A ls+ cnn combined prediction method of ut1-utc considering angular momentum

Abstract

The application discloses a LS+CNN combined UT1-UTC prediction method considering effective angular momentum, relates to the field of geophysics and related computer science technology, and comprises the following steps: acquiring a UT1-UTC basic data set and an effective angular momentum data set, constructing an LS model and a CNN model, combining an LS fitting extrapolation result with an output result of the CNN model, and obtaining a high-precision UT1-UTC prediction value. The model has high precision and reliability in long-term UT1-UTC prediction, can overcome the shortcomings of a traditional linear model in UT1-UTC parameter prediction, and improves the prediction precision and stability of UT1-UTC compared with the traditional method. The method realizes high-precision long-term UT1-UTC prediction, and provides a new UT1-UTC parameter prediction method for the fields of navigation satellite orbit determination, astronomical observation and geophysical research.

Description

Technical Field

[0001] This application relates to the fields of geophysics and related computer science and technology, and in particular to a method for predicting UT1-UTC by combining LS+CNN with consideration of effective angular momentum. Background Technology

[0002] UT1-UTC (Universal Time to Coordinated Universal Time) is a key indicator in Earth Rotation Parameters (ERPs), reflecting the non-uniformity of the Earth's rotation rate. UT1 represents astronomical time based on the Earth's rotation, while UTC is the International Standard Time based on atomic time. Variations in UT1-UTC directly affect the accuracy and reliability of fields such as the Global Positioning System (GPS), satellite navigation, astronomical observation, and geophysical research. For example, in satellite navigation, the accuracy of UT1-UTC directly affects the precise calculation of satellite orbits and positioning accuracy; in geophysical research, variations in UT1-UTC can be used to analyze the dynamic processes within the Earth and the impacts of climate change.

[0003] Currently, UT1-UTC forecasts primarily rely on linear models and numerical analysis methods. Linear models typically predict future values ​​by fitting long-term trends and periodic variations based on historical observation data. For example, the UT1-UTC prediction model published by the International Earth Rotation and Reference Systems Service (IERS) is mainly based on the least squares (LS) method and time series analysis. Earth's rotation is influenced by multiple factors, including angular momentum exchange within the atmosphere, ocean, mantle, and Earth's interior. However, existing UT1-UTC prediction models often neglect the combined effects of these angular momentum exchanges. Although some studies have attempted to incorporate atmospheric angular momentum (AAM) and oceanic angular momentum (OAM) into their models, these studies typically consider only single factors and lack a comprehensive analysis of angular momentum exchange. Summary of the Invention

[0004] The purpose of this application is to provide an LS+CNN combined prediction method for UT1-UTC that takes into account the effective angular momentum. This method can fully consider the combined effect between various angular momentum factors and effectively improve the accuracy of UT1-UTC prediction.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] This application provides a method for predicting UT1-UTC by combining LS+CNN with consideration of effective angular momentum, including:

[0007] Obtain the UT1-UTC basic dataset and the effective angular momentum dataset; the effective angular momentum dataset includes: AAM data, OAM data, hydrospheric angular momentum (HAM) data and sea-level angular momentum (SLAM) data;

[0008] The UT1-UTC basic dataset is converted to the excitation domain to obtain the geodetic angular momentum sequence;

[0009] The geodetic angular momentum residual is obtained by subtracting the effective angular momentum dataset from the geodetic angular momentum sequence.

[0010] Least squares fitting is performed on the geodetic angular momentum residual and the effective angular momentum dataset respectively to obtain the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence;

[0011] The geodetic residual difference is obtained by subtracting the geodetic angular momentum residual fitted sequence from the geodetic angular momentum residual.

[0012] The effective angular momentum residual sequence is obtained by subtracting the effective angular momentum fitting sequence from the effective angular momentum dataset;

[0013] The geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence are extrapolated by the least squares method to obtain the geodetic angular momentum residual fitting extrapolation sequence and the effective angular momentum fitting extrapolation sequence.

