Data assimilation analysis system based on object-oriented framework

Abstract

The invention discloses a data assimilation analysis system based on an object-oriented framework, and the system comprises the steps: obtaining assimilation parameters, defining an analysis variable space for the assimilation parameters through an unstructured grid framework module, and carrying out the distribution collection and boundary exchange calculation through a geometric grid and parallel algorithm communication module MPDD, a control variable, an observation variable, a covariance matrix, various transformation operators and a solver module are constructed, a multivariable price function and a price function gradient are established, the system transmits the multivariable price function and the price function gradient to a universal minimization solver, an analysis field is made to approach an observation result through iterative solution, and the price function gradient is obtained. Other points except the observation position are diffused outwards in the mode of the covariance matrix to influence other mode grid points, and an assimilation analysis field result is obtained. According to the method, rapid expansion and easy replacement of various different space variables, observation operators, covariance matrix schemes, multivariable solvers and the like are realized, and the flexibility and adaptability of a data assimilation analysis system are improved.

Description

Technical Field

[0001] The invention relates to the field of data assimilation, and in particular to a meteorological data assimilation analysis system with an object-oriented framework structure. Background Art

[0002] The meteorological data assimilation and analysis system is a system used to integrate and analyze meteorological observation data and numerical model forecast data. Its main purpose is to accurately characterize atmospheric conditions or provide accurate initial fields for numerical forecast models to improve the accuracy of meteorological forecasts by combining observation data with numerical model forecast data through a series of mathematical and statistical methods. The meteorological data assimilation and analysis system uses observation data (such as meteorological station observation data, radar data, satellite data, etc.) and numerical model forecast data as input, and creates an optimal initial field through data assimilation technology. This initial field contains the actual state of meteorological phenomena and can better reflect the current weather conditions. By analyzing the assimilated data, the meteorological data assimilation and analysis system can provide accurate weather condition analysis, or provide accurate initial fields for the forecast system to make forecasts more accurate, including short-term weather forecasts and long-term climate forecasts. It can help meteorological forecasters better understand the trends and characteristics of weather changes, provide more accurate forecast information, and help reduce dependence on manual experience and improve the accuracy and reliability of forecasts. The meteorological data assimilation and analysis system has important application value in the fields of weather forecasting, climate research, meteorological disaster warning, etc. It can help improve society's ability to respond to disastrous weather changes, and also plays an important guiding role in decision-making and planning in industries such as agriculture, transportation, and energy.

[0003] The meteorological data assimilation and analysis system is a very complex software system that needs to process a large amount of real-time observation data, including radar, satellite, etc. At the same time, there are many choices for assimilation methods to deal with specific assimilation problems. Traditional meteorological data assimilation systems are process-oriented. In the process of increasing observations, improving assimilation methods and maintaining business operations, these systems are very difficult to respond flexibly. The software transformation of meteorological data assimilation has become an urgent problem that needs to be solved by various business centers around the world.

[0004] Object-Oriented Programming (OOP) is a software design method and programming paradigm with objects as the core concept. This method imitates objects in the real world and encapsulates data (attributes) and methods (behaviors) for operating these data in objects. The functions of the entire system are completed through the interaction between objects.

[0005] In object-oriented programming, an object represents an entity with a specific state and behavior. The state is represented by attributes, while the behavior is implemented by methods. Objects are defined by classes, which are blueprints of objects and specify their properties and methods. Object-oriented programming languages ​​provide mechanisms such as encapsulation, inheritance, and polymorphism to support OOP.

[0006] Encapsulation means hiding the implementation details of an object and exposing only limited interfaces to the outside world, thereby protecting the internal state of the object from external interference.

[0007] Inheritance allows a class (subclass) to inherit the properties and methods of another class (parent class or superclass), which not only supports code reuse but also makes code expansion easier.

[0008] Polymorphism means that objects can exhibit different behaviors depending on the context, which increases the flexibility of the code.

[0009] Object-oriented design brings many benefits, including improved code reusability, maintainability, scalability and understandability. It allows complex systems to be decomposed into a series of smaller and easier to manage objects, making program design and implementation more modular and flexible.

