CN112883636A - Acceleration method and device for parameter model of numerical mode - Google Patents
Acceleration method and device for parameter model of numerical mode Download PDFInfo
- Publication number
- CN112883636A CN112883636A CN202110121680.2A CN202110121680A CN112883636A CN 112883636 A CN112883636 A CN 112883636A CN 202110121680 A CN202110121680 A CN 202110121680A CN 112883636 A CN112883636 A CN 112883636A
- Authority
- CN
- China
- Prior art keywords
- numerical
- mode
- input
- main function
- output
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 230000001133 acceleration Effects 0.000 title claims description 19
- 230000006870 function Effects 0.000 claims abstract description 72
- 238000004088 simulation Methods 0.000 claims abstract description 56
- 238000010801 machine learning Methods 0.000 claims abstract description 40
- 238000004364 calculation method Methods 0.000 claims abstract description 17
- 230000008878 coupling Effects 0.000 claims abstract description 12
- 238000010168 coupling process Methods 0.000 claims abstract description 12
- 238000005859 coupling reaction Methods 0.000 claims abstract description 12
- 239000002245 particle Substances 0.000 claims description 8
- 239000012071 phase Substances 0.000 claims description 8
- 239000007791 liquid phase Substances 0.000 claims description 5
- 238000003062 neural network model Methods 0.000 claims description 4
- 239000007788 liquid Substances 0.000 claims description 3
- 238000001556 precipitation Methods 0.000 claims description 3
- 230000015654 memory Effects 0.000 description 15
- 230000008569 process Effects 0.000 description 11
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 9
- 238000003491 array Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000005291 magnetic effect Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000005442 atmospheric precipitation Substances 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000011165 process development Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 230000003936 working memory Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The method comprises the steps of splitting a main function in a parameter model of a numerical mode into a single-column mode, and splitting a sub-function called by the main function into the single-column mode; modifying the output of the parameter model into an input and output variable of a main function in a single column mode; performing numerical simulation within a preset time span, and storing all input and output variables obtained by numerical simulation; and constructing a machine learning model based on the input and output variables obtained by all numerical simulations, and coupling the machine learning model to the main function in the single column mode. Therefore, the calculation speed of the parameterization scheme is improved, the overall operation speed of the numerical mode is effectively improved, meanwhile, the calculation resources can be effectively saved, and the operation cost of the numerical mode is reduced.
Description
Technical Field
The present application relates to the field of computers, and in particular, to a method and an apparatus for accelerating a parametric model for a numerical model.
Background
The numerical weather forecast (numerical weather prediction) refers to a method for predicting an atmospheric motion state and a weather phenomenon in a future period by performing numerical calculation through a large-scale computer under certain initial value and side value conditions according to an atmospheric actual condition, solving a fluid mechanics and thermodynamics equation set describing a weather evolution process and predicting the atmospheric motion state and the weather phenomenon in the future period. Through decades of development of computer technologies, atmospheric sounding technologies, atmospheric sciences and related disciplines, the numerical forecasting system has been able to provide weather forecasts with high reliability, and forecasting of specific meteorological elements (such as air temperature) has also reached a considerable accuracy.
In numerical weather forecasting, besides the dynamic process of solving the equation set consisting of continuous equations, thermodynamic equations, water vapor equations, state equations and 3 aerodynamic equations, many physical and chemical parameterization schemes are included. The physical description of the numerical weather forecast mode on the weather process development is mainly realized through the parameterization of the physical process. Thus. The development of the parameterization scheme of the physical process is always an international research hotspot and is also a difficulty for improving the numerical mode.
In the numerical calculation process of numerical weather forecast, the calculation process of the parameterization scheme, especially the radiation parameterization scheme and the micro-physical parameterization scheme is complex, and the calculation amount is large, so that the consumed time is long and the calculation resources are large.
Disclosure of Invention
An object of the present application is to provide an acceleration method, an acceleration apparatus, a parameter model method and an acceleration apparatus for a numerical model, which solve the problems of complex overall computation process, large computation amount and large computation resource consumption of a parameterization scheme of a numerical model in the prior art.
According to an aspect of the present application, there is provided an acceleration method for a parametric model of numerical patterns, the method comprising:
splitting a main function in a parameter model of a numerical mode into a single-column mode, and splitting a sub-function called by the main function into the single-column mode;
modifying the output of the parameter model into an input and output variable of a main function in a single column mode;
performing numerical simulation within a preset time span, and storing all input and output variables obtained by numerical simulation;
and constructing a machine learning model based on the input and output variables obtained by all numerical simulations, and coupling the machine learning model to the main function in the single column mode.