[0014] The effective angular momentum residual sequence and the geodetic residual difference are respectively input into a trained convolutional neural network (CNN) model to obtain effective angular momentum residual prediction data and geodetic residual difference prediction data;

[0015] The effective angular momentum fitting extrapolation sequence is added to the effective angular momentum residual prediction data to obtain the effective angular momentum prediction result;

[0016] The geodetic angular momentum residual fitting extrapolation sequence and the geodetic residual difference prediction data are summed to obtain the geodetic angular momentum residual prediction result.

[0017] The effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain the UT1-UTC prediction result.

[0018] Optionally, the UT1-UTC basic dataset is transformed into the excitation domain to obtain a geodetic angular momentum sequence, specifically including:

[0019] Leap seconds are removed from the UT1-UTC basic dataset to obtain the deducted dataset;

[0020] The Earth solid tide correction is applied to the deducted dataset to obtain the UT1R-TAI parameter sequence;

[0021] The UT1R-TAI parameter sequence is converted to the excitation domain to obtain the geodetic angular momentum sequence.

[0022] Optionally, the acquisition of geodetic angular momentum residuals specifically includes:

[0023] Subtracting the effective angular momentum dataset from the geodetic angular momentum sequence yields the geodetic angular momentum residual.

[0024] Optionally, least squares fitting is performed on the geodetic angular momentum residuals and the effective angular momentum datasets respectively to obtain geodetic angular momentum residual fitting sequences and effective angular momentum fitting sequences, and to obtain the difference in geodetic angular momentum residuals and the effective angular momentum residual sequences, specifically including:

[0025] The effective angular momentum dataset is interpolated to an interpolated dataset with the same time interval as the UT1-UTC base dataset;

[0026] The effective angular momentum dataset is interpolated to an interpolated dataset with the same time interval as the UT1-UTC base dataset;

[0027] Least squares fitting is performed on the geodetic angular momentum residuals and the interpolation dataset respectively to obtain the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence.

[0028] The geodetic angular momentum residual is subtracted from the fitted sequence of geodetic angular momentum residuals to obtain the difference in geodetic angular momentum residuals. The effective angular momentum data is then subtracted from the fitted sequence of effective angular momentum to obtain the effective angular momentum residual sequence.

[0029] Optionally, the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence are subjected to least squares extrapolation to obtain the extrapolated sequence of the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence. This mainly includes the construction of the least squares extrapolation model:

[0030]

[0031] Where χ(t) represents the least squares fitting result, ak a represents a constant. k+1 Let t represent the linear term, t represent the UTC time, k be [0,7], i represent the segment number of the least squares fit, and C represent the linear term. i and D i R represents the coefficient corresponding to the periodic term. i e represents a periodic term i This represents zero-mean white noise.

[0032] Extrapolation was performed using the aforementioned geodetic angular momentum residual fitting sequence and effective angular momentum fitting sequence.

[0033] Optionally, the difference in geodetic angular momentum residuals and the geodetic angular momentum residual sequence can be predicted, specifically including:

[0034] Using the neural network model described above, which has 3 convolutional layers and 3×3 kernels in each layer, the difference in geodetic angular momentum residuals and the geodetic angular momentum residual sequence are predicted.

[0035] Optionally, the effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain the UT1-UTC prediction result, specifically including:

[0036] The effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain the geodetic angular momentum prediction result;

[0037] The geodetic angular momentum prediction results were converted into UT1R-TAI prediction results using the Liouville equation;

[0038] The effects of leap seconds and Earth solid tides are added to the UT1R-TAI forecast results to obtain the final UT1-UTC forecast results.

