An adaptive optimization method for collaborative inversion of multi-scale control variables

Abstract

The present invention provides an adaptive optimization method for collaborative inversion of multi-scale control variables, belonging to the field of ocean internal tide technology. The method includes: driving a forward internal tide model based on preset open boundary conditions and a bottom friction coefficient as model input parameters; optimizing control variables at independent points by assimilating observational data sets: using cubic spline interpolation technology to perform nonlinear fitting of the spatial distribution of the open boundary conditions, and using surface spline interpolation methods to perform multidimensional smoothing optimization of the field structure of the bottom friction coefficient; and continuously optimizing the independent point values ​​of the open boundary conditions and the bottom friction coefficient until the deviation index between the simulation result and the observational data set falls within a preset error threshold. The present invention solves the problems of current internal tide model control variable optimization, such as the inability to characterize the nonlinear physical characteristics of the actual ocean and the inability to achieve high-precision simulation of internal tides.

Description

Technical Field

[0001] The present invention belongs to the technical field of ocean internal tides, and in particular relates to an adaptive optimization method for collaborative inversion of multi-scale control variables. Background Art

[0002] In the dynamics of ocean internal tides, high-precision numerical models are key tools for revealing the mechanisms of internal tide generation, propagation, and energy dissipation. The simulation accuracy of internal tide models is highly dependent on the accuracy of the control variables. The open boundary conditions are based on boundary inputs such as tide level and flow velocity, and the bottom friction coefficient is based on the resistance of the seabed topography to tidal flow. Both are core parameters that affect model performance. However, existing internal tide models face the following problems in optimizing control variables:

[0003] The models of related technologies mostly use linear interpolation or empirical formulas to set control variables, which makes it difficult to accurately describe the tidal wave deformation and energy dispersion of internal tide fluctuations in the actual ocean, resulting in significant deviations between the simulation results and the measured data.

[0004] The optimization of open boundary conditions and bottom friction coefficient is usually treated independently, lacking the coordinated modeling of their spatial correlation and scale effects.

[0005] The inversion algorithm of related technologies is prone to fall into local optimal solutions when dealing with high-dimensional control variables, and the optimization process lacks constraints on spatial continuity, resulting in slow convergence of simulation errors. Summary of the Invention

[0006] The present invention provides an adaptive optimization method for collaborative inversion of multi-scale control variables. The method aims to solve the problems existing in the optimization of control variables of the current internal tide model, such as the inability to characterize the nonlinear physical characteristics of the actual ocean and the inability to achieve high-precision simulation of the internal tide.

[0007] Methods include:

[0008] S101, based on the preset open boundary condition and bottom friction coefficient as model input parameters, driving the forward internal tide model to run, and extracting surface flow velocity data from the preset observation points as an observation data set;

[0009] S102. Initial parameter values ​​are set based on the open boundary condition and the bottom friction coefficient, and the control variables at the independent points are optimized by assimilating the observation data set: the spatial distribution of the open boundary condition is nonlinearly fitted using the cubic spline interpolation technique, and the field structure of the bottom friction coefficient is multidimensionally smoothed and optimized using the surface spline interpolation method;

[0010] S103, looping through S101 and S102, adjusting the model input parameters, and continuously optimizing the open boundary conditions and the independent point values ​​of the bottom friction coefficient until the deviation index between the simulation results and the observation data set falls within a preset error threshold.

[0011] It should be further explained that the specific steps of S101 are: performing forward simulation based on the preset internal tide assimilation model, using the given open boundary conditions and bottom friction coefficient as model input parameters, running the forward model to obtain simulation results, and extracting surface flow velocity data from the preset observation points as the observation data set.

[0012] It should be further explained that the specific steps of S102 are: setting the initial values ​​of the control variables, the initial value of the open boundary condition is 0, the initial value of the bottom friction coefficient is 0.002, and optimizing the control variables at the independent points based on the observation data set obtained in S101:

[0013] The cubic spline interpolation method is used to optimize the spatial distribution of the open boundary conditions; the surface spline interpolation method is used to optimize the multi-dimensional field structure of the bottom friction coefficient.