Further, splitting the main function in the parametric model of the numerical mode into single-column modes includes:
the main function in the WSM6 parametric model of the WRF numerical pattern is split into single column patterns based on longitude and latitude.
Further, the numerical simulation within the preset time span is performed, which includes:
and operating a WRF numerical mode, and inputting weather parameters in a preset time span into the WRF numerical mode for numerical simulation.
Further, modifying the output of the parametric model to the input-output variables of the primary function in the single-column mode comprises:
modifying the output data of the parameter model according to the latitude, and storing the modified data into a text according to a preset structure;
and taking the data stored in the text as input and output variables of the main function in the single column mode.
Further, constructing a machine learning model based on the input and output variables obtained by the numerical simulation, including:
constructing a gradient descent tree model by using spark and the input and output variables obtained by all numerical simulations,
or constructing a multilayer neural network model based on the input and output variables obtained by the numerical simulation.
Further, coupling the machine learning model to the master function in single column mode comprises:
coupling the machine learning model to the main function in the single column mode to replace a sub-function in the single column mode, so as to obtain a numerical mode comprising a program calling the machine learning model;
compiling and running the numerical pattern comprising the program calling the machine learning model.
Further, the input and output variables obtained based on the numerical simulations include: an input variable and an output variable,
wherein the input variables include temperature, humidity, air pressure, liquid phase and ice phase particle concentration at the current time step;
the output variables include temperature, humidity, air pressure, liquid and ice phase particle concentrations and precipitation information for the next time step.
Further, inputting weather parameters within a preset time span into the WRF numerical mode for numerical simulation, including:
determining a target area and a preset time span of the target area according to the calculation power and the storage space;
and acquiring weather parameters in the preset time span of the target area, and inputting the acquired weather parameters into the WRF numerical mode for numerical simulation.
According to another aspect of the present application, there is also provided an acceleration apparatus for a parametric model of numerical patterns, the acceleration apparatus including:
the splitting device is used for splitting a main function in a parameter model of a numerical mode into a single-column mode and splitting a sub-function called by the main function into the single-column mode;
modifying means for modifying an output of the parametric model to an input-output variable of a primary function in a single column mode;
the simulation device is used for carrying out numerical simulation in a preset time span and storing all input and output variables obtained by numerical simulation;
and the constructing device is used for constructing a machine learning model based on the input and output variables obtained by the numerical simulation and coupling the machine learning model to the main function in the single column mode.
According to yet another aspect of the present application, there is also provided a computer readable medium having computer readable instructions stored thereon, the computer readable instructions being executable by a processor to implement the method as described above.
Compared with the prior art, the method and the device have the advantages that the main function in the parameter model of the numerical mode is split into the single-column mode, and the sub-function called by the main function is split into the single-column mode; modifying the output of the parameter model into an input and output variable of a main function in a single column mode; performing numerical simulation within a preset time span, and storing all input and output variables obtained by numerical simulation; and constructing a machine learning model based on the input and output variables obtained by all numerical simulations, and coupling the machine learning model to the main function in the single column mode. Therefore, the calculation speed of the parameterization scheme is improved, the overall operation speed of the numerical mode is effectively improved, meanwhile, the calculation resources can be effectively saved, and the operation cost of the numerical mode is reduced.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
FIG. 1 illustrates a flow diagram of an acceleration method for a parametric model of numerical patterns provided in accordance with an aspect of the present application;
fig. 2 shows a schematic diagram of an acceleration apparatus for a parametric model of numerical modes according to yet another aspect of the present application.
The same or similar reference numbers in the drawings identify the same or similar elements.
Detailed Description
The present application is described in further detail below with reference to the attached figures.
In a typical configuration of the present application, the terminal, the device serving the network, and the trusted party each include one or more processors (e.g., Central Processing Units (CPUs)), input/output interfaces, network interfaces, and memory.
The Memory may include volatile Memory in a computer readable medium, Random Access Memory (RAM), and/or nonvolatile Memory such as Read Only Memory (ROM) or flash Memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, Phase-Change RAM (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 Disc Read-Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, magnetic cassette tape, magnetic tape storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include non-transitory computer readable media (transient media), such as modulated data signals and carrier waves.
