in claim 1, wherein, The test sequence of the air gap of the semi-submersible platform changing with time obtained through the wave pool test specifically includes: constructing a physical model of the semi-submersible platform; generating a wave field according to the wave parameters of the target water area; placing the physical model in the wave field, and measuring the change of the wave surface elevation around the semi-submersible platform through a wave height sensor to obtain the test sequence of the air gap of the semi-submersible platform changing with time.

8. A numerical simulation based semi-submersible platform air gap extreme prediction system for performing a numerical simulation based semi-submersible platform air gap extreme prediction method according to any one of claims 1 to 7, characterized by, The prediction system includes: a numerical simulation module configured to obtain an air gap response sequence of the semi-submersible platform in the target water area through numerical simulation, the air gap response sequence including a wave frequency response and a low frequency response; a low frequency correction module configured to extract the low frequency response from the air gap response sequence, correct the low frequency response based on a data-driven model, and obtain a corrected low frequency response sequence; a response reconstruction module configured to reconstruct the corrected low frequency response sequence and the wave frequency response to obtain a composite air gap response sequence; an extreme value prediction module configured to perform non-Gaussian extreme value statistics based on the composite air gap response sequence to obtain the air gap extreme value prediction result of the target water area; a model optimization module configured to obtain a test sequence of the air gap of the semi-submersible platform changing with time through a wave pool test, and iteratively optimize the data-driven model based on the difference between the test sequence and the air gap extreme value prediction result.

Images