[0010] Object-oriented methodology has been widely used in various fields of software development, including desktop applications, Web applications, mobile applications, etc., and is one of the cornerstones of modern software engineering.

[0011] The data assimilation and analysis system based on the object-oriented framework is a meteorological data assimilation and analysis system designed and implemented using the object-oriented programming concept. It takes meteorological observation data and numerical model forecast data as the attributes of objects, encapsulates the methods of data assimilation and analysis as the methods of objects, and completes the task of meteorological data assimilation and analysis through the interaction between objects. In the data assimilation and analysis system based on the object-oriented framework, observation variables, pattern variables, control variables, observation operators, pattern operators, input and output modules and minimization algorithms are encapsulated as objects, including temperature, humidity, wind speed, precipitation, etc. At the same time, the system interactively combines and defines some methods based on objects to perform data assimilation and analysis, such as Kalman filtering, variational assimilation and other methods. The data assimilation and analysis system based on the object-oriented framework completes the process of data assimilation and analysis through the interaction between objects. The data assimilation and analysis system based on the object-oriented framework has the advantages of object-oriented programming, including code reusability, maintainability and scalability. It can encapsulate data and methods as objects to improve the modularity and flexibility of the system. At the same time, the object-oriented design concept can also better reflect the actual needs of meteorological data assimilation and analysis, and improve the efficiency and accuracy of the system.

[0012] This approach may have some potential design flaws: process-oriented design usually organizes code in linear steps, which makes it difficult to cope with increasing complexity. Such a code structure may make it difficult to maintain and modify, especially when the project becomes larger and more complex. Process-oriented design may lack flexibility in dealing with changes and new features. When new features need to be added or existing features need to be adjusted, this design approach may require the refactoring of a large amount of code. Process-oriented code is often centered on specific tasks and steps, making it difficult to reuse. This may lead to code duplication and difficulty in sharing and reusing code segments between different parts. Process-oriented design often leads to high coupling between codes, that is, strong dependencies between different parts. This may increase the fragility of the code, and one change may affect the functionality of the entire system. Therefore, there is an urgent need for a data assimilation analysis system based on an object-oriented framework. Summary of the invention

[0013] The purpose of the present invention is to provide a data assimilation analysis system based on an object-oriented framework to solve the problems existing in the existing assimilation analysis system. Through the multi-module design assimilation analysis framework proposed in the technical solution, the rapid construction of each module in the assimilation application and the rapid construction of multiple assimilation schemes can be realized. In addition, the technical solution also splits each module and proposes a polymorphic design of a unified interface, which reduces the coupling between modules and improves the scalability of the system.

[0014] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0015] The present invention comprises a plurality of general modules, and the general modules are constructed by an object-oriented framework.

[0016] The specific steps include:

[0017] Obtain assimilation parameters, which use the Geometry grid framework module to define the geometry or physical space, and use the geometry grid and parallel algorithm and the MPDD parallel communication module to distribute and collect and exchange boundary calculations in parallel. The assimilation parameters include the control variables to be solved and analyzed, the input observation variables, the background error covariance matrix, the observation error covariance matrix, the spatial interpolation operator from the pattern space to the observation space, the transformation operator from the pattern space to the observation space, and the transformation operator in the control variable space.

[0018] The multivariable price function and the price function gradient are passed to a general minimization solver, and the assimilated analysis field result is obtained through iterative solution and then output to the initial field file of the numerical prediction model through a model coupler;

[0019] The price function and the price function gradient are constructed as follows:

[0020]

[0021] When using 4D variational data assimilation, the price function and the price function gradient can be constructed as follows:

[0022]

[0023]

[0024] When using the three-dimensional ensemble variation assimilation scheme, the price function and the price function gradient can be constructed as follows:

[0025]

[0026] The minimization algorithm uses the gradient of the price function in the above formula to make the analysis field converge to the observation result. At the same time, the multivariate price function makes other points outside the observation position diffuse outward in the mode of the covariance matrix to affect other pattern grid points.

[0027] Furthermore, the spatial information is encapsulated as a geometric algorithm library, and the parallel computing operation is encapsulated as a parallel algorithm library. When the pattern integral transformation is needed to deduce the time when the independent variable is located to the time when the observed data is located, the pattern integral operator is expressed as Its forward nonlinear integral operator is expressed as The tangent operator is MX and the adjoint operator is M T X, so the mode variable X0 at time 0 can be transformed to time 1 through the mode transformation operator:

[0028]

[0029] Then it can be transformed to time n:

[0030]

[0031] Thus, the observation at time n is minimized:

[0032]

[0033] Taking the X0 model variable as the independent variable, the model forecast variable at the nth moment is made to approach the optimal solution form of the observation at the nth moment.

[0034] Furthermore, the price function part can switch to use 3DVar, 4DVar, or 3DEnVar solutions.

[0035] Furthermore, the solver can adopt a FRCG solver or a LBFGS solver.

[0036] Furthermore, the correlation covariance matrix can be selected to use a recursive filtering covariance matrix, a Laplace covariance matrix, or a collective covariance matrix.

[0037] Furthermore, the mode integral operator, etc., can choose to use a static equilibrium mode or a non-static equilibrium mode.

[0038] Furthermore, the observation input module inputs the data into the container of the observation data according to the dimensions of longitude, latitude, altitude, time and value.

[0039] Furthermore, the observation operator module transforms the variable data in the system into the observation space to ensure the spatial consistency of the observation data.

[0040] Furthermore, the mode coupler converts the mode input variables into grid standard variables generated by Geometry to ensure the spatial consistency of the analysis data.

[0041] In a second aspect, an embodiment of the present application further provides an electronic device, including:

[0042] A processor; and a memory arranged to store computer executable instructions, which when executed cause the processor to perform the method steps described in the first aspect.

[0043] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0044] The beneficial effects of the present invention are:

[0045] The present invention is a data assimilation analysis system based on an object-oriented framework. Compared with the prior art, the present invention has the following technical effects:

[0046] 1. Through the technical solution of the present invention, it is possible to easily realize rapid expansion and easy replacement of various spatial variables, observation operators, covariance matrix schemes and multivariate solvers, and the present invention improves the flexibility and adaptability of the data assimilation analysis system.

[0047] 2. According to the design of the mode coupler of the present invention, it is possible to connect with a variety of different forecast modes, thereby improving the universality of the data assimilation and analysis system.

[0048] 3. This technical solution also realizes the separation of module design and assimilation method, so that the assimilation system can quickly realize various assimilation analysis applications as needed, thereby bringing higher system efficiency and better application effect.

[0049] 4. The present invention adopts a polymorphic design of a unified interface, so that each module can quickly compare and evaluate with other similar solutions while realizing its own functions, thereby enhancing the integrated coordination and robustness of the system.

[0050] In general, the technical solution of the present invention can make the data assimilation analysis system more scalable, improve the ease of use and research and development potential of various data assimilation applications, and has great application value and practical significance in fields such as numerical weather forecasting. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A flow chart of the steps of a data assimilation analysis system based on an object-oriented framework of the present invention;

[0052] Figure 2 A schematic diagram of the application process of an object-oriented data assimilation analysis system based on an object-oriented framework in a regional model forecast process of the present invention;

[0053] Figure 3 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0054] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0055] The present invention provides a data assimilation analysis system based on an object-oriented framework, comprising the following steps:

[0056] By constructing a series of general modules and related interface structures such as geometric grids, regional decomposition parallelism, control variables, observation variables, covariance matrices, various transformation operators, solvers, etc., a flexible and replaceable data assimilation system is constructed. The general operator structures are as follows:

[0057]

[0058]

[0059]

[0060] in:

[0061] Geometric grid and parallel construction module: The underlying encapsulation of geometric grid and parallel algorithm is the Geometry underlying library and the MPDD software package.

[0062] Geometry defines the physical grid characteristics of the assimilation analysis, mainly including the type of geographic grid used and the grid configuration, including: the horizontal grid configuration used (longitude and latitude grid, projection grid, triangular irregular grid); vertical grid configuration (uniform, non-uniform, terrain coordinate system information); the location of each grid point; the preset coefficient information required for the calculation of the grid point derivative, etc. The Geometry class fully defines the physical space where the assimilation analysis is located, and is also a member variable of all variable containers.