Fig. 1 shows a flow diagram of an acceleration method for a parametric model of numerical patterns according to an aspect of the present application, the method comprising: step S11 to step S14,
in step S11, splitting a main function in a parameter model of a numerical mode into a single-column mode, and splitting a sub-function called by the main function into the single-column mode; the numerical mode is preferably a numerical mode of a weather forecast, the parameter model is a parameterized scheme of the numerical mode of the weather forecast, the parameterized scheme comprises a main function and a sub-function called by the main function, when the parameterized scheme is optimized, the main function can be firstly split into a single-column mode, and the sub-function can also be split into the single-column mode, and the single-column mode is a calculation mode based on a single air column cycle.
In step S12, modifying the output of the parametric model to the input-output variables of the main function in the single-column mode; the main function is split and then the program is modified, so that the input and output variables of the modified main function can be output, namely the output of the parameter model is modified into the input and output variables of the main function in the calculation mode of single air column circulation, and therefore the input and output of the main function are the same as those of the previous main function, and the program can run normally.
In step S13, performing numerical simulation within a preset time span, and storing all input/output variables obtained by the numerical simulation; the numerical simulation with the time span of years is carried out in a fixed simulation range, all input and output variables are saved, and the size of the simulation range influences the generalization capability of a subsequent machine learning model, so that the preset time span can be adjusted according to actual conditions, and data in the preset time span is collected to carry out the numerical simulation.
In step S14, a machine learning model is constructed based on the input and output variables obtained by the numerical simulation, and the machine learning model is coupled to the main function in the single column mode. Here, the machine learning model is constructed by using the obtained input and output variables, so that the machine learning model is coupled to an original parameterization scheme, namely, a main function of a single column mode, a numerical mode is operated, and an acceleration effect can be achieved.
In an embodiment of the present application, in step S11, the master function in the WSM6 parametric model of the WRF numerical mode is split into single-column modes based on longitude and latitude. In the numerical weather mode, the original longitude-based (i-dimension) cyclic WSM62D function in the numerical mode is split by using a parameterization scheme of the WSM6 in a wrf (weather Research and Forecasting model) mode, and is modified into a single-column mode based on longitude and latitude (i-dimension and j-dimension), namely, the WSM61D, and meanwhile, the sub-function of the WSM62D is also subjected to single-column transformation. The WSM62D function is a main function of a WRF mode and is used for calculating interconversion processes between atmospheric precipitation and various cloud ice and cloud water at a fixed latitude. In the calculation process, various cloud micro physical processes on different latitude cross sections are circularly and sequentially calculated for the latitude. The sub-functions of the WSM62D function call are also cycled based on fixed latitude, and are similarly split into single-column modes.
In an embodiment of the present application, in step S13, a WRF numerical mode is operated, and weather parameters within a preset time span are input into the WRF numerical mode for numerical simulation. Here, in the weather forecast mode of the WRF, the corresponding file and the parameter are input to perform the operation, where the parameter is a weather parameter within a preset time span, for example, a weather parameter within a certain fixed time span of a certain area.
In an embodiment of the present application, in step S12, output data of the parametric model is modified according to latitude, and the modified data is stored in a text according to a preset structure; and taking the data stored in the text as input and output variables of the main function in the single column mode. The output data of the parametric model is a plurality of multidimensional arrays, the arrays are modified into one dimension or two dimensions according to the dimensions each time, the arrays are modified according to the latitudes by the main function of the single column mode, one element or one dimension in the arrays is modified, and then two layers of nested loops are modified. Before all input variables are read by the main function of the single column mode, the input variables are stored in a text according to a certain structure, and the returned output variables are output after calculation is finished.
In an embodiment of the present application, in step S14, a gradient descent tree model is constructed using spark and the input and output variables obtained by all numerical simulations, or a multi-layer neural network model is constructed based on the input and output variables obtained by all numerical simulations. When the machine learning model is constructed, the obtained file amount is huge, so that a gradient descent tree model can be constructed by using spark, or a multilayer neural network model can be constructed under the condition of no spark cluster; the gradient descent tree model comprises xgboost, lightgbm and the like.
In an embodiment of the present application, in step S14, the machine learning model is coupled to the main function in the single column mode to replace the sub-function in the single column mode, so as to obtain a numerical value pattern including a program calling the machine learning model; compiling and running the numerical pattern comprising the program calling the machine learning model. When coupling is carried out, the constructed machine learning model is used for replacing a main function of a single column mode in a parameterization scheme, namely, the WSM61D function part in the model is replaced, so that a mode of a calling program is obtained, the program can call the machine learning model, for example, a sub-function of the WSM61D is replaced by a C + + machine learning model function, and input and output variables and the dimension of the sub-function are consistent with the dimension of the C + + machine learning model, so that the whole WRF model can normally operate; and then compiling and operating the numerical mode, so that an acceleration effect can be achieved, in the numerical weather forecast, the calculation resources can be effectively saved, the operation cost of the numerical weather mode is reduced, and the method has great significance for timely early warning, prevention and control of disasters.
