CN110705115B

CN110705115B - Weather forecast method and system based on deep belief network

Info

Publication number: CN110705115B
Application number: CN201910960395.2A
Authority: CN
Inventors: 赵琉涛
Original assignee: Beijing Beike Rongzhi Cloud Computing Technology Co ltd
Current assignee: Beijing Beike Rongzhi Cloud Computing Technology Co ltd
Priority date: 2019-10-10
Filing date: 2019-10-10
Publication date: 2023-05-05
Anticipated expiration: 2039-10-10
Also published as: CN110705115A

Abstract

The application discloses a weather forecast method and a system based on a deep confidence network, wherein the method comprises the steps of obtaining a plurality of historical observation data according to site information of a target area; constructing an enhanced decision tree to train the historical observation data to obtain first prediction data; inputting the first prediction data into a CFD computational fluid dynamics model, performing steady state numerical simulation to obtain second prediction data, and obtaining flow field distribution under the incoming flow condition according to the second prediction data; interpolating the second predicted data at any position in the target area based on a kriging interpolation method to extract an analog wind speed value, thereby obtaining a training data set; inputting training samples of a training data set into a DBN deep neural network, training the DBN deep neural network, and carrying out weather prediction on data to be observed by utilizing a trained weather forecast model so as to generate a prediction result. The accuracy of the forecasted weather can be provided by the method and the system of the invention.

Description

Weather forecast method and system based on deep belief network

Technical Field

The application relates to the field of weather service information guarantee, in particular to a weather forecasting method and system based on a deep confidence network.

Background

The weather forecast guarantee has great influence on the safe and smooth holding of a plurality of activities carried out by the user. The traditional weather information guarantee is constructed on a weather data service platform of an HPC and cloud platform hybrid architecture, and by carrying out information technologies based on numerical forecasting, statistical learning, high-performance calculation and the like, the application research of multi-source data assimilation short-term forecasting on the aspect of weather forecasting is mainly surrounded, the weather forecasting speed and accuracy are improved, the weather service is provided for different target groups, the smart weather service is provided for fine and specialized places, and the weather service guarantee of key places is supported.

However, the weather forecast method constructed on the weather data service platform of the HPC and cloud platform hybrid architecture has the defect of low forecast accuracy due to initial error and mode error.

Aiming at the numerical forecasting problem, the initial value of the mode has objective factors, and the initial condition of the mode is only an approximation of the real state of the atmosphere, so that the initial error is permanently existing and mainly is the error generated in the observation process, namely the observation error. When the weather satellite is used for observation, the scattering of the visible light by the atmosphere causes certain deviation between the detected ground albedo and cloud top data, and certain error exists in the process of inverting the cloud top and ground temperature by directly observed infrared radiation data. And (3) observing by a weather radar, and presuming that the precipitation has errors according to the empirical relation between the precipitation and the radar echo intensity, wherein the errors generated during remote detection are larger.

Pattern errors are errors that result from inaccuracy in the description of a pattern to a physical process. The mechanism of the small-scale system in the mode is difficult to describe accurately due to the turbulent flow process in the atmosphere. Parameterization is typically employed in the pattern to take into account the effects of these processes, and the averaging process represents the physical and dynamic processes of the subgrid as an average effect on the scale above the grid. Errors exist between this description and the actual subgrid process; thus, there is a defect of low accuracy in forecasting weather.

Disclosure of Invention

The present application aims to overcome or at least partially solve or alleviate the above-mentioned problems.

According to one aspect of the present application, there is provided a weather forecast method based on a deep belief network, the weather forecast method comprising the steps of:

obtaining a plurality of historical observation data according to site information of a target area;

constructing an enhanced decision tree to train the historical observation data to obtain first prediction data;

inputting the first prediction data into a CFD computational fluid dynamics model, performing steady state numerical simulation to obtain second prediction data, and obtaining flow field distribution under the incoming flow condition according to the second prediction data;

interpolating the second predicted data at any position in the target area based on a kriging interpolation method to extract an analog wind speed value, thereby obtaining a training data set;

inputting training samples of a training data set into a DBN deep neural network, training the DBN deep neural network, and carrying out weather prediction on data to be observed by utilizing a trained weather forecast model so as to generate a prediction result.

Optionally, the weather forecast method compares and analyzes the forecast result obtained by the DBN deep neural network with the forecast result obtained by the traditional physical model so as to adjust and optimize the parameters of the first forecast data.

Optionally, the feature echo recognition algorithm is adopted to tune parameters of the first prediction data of the CFD model, and specifically includes the following substeps:

acquiring basic data of a radar, and acquiring echo characteristic parameters according to the basic data;

analyzing the echo characteristic parameters by using a statistical method to determine the difference between the feature echo and the rainfall echo characteristic parameters;

respectively setting membership functions according to the difference of characteristic parameters of each ground object echo and precipitation echo, and converting the membership functions into identification results, wherein the identification results are in the range of [0,1 ];

and carrying out weighted average on the identification result to obtain a criterion value of a certain point, and judging the point as the ground object echo when the criterion value of the certain point exceeds a preset threshold value.

Optionally, the traditional physical model solves an atmospheric evolution model, a plurality of dynamic cores and a physical option solving model for a mode equation set.

Optionally, the enhanced decision tree is a CART decision tree model.

Optionally, constructing the CART decision tree model includes the following sub-steps:

obtaining meteorological features based on the base Ny index and by using a feature selection method;

dividing according to the meteorological features and regression variances to obtain regression trees;

pruning the regression tree, correcting the regression tree by adopting a post pruning method, and constructing the CART decision tree model according to the corrected regression tree.

Optionally, the obtaining the flow field distribution under the incoming flow condition according to the second prediction data includes the following substeps:

performing CFD pre-calculation on the flow field under the place incoming flow condition of the first prediction data to obtain key flow field data, storing the key flow field data into a database, and establishing a scene flow field characteristic database;

wind speed prediction is carried out by utilizing a scene flow field characteristic database, and second prediction data are obtained;

and taking the second prediction data as an input boundary condition of the CFD model, inquiring the incoming flow condition of which the second prediction data is similar, and calling flow field distribution data of a scene flow field characteristic database to obtain flow field distribution under the incoming flow condition.

Optionally, the historical observation data, the first prediction data or the second prediction data are optimized to obtain standardized prediction data, and the training data set is constructed based on the standardized prediction data.

According to another aspect of the application, there is provided a weather forecast system based on a deep belief network, the system including an acquisition module, a first training module, a simulation module, an interpolation module, and a second training module;

the acquisition module is used for acquiring a plurality of historical observation data according to the site information of the target area;

the first training module is used for constructing an enhanced decision tree to train the historical observation data to obtain first prediction data;

the simulation module is used for inputting the first prediction data into a CFD computational fluid dynamics model, performing steady state numerical simulation to obtain second prediction data, and obtaining flow field distribution under the incoming flow condition according to the second prediction data;

the interpolation module interpolates the second predicted data at any position in the target area based on a kriging interpolation method to extract a simulated wind speed value, and a training data set is obtained;

the second training module is used for inputting training samples of the training data set into the DBN deep neural network, training the DBN deep neural network, and weather prediction is carried out on data to be observed by utilizing the trained weather forecast model so as to generate a prediction result.

Optionally, the simulation module performs the following operations:

According to another aspect of the present application there is provided a computer device comprising a memory, a processor and a computer program stored in said memory and executable by said processor, the computer program being stored in space for program code in the memory, which computer program, when executed by the processor, is adapted to carry out any of the weather forecast method steps according to the invention.

According to another aspect of the present application, there is provided a computer readable storage medium comprising a storage unit for program code, the storage unit being provided with a program for performing the steps of the weather forecast method according to the present invention, the program being executed by a processor.

According to another aspect of the present application, there is provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the weather forecast method steps according to the present invention.

According to the embodiment, the historical observation data and the real-time observation data are predicted through the CFD computational fluid dynamics model, so that the fine prediction can be realized, the mining observation data of the CFD computational fluid dynamics model in the time dimension identifies and learns the weather patterns repeatedly appearing in the history, and the atmospheric movement process of the weather phenomenon can be predicted;

in addition, the second prediction data is expanded from the position of the observation point to any point in the range of the target area by using a Kriging interpolation method; meanwhile, the correlation among meteorological elements is considered, and the DBN deep neural network is utilized to train the simulated wind speed value, so that the data to be observed are forecast, and the physical correlation among different elements of each observed data is reflected;

the weather forecast method of the embodiment realizes multi-dimensional display analysis, information monitoring and management of original data, intermediate data, forecast data and evaluation analysis data which are contained in multi-source data assimilation, mesoscale numerical forecast and deep learning-based refined weather forecast.

The above, as well as additional objectives, advantages, and features of the present application will become apparent to those skilled in the art from the following detailed description of a specific embodiment of the present application when read in conjunction with the accompanying drawings.

Drawings

Some specific embodiments of the present application will be described in detail hereinafter by way of example and not by way of limitation with reference to the accompanying drawings. The same reference numbers will be used throughout the drawings to refer to the same or like parts or portions. It will be appreciated by those skilled in the art that the drawings are not necessarily drawn to scale. In the accompanying drawings:

FIG. 1 is a flow diagram of a deep belief network-based weather forecast method in accordance with an embodiment of the present application;

FIG. 2 is a schematic diagram of a subscription message model provided in accordance with an embodiment of the present application;

FIG. 3 is a schematic diagram of a structure of a DBN deep neural network provided according to an embodiment of the present application;

FIG. 4 is a schematic diagram of optimizing data provided in accordance with an embodiment of the present application;

FIG. 5 is a schematic diagram of a deep belief network-based weather forecast system in accordance with an embodiment of the present application;

FIG. 6 is a schematic diagram of a computing device according to an embodiment of the present application;

fig. 7 is a schematic diagram of a computer-readable storage medium according to an embodiment of the present application.

Detailed Description

FIG. 1 is a flow diagram of a deep belief network-based weather forecast method in accordance with an embodiment of the present application; as shown in fig. 1, the weather forecast method provided in the embodiment of the present application may include the following steps:

s100, obtaining a plurality of historical observation data based on site information of a target area;

s200, constructing an enhanced decision tree and training historical observation data to obtain first prediction data;

s300, inputting the first prediction data into a CFD computational fluid dynamics model, performing steady state numerical simulation to obtain second prediction data, and obtaining flow field distribution under the incoming flow condition according to the second prediction data;

s400, interpolating the second predicted data at any position in a target area based on a Kriging interpolation method to extract an analog wind speed value, and obtaining a training data set;

s500, inputting training samples of a training data set into a DBN deep neural network, training the DBN deep neural network, and carrying out weather prediction on data to be observed by using a trained weather forecast model so as to generate a prediction result.

Optionally, the historical observation data in the present embodiment includes surface roughness data and the like. The conventional observation data can acquire the fixed-point weather information which is more accurate than the remote sensing data in a direct measurement mode. According to actual requirements, a ground observation station is established at key points of a specific scene, conventional observation data such as wind speed, wind direction, air temperature, air pressure, relative humidity and the like are collected to be used for constructing a multisource data assimilation model and a training data set of a small area of the specific scene, and meanwhile quality control is carried out on data stored in an application center database to ensure that the quality of the data is over-related and the availability and reliability of the data are improved. The quality control scheme primarily examines and marks data products in terms of climate extremum range, temporal consistency, spatial consistency, and inter-element consistency.

Optionally, the enhanced decision tree of S200 in this embodiment is a CART decision tree model. Generally, when the dependent variable of the dataset is a continuity value, a regression tree is generated, and the mean value observed by the leaf nodes can be used as a predicted value.

Wherein, the construction of the CART decision tree model comprises the following substeps:

s210: obtaining meteorological features based on the base Ny index and by using a feature selection method; the meteorological features can be selected by adopting methods such as meteorological information gain, gain rate, coefficient of base and chi-square test;

s220: dividing according to the meteorological features and regression variances to obtain regression trees; the larger the regression variance is, the more scattered (impure) the node data is, so the embodiment adopts a minimization method to divide;

s230: pruning the regression tree, correcting the regression tree by adopting a post pruning method, and constructing the CART decision tree model according to the corrected regression tree;

specifically, after the regression tree is sufficiently grown, pruning is performed based on a loss matrix or a complexity method, and the regression tree is corrected by adopting a post pruning method.

Specifically, in the prior art, a forecast result of a numerical value of observed data is generally taken as a boundary condition, and a computational fluid dynamics (CFD, computational Fluid Dynamics) model is adopted to simulate the observed data such as terrain, surface roughness and the like so as to obtain finer spatial flow field distribution in a field under a certain incoming flow condition; however, in the CFD model in the prior art, a Navier-Stokes equation (N-S equation) needs to be solved to obtain spatial flow field distribution during each prediction, so that the calculated amount is huge, and the requirement of prediction timeliness is difficult to meet.

In this embodiment, assuming that the historical observation data is a steady-state incoming flow condition, the spatial flow field distribution corresponding to the incoming flow condition in the target area is simulated to be unique; therefore, the embodiment adopts the CFD model to perform steady state numerical simulation on the spatial flow field, so as to obtain the spatial flow field distribution under the incoming flow condition.

In this embodiment, if several predicted moments have the same or similar incoming flow conditions, the flow fields in the field at these moments will also exhibit the same or similar distribution. Therefore, when wind speed is predicted at a certain moment, the flow field data of the same or similar incoming flow conditions at the previous moment (for example, the flow field data of the incoming flow conditions at intervals of 30S within 1 year) can be referred to, that is, the calculation is performed from several factors such as the size of the numerical simulation calculation domain, the interpolation order of the flow terms, the calculation grid, the judgment of the numerical simulation result, and the like, without re-solving the Navier-Stokes equation at each predicted point to obtain the spatial flow field distribution, so that the embodiment can rapidly predict the adjacent wind speed.

Wherein, step S300 comprises the following sub-steps;

s310: performing CFD pre-calculation on the flow field under the place incoming flow condition of the first prediction data to obtain key flow field data, storing the key flow field data into a database, and establishing a scene flow field characteristic database;

s320: wind speed prediction is carried out by utilizing a scene flow field characteristic database, and second prediction data are obtained;

s330: and taking the second prediction data as an input boundary condition of the CFD model, inquiring the incoming flow condition of which the second prediction data is similar, and calling flow field distribution data of a scene flow field characteristic database to obtain flow field distribution under the incoming flow condition.

The prediction method completes the complex CFD flow field numerical simulation before wind speed prediction, and reasonably solves the timeliness problem of the CFD model.

According to the embodiment, the CFD computational fluid dynamics model is utilized to calculate the mesoscale forecast data (namely the first forecast data), the small-scale forecast data (namely the second forecast data) is finally output, the small-scale forecast data and the mesoscale data are analyzed and compared, and the small-scale forecast data are continuously optimized, so that the weather forecast capability in a small-scale range is improved.

According to the embodiment, a DBN deep neural network is constructed, namely current observation data, mesoscale forecast data, an assimilation method, small-scale forecast data and the like are added into the DBN neural network, and training is carried out by using a deep learning method and various input data; referring to fig. 3, by setting a comparison test, the optimal configuration of parameters such as the hidden layer number of the neural network, the number of neurons of each layer, the training learning rate, the iteration number and the like can be determined. The matching between the small-scale forecasting result output by the deep learning and the actual observation data acquired by the fixed-context weather data sensor and the laser radar is ensured, and the timeliness and the accuracy of the small-scale forecasting result are improved.

Optionally, referring to fig. 4, in S500, the historical observation data, the first prediction data, or the second prediction data is optimized to obtain normalized prediction data, and the training data set is constructed based on the normalized prediction data.

Specifically, data such as mesoscale numerical forecast data, flow field distribution characteristics simulated by a hydrodynamic model, historical observation data and the like are optimized (namely, a format conversion method and the like are adopted), standardized data obtained after optimization are used as true values to construct a training data set, and training samples in the training data set are input into a DBN deep neural network to perform training.

In this example, the historical observation data, the first prediction data or the second prediction data and other weather data files which do not conform to the file name naming standards can be converted into standard file names, and the following will exemplify that the data of various types are subjected to unified format and stored classification so as to obtain the training data set;

the text file can be classified and stored in a standard theodolite format, namely, according to the sequence of the district station number, year, month, day, time, latitude, longitude, station height, station type, variable and quality control code, and clear missing measured value information is given;

for the binary file, a standard digital file format can be adopted, corresponding year, month, day, time and variable information is given for the first behavior description information of the binary file, and the main body part of the digital file is stored according to the order of behavior latitude and longitude;

the HDF file can adopt a standard HDF4 format, and related time, variable, attribute, longitude and latitude information, missing measurement information and the like are given in the HDF4 format file;

the NC file can adopt a standard Netcdf file format, and related time, variable, attribute, longitude and latitude information, missing measurement and other information are given in the Netcdf file.