[0063] The parallel software class MPDD mainly defines the parallel computing partition in the physical space. Generally speaking, various computing functions in the physical space need to be evenly distributed to each computing node for parallel computing, and adjacent area calculations such as derivatives involved in the calculation on the boundary need to rely on parallel computing functions with specific functions. Therefore, the regional decomposition parallel computing library mainly includes three member functions: 1. Data distribution, data is distributed from a single process to each distributed process; 2. Data aggregation, data is aggregated from each distributed process to the main process; 3. Data exchange in the boundary area, updating the data in the buffer area set up at the edge of the data to achieve parallel computing.

[0064] A container for storing variables (control variables, observation variables, model variables, etc.); the variables for assimilation analysis are the state variable information of the atmospheric space. There are two main types of variables: model state variables and observation variables.

[0065] The most basic combination of mode state variables is the combination of basic atmospheric quantities composed of wind speed, temperature, surface pressure, and specific humidity (u, v, T, ps, q). The construction of the container is based on the physical space information given by Geometry to create a given type of atmospheric state variable. The mode state variable is defined as: X, X itself is also the independent variable analysis field information that needs to be solved. X is a packaged collection of multiple state variables. When solving the price function, X participates in numerical calculations as a multi-variable vector. Therefore, X itself has basic operations and can perform interface encapsulation of operators such as addition, subtraction, dot multiplication, and division by scalars.

[0066] Observed variables are direct variables obtained from atmospheric observation instruments. Atmospheric observations include some conventional direct observed variables, such as wind field, temperature, water vapor, air pressure and other information on the surface obtained by ground station observations. They also usually include some remote sensing-related brief observations, such as radar combined reflectivity, Doppler radial wind speed, etc. Represented as Y, the observed variables have the characteristics of random and irregular distribution. Therefore, the observed data is saved in a single sequence. In addition to the observed values, it also includes the latitude, longitude and altitude information of the observation location. Observed variables also participate in numerical calculations as vectors, and also have interface encapsulation of operators such as addition, subtraction, dot multiplication and division by scalar.

[0067] Observation operator; the transformation involved in the observation operator is better than the observation variable has the characteristics of random, irregular and sparse distribution, and some remote sensing observations are not direct observations. Therefore, the transformation from pattern space to observation variable space needs to rely on transformation operators to achieve. The pattern variable X and the observation variable Y are vectors of different dimensions, so the transformation operator can be expressed as a matrix left multiplication form. The transformation operators related to the observation variable transformation can usually be divided into two categories: interpolation transformation operators and observation forward transformation operators.

[0068] The interpolation transformation operator is a direct interpolation operator that interpolates from the pattern space to the observation space. This operator is suitable for conventional observations such as ground stations and soundings. By interpolating from the adjacent grid points in the pattern space to the location of the observation station, the optimal estimation information of the pattern variable in the observation variable space can be obtained. The forward operator of the interpolation operator can be expressed as The interpolation operator has two interfaces: left multiplication and concomitant left multiplication, expressed as and H T X.

[0069] If the observed spatial variable is an indirect remote sensing variable, such as radar reflectivity, then the model variable (such as q r The transformation of the rainwater mixing ratio to the observed variable requires a positive evolution operator to convert the physical variable to the instrument's observed variable, that is, from q r The reflectivity dBZ is obtained by forward calculation, and an interpolation transformation operator is superimposed to transform from the model variable space to the observation variable space. This forward modeling process can usually be expressed as Therefore, the observed forward transformation operator has three interfaces: forward calculation And the linearized forward calculation KX, and the forward adjoint calculation K T X. When the forward transformation operator is superimposed with the interpolation operator for calculation, it can be nested as follows: and K T H T X.