In an embodiment of the present application, the input and output variables obtained based on the numerical simulations include: an input variable and an output variable, wherein the input variable comprises a temperature, a humidity, an air pressure, a liquid phase and an ice phase particle concentration of a current time step; the output variables include temperature, humidity, air pressure, liquid and ice phase particle concentrations and precipitation information for the next time step. When the machine learning mode is constructed by using variables, the variables are input and output variables obtained based on all numerical simulations, and include output variables and input variables, the input variables include temperature, humidity, air pressure, various liquid phase and ice phase particle concentrations, and the like, and the corresponding output variables include temperature, humidity, air pressure, various liquid phase and ice phase particle concentrations, and the like of the next time step, where the time step is a period duration of a weather forecast, for example, the time step is set to be 7 days, and the input variables are weather parameters of the current 7 days, and then the weather parameters of the next 7 days are output.
In an embodiment of the present application, in step S13, the target area and the preset time span of the target area are determined according to the calculation power and the storage space; and acquiring weather parameters in the preset time span of the target area, and inputting the acquired weather parameters into the WRF numerical mode for numerical simulation. In the actual operation process, a target area to be forecasted and a preset time span are determined according to the calculated force and the storage space, for example, if the target area is the east China area and the preset time span is 1 year, weather parameters in 1 year of the east China area are acquired, and the acquired weather parameters are input into the WRF for numerical simulation.
Furthermore, the embodiment of the present application also provides a computer readable medium, on which computer readable instructions are stored, the computer readable instructions being executable by a processor to implement the aforementioned acceleration method for the parametric model of the numerical mode.
In correspondence with the method described above, the present application also provides a terminal, which includes modules or units capable of executing the steps of the method described in fig. 1 or each embodiment, and these modules or units can be implemented by hardware, software or a combination of hardware and software, and this application is not limited thereto. For example, in an embodiment of the present application, there is also provided an acceleration apparatus for a parametric model of numerical patterns, the acceleration apparatus including:
one or more processors; and
a memory storing computer readable instructions that, when executed, cause the processor to perform the operations of the method as previously described.
For example, the computer readable instructions, when executed, cause the one or more processors to:
splitting a main function in a parameter model of a numerical mode into a single-column mode, and splitting a sub-function called by the main function into the single-column mode;
modifying the output of the parameter model into an input and output variable of a main function in a single column mode;
performing numerical simulation within a preset time span, and storing all input and output variables obtained by numerical simulation;
and constructing a machine learning model based on the input and output variables obtained by all numerical simulations, and coupling the machine learning model to the main function in the single column mode.
Fig. 2 shows a schematic structural diagram of an acceleration apparatus for a parametric model of numerical modes according to yet another aspect of the present application, the acceleration apparatus comprising: the device comprises a splitting device 11, a modifying device 12, a simulating device 13 and a constructing device 14, wherein the splitting device 11 is used for splitting a main function in a parameter model of a numerical mode into a single-column mode and splitting a sub-function called by the main function into the single-column mode; the modifying device 12 is used for modifying the output of the parameter model into an input and output variable of the main function in the single column mode; the simulation device 13 is used for performing numerical simulation within a preset time span and storing all input and output variables obtained by numerical simulation; the constructing device 14 is configured to construct a machine learning model based on the input and output variables obtained by the numerical simulation, and couple the machine learning model to the main function in the single-column mode.
It should be noted that the content executed by the splitting apparatus 11, the modifying apparatus 12, the simulating apparatus 13 and the constructing apparatus 14 is the same as or corresponding to the content executed in the above steps S11, S12, S13 and S14, and for brevity, the description is omitted here.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
It should be noted that the present application may be implemented in software and/or a combination of software and hardware, for example, implemented using Application Specific Integrated Circuits (ASICs), general purpose computers or any other similar hardware devices. In one embodiment, the software programs of the present application may be executed by a processor to implement the steps or functions described above. Likewise, the software programs (including associated data structures) of the present application may be stored in a computer readable recording medium, such as RAM memory, magnetic or optical drive or diskette and the like. Additionally, some of the steps or functions of the present application may be implemented in hardware, for example, as circuitry that cooperates with the processor to perform various steps or functions.