[0039] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0040] This application provides an LS+CNN combined prediction method for UT1-UTC that takes into account effective angular momentum. The method includes: acquiring a UT1-UTC basic dataset and an effective angular momentum dataset; the effective angular momentum dataset includes atmospheric angular momentum data, ocean angular momentum data, land water angular momentum data, and sea level angular momentum data; converting the UT1-UTC basic dataset to an excitation domain to obtain a geodetic angular momentum sequence; subtracting the effective angular momentum dataset from the geodetic angular momentum sequence to obtain a geodetic angular momentum residual; performing least squares fitting on the geodetic angular momentum residual and the effective angular momentum dataset respectively to obtain a geodetic angular momentum residual fitting sequence and an effective angular momentum fitting sequence; subtracting the geodetic angular momentum residual fitting sequence from the geodetic angular momentum residual to obtain a geodetic residual difference; and subtracting the effective angular momentum dataset from the effective angular momentum dataset... The effective angular momentum fitting sequence is used to obtain an effective angular momentum residual sequence. The geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence are extrapolated using the least squares method to obtain a geodetic angular momentum residual fitting extrapolation sequence and an effective angular momentum fitting extrapolation sequence. The effective angular momentum residual sequence and the geodetic residual difference are input into a trained CNN model to obtain effective angular momentum residual prediction data and geodetic residual difference prediction data. The effective angular momentum fitting extrapolation sequence is added to the effective angular momentum residual prediction data to obtain an effective angular momentum prediction result. The geodetic angular momentum residual fitting extrapolation sequence and the geodetic residual difference prediction data are added to obtain a geodetic angular momentum residual prediction result. The effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain a UT1-UTC prediction result. Because this application transforms the UT1-UTC basic dataset to the excitation domain, it can reflect all the influencing factors. Moreover, the geodetic angular momentum sequence obtained by transforming it to the excitation domain and the effective angular momentum dataset are on the same level. This allows the comprehensive effect between various angular momentum factors to be fully considered in the subsequent least squares extrapolation and convolutional neural network prediction processes, effectively improving the accuracy of UT1-UTC prediction. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A flowchart illustrating an LS+CNN combined prediction UT1-UTC method considering effective angular momentum, provided as an embodiment of this application;

[0043] Figure 2 A schematic diagram of the basic sequence and preprocessing process of the UT1-UTC basic dataset from EOP 20C04 provided for an embodiment of this application;

[0044] Figure 3 The interpolation results of the axial components of AAM, OAM, HAM, and SLAM and the total EAM are shown in a schematic diagram according to an embodiment of this application.

[0045] Figure 4 A schematic diagram of CNN model parameter training for residual difference sequence of geodetic angular momentum sequence provided in an embodiment of this application;

[0046] Figure 5 A statistical diagram illustrating the training accuracy of CNN model parameters for residual difference sequences of geodetic angular momentum sequences provided in an embodiment of this application;

[0047] Figure 6 A schematic diagram of CNN model parameter training for residual sequences of an effective angular momentum dataset provided in an embodiment of this application;

[0048] Figure 7 This is a statistical diagram illustrating the training accuracy of CNN model parameters for the residual sequence of an effective angular momentum dataset provided in an embodiment of this application. Detailed Implementation

[0049] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0050] The influence of land-water angular momentum (HAM) and sea-level angular momentum (SLAM) on UT1-UTC has not been adequately considered in existing technologies. This incomplete consideration of angular momentum exchange limits the accuracy and reliability of UT1-UTC forecasts.

[0051] With the continuous advancement of Earth rotation observation technology, the amount and accuracy of UT1-UTC observation data have significantly improved. However, existing forecasting methods have obvious shortcomings in adapting to high-precision, high-frequency observation data. For example, while the least squares method can effectively extract long-term trends and periodic changes, it performs poorly when dealing with nonlinear characteristics and short-term fluctuations. Furthermore, traditional time series analysis methods (such as the ARIMA model) struggle to accurately capture the changing patterns of UT1-UTC when faced with complex Earth rotation dynamics, leading to large forecast errors.

[0052] In recent years, deep learning technology has made significant progress in the field of time series forecasting, especially in handling complex nonlinear relationships and large-scale data. Convolutional Neural Networks (CNNs), as a powerful feature extraction tool, can automatically learn hidden patterns and regularities in input data and have been successfully applied to time series forecasting in various fields such as meteorology and finance. However, research on applying CNN technology to UT1-UTC forecasting is still in its early stages and has not yet fully incorporated the physical background of Earth's rotation dynamics.