[0070] Covariance matrix; In the transformations involved in the background covariance matrix and the observation error covariance matrix, the covariance matrix provides statistical normalized weights for the input background field in the assimilation analysis. The covariance matrix is ​​expressed as B -1 , further construction, the operator can be decomposed into Therefore, the covariance matrix provides two standard interface interfaces: left multiplication and with left multiplication In the case where it is difficult to invert the covariance matrix, we can simultaneously construct a direct left multiplication and directly followed by left multiplication X is used as a preconditioner for matrix solving.

[0071] For the covariance matrix of the observed data, write it as R -1 , the observed data are discrete and independent, and their covariance matrix is ​​usually written as a diagonal matrix, so the left multiplication and with left multiplication It can meet the structural requirements.

[0072] Control variable transformation (model integration, constraints); When the numerical forecast model involves transformation, the optimized independent variable of the assimilation analysis is the atmospheric state variable at the forecast start time, and a large amount of observation data is usually distributed at non-forecast start time. For example, the model forecast initial field is 00UTC, and a large amount of radar satellite data may be distributed in the 6-hour time window of 18UTC-00UTC. In order to make full use of these advanced or delayed observation data information, it is necessary to use the model integral transformation to deduce the time when the independent variable is located to the time when the observation data is located. The model integral operator is expressed as Its forward nonlinear integral operator is expressed as The tangent operator is MX and the adjoint operator is M T X, so the mode variable X0 at time 0 can be transformed to time 1 through the mode transformation operator:

[0073]

[0074] Then it can be transformed to time n:

[0075]

[0076] Thus, the observation at time n is minimized:

[0077]

[0078] It becomes the optimal solution form that takes the X0 model variable as the independent variable and makes the model prediction variable at the nth moment close to the observation at the nth moment.

[0079] Mode coupler, and input and output; interpolation operator, the transformation and input and output system involved in the mode coupler can be connected to a variety of different forecast modes in terms of structure, such as the most typical NCAR WRF forecast mode, and the CMA-MESO forecast mode of the Earth System Numerical Prediction Center of the China Meteorological Administration. For different forecast modes, the system uses interpolation operators and mode couplers to connect. The interpolation operator uses an irregular grid interpolation algorithm based on the KDTree algorithm to adapt to different types of projections and irregular grids. When the initial fields of different modes enter the system, the interpolation operator will interpolate them to the assimilation analysis physical space defined by Geometry. At the same time, when the atmospheric state variables given by the assimilation system do not match the state variables of the model input, the mode coupler will convert variables for different modes, such as converting the potential temperature θ and dimensionless pressure Π used by the GRAPES forecast system into universal analysis variables temperature T and pressure P. Through the interpolation operator and mode coupler transformation, various forecast modes become standard atmospheric variables on a unified given grid structure.

[0080] The observation data input part is designed according to a unified structure. The observation data is configured and input into the observation data container according to the dimensions of longitude, latitude, altitude, time, and value. Different types of observation input classes inherit from the basic class of ObsBase. For example, radar radial wind observation is configured as longitude, latitude, altitude, time, and radial wind value; ground temperature observation is configured as longitude, latitude, altitude, time, and ground temperature value, so as to construct a unified observation data input class.

[0081] Price function: The above operator encapsulation modules can be assembled and integrated through price functions. For example, when using three-dimensional variational data assimilation (3DVar), the price function and the price function gradient can be constructed as follows:

[0082]

[0083] When using four-dimensional variational data assimilation (4DVar), the price function and the price function gradient can be constructed as follows:

[0084]

[0085] When using the three-dimensional ensemble variation assimilation scheme (3DEnVar), the price function and the price function gradient can be constructed as follows:

[0086]

[0087] Universal solver: For a multivariable optimization problem with a given price function, if you need to solve it, you need at least two universal computing interfaces: 1. The calculated value of the price function; 2. The gradient of the price function with respect to the independent variable. When two interfaces are available for calculation, you can use the L-BFGS quasi-Newton method or the FRCG conjugate gradient method solver to solve it.