In addition, some of the present application may be implemented as a computer program product, such as computer program instructions, which when executed by a computer, may invoke or provide methods and/or techniques in accordance with the present application through the operation of the computer. Program instructions which invoke the methods of the present application may be stored on a fixed or removable recording medium and/or transmitted via a data stream on a broadcast or other signal-bearing medium and/or stored within a working memory of a computer device operating in accordance with the program instructions. An embodiment according to the present application comprises an apparatus comprising a memory for storing computer program instructions and a processor for executing the program instructions, wherein the computer program instructions, when executed by the processor, trigger the apparatus to perform a method and/or a solution according to the aforementioned embodiments of the present application.
It will be evident to those skilled in the art that the present application is not limited to the details of the foregoing illustrative embodiments, and that the present application may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned. Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the apparatus claims may also be implemented by one unit or means in software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.
Claims (10)
1. A method for accelerating a parametric model of numerical patterns, the method comprising:
splitting a main function in a parameter model of a numerical mode into a single-column mode, and splitting a sub-function called by the main function into the single-column mode;
modifying the output of the parameter model into an input and output variable of a main function in a single column mode;
performing numerical simulation within a preset time span, and storing all input and output variables obtained by numerical simulation;
and constructing a machine learning model based on the input and output variables obtained by all numerical simulations, and coupling the machine learning model to the main function in the single column mode.
2. The method of claim 1, wherein splitting the master function in the parametric model of numerical patterns into single-column patterns comprises:
the main function in the WSM6 parametric model of the WRF numerical pattern is split into single column patterns based on longitude and latitude.
3. The method of claim 2, wherein performing the numerical simulation over the predetermined time span comprises:
and operating a WRF numerical mode, and inputting weather parameters in a preset time span into the WRF numerical mode for numerical simulation.
4. The method of claim 1, wherein modifying the output of the parametric model to an input-output variable of the master function in single-column mode comprises:
modifying the output data of the parameter model according to the latitude, and storing the modified data into a text according to a preset structure;
and taking the data stored in the text as input and output variables of the main function in the single column mode.
5. The method of claim 1, wherein constructing a machine learning model based on the all numerical modeled input and output variables comprises:
constructing a gradient descent tree model by using spark and the input and output variables obtained by all numerical simulations,
or constructing a multilayer neural network model based on the input and output variables obtained by the numerical simulation.
6. The method of claim 1, wherein coupling the machine learning model to the master function in single-column mode comprises:
coupling the machine learning model to the main function in the single column mode to replace a sub-function in the single column mode, so as to obtain a numerical mode comprising a program calling the machine learning model;
compiling and running the numerical pattern comprising the program calling the machine learning model.
7. The method of claim 1, wherein the input-output variables derived based on the all numerical simulations comprise: an input variable and an output variable,
wherein the input variables include temperature, humidity, air pressure, liquid phase and ice phase particle concentration at the current time step;
the output variables include temperature, humidity, air pressure, liquid and ice phase particle concentrations and precipitation information for the next time step.
8. The method of claim 3, wherein inputting weather parameters for a predetermined time span into the WRF numerical mode for numerical simulation comprises:
determining a target area and a preset time span of the target area according to the calculation power and the storage space;
and acquiring weather parameters in the preset time span of the target area, and inputting the acquired weather parameters into the WRF numerical mode for numerical simulation.
9. An acceleration apparatus for a parametric model of numerical patterns, characterized in that it comprises:
the splitting device is used for splitting a main function in a parameter model of a numerical mode into a single-column mode and splitting a sub-function called by the main function into the single-column mode;
modifying means for modifying an output of the parametric model to an input-output variable of a primary function in a single column mode;
the simulation device is used for carrying out numerical simulation in a preset time span and storing all input and output variables obtained by numerical simulation;
and the constructing device is used for constructing a machine learning model based on the input and output variables obtained by the numerical simulation and coupling the machine learning model to the main function in the single column mode.