[0014] It should be further explained that the specific steps of optimizing the inversion of the open boundary condition using cubic spline interpolation are as follows: selecting a preset point as an independent point from all the open boundary points in the calculation area and evaluating the value using the optimization algorithm;

[0015] The values ​​of the remaining grid points are obtained by cubic spline interpolation of the values ​​of the independent points. Let N is the total number of open boundary points, is the value of the boundary condition at the independent point, is the number of independent points, It is the use of The result obtained by cubic spline interpolation is

[0016] (1)

[0017] in, are the cubic spline interpolation coefficients.

[0018] It should be further explained that in the method, the interval Upper cubic spline interpolation function Exist and use To express At the point The derivative value at , then in each small interval Satisfaction on:

[0019] (2).

[0020] It should be further explained that In the interval The above expression is as follows:

[0021] (3)

[0022] in, ,right Taking the second derivative we can get The expression is as follows:

[0023] (4).

[0024] It should be further explained that the open boundary conditions are as follows:

[0025] The first type of open boundary condition is: the curve is at the two end points and The slope of the tangent at is known, that is, and is known;

[0026] The second type of open boundary condition: function At both endpoints and The second-order derivative at is known, that is and is known;

[0027] Periodic Boundary Conditions: Functions is a periodic function, a spline function It should also be a periodic function, that is, it satisfies the condition at the endpoints and .

[0028] It should be further explained that the optimization inversion of the bottom friction coefficient using surface spline interpolation in S102 specifically includes:

[0029] Select some grid points in the bottom friction coefficient field as independent points, and set For the l The value of the bottom friction coefficient at each independent point, the value of the bottom friction coefficient at each grid point in the bottom friction coefficient field Value is through Obtained through surface spline interpolation, and Satisfies the following relationship:

[0030] (13)

[0031] in, is the surface spline interpolation coefficient.

[0032] It should be further explained that in the method, the expression of the surface spline interpolation coefficient is extracted:

[0033] (18)

[0034] Get the cost function J Gradient of the bottom friction coefficient at an independent point:

[0035] (19);

[0036] pass The gradient of the cost function for the bottom friction coefficient field is transferred to independent points, the influence of interpolation and distance on the gradient is clarified, and the gradient is used to drive parameter updates. Combined with the smoothness of the surface spline, the simulation and observation errors are reduced, and the goal of data assimilation is ultimately achieved.

[0037] It can be seen from the above technical solutions that the present invention has the following advantages:

[0038] The adaptive optimization method of collaborative inversion of multi-scale control variables provided in this application uses open boundary conditions and bottom friction coefficient as key parameters of the driving model. In view of the defects of independent processing of control variables and insufficient coupling in the existing technology, the overall characterization ability of the model for the physical process of ocean internal tides is improved through multi-variable collaborative driving. The surface velocity generated by simulation is used as observation data to establish a direct relationship between the model output and the actual ocean observation, and solve the problem of large deviation between the simulation results and the actual measurement caused by the break of the traditional model data assimilation link. Through piecewise smooth functions and second-order derivative continuity constraints, the curve mutations and high-frequency noise caused by traditional linear interpolation are avoided, and the nonlinear spatial distribution of tidal fluctuations is more realistically reflected. By cyclically assimilating observation data, the open boundary conditions and the bottom friction coefficient are adjusted synchronously to achieve the coupling of the spatial distribution and physical effects of the two. The optimized control variable field can be directly used in marine engineering design, avoiding the design risks caused by traditional empirical parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 Flowchart of the adaptive optimization method for collaborative inversion of multi-scale control variables;

[0041] Figure 2 Spatial distribution of inversion results for given open boundary conditions and two interpolation schemes. DETAILED DESCRIPTION

[0042] The adaptive optimization method for collaborative inversion of multi-scale control variables provided by the present invention uses an independent point scheme to optimize model parameters. Some specific points in the calculation area are selected as independent points, and the spatial distribution of the control variables is obtained from the values ​​of these independent points through the spline interpolation method. Different methods are used for different parameters in the model, such as the cubic spline interpolation method for the optimization of open boundary conditions and the surface spline interpolation method for the optimization of the bottom friction coefficient. This solves the problems that exist in the optimization of control variables of the current internal tide model, such as the inability to characterize the nonlinear physical characteristics of the actual ocean and the inability to achieve high-precision simulation of the internal tide.