[0053] Effective angular momentum (EAM) is a comprehensive reflection of the angular momentum exchange between the atmosphere, ocean, and mantle within the Earth system. It reflects the influence of the interactions between the Earth's spheres on the Earth's rotation rate. Studies have shown that changes in atmospheric angular momentum (AAM) are mainly caused by atmospheric circulation and pressure distribution, changes in ocean angular momentum (OAM) are closely related to ocean circulation and tidal processes, while terrestrial water angular momentum (HAM) and sea-level angular momentum (SLAM) are related to rheological processes within the Earth. These angular momentum exchange processes have a significant driving effect on changes in UT1-UTC.

[0054] Although existing research recognizes the impact of angular momentum exchange on UT1-UTC, most studies focus only on single angular momentum factors (such as AAM or OAM), neglecting their combined effects. Furthermore, current research lacks a systematic analysis and optimization of angular momentum weighting coefficients when incorporating angular momentum exchange into UT1-UTC prediction models. This incomplete consideration of angular momentum prevents existing models from accurately reflecting Earth's rotational dynamics, thus affecting the accuracy of UT1-UTC predictions.

[0055] To overcome the shortcomings of existing technologies, this application proposes a combined LS+CNN method for predicting UT1-UTC that takes into account effective angular momentum. This method combines least squares (LS) and convolutional neural networks (CNNs), and fully considers the combined influences of the atmosphere, ocean, land water, and sea level to establish a more accurate UT1-UTC prediction model. Specifically, this application incorporates the physical background of angular momentum exchange within the Earth system into the model by calculating effective angular momentum (EAM), while utilizing the nonlinear feature extraction capabilities of CNNs to uncover hidden patterns in the UT1-UTC time series. Through this combined approach, this application aims to improve the prediction accuracy of UT1-UTC and provide a more reliable Earth rotation parameter prediction technology for related applications.

[0056] This application provides a method for predicting the parameters of Earth's rotation rate variation (UT1-UTC) using a combination of least squares (LS) and convolutional neural network (CNN) models involving effective angular momentum, aiming to improve the prediction accuracy of UT1-UTC. This method improves the prediction accuracy and reliability of UT1-UTC parameters by combining LS and CNN models and fully considering the influence of EAM on UT1-UTC. The prediction method includes: acquiring UT1-UTC parameters and EAM sequences; preprocessing the UT1-UTC sequences, including subtracting jump seconds and tidal terms, to obtain the UT1R-TAI base sequence; subsequently, constructing LS and CNN models; and finally, combining the LS fitting results with the output of the CNN model to obtain high-precision UT1-UTC prediction values. The LS+CNN combined model described in this application has high accuracy and reliability in long-term UT1-UTC prediction. Compared with traditional methods, it effectively overcomes the shortcomings of traditional linear models in predicting UT1-UTC parameters and effective angular momentum, while improving the prediction accuracy and stability of UT1-UTC. This method enables high-precision long-term forecasting of UT1-UTC, providing a novel method for forecasting UT1-UTC parameters in fields such as navigation satellite orbit determination, astronomical observation, and geophysical research.

[0057] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] In one exemplary embodiment, such as Figure 1 As shown, an LS+CNN combined prediction method for UT1-UTC that takes into account effective angular momentum is provided, including the following steps S1 to S11. Wherein:

[0059] S1. Obtain the UT1-UTC basic dataset and the effective angular momentum dataset; the effective angular momentum dataset includes: atmospheric angular momentum data, ocean angular momentum data, land water angular momentum data and sea level angular momentum data.

[0060] In this embodiment, the UT1-UTC basic dataset and related effective angular momentum datasets were collected and organized. The UT1-UTC sequence was obtained from the IERS website, and the effective angular momentum was obtained from the GFZ website. It mainly includes atmospheric angular momentum (AAM), ocean angular momentum (OAM), land water angular momentum (HAM), and sea level angular momentum (SLAM) data.

[0061] First, effective angular momentum data for AAM, OAM, HAM, and SLAM were collected, and UT1-UTC sequences with a time interval of 1 day (UT1-UTC basic dataset) were downloaded from the International Earth Rotation and Reference System Service (IERS) website.