[0088] Based on each module construction, start to obtain assimilation parameters. The assimilation parameters use the Geometry grid framework module to define the geometric or physical space, and use the geometric grid and parallel algorithm and the MPDD parallel communication module to distribute the collection and boundary exchange calculation in parallel. The assimilation parameters include the control variables to be solved and analyzed, the input observation variables, the background error covariance matrix, the observation error covariance matrix, the spatial interpolation operator from the pattern space to the observation space, the transformation operator from the pattern space to the observation space, and the transformation operator in the control variable space.

[0089] Construct a multivariable price function and price function gradient:

[0090]

[0091] Among them, H T is the adjoint operator of the interpolation operator, K T is the adjoint operator of the observation operator;

[0092] The minimization algorithm uses the gradient of the price function in the above formula to make the analysis field converge to the observation result. At the same time, the multivariate price function makes other points outside the observation position diffuse outward in the mode of the covariance matrix to affect other pattern grid points.

[0093] The multivariable price function and the price function gradient are passed to a general minimization solver, and the assimilation analysis field result is obtained through iterative solution and then output to the initial field file of the numerical prediction model through a model coupler.

[0094] After the above modules are constructed, based on the above object-oriented module construction, the assimilation system structure based on full-quantity analysis is constructed as follows Figure 1 As shown:

[0095] Here X is the control variable to be analyzed, Y is the input observation variable, B is the background error covariance matrix, and R is the observation error covariance matrix. is the spatial interpolation operator from pattern space to observation space, is the transformation operator from pattern space to observation space, It is a transformation operator in the control variable space, which can be a transformation of the control variable, or an integral operator related to four-dimensional variation.

[0096] All of the above modules use the Geometry grid framework module to define geometric or physical space, and use the MPDD parallel communication module to implement parallel distribution, aggregation, and boundary exchange calculations to achieve large-scale parallel computing.

[0097] The mode coupler is used to read the initial values ​​of the control variables from the given mode background field, and the observation input module reads all the observation values.

[0098] Take the above modules as input to construct the multivariable price function J(X) to be solved and the price function gradient It is passed to the general minimization solver, and through iterative solution, the assimilated analysis field results can be obtained, and finally output to the initial field file of the numerical prediction model through the model coupler.

[0099] Based on actual application scenarios, such as Figure 2 The schematic diagram of the process of connecting the assimilation system to the regional forecast system is divided into a cold start cycle and a hot start cycle. The cold start cycle uses the global model as the initial field to drive the model forecast, while the hot start cycle uses the regional forecast field as the initial field to drive the forecast model. In the cold start and hot start cycle processes, the assimilation analysis system generates the analysis field that ultimately drives the regional forecast model through observation data based on the assimilation analysis method.

[0100] The numerical weather forecast model needs to use the current weather conditions as the starting point and integrate backward. The weather conditions here refer to the accurate and complete gridded three-dimensional atmospheric element field (wind speed, wind direction, air pressure, temperature, water vapor, etc.). The quality of the analysis field output by the assimilation system, that is, the starting point of the model weather forecast, affects the actual accuracy of the subsequent day forecast. The higher the level of assimilation technology, the more accurate the calculated weather conditions will be, and the more accurate the subsequent model forecast will be.

[0101] The assimilation system itself has many different paradigms, using different price function structures and different solvers. It may connect to many different modes and have many different observations. The object-oriented assimilation analysis system design proposed in this patent abstractly encapsulates the various parts of the assimilation system according to object-oriented standards, and can realize the unified encapsulation of various modules and various different solutions to quickly realize the assimilation application of different scenarios; it can also compare the assimilation effects of different technical solutions by controlling variables.

[0102] As above Figure 2As shown in the figure, in the actual scenario: the global forecast field can be selected from the European Center product (ECMWF), the China Meteorological Administration product (CMA-GFS), etc.; the subsequent standardization module converts the global forecast field into the three-dimensional initial field of the corresponding region. For different regional forecast modes, different standardization modules are used here to generate the corresponding model initial field. The selectable regional forecast modes include WRF, CMA-MESO, etc.

[0103] For the subsequent assimilation system, WRF-DA is generally used for WRF forecast, and CMA-MESO uses the specially developed GRAPES-3DVAR assimilation system. The object-oriented assimilation system proposed in this patent can be compatible with different forecast system initial fields through the design of switching couplers, access various different observation data, and obtain the analysis fields corresponding to each mode;

[0104] The analysis and boundary fields are then input into the regional forecast model, which outputs regional numerical forecast products;

[0105] Based on the object-oriented assimilation system design, the J function part can switch to the 3DVar scheme or the 4DVar scheme; the solver part can choose to use different solvers, such as FRCG, LBFGS, etc.; the control variable transformation part can adapt to various strong constraints; the covariance matrix part can choose different covariance matrix schemes, such as Gaussian filter, set scheme, etc.; the observation operator scheme, for different observation data types, different operator schemes can be designed or selected.

[0106] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 3 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0107] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0108] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0109] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a coal seam porosity calculation device at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the aforementioned data assimilation analysis systems based on the object-oriented framework.

[0110] The above application Figure 1The data assimilation analysis system based on an object-oriented framework disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or instructions in the form of software. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0111] The electronic device may also perform Figure 1 A data assimilation analysis system based on an object-oriented framework is proposed and implemented Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0112] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple application programs, any one of the aforementioned data assimilation analysis systems based on the object-oriented framework is executed.

[0113] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0114] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, 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 generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0117] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0118] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0119] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0120] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0121] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0122] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A data assimilation analysis system based on an object-oriented framework, characterized in that: The system comprises a plurality of general modules, and the general modules are constructed by an object-oriented framework.

2. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The general module is constructed by an object-oriented framework and specifically includes the following steps: Obtain assimilation parameters, which use the Geometry grid framework module to define the geometry or physical space, and use the geometry grid and parallel algorithm and the MPDD parallel communication module to distribute and collect and exchange boundary calculations in parallel. The assimilation parameters include the control variables to be solved and analyzed, the input observation variables, the background error covariance matrix, the observation error covariance matrix, the spatial interpolation operator from the pattern space to the observation space, the transformation operator from the pattern space to the observation space, and the transformation operator in the control variable space. The multivariable price function and the price function gradient are passed to a general minimization solver, and the assimilated analysis field result is obtained through iterative solution and then output to the initial field file of the numerical prediction model through a model coupler; The price function and the price function gradient are constructed as follows: When using 4D variational data assimilation, the price function and the price function gradient can be constructed as follows: When using the three-dimensional ensemble variation assimilation scheme, the price function and the price function gradient can be constructed as follows: The minimization algorithm uses the gradient of the price function in the above formula to make the analysis field converge to the observation result. At the same time, the multivariate price function makes other points outside the observation position diffuse outward in the mode of the covariance matrix to affect other pattern grid points.

3. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The spatial information is encapsulated as a geometric algorithm library, and the parallel computing operation is encapsulated as a parallel algorithm library. When the pattern integral transformation is needed to deduce the time when the independent variable is located to the time when the observed data is located, the pattern integral operator is represented as M, its forward nonlinear integral operator is represented as M(X), the tangent linear operator is MX, and the adjoint operator is M T X, so the mode variable X0 at time 0 can be transformed to time 1 through the mode transformation operator: Then it can be transformed to time n: Thus, the observation at time n is minimized: Taking the X0 model variable as the independent variable, the model forecast variable at the nth moment is made to approach the optimal solution form of the observation at the nth moment.

4. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The price function part can switch to use 3DVar, 4DVar, or 3DEnVar solutions.

5. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The solver can adopt FRCG solver or LBFGS solver.

6. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The correlation covariance matrix can be selected to use a recursive filter covariance matrix, a Laplace covariance matrix, or a collective covariance matrix.

7. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The mode integral operator and the like can be selected to use a static equilibrium mode or a non-static equilibrium mode.

8. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The observation input module inputs the data into the container of the observation data according to the dimensions of longitude, latitude, altitude, time and value.

9. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The observation operator module transforms the variable data in the system into the observation space to ensure the spatial consistency of the observation data.

10. The data assimilation analysis system based on an object-oriented framework according to claim 1, characterized in that: The mode coupler converts mode input variables into grid standard variables generated by Geometry to ensure the spatial consistency of the analysis data.

Images