CN114330923A - Photovoltaic power generation power prediction method based on public meteorological data - Google Patents
Photovoltaic power generation power prediction method based on public meteorological data Download PDFInfo
- Publication number
- CN114330923A CN114330923A CN202210011626.7A CN202210011626A CN114330923A CN 114330923 A CN114330923 A CN 114330923A CN 202210011626 A CN202210011626 A CN 202210011626A CN 114330923 A CN114330923 A CN 114330923A
- Authority
- CN
- China
- Prior art keywords
- data
- meteorological data
- photovoltaic power
- power generation
- historical
- 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
- 238000010248 power generation Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 title claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000012795 verification Methods 0.000 claims description 26
- 238000012360 testing method Methods 0.000 claims description 19
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 abstract 1
- 229910052799 carbon Inorganic materials 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000000556 factor analysis Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Photovoltaic Devices (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a photovoltaic power generation power prediction method based on public meteorological data, which comprises the steps of firstly obtaining historical meteorological data disclosed in an hour level and predicted meteorological data in the future for 24 hours from a public meteorological data provider, and obtaining historical photovoltaic power generation power data in a 5-minute level from a photovoltaic power station, and then aligning the three data according to time; secondly, extracting characteristics of historical photovoltaic power generation power data, historical meteorological data and future meteorological data, and training a prediction model based on a regression prediction algorithm, so that the photovoltaic power generation power of the future 24 hours and 15 minutes is predicted, and the photovoltaic power generation power value to be predicted finally is obtained. On the basis of a photovoltaic power generation power prediction algorithm, the method for data alignment and feature extraction aiming at public meteorological data is added, so that the photovoltaic power generation power prediction without professional meteorological data can be realized.
Description
Technical Field
The invention relates to a machine learning method for ultra-short term prediction of photovoltaic power generation power, in particular to a photovoltaic power generation power prediction method based on