[0062] It is worth noting that historical data was used when training the model.

[0063] As an optional implementation, this embodiment can use the following method to collect data, gathering the UT1-UTC basic dataset and the effective angular momentum dataset:

[0064] The UT1-UTC basic dataset was obtained from the International Earth Rotation and Reference Systems Service (IERS), covering the period from January 1999 to April 2023, with a sampling interval of once per day.

[0065] Atmospheric angular momentum (AAM) data: sourced from the ERA5 global atmospheric reanalysis product, with the same time span as the UT1-UTC data.

[0066] Ocean angular momentum (OAM) data: obtained from the GOCE ocean circulation model, with the same time span as the UT1-UTC data.

[0067] Land water angular momentum (HAM) data and sea level angular momentum (SLAM) data: calculated using geophysical models, with a time span consistent with the UT1-UTC base dataset. Figure 2 and Figure 3 As shown.

[0068] Then, data preprocessing steps are performed, including subtracting leap seconds from the UT1-UTC base dataset and calculating the effects of Earth solid tides, as well as interpolation processing of the effective angular momentum dataset, specifically including the following steps:

[0069] Step ①: Before forecasting the UT1-UTC sequence, it is necessary to subtract leap seconds and correct for Earth solid tides. Leap seconds are corrected using values ​​provided by the Earth's rotation center. The Earth solid tide correction model is as follows:

[0070]

[0071] Wherein, δUT1 is the Earth solid tidal correction term. B i C i α is given in columns 7 and 8 of Table 8.1 of the 2010 IERS Convention Standard. ij It is the i-th tide α in Table 8.1 j Integer multiples of α ij α represents the fundamental nutation angle coefficient. j These are the parameters for the symmetry of the sun and moon, with j ranging from column 1 to 5.

[0072] Step 2: Interpolate the collected effective angular momentum dataset to the same intervals and times as the UT1-UTC base dataset.

[0073] Interpolation processing of the effective angular momentum dataset includes: interpolating the axial components of AAM, OAM, HAM, and SLAM to time intervals consistent with the UT1-UTC base dataset; and merging the interpolated angular momentum components to form a unified time series. Preprocessing of the UT1-UTC base dataset includes: subtracting leap seconds and applying Earth solid tide correction.

[0074] S2. Convert the UT1-UTC basic dataset to the excitation domain to obtain the geodetic angular momentum sequence.

[0075] In this embodiment, the preprocessed UT1R-TAI parameter sequence (preprocessed result of the UT1-UTC basic dataset; UT1R-TAI parameter = (UT1-UTC) - leap second - Earth solid tidal correction) is transformed into the excitation domain to obtain the geodetic angular momentum sequence. The UT1R-TAI parameter sequence can be transformed into the excitation domain using the Liouville equation, as shown in the following formula:

[0076]

[0077] Among them, UT1R is the world time after further correction of periodic tidal changes based on UT1, TAI represents atomic time, UT1R-TAI is the sequence obtained by preprocessing UT1-UTC, Ψ is the geodetic angular momentum function, Ω is the average angular velocity of the earth with a value of 7.292115×10-5rands-1; t represents UTC time.

[0078] S3. Subtract the effective angular momentum dataset from the geodetic angular momentum sequence to obtain the geodetic angular momentum residual.

[0079] S4. Perform least squares fitting on the geodetic angular momentum residual and the effective angular momentum dataset respectively to obtain the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence.

[0080] Construction based on a linear model: In this embodiment, the interpolated datasets of the geodetic angular momentum residuals and the effective angular momentum datasets are respectively subjected to least squares fitting to obtain the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence, as shown in the formula:

[0081]

[0082] Where χ(t) represents the least squares fitting result, a k a represents a constant. k+1Let t represent the linear term, t represent the UTC time, k be [0,7], i represent the segment number of the least squares fit, and C represent the linear term. i and D i R represents the coefficient corresponding to the periodic term. i The periodic term, considered in the fitting process, includes annual oscillations, semi-annual oscillations, 1 / 3-year oscillations, and 13.7-day oscillations, typically taking values ​​of 1 year, 1 / 2 year, 1 / 3 year, and 13.7 / 365 years, e. i This represents zero-mean white noise.