[0043] The following describes in detail the specific steps of the adaptive optimization method for collaborative inversion of multi-scale control variables involved in this application. Specific details such as specific system structures and techniques are provided for illustrative purposes, not for limitation, to facilitate a thorough understanding of the embodiments of this application. However, it should be apparent to those skilled in the art that this application may also be implemented in other embodiments without these specific details.

[0044] It should be understood that when used in this specification, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their collections. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0045] The phrases "one embodiment" or "some embodiments" described in this application mean that the specific features, structures, or characteristics described in the embodiment are included in one or more embodiments of the application. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in other embodiments," etc. that appear in different places in this application do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized.

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0047] See also Figure 1 FIG. 1 is a flow chart of an adaptive optimization method for collaborative inversion of multi-scale control variables in a specific embodiment, the method comprising:

[0048] S101: The forward model is driven by given control variables, where the control variables are open boundary conditions and bottom friction coefficient, and the surface velocity generated by the simulation results at a given observation point is used as observation data.

[0049] In this embodiment, the forward model uses control variables as driving conditions. Open boundary conditions specify the input conditions for physical quantities such as tides and currents at the edges of the model region. For example, they set the tidal level, flow velocity, and direction at the model's sea boundary. The bottom friction coefficient reflects the frictional resistance of the seabed to the flow, and its value affects the degree of velocity attenuation near the seabed. After the forward model is run, surface flow velocity data is recorded at pre-set observation points.

[0050] It's understandable that by constructing a forward model based on relevant theories such as fluid mechanics and ocean dynamics, and solving equations such as the Navier-Stokes equation and the continuity equation through finite difference or finite element methods, the tidal motion process within the ocean can be simulated under given control variables, thereby obtaining the surface velocity at the observation point. This clarifies the input-output relationship of the simulation process, and the observation data provides a reliable basis for comparison for subsequent optimization, ensuring that the optimization direction is centered around actual ocean phenomena and making the model simulation results closer to reality.

[0051] S102: attaching initial values ​​to the control variables, optimizing the control variables at the independent points by assimilating the observation data, optimizing the open boundary conditions by using cubic spline interpolation, and optimizing the bottom friction coefficient by using surface spline interpolation.

[0052] In this embodiment, initial values ​​are assigned to the control variables, which can be determined based on existing research experience, historical data, or simple estimation. Using data assimilation technology, the observation data obtained in step S101 are integrated with the model simulation results to optimize the control variables at the independent points.

[0053] For open boundary conditions, this embodiment adopts the cubic spline interpolation method. This method constructs a piecewise cubic polynomial to ensure that the interpolation function has continuous first-order and second-order derivatives at the nodes, thereby smoothly fitting the changing curve of the open boundary conditions. For the bottom friction coefficient, surface spline interpolation is used. By constructing a smooth surface in space, the bottom friction coefficient value of the unknown point is estimated based on the known data points. This method is suitable for parameter optimization in two-dimensional space.

[0054] Data assimilation in this embodiment leverages the available information from observational data to modify model parameters, minimizing the error between simulation results and actual observations. Cubic spline interpolation and surface spline interpolation, through mathematical fitting, use limited known data points to rationally infer control variable values ​​at other locations, ensuring that the distribution of control variables more closely matches the characteristics of the actual marine environment. In this way, data assimilation combined with interpolation methods fully exploits the value of observational data, mathematically optimizes control variables, and improves the accuracy and rationality of model parameters.

[0055] S103: Repeat steps S101 to S102 to continuously optimize the model control variables so that the error between the simulation results and the observed data falls within a preset threshold.

[0056] This embodiment repeatedly repeats steps S101 and S102. Each iteration reruns the forward model based on the previously optimized control variables, obtains new simulation results, compares them with the observed data, and calculates the error. When the error between the simulation results and the observed data falls within a pre-set threshold, the model control variables are considered optimized to an appropriate level, and the iteration stops.

[0057] As can be seen, the method of this embodiment uses iterative optimization to gradually adjust the control variables, bringing the model simulation results closer to the actual observed data. Using an error feedback mechanism, reaching a preset error threshold as a convergence condition achieves the optimal solution for the control variables. This ensures that the model control variables are fully optimized to meet the accuracy requirements for internal tide simulation and prediction in practical applications.