10. A computer readable medium having computer readable instructions stored thereon which are executable by a processor to implement the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110121680.2A CN112883636A (en) | 2021-01-28 | 2021-01-28 | Acceleration method and device for parameter model of numerical mode |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110121680.2A CN112883636A (en) | 2021-01-28 | 2021-01-28 | Acceleration method and device for parameter model of numerical mode |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112883636A true CN112883636A (en) | 2021-06-01 |
Family
ID=76053281
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110121680.2A Pending CN112883636A (en) | 2021-01-28 | 2021-01-28 | Acceleration method and device for parameter model of numerical mode |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112883636A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10613252B1 (en) * | 2015-10-02 | 2020-04-07 | Board Of Trustees Of The University Of Alabama, For And On Behalf Of The University Of Alabama In Huntsville | Weather forecasting systems and methods |
| WO2020131187A2 (en) * | 2018-09-26 | 2020-06-25 | Sofar Ocean Technologies, Inc. | Ocean weather forecasting system |
| CN111815098A (en) * | 2019-12-11 | 2020-10-23 | 北京嘀嘀无限科技发展有限公司 | Traffic information processing method and device based on extreme weather, storage medium and electronic equipment |
| US20210019456A1 (en) * | 2017-11-27 | 2021-01-21 | Siemens Industry Software Nv | Accelerated simulation setup process using prior knowledge extraction for problem matching |
-
2021
- 2021-01-28 CN CN202110121680.2A patent/CN112883636A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10613252B1 (en) * | 2015-10-02 | 2020-04-07 | Board Of Trustees Of The University Of Alabama, For And On Behalf Of The University Of Alabama In Huntsville | Weather forecasting systems and methods |
| US20210019456A1 (en) * | 2017-11-27 | 2021-01-21 | Siemens Industry Software Nv | Accelerated simulation setup process using prior knowledge extraction for problem matching |
| WO2020131187A2 (en) * | 2018-09-26 | 2020-06-25 | Sofar Ocean Technologies, Inc. | Ocean weather forecasting system |
| CN111815098A (en) * | 2019-12-11 | 2020-10-23 | 北京嘀嘀无限科技发展有限公司 | Traffic information processing method and device based on extreme weather, storage medium and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| 梅海霞: "双参数微物理方案在WRF单柱模式中的模拟检验和对比研究", 高原气象, vol. 34, no. 4, 31 August 2015 (2015-08-31), pages 890 - 909 * |
| 梅海霞;沈新勇;王卫国;黄小梅;: "WRF单柱模式中单参数方法对热带对流模拟能力的影响", 气候与环境研究, no. 06, 20 November 2015 (2015-11-20), pages 685 - 704 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20170124461A1 (en) | Air quality forecast by adapting pollutant emission inventory | |
| Evans et al. | Enabling spatial big data via CyberGIS: Challenges and opportunities | |
| CN109541730A (en) | A kind of method and apparatus of pollutant prediction | |
| CN112785034B (en) | Typhoon path forecasting method, system, medium and terminal based on fusion neural network | |
| Zhang et al. | Enabling point pattern analysis on spatial big data using cloud computing: Optimizing and accelerating Ripley’s K function | |
| CN114580263A (en) | Knowledge graph-based information system fault prediction method and related equipment | |
| CN113516247A (en) | Parameter calibration method, quantum chip control method, device and system | |
| WO2021169386A1 (en) | Graph data processing method, apparatus and device, and medium | |
| CN116450486B (en) | Modeling method, device, equipment and medium for nodes in multi-element heterogeneous computing system | |
| CN115457492A (en) | Target detection method and device, computer equipment and storage medium | |
| Geyer et al. | Limits of reproducibility and hydrodynamic noise in atmospheric regional modelling | |
| CN113791768A (en) | Code generation method and device, storage medium and terminal | |
| CN113537614A (en) | Construction method, system, equipment and medium of power grid engineering cost prediction model | |
| Iquebal et al. | Emulating the evolution of phase separating microstructures using low-dimensional tensor decomposition and nonlinear regression | |
| Wang et al. | A distributed load balancing algorithm for climate big data processing over a multi‐core CPU cluster | |
| CN116108645A (en) | Engineering digital twin method based on meta universe | |
| CN111814960A (en) | Method and device for identifying weather climate system, storage medium and processor | |
| Lv et al. | Research on ship dynamic map based on AIS | |
| CN112883636A (en) | Acceleration method and device for parameter model of numerical mode | |
| CN115983104A (en) | Wind speed prediction method and device, storage medium and electronic equipment | |
| Song et al. | Spatial-temporal transformer network for multi-year ENSO prediction | |
| CN108267613B (en) | Wind measurement data processing method and related equipment | |
| CN111210513A (en) | Information processing method, device, equipment and storage medium | |
| Baart et al. | A comparison between WCS and OPeNDAP for making model results and data products available through the internet | |
| Yang et al. | Probabilistic modeling of renewable energy source based on Spark platform with large‐scale sample data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |