CN114661700A - Artificial influence weather operation effect inspection method based on AI - Google Patents
Artificial influence weather operation effect inspection method based on AI Download PDFInfo
- Publication number
- CN114661700A CN114661700A CN202210168886.5A CN202210168886A CN114661700A CN 114661700 A CN114661700 A CN 114661700A CN 202210168886 A CN202210168886 A CN 202210168886A CN 114661700 A CN114661700 A CN 114661700A
- Authority
- CN
- China
- Prior art keywords
- data
- weather
- model
- features
- convection
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/26—Visual data mining; Browsing structured data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Abstract
The invention discloses an artificial weather modification operation effect inspection method based on AI, which comprises the steps of acquiring artificial weather modification operation effect data based on dual-polarization radar data, selecting the weather process generated by the convection cloud, cleaning the data after obtaining the weather related data, carrying out AI model training by using the radar data of the convection cloud process, by selecting the process case of artificially influencing weather, intercepting the historical data before influence, using an AI model for prediction, and comparing with the actually occurred weather process, selecting typical case to perform effect evaluation to obtain report, the technology can be combined with the evaluation of weather effect artificially influenced, so that the forecasting accuracy is improved to a certain extent, the forecasting result is closer to the real occurrence result, in the effect evaluation of artificial weather influence, the prediction result without operation is closer to the real situation, so that the evaluation result is convincing.
Description
Technical Field
The invention belongs to the technical field of weather operation effect inspection, and particularly relates to an artificial influence weather operation effect inspection method based on AI.
Background
As AI technology has begun to enter the field of weather forecasting, areas in the shanghai, guangdong, etc. have begun to use AI technology to assist in manual weather forecasting. As the amount of data accumulates, AI-based technology forecasts may exceed human forecasters in the foreseeable future. However, based on the confidence of AI prediction, the combination of artificial weather effect evaluation and AI technology is not required at present, the prediction accuracy is not obviously improved, and the prediction result deviates from the real occurrence result. Therefore, in the effect evaluation of artificial weather influence, a certain difference exists between the prediction result of non-operation and the real situation, so that the evaluation result cannot be convincing, and therefore an AI-based artificial weather influence operation effect inspection method is urgently needed.
Disclosure of Invention
The invention aims to provide an AI-based artificial weather effect inspection method, which can combine the evaluation of artificial weather effect with the technology, improve the forecast accuracy to a certain extent, and make the forecast result and the real occurrence result closer and closer to each other, so as to solve the problems in the background technology.
In order to achieve the purpose, the invention adopts the following technical scheme:
an AI-based method for inspecting the effect of weather modification work, which comprises the following steps:
s1, selecting a weather process generated by convection cloud based on the dual-polarization radar data to obtain weather related data;
s2, carrying out data cleaning on weather related data, and carrying out AI model training by using radar data of a convection cloud process;
s3, selecting a process case artificially influencing weather, intercepting historical data before influence, predicting by using an AI model, and comparing with an actually occurring weather process;
and S4, selecting a typical case for effect evaluation to obtain a report.
Preferably, the dual-polarization radar data in step 1 is transmitted in a dual-transmitting and dual-receiving mode, the transmitter outputs signals which are divided into two paths and are transmitted simultaneously in a horizontal polarization mode and a vertical polarization mode, and the data are processed in parallel by the dual-channel receiver and the digital intermediate frequency, wherein the amplitude difference of the two channels is less than 0.2dB, and the phase difference of the two channels is less than 2 degrees.
Preferably, the weather process generated by convection clouds is caused by convection unstable stratification in the atmosphere and is accompanied by weather phenomena of rainstorm, strong wind, hail and tornado, the characteristics of the convection weather comprise that the convection weather is a product of accumulated rain clouds with vigorous convection, the range is small, the development is fast, and when the selection is performed, data related to manual operation influence areas are selected, wherein the data comprises X-axis and Y-axis horizontal wind directions, a coordinate origin point acquisition operation point and a rectangular coordinate system.
Preferably, in the step 2, the data cleaning is to visually check the characteristics of the data in a data visualization manner, and process the data to solve the problems of data missing, abnormal values, data imbalance and dimension inconsistency, where the data imbalance is caused by data deviation, and the model trained later is over-fit or under-fit.
Preferably, the AI model training includes feature extraction, feature selection and a training model, the feature extraction includes numerical feature data, label or description class data, unstructured data and network relation data, the feature selection is to set corresponding coverage and IV indexes for the features entered into the model, then to screen the features according to the indexes and a threshold determined according to experience, to confirm the stability of the features, if the features are stable, the features are retained, if the features are unstable, the features are removed.
Preferably, the data in the numerical characteristic data includes a large number of numerical characteristics, and the numerical characteristics can be directly obtained from the number bins, specifically, the main characteristic is extracted first, and then other dimensional characteristics are extracted; the label or description class data contains low category relevance, three categories are converted into features, and each feature value is represented by 0 or 1; the unstructured data exist in UGC content data, text data are cleaned and mined when unstructured features are extracted, and features reflecting user attributes are mined; network relational data is data that is numerical, characteristic, label, or describes class data and relationships around unstructured data.
Preferably, when the AI model is used for prediction in step 3, prediction is performed according to historical data before a case, the predicted data is compared with data in an actual weather process, when the error of the comparison result is less than 1%, S4 is performed, when the error of the comparison result is greater than 1%, the AI model needs to be trained again, and after the training is completed, the AI model is predicted again and compared with the actual weather process.
Preferably, before the AI model is retrained, the data information stored in the previous training needs to be formatted, new specific polarization radar-based data is reselected, and the AI model training is performed again after the data is cleaned.
Preferably, when the effect evaluation is performed on the typical case in the step 4, if the error between the predicted weather result and the data result of the actual weather occurrence process is within 2%, the effect of the weather modification operation is not obvious, and if the error is greater than 2%, the effect of the weather modification operation is obvious.
Compared with the prior art, the method for testing the weather modification operation effect based on the AI has the following advantages that:
according to the method, the weather process generated by the convection cloud is selected based on the dual-polarization radar data, the weather related data is obtained and then is subjected to data cleaning, the radar data in the convection cloud process is used for AI model training, the process case of artificially influencing the weather is selected, the historical data before influence is intercepted, the AI model is used for prediction and is compared with the actually occurring weather process, the artificially influencing weather effect evaluation and the technology can be combined, the prediction accuracy is improved to a certain extent, the prediction result is closer to the actually occurring result, the prediction result without operation is closer to the actual situation in the artificially influencing weather effect evaluation, and the evaluation result is convincing.
Drawings
FIG. 1 is a block flow diagram of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. The specific embodiments described herein are merely illustrative of the invention and do not delimit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention provides an AI-based method for checking the effect of weather modification operation, which comprises the following steps:
s1, selecting a weather process generated by convection cloud based on the dual-polarization radar data to obtain weather related data;
the dual-polarization radar data adopts a dual-transmitting and dual-receiving mode, an output signal of a transmitter is divided into two paths and is simultaneously transmitted in a horizontal and vertical polarization mode, a dual-channel receiver and a digital intermediate frequency carry out parallel processing on the data, wherein the amplitude difference of the two channels is less than 0.2dB, and the phase difference of the two channels is less than 2 degrees.
The weather process generated by convection clouds is caused by unstable convection stratification in the atmosphere and is accompanied by weather phenomena of rainstorm, strong wind, hail and tornado, the characteristics of the convection weather comprise that the convection weather is a product of accumulated rain clouds with vigorous convection, the characteristics of small range and rapid development are realized, and when the selection is carried out, data related to manual operation influence areas are selected, including X-axis and Y-axis horizontal wind directions, and a coordinate origin point acquisition operation point and a rectangular coordinate system are included.
S2, carrying out data cleaning on weather related data, and carrying out AI model training by using radar data of a convection cloud process;
the data cleaning is to visually check the characteristics of data in a data visualization mode and process the data to solve the problems of data loss, abnormal values, data imbalance and inconsistent dimensions of the data, wherein the data imbalance is caused by overfitting or underfitting of a model trained later due to data deviation.
The AI model training comprises feature extraction, feature selection and a training model, wherein the feature extraction comprises numerical feature data, label or description class data, unstructured data and network relation data, the feature selection is to set corresponding coverage and IV indexes for the features entering the model, then to screen the features according to the indexes and a threshold value determined by experience, to confirm the stability of the features, if the features are stable, to keep the features, and if the features are unstable, to remove the features.
The data in the numerical characteristic data comprises a large number of numerical characteristics, the numerical characteristics can be directly obtained from the number bins together, and specifically, the main characteristic is extracted firstly, and then other dimensional characteristics are extracted; the label or description class data contains low category relevance, three categories are converted into features, and each feature value is represented by 0 or 1; the unstructured data exist in UGC content data, text data are cleaned and mined when unstructured features are extracted, and features reflecting user attributes are mined; network relational data is data that is numerical, characteristic, label, or describes class data and relationships around unstructured data.
S3, selecting a process case artificially influencing weather, intercepting historical data before influence, predicting by using an AI model, and comparing with an actually occurring weather process;
when the AI model is used for prediction, prediction is carried out according to historical data before a case, the predicted data is compared with data in the actual weather process, S4 is carried out when the error of the comparison result is less than 1%, the AI model needs to be trained again when the error of the comparison result is more than 1%, and the AI model is predicted again after the training is finished and compared with the actual weather process.
Before the AI model is retrained, the data information stored in the previous training needs to be formatted, new specific polarization radar-based data is reselected, and the AI model training is carried out again after the data is cleaned.
S4, selecting a typical case for effect evaluation to obtain a report;
when effect evaluation is carried out with a typical case, the effect of the weather modification operation is not obvious according to the fact that the error between the predicted weather result and the data result of the actual weather generation process is within 2%, and when the error is larger than 2%, the effect of the weather modification operation is obvious.
In conclusion, by combining the artificial weather effect evaluation with the AI technology, the forecasting accuracy is improved to a certain extent, the forecasting result is closer to the real occurrence result, so that the forecasting result without operation is closer to the real situation in the artificial weather effect evaluation, and the evaluation result is convincing.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments or portions thereof without departing from the spirit and scope of the invention.
Claims (9)
1. An AI-based method for inspecting the effect of weather modification operation is characterized in that: the method comprises the following steps:
s1, selecting a weather process generated by convection cloud based on the dual-polarization radar data to obtain weather related data;
s2, carrying out data cleaning on weather related data, and carrying out AI model training by using radar data of a convection cloud process;
s3, selecting a process case artificially influencing weather, intercepting historical data before influence, predicting by using an AI model, and comparing with an actually occurring weather process;
and S4, selecting a typical case for effect evaluation to obtain a report.
2. The AI-based weather modification work effect verification method as in claim 1, wherein: in the step 1, the dual-polarization radar data adopts a dual-transmitting and dual-receiving mode, an output signal of a transmitter is divided into two paths and is simultaneously transmitted in a horizontal and vertical polarization mode, a dual-channel receiver and a digital intermediate frequency carry out parallel processing on the data, wherein the amplitude difference of the two channels is less than 0.2dB, and the phase difference of the two channels is less than 2 degrees.
3. The AI-based weather modification work effect verification method of claim 2, wherein: the weather process generated by convection clouds is caused by unstable convection stratification in the atmosphere and is accompanied by weather phenomena of rainstorm, strong wind, hail and tornado, the characteristics of the convection weather comprise that the convection weather is a product of accumulated rain clouds with vigorous convection, the characteristics of small range and rapid development are realized, and when the selection is carried out, data related to manual operation influence areas are selected, including X-axis and Y-axis horizontal wind directions, and a coordinate origin, namely an operation point and a rectangular coordinate system are taken.
4. The AI-based weather modification work effect verification method as in claim 1, wherein: in the step 2, the data cleaning is to visually check the characteristics of the data in a data visualization mode, and process the data to solve the problems of data loss, abnormal values, data imbalance and inconsistent dimensions of the data, wherein the data imbalance is caused by over-fitting or under-fitting of a model trained later due to data deviation.
5. The AI-based weather modification work effect verification method of claim 4, wherein: the AI model training comprises feature extraction, feature selection and a training model, wherein the feature extraction comprises numerical feature data, label or description class data, unstructured data and network relation data, the feature selection is to set corresponding coverage and IV indexes for the features entering the model, then to screen the features according to the indexes and a threshold value determined by experience, to confirm the stability of the features, if the features are stable, to keep the features, and if the features are unstable, to remove the features.
6. The AI-based weather modification work effect verification method of claim 5, wherein: the data in the numerical characteristic data comprises a large number of numerical characteristics, the numerical characteristics can be directly obtained from a plurality of bins together, and specifically, the main characteristics are extracted firstly, and then other dimensional characteristics are extracted; the label or description class data contains low category relevance, three categories are converted into features, and each feature value is represented by 0 or 1; the unstructured data exist in UGC content data, text data are cleaned and mined when unstructured features are extracted, and features reflecting user attributes are mined; network relational data is data that is numerical, characteristic, label, or describes class data and relationships around unstructured data.
7. The AI-based weather modification work effect verification method of claim 6, wherein: when the AI model is used for prediction in the step 3, prediction is carried out according to historical data before a case, the predicted data is compared with data in the actual weather process, S4 is carried out when the error of the comparison result is less than 1%, the AI model needs to be trained again when the error of the comparison result is more than 1%, and the AI model is predicted again after the training is finished and compared with the actual weather process.
8. The AI-based weather modification work effect verification method of claim 7, wherein: before the AI model is retrained, the data information stored in the previous training needs to be formatted, new specific polarization radar-based data is reselected, and the AI model training is performed again after the data is cleaned.
9. The AI-based weather modification work effect verification method of claim 1, wherein: and 4, when effect evaluation is carried out on the typical case in the step 4, if the error between the predicted weather result and the data result of the actual weather occurrence process is within 2%, the effect of the weather modification operation is not obvious, and if the error is more than 2%, the effect of the weather modification operation is obvious.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210168886.5A CN114661700A (en) | 2022-02-23 | 2022-02-23 | Artificial influence weather operation effect inspection method based on AI |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210168886.5A CN114661700A (en) | 2022-02-23 | 2022-02-23 | Artificial influence weather operation effect inspection method based on AI |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114661700A true CN114661700A (en) | 2022-06-24 |
Family
ID=82026608
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210168886.5A Pending CN114661700A (en) | 2022-02-23 | 2022-02-23 | Artificial influence weather operation effect inspection method based on AI |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114661700A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117329830A (en) * | 2023-11-28 | 2024-01-02 | 北京万通益生物科技有限公司 | Automatic monitoring control system of lactobacillus powder heating and drying equipment |
-
2022
- 2022-02-23 CN CN202210168886.5A patent/CN114661700A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117329830A (en) * | 2023-11-28 | 2024-01-02 | 北京万通益生物科技有限公司 | Automatic monitoring control system of lactobacillus powder heating and drying equipment |
| CN117329830B (en) * | 2023-11-28 | 2024-04-05 | 北京万通益生物科技有限公司 | Automatic monitoring control system of lactobacillus powder heating and drying equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110019595B (en) | Multi-source meteorological data integration method and system | |
| CN112199421B (en) | Multi-source heterogeneous data fusion and measurement data multi-source mutual verification method and system | |
| CN108470022B (en) | Intelligent work order quality inspection method based on operation and maintenance management | |
| CN112415331B (en) | Power grid secondary system fault diagnosis method based on multi-source fault information | |
| CN108632832B (en) | Network coverage analysis method and system | |
| CN110134907B (en) | Rainfall missing data filling method and system and electronic equipment | |
| CN105425320A (en) | Probabilistic forecasting method and system of coastal gale caused by tropical cyclone | |
| CN111160612A (en) | Off-site flight delay analysis and prediction method based on weather influence | |
| CN112243249A (en) | LTE new access anchor point cell parameter configuration method and device under 5G NSA networking | |
| CN110348683A (en) | The main genetic analysis method, apparatus equipment of electrical energy power quality disturbance event and storage medium | |
| CN116388402B (en) | Abnormality alarm analysis method applied to power transformation equipment | |
| CN109993390B (en) | Alarm association and order dispatching optimization method, device, equipment and medium | |
| CN114661700A (en) | Artificial influence weather operation effect inspection method based on AI | |
| CN103942251A (en) | Method and system for inputting high altitude meteorological data into database based on multiple quality control methods | |
| CN113869633A (en) | Power distribution network multi-source data quality control method | |
| CN113644741A (en) | Mass fault recording data processing method based on edge calculation | |
| CN111427101B (en) | Thunderstorm strong wind grading early warning method, system and storage medium | |
| CN110751201B (en) | SAR equipment task failure cause reasoning method based on textural feature transformation | |
| CN117034149A (en) | Fault processing strategy determining method and device, electronic equipment and storage medium | |
| CN115600142B (en) | Different-method-based strong wind fusion integration forecasting method | |
| CN108596450B (en) | Power grid risk early warning method and system | |
| CN116151799A (en) | BP neural network-based distribution line multi-working-condition fault rate rapid assessment method | |
| CN116663393A (en) | Random forest-based power distribution network continuous high-temperature fault risk level prediction method | |
| CN116205631A (en) | VMD-CEEMD wind turbine generator gearbox vibration fault analysis method based on dragonfly algorithm | |
| CN115882451A (en) | Method, device and equipment for predicting generated power of new energy power station |
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 |