[0058] On the basis of the above embodiment, in order to further improve the reliability of the adaptive optimization method for collaborative inversion of multi-scale control variables provided in the above embodiment, the following is a more specific implementation method. In the following embodiment, the following steps are specifically included:

[0059] S201: The forward model is driven by given control variables, where the control variables are the open boundary conditions and the bottom friction coefficient, and the surface flow velocity generated by the simulation results at a given observation point is used as observation data.

[0060] This embodiment uses the isopycnic internal tide assimilation model to simulate the internal tide, drives the forward model with given control variables (open boundary conditions, bottom friction coefficient), and uses the surface flow velocity generated by the simulation results at a given observation point as the "observation data".

[0061] S202: attaching initial values ​​to the control variables, optimizing the control variables at the independent points by assimilating the observation data, optimizing the open boundary conditions by using cubic spline interpolation, and optimizing the bottom friction coefficient by using surface spline interpolation.

[0062] In some embodiments, initial values ​​are assigned to the control variables, wherein the open boundary condition is assigned an initial value of 0, and the bottom friction coefficient is assigned an initial value of 0.002. The control variables at independent points are optimized by assimilating the "observation data" in S101. The cubic spline interpolation method is selected for the optimization of the open boundary condition, and the surface spline interpolation method is selected for the optimization of the bottom friction coefficient.

[0063] This embodiment uses the cubic spline interpolation method to optimize the open boundary condition of one of the model parameters. The method is as follows: select some specific points from all the open boundary points in the calculation area as independent points, and calculate their values ​​using the optimization algorithm, while the values ​​of other grid points are obtained by cubic spline interpolation of the values ​​of the independent points. Specifically, let N is the total number of open boundary points, is the value of the boundary condition at the independent point, is the number of independent points, It is the use of The result obtained by cubic spline interpolation is

[0064] (1)

[0065] in, is the cubic spline interpolation coefficient, and its specific derivation process is as follows: Assume that in the interval Upper cubic spline interpolation function Exist and use To express At the point The derivative value at i =0,1,2,…, n ,in n +1= N P , then in each small interval Satisfaction on:

[0066] (2)

[0067] In the interval The above expression is as follows:

[0068] (3)

[0069] in, , , next Taking the second derivative we can get The expression is as follows:

[0070] (4)

[0071] To ensure At the node There is a continuous second-order derivative at ,Right now

[0072] (5)

[0073] Simplifying the above formula, we can get:

[0074] (6)

[0075] in:

[0076] Equation (6) is about n +1 unknown quantity n -1 linear equations, so there are infinite solutions to this system of equations. However, in practical applications, usually only one specific solution can be selected. At this time, boundary conditions need to be given according to the specific situation. Common boundary conditions are:

[0077] The first type of boundary condition: the curve is at two endpoints and The slope of the tangent at is known, that is, and is known (given as 1 for optimization with open boundary conditions);

[0078] The second type of boundary condition: function At both endpoints and The second-order derivative at is known, that is and is known and is given as 0 for the optimization of open boundary conditions;

[0079] Periodic Boundary Conditions: Functions is a periodic function, and accordingly, the spline function It should also be a periodic function, that is, it satisfies the condition at the endpoints and .

[0080] After the three boundary conditions are given, the cubic spline interpolation coefficients are deduced. Since the derivation process of the three boundary conditions is similar, only the periodic boundary condition is used as an example for derivation. In order to easily apply the cubic spline interpolation method to the optimization inversion of open boundary conditions, the equal spacing (the distance is set to h ) method to select independent points on the open boundary, assuming , then we have According to the characteristics of periodic boundary conditions, we have

[0081] (7) Combining equations (6) and (7), we can obtain:

[0082] (8) Among them:

[0083] (9)

[0084] Combining (8) and (9), the value of the point on the open boundary can be obtained by calculation:

[0085] (10)

[0086] The above formula is the specific expression of formula (1), so the next step is to extract the cubic spline interpolation coefficient from formula (10): The specific expression of .

[0087] make ,like j exist On, then j The cubic spline interpolation coefficient of the lth independent point corresponding to the grid point is:

[0088] (11)

[0089] The cost function can be derived J Gradient of open boundary conditions at independent points:

[0090] (12)

[0091] For the optimized inversion of open boundary conditions, Figure 2 As shown in the figure, taking the periodic boundary condition as an example, the Fourier coefficient a in the open boundary condition is inverted. The red solid line in the figure represents the given open boundary condition, that is, the spatial distribution of the Fourier coefficient a. The black solid line and the black dashed line represent the results after cubic spline interpolation and linear interpolation inversion, respectively. Compared with the independent point scheme based on linear interpolation given in the figure, the open boundary curve obtained by the cubic spline interpolation scheme is smoother and fits the given spatial distribution better, has smaller error, and has stronger physical significance.

[0092] S203: Repeat steps S201 to S202 to continuously optimize the model control variables so that the error between the simulation results and the observed data falls within a preset threshold.

[0093] In step S203 of this embodiment, the method for optimizing the bottom friction coefficient, one of the model parameters, by using the surface spline interpolation method is as follows: select some grid points (a total of N ) as independent points, let For the l The value of the bottom friction coefficient at each independent point, the value of the bottom friction coefficient at each grid point in the bottom friction coefficient field Value is through Obtained through surface spline interpolation, and Satisfies the following relationship:

[0094] (13)

[0095] in, is the surface spline interpolation coefficient, and its specific derivation process is as follows:

[0096] According to the expression of the surface spline interpolation function, Equation (13) can be rewritten as:

[0097] (14)

[0098] in, R is the influence radius, Represents the grid points in the bottom friction coefficient field With the l The distance between independent points. In order to determine the coefficient matrix The specific expression of (14) is rewritten into matrix form:

[0099] (15)

[0100] Specifically,

[0101]

[0102] in,

[0103]

[0104] set up , then (15) can be written as , you can get

[0105] (16)

[0106] Substituting (16) into (14) we can obtain

[0107] (17)

[0108] Therefore, the expression of the surface spline interpolation coefficient is extracted from formula (17):

[0109] (18)

[0110] The cost function can be derived J Gradient of the bottom friction coefficient at an independent point:

[0111] (19).

[0112] The gradient involved in this embodiment indicates the cost function J The rate of change of the bottom friction coefficient at the independent point. In the optimization algorithm, the parameter update direction is the negative gradient direction, and the step size is controlled by the learning rate. The gradient size reflects the contribution of adjusting the bottom friction coefficient at a certain independent point to reducing the error between simulation and observation. For example, points with larger gradients have a more significant impact on the error and need to be adjusted first. By iteratively calculating the gradient and updating the bottom friction coefficient, the control variable gradually approaches the optimal value, reducing the error between the simulated flow rate and the observed data. The smoothness of the surface spline interpolation ensures that the gradient propagation has spatial continuity, avoids local overfitting, and thus the global error converges more evenly. When the gradient norm is lower than the preset threshold, it means that the error has been sufficiently reduced and the optimization process can be terminated.

[0113] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0114] As an example of this application, a port (hereinafter referred to as Port X) needs to improve the accuracy of surface tide predictions due to channel expansion to ensure the safety of ships entering and leaving the port. The tidal model previously used by the port was inaccurate due to the input of sea boundary tide levels and the setting of seabed resistance parameters. This resulted in a significant deviation between the simulated surface current velocity and the measured value (with an average error of approximately 0.4 m / s), affecting ship berthing scheduling.

[0115] To this end, the multi-scale control variable collaborative inversion method proposed in this paper is used to optimize the model control variables, with the goal of reducing the error between the simulated flow velocity and the measured value to within 0.15 m / s.

[0116] Step S501: Build a forward model and generate observation data.

[0117] The sea area of ​​Port X covers approximately 50 km², with an average water depth of 8 m. The forward model uses "open boundary conditions" and "bottom friction coefficient" as core control variables to drive the model's simulation of tidal motion.

[0118] The tidal time series at the model boundary is determined based on historical tidal data for the port's area, using average values ​​from the same period. For example, the tidal level at point A at the bay's mouth follows a semidiurnal pattern, rising and falling twice daily, with an average tidal level of 2.0 m and a maximum tidal range of 3.5 m. At point B at the bay's top, the tidal amplitude decays to 2.5 m due to shallower water depth, with a phase lag of approximately two hours. Based on historical research experience in this area, the initial bottom friction coefficient was uniformly set at 0.0025, with no spatial variation, assuming uniform seabed resistance.

[0119] Five key observation points were selected within Port X (C: 6 m water depth near the bay entrance, D: 10 m water depth at the center of the channel, E: 4 m water depth at the bay top shoal, F: 7 m water depth outside the breakwater, and G: 5 m water depth at the port entrance). The surface current velocity at a depth of 1 m was measured in real time using an acoustic Doppler current profiler (ADCP) for 30 consecutive days. This data was used as real-world data for subsequent optimization.

[0120] Step S502: Initialize control variables and perform collaborative optimization.

[0121] Open boundary conditions were set, initially using the historical average tide level; the initial value for the entire field was uniformly 0.0025. Taking the surface flow velocity at the observation point as the target, the control variables were adjusted by comparing the model simulation values ​​with the measured values. After the model was run, it was found that the simulated flow velocity was generally 0.2-0.3 m / s lower than the measured values ​​at the beginning of high tide, inferring that the initial phase of the tide level at the bay mouth was slow. The open boundary tide level time series was adjusted using cubic spline interpolation—the tide level time curve at point A at the bay mouth was "advanced" by 1 hour. For example, the original tide level of -1.5 m at t=0 was adjusted to -1.3 m, and the high tide level of 1.5 m at t=6 was adjusted to 1.7 m, so that the interpolated tide level series more closely matched the actual rise and fall rhythm.

[0122] Further analysis found that at the nearshore shallow observation point E, the simulated flow velocity at a water depth of 4 m was significantly lower than the measured value, while at the deepwater observation point D, the error was as small as only 0.1 m / s at a water depth of 10 m.

[0123] It is speculated that seafloor friction is stronger in shallow areas, necessitating an increase in the bottom friction coefficient. Using surface spline interpolation, a spatial distribution of the bottom friction coefficient was constructed in a two-dimensional ocean area. The bottom friction coefficient around point E, nearshore in water depths <5 m, was increased from 0.0025 to 0.0035. Point D remained unchanged at 0.002 for water depths >8 m. A smooth transition was achieved in the transition zone, between 5 and 8 m water depths, to avoid sudden changes in the parameter.

[0124] Step S503: Iterate and optimize until the error reaches the target.

[0125] In the first iteration, the boundary tide level was opened 1 hour earlier, the nearshore bottom friction coefficient was increased, and the average error between the simulated and measured surface velocity values ​​at the five observation points was calculated to be 0.28 m / s.

[0126] It was found that the simulated flow velocity at point B at the top of the bay was still too low, with an error of 0.25 m / s. The tidal amplitude of the bay mouth was fine-tuned again by cubic spline interpolation, increasing it from 3.5 m to 3.7 m, shortening the tidal lag time at the top of the bay to 1.5 hours.

[0127] The error of the simulated flow velocity near point E is still 0.2 m / s. The bottom friction coefficient in this area is further increased to 0.0038 through surface spline interpolation, which is closer to the actual shoal resistance.

[0128] Second Iteration: The model was rerun, and the average error decreased to 0.18 m / s. The simulated velocity error at the center of the channel was 0.12 m / s, approaching the threshold. The tidal phase at the bay mouth was fine-tuned and brought forward by 0.5 hours to synchronize the tide level at point D with the measured level. At point F, outside the breakwater, at a depth of 7 m, the simulated velocity was 0.1 m / s higher due to the flow being blocked by the breakwater. Surface spline interpolation was used to slightly reduce the bottom friction coefficient in this area to 0.0022.

[0129] The third iteration: After running, the average error of the five observation points dropped to 0.12 m / s, which is lower than the 0.15 m / s threshold, meeting the accuracy requirement, and the optimization was stopped.

[0130] The final result, using cubic spline interpolation to correct the tide time series, shows an amplitude of 3.7 m at the bay entrance, a period of 12.4 hours, and a phase completely synchronized with the measured values. The spatial distribution field, obtained using surface spline interpolation, shows a range of 0.0035–0.0038 in areas with water depths less than 5 m nearshore, 0.0022 in areas with a depth of 7 m outside the breakwater, and 0.002 in deepwater, with smooth variations in the transition zone. The average error between the simulated surface velocity and the measured data after the optimization model is only 0.12 m / s. Port dispatchers report that the model significantly improves the accuracy of high tide and strong current forecasts for ships entering the port, effectively reducing berthing delays caused by tidal misjudgments and demonstrating its outstanding practical application value.

[0131] The adaptive optimization method for collaborative inversion of multi-scale control variables of the present application is a combination of the units and algorithm steps of each example described in the embodiments disclosed herein, and can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0132] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An adaptive optimization method for collaborative inversion of multi-scale control variables, characterized in that: Methods include: S101, based on the preset open boundary condition and bottom friction coefficient as model input parameters, driving the forward internal tide model to run, and extracting surface flow velocity data from the preset observation points as an observation data set; S102. Initial parameter values ​​are set based on the open boundary condition and the bottom friction coefficient, and the control variables at the independent points are optimized by assimilating the observation data set: the spatial distribution of the open boundary condition is nonlinearly fitted using the cubic spline interpolation technique, and the field structure of the bottom friction coefficient is multidimensionally smoothed and optimized using the surface spline interpolation method; The specific steps of optimizing the inversion of the open boundary condition using cubic spline interpolation are as follows: selecting a preset point as an independent point from all the open boundary points in the calculation area and evaluating the value using the optimization algorithm; The values ​​of the remaining grid points are obtained by cubic spline interpolation of the values ​​of the independent points. Let N is the total number of open boundary points, is the value of the boundary condition at the independent point, is the number of independent points, It is the use of The result obtained by cubic spline interpolation is （1） in, is the cubic spline interpolation coefficient; Method, defined in the interval Upper cubic spline interpolation function Exist and use To express At the point The derivative value at , then in each small interval Satisfaction on: （2） In the interval The above expression is as follows: （3） in, ,right Taking the second derivative we can get The expression is as follows: （4）； S103, looping through S101 and S102, adjusting the model input parameters, and continuously optimizing the open boundary conditions and the independent point values ​​of the bottom friction coefficient until the deviation index between the simulation results and the observation data set falls within a preset error threshold.

2. The adaptive optimization method for collaborative inversion of multi-scale control variables according to claim 1, characterized in that: The specific steps of S101 are: performing forward simulation based on a preset internal tide assimilation model, using the given open boundary conditions and bottom friction coefficient as model input parameters, running the forward model to obtain simulation results, and extracting surface flow velocity data from preset observation points as an observation data set.

3. The adaptive optimization method for collaborative inversion of multi-scale control variables according to claim 1, characterized in that: The specific steps of S102 are: setting the initial values ​​of the control variables, the initial value of the open boundary condition is 0, the initial value of the bottom friction coefficient is 0.002, and optimizing the control variables at the independent points based on the observation data set obtained in S101: The cubic spline interpolation method is used to optimize the spatial distribution of the open boundary conditions; the surface spline interpolation method is used to optimize the multi-dimensional field structure of the bottom friction coefficient.

4. The adaptive optimization method for collaborative inversion of multi-scale control variables according to claim 1, characterized in that: The open boundary conditions are as follows: The first type of open boundary condition is: the curve is at the two end points and The slope of the tangent at is known, that is, and is known; The second type of open boundary condition: function At both endpoints and The second-order derivative at is known, that is and is known; Periodic Boundary Conditions: Functions is a periodic function, a spline function It should also be a periodic function, that is, it satisfies the condition at the endpoints and .

5. The adaptive optimization method for collaborative inversion of multi-scale control variables according to claim 1, characterized in that: In S102, surface spline interpolation is used to optimize the inversion of the bottom friction coefficient, specifically including: Select some grid points in the bottom friction coefficient field as independent points, and set For the l The value of the bottom friction coefficient at each independent point, the value of the bottom friction coefficient at each grid point in the bottom friction coefficient field Value is through Obtained through surface spline interpolation, and Satisfies the following relationship: （13） in, is the surface spline interpolation coefficient.

6. The adaptive optimization method for collaborative inversion of multi-scale control variables according to claim 5, characterized in that: In the method, the expression of the surface spline interpolation coefficient is extracted: （18） Get the cost function J Gradient of the bottom friction coefficient at an independent point: （19）； pass The gradient of the cost function for the bottom friction coefficient field is transferred to independent points, the influence of interpolation and distance on the gradient is clarified, and the gradient is used to drive parameter updates. Combined with the smoothness of the surface spline, the simulation and observation errors are reduced, and the goal of data assimilation is ultimately achieved.

Images