[0083] In this embodiment, when constructing a linear model, the periodic terms of the constructed linear model when using LS to fit the effective angular momentum include annual oscillation, semi-annual oscillation, 1 / 3 annual oscillation, and 13.7-day oscillation, typically taking values ​​of 1 year, 1 / 2 year, 1 / 3 year, and 13.67 / 365 years.

[0084] S5. The geodetic residual difference is obtained by subtracting the geodetic angular momentum residual fitted sequence from the geodetic angular momentum residual.

[0085] S6. Subtract the effective angular momentum fitting sequence from the effective angular momentum dataset to obtain the effective angular momentum residual sequence.

[0086] S7. Extrapolate the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence using the least squares method to obtain the geodetic angular momentum residual fitting extrapolation sequence and the effective angular momentum fitting extrapolation sequence.

[0087] S8. Input the effective angular momentum residual sequence and the geodetic residual difference into the trained neural network model respectively to obtain the effective angular momentum residual prediction data and the geodetic residual difference prediction data.

[0088] The training process of the convolutional neural network model includes:

[0089] Obtain the training UT1-UTC basic dataset and the training effective angular momentum dataset; transform the training UT1-UTC basic dataset to the excitation domain to obtain the geodetic angular momentum sequence; subtract the geodetic angular momentum sequence from the effective angular momentum dataset to obtain the geodetic angular momentum residual sequence. Fit the geodetic angular momentum residual sequence and the effective angular momentum sequence using least squares to obtain the corresponding fitting term and residual term, respectively; use the obtained residual term as the input sequence of the CNN model; normalize the training residual sequence to obtain normalized data; train the CNN model using the normalized data to obtain a suitable CNN model.

[0090] In this embodiment, a CNN model is constructed using residual sequences, and the parameters are trained. The specific training process of the CNN model is as follows:

[0091] (1) Normalization: Normalize the residual sequence to a suitable range (such as [0, 1] or [-1, 1]) to improve the convergence speed of the model.

[0092] (2) Construct a CNN model. Based on the characteristics of geodetic angular momentum and effective angular momentum fitting residual sequences, a suitable CNN model structure is constructed through network training. The model typically includes the following parts:

[0093] Convolutional layers: used to extract local features from the residual sequence. Multiple convolutional layers can be set, each using a small-sized convolutional kernel (e.g., 3×3).

[0094] Pooling layers: Used to reduce the dimensionality of features and decrease computational cost. Max pooling or average pooling are typically used.

[0095] Fully connected layer: Combines the features extracted by the convolutional layer and pooling layer to output the final prediction result.

[0096] In this embodiment, the CNN model has three convolutional layers, each with a 3×3 kernel size. Max pooling is used in the pooling layers, and the number of neurons in the fully connected layers is adjusted based on cross-validation results. The CNN model is trained using historical data, and network parameters are optimized through backpropagation. Mean squared error (MSE) is used as the loss function, and model performance is evaluated through cross-validation. Hyperparameters (such as learning rate, number of convolutional kernels, and number of layers) are adjusted to improve the model's generalization ability. The error statistics after training in this embodiment are as follows: Figure 4 , Figure 5 , Figure 6 and Figure 7 As shown.

[0097] S9. Add the effective angular momentum fitting extrapolation sequence to the effective angular momentum residual prediction data to obtain the effective angular momentum prediction result.

[0098] S10. The geodetic angular momentum residual fitting extrapolation sequence and the geodetic residual difference prediction data are summed to obtain the geodetic angular momentum residual prediction result.

[0099] S11. Add the effective angular momentum prediction result to the geodetic angular momentum residual prediction result to obtain the geodetic angular momentum prediction result; use the Liouville equation and the effects of adding leap seconds and Earth solid tides to obtain the final prediction result of UT1-UTC.

[0100] Finally, the accuracy of the forecast results was evaluated by comparing the mean absolute error (MAE) with the international standard sequence Bulletin A. The accuracy evaluation standard for UT1-UTC parameters is based on the following formula:

[0101]

[0102] Among them, MAE j Let P be the precision of the j-th time series, n be the length of the time series, i be the length of the predicted time series, and P be the precision of the predicted time series. i For the i-th forecast value, O i Let i be the i-th truth value.

[0103] This embodiment, by employing a combined LS+CNN model, fully leverages the advantages of both, utilizing the interpretability of linear models while introducing the powerful modeling capabilities of deep learning for complex time-series data. Specific effects and benefits include:

[0104] Improving UT1-UTC Forecast Accuracy: This embodiment significantly improves the forecast accuracy of UT1-UTC by using a combination of LS+CNN that takes into account effective angular momentum. Traditional methods (such as least squares and time series analysis) have limitations in handling complex nonlinear changes and short-term fluctuations. This embodiment combines the linear fitting capability of least squares with the nonlinear feature extraction capability of convolutional neural networks, enabling it to more accurately capture the complex changing patterns in the UT1-UTC time series. Experimental results show that the method in this embodiment performs excellently in medium- to long-term (e.g., quarterly to annual) forecasts, with a significantly reduced forecast error compared to traditional methods.

[0105] Enhancing the model's generalization ability: This embodiment significantly enhances the model's generalization ability through techniques such as data preprocessing, model training optimization, and cross-validation. Specifically: Data normalization eliminates the dimensional differences between different data sources, improving the model's training efficiency and stability. CNN's convolutional and pooling layers can automatically extract local and global features from the input data, effectively capturing short-term fluctuations and long-term trends in time series. Cross-validation optimizes hyperparameters (such as learning rate, number of convolutional kernels, and number of layers) to ensure good model performance on independent test sets. Experimental results show that the method in this embodiment outperforms existing techniques on independent test sets and can adapt to changes in different time scales and data characteristics.

[0106] Advancing the Development of Earth Rotation Parameter Prediction Technology: This embodiment proposes a novel UT1-UTC prediction method, combining the advantages of traditional numerical analysis and deep learning techniques, providing a new technical means for high-precision prediction of Earth rotation parameters. This method is not only theoretically innovative but also performs excellently in practical applications. By introducing the concept of effective angular momentum, this embodiment fully considers the impact of angular momentum exchange within the Earth system on UT1-UTC, overcoming the shortcomings of existing technologies that do not comprehensively consider angular momentum exchange. Furthermore, this embodiment provides new ideas and methods for research in related fields, promoting the development of Earth rotation parameter prediction technology.

[0107] In summary, this invention proposes a combined LS+CNN method for predicting UT1-UTC that takes into account effective angular momentum. By combining the advantages of traditional linear models and deep learning techniques, it significantly improves the prediction accuracy of UT1-UTC and enhances the model's generalization ability. This invention not only provides high-precision Earth rotation parameter prediction technology support for related fields but also promotes the development of Earth rotation parameter prediction technology, possessing significant scientific value and practical application value.

[0108] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0109] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0110] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0111] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0112] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A combined LS+CNN prediction method for UTI-UTC considering effective angular momentum, characterized in that, include: Obtain the UT1-UTC basic dataset and the effective angular momentum dataset; The effective angular momentum dataset includes: atmospheric angular momentum data, ocean angular momentum data, land water angular momentum data, and sea level angular momentum data; The UT1-UTC basic dataset is converted to the excitation domain to obtain the geodetic angular momentum sequence; The geodetic angular momentum residual is obtained by subtracting the effective angular momentum dataset from the geodetic angular momentum sequence. Least squares fitting is performed on the geodetic angular momentum residual and the effective angular momentum dataset respectively to obtain the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence; The geodetic residual difference is obtained by subtracting the geodetic angular momentum residual fitted sequence from the geodetic angular momentum residual. The effective angular momentum residual sequence is obtained by subtracting the effective angular momentum fitting sequence from the effective angular momentum dataset; The geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence are extrapolated by the least squares method to obtain the geodetic angular momentum residual fitting extrapolation sequence and the effective angular momentum fitting extrapolation sequence. The effective angular momentum residual sequence and the geodetic residual difference are respectively input into the trained CNN model to obtain the effective angular momentum residual prediction data and the geodetic residual difference prediction data. The effective angular momentum fitting extrapolation sequence is added to the effective angular momentum residual prediction data to obtain the effective angular momentum prediction result; The geodetic angular momentum residual fitting extrapolation sequence and the geodetic residual difference prediction data are summed to obtain the geodetic angular momentum residual prediction result. The effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain the UT1-UTC prediction result.

2. The LS + CNN combined UT1-UTC prediction method considering the angular momentum according to claim 1, wherein, The UT1-UTC basic dataset is transformed into the excitation domain to obtain the geodetic angular momentum sequence, specifically including: Leap seconds are removed from the UT1-UTC basic dataset to obtain the deducted dataset; The Earth solid tide correction is applied to the deducted dataset to obtain the UT1R-TAI parameter sequence; The UT1R-TAI parameter sequence is converted to the excitation domain to obtain the geodetic angular momentum sequence.

3. The LS + CNN combined UT1-UTC prediction method considering the angular momentum according to claim 1, wherein, Least squares fitting is performed on the geodetic angular momentum residuals and the effective angular momentum dataset to obtain a geodetic angular momentum residual fitting sequence and an effective angular momentum fitting sequence, respectively. The geodetic residual difference is obtained by subtracting the geodetic angular momentum residual fitting sequence from the geodetic angular momentum residuals. The effective angular momentum residual sequence is obtained by subtracting the effective angular momentum fitting sequence from the effective angular momentum dataset, specifically including: The effective angular momentum dataset is interpolated to an interpolated dataset with the same time interval as the UT1-UTC base dataset; Least squares fitting is performed on the geodetic angular momentum residuals and the interpolation dataset respectively to obtain the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence. Subtract the geodetic angular momentum residual from the fitted sequence of geodetic angular momentum residuals to obtain the difference in geodetic angular momentum residuals. Subtracting the effective angular momentum data from the effective angular momentum fitting sequence yields the effective angular momentum residual sequence.

4. The LS + CNN combined UT1-UTC prediction method considering the angular momentum according to claim 1, wherein, The process of extrapolating the geodetic angular momentum residual fitting sequence and the effective angular momentum fitting sequence using the least squares method to obtain the geodetic angular momentum residual fitting extrapolation sequence and the effective angular momentum fitting extrapolation sequence includes: Construction of the least squares extrapolation model: Where χ(t) represents the least squares fitting result, a k a represents a constant. k+1 Let t represent the linear term, t represent the UTC time, k be [0,7], i represent the segment number of the least squares fit, and C represent the linear term. i and D i R represents the coefficient corresponding to the periodic term. i e represents a periodic term i This represents zero-mean white noise; Extrapolation was performed using the aforementioned geodetic angular momentum residual fitting sequence and effective angular momentum fitting sequence.

5. The LS + CNN combined UT1-UTC prediction method considering the angular momentum according to claim 1, wherein, The effective angular momentum residual sequence and the geodetic residual difference are respectively input into the trained CNN model to obtain effective angular momentum residual prediction data and geodetic residual difference prediction data, specifically including: The neural network model has three convolutional layers, with each layer having a kernel size of 3×3, and is used to predict the above geodetic residual difference and effective angular momentum residual sequence.

6. The LS + CNN combined UT1-UTC prediction method considering the angular momentum according to claim 1, wherein, The effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain the UT1-UTC prediction result, which specifically includes: The effective angular momentum prediction result is added to the geodetic angular momentum residual prediction result to obtain the geodetic angular momentum prediction result; The geodetic angular momentum prediction results were converted into UT1R-TAI prediction results using the Liouville equation, and the effects of leap seconds and Earth solid tides were added to the UT1R-TAI prediction results to obtain the final UT1-UTC prediction results.

Images