In another embodiment, the weather forecast method based on the deep belief network further includes step S600 of comparing the forecast result obtained by the DBN deep neural network with the forecast result obtained by the traditional physical model, so as to further tune the parameters of the first forecast data of the statistical model (i.e. CFD model);

wherein the traditional physical model is, for example, a model of mode equation system solving atmospheric evolution model or a plurality of dynamic core and physical option solving models.

Because ground object echoes generated by the super refraction of radar pulses in the atmosphere belong to non-meteorological factors, radar detection results can be polluted, and the quantitative application of radar data is seriously influenced; therefore, the feature echo recognition algorithm can be adopted to tune the parameters of the first prediction data of the CFD model.

The ground object echo recognition algorithm of the embodiment is based on a fuzzy logic method. The fuzzy logic method is to analyze various radar echo characteristics by using a statistical method, find out characteristics of ground object echoes different from precipitation echoes, give the same weight to the characteristics, obtain a quantized value indicating the influence possibility of the ground object echoes in each distance library, and finally identify the ground object echo information exceeding a certain threshold value.

Because the ground target is generally static, the echo intensities are irregularly distributed, the echo intensities of the ground target are frequently present in a low elevation layer, and the echo intensity of a layer higher than the echo intensity of the ground target is far less than the echo intensity of the layer; the layered cloud precipitation level range is wider and the intensity is uniform; the convection cloud precipitation structure is compact, the height is high, and the echo intensity is strong. Therefore, the present example of optimizing parameters of the first prediction data of the CFD model using the clutter recognition algorithm comprises the following sub-steps:

s610: acquiring basic data of a radar, and acquiring echo characteristic parameters according to the basic data;

specifically, the radar base data includes: basic data such as echo intensity, radial velocity, velocity spectrum width and the like; the echo characteristic parameters include: horizontal texture, vertical variation of echo intensity, variation in radial direction, area average of radial velocity, area variance of radial velocity, area average of velocity spectrum width, and the like.

S620: analyzing the echo characteristic parameters by using a statistical method to determine the difference between the feature echo and the rainfall echo characteristic parameters;

s630: respectively setting membership functions according to the difference of characteristic parameters of each ground object echo and precipitation echo, and converting the membership functions into identification results, wherein the identification results are in the range of [0,1 ];

and S640, carrying out weighted average on the identification result to obtain a criterion value of a certain point, and judging the point as the ground object echo when the criterion value of the certain point exceeds a preset threshold value.

According to the weather forecast method, real-time data of weather and the generated forecast results are continuously displayed to a user through a display device (e.g. a web page), and early warning information is sent to the user in real time for the forecast results which are forecast to influence normal holding activities, such as strong wind, rain and snow.

In one embodiment, referring to fig. 2, a method for displaying real-time data and prediction results of weather by using a message queue background is used to realize the requirement of a user for synchronously pushing service requests by customized business data.

The message queue background includes a point-to-point message queue model including a message Producer (Producer) and a message Consumer (Consumer); the message Producer sends a message to the message consumer via a point-to-point message queue model (i.e., the message Producer sends a message to a particular queue from which the message consumer receives a message), and the message consumer signs an Acknowledge (Acknowledge) successfully processed message.

Wherein the message Producer is configured to send a message to a message queue; the message Consumer Consumer is used for receiving the message from the message queue; it should be noted that the producer and consumer of the message may not be in operation at the same time.

Optionally, the message queue background further includes a publish-subscribe message model, the publish-subscribe model including publishers and subscribers, supporting publishers to publish messages to a specific Topic (Topic), and 0 or more subscribers to receive messages from the specific Topic (Topic); referring to FIG. 2, under this model, publishers and subscribers are unaware of each other. In the actual operation process, the subscription is needed, then the message is sent, and then the subscribed message is received.

The subscription message model in this embodiment supports message persistence, which is very important for some key core services, and after message persistence is enabled, after message queue downtime is restarted, the message can be recovered from persistence storage, the message is not lost, and consumption processing can be continued. The sending and receiving of messages also supports asynchronous modes: asynchronous message sending, no need to wait for the confirmation of message queue; the message is received asynchronously, and the message consumer is triggered to receive the message in a Push mode.

In addition, the real-time data of the weather and the generated prediction result are displayed continuously through the display device, so that a data display platform is constructed, the platform supports the functions of image-text combination, a geographic information system and the like, and data graphs such as current observation data, mesoscale prediction data, small-scale prediction and the like are displayed to a user in parallel, so that data support is provided for small-range, high-frequency and accurate short-term weather prediction.

According to the embodiment, the historical observation data and the real-time observation data are predicted through the CFD computational fluid dynamics model, so that the fine prediction can be realized, the mining observation data of the CFD computational fluid dynamics model in the time dimension identifies and learns the weather patterns repeatedly appearing in the history, and the atmospheric movement process of the weather phenomenon can be predicted; in addition, the second prediction data is expanded from the position of the observation point to any point in the range of the target area by using a Kriging interpolation method; meanwhile, the correlation among meteorological elements is considered, and the DBN deep neural network is utilized to train the simulated wind speed value, so that the data to be observed are forecast, and the physical correlation among different elements of each observed data is reflected;

According to the embodiment, the data collected by the weather data sensor and the laser radar for constructing the fixed scene are taken as historical observation data, the forecast data is output from the middle scale forecast, the data is assimilated, and the weather forecast capability in the middle scale range is improved.

Based on the same inventive concept, as shown in fig. 5, the embodiment of the application further provides a weather forecast system based on a depth confidence network, where the weather forecast system includes an acquisition module, a first training module, a simulation module, an interpolation module and a second training module;

Optionally, the simulation module performs the following operations:

The foregoing identification system provided in this embodiment may perform the method provided in the foregoing embodiment of any one of the foregoing weather forecast methods based on the deep belief network, and detailed processes are detailed in the description of the method embodiment, which is not repeated herein.

The present embodiment also provides a computing device comprising a memory 520, a processor 510 and a computer program stored in said memory 520 and executable by said processor 510, which computer program is stored in a space 530 for program code in the memory 520, which computer program, when being executed by the processor 510, is realized for performing any of the identification method steps 531 according to the invention.

Embodiments of the present application also provide a computer-readable storage medium. Referring to fig. 7, the computer-readable storage medium includes a storage unit for program code provided with a program 531' for executing the steps of the identification method according to the present invention, the program being executed by a processor.

Embodiments of the present application also provide a computer program product comprising instructions. The computer program product, when run on a computer, causes the computer to perform the method steps according to the invention.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed by a computer, produces a flow or function in accordance with embodiments of the present application, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, by wired (e.g., coaxial cable, optical fiber, digital Subscriber Line (DSL)), or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.

Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of function in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

Those of ordinary skill in the art will appreciate that all or some of the steps in implementing the methods of the above embodiments may be implemented by a program that instructs a processor to perform the steps, and the program may be stored in a computer readable storage medium, where the storage medium is a non-transitory (english) medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid state disk, a magnetic tape (english), a floppy disk (english), an optical disc (english), and any combination thereof.

The foregoing is merely a preferred embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions easily contemplated by those skilled in the art within the technical scope of the present application should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A weather forecast method based on a deep belief network, the weather forecast method comprising the steps of:

inputting training samples of a training data set into a DBN deep neural network, training the DBN deep neural network, and carrying out weather prediction on data to be observed by utilizing a trained weather forecast model so as to generate a prediction result;

the flow field distribution under the incoming flow condition obtained according to the second predicted data comprises the following substeps:

2. The weather forecast method of claim 1, wherein the weather forecast method compares the forecast results obtained by the DBN deep neural network with the forecast results obtained by the conventional physical model for optimizing parameters of the first forecast data.

3. The weather forecast method according to claim 2, characterized in that the optimization of the parameters of the first prediction data of the CFD model is performed by using a ground object echo identification algorithm, comprising the following sub-steps:

4. The weather forecast method of claim 2, wherein the traditional physical model solves for a system of mode equations an atmospheric evolution model, a plurality of dynamic cores, and a physical options solving model.

5. The weather forecast method of claim 1, wherein the enhanced decision tree is a CART decision tree model.

6. The weather forecast method of claim 5, wherein constructing the CART decision tree model comprises the sub-steps of:

7. The weather forecast method of any of claims 1-6, wherein historical observation data, first prediction data, or second prediction data is optimized to obtain normalized prediction data, and the training data set is constructed based on the normalized prediction data.

8. A weather forecast system based on a depth confidence network comprises an acquisition module, a first training module, a simulation module, an interpolation module and a second training module;

the second training module is used for inputting training samples of the training data set into the DBN deep neural network, training the DBN deep neural network, and carrying out weather prediction on data to be observed by utilizing a trained weather forecast model so as to generate a prediction result;

the simulation module performs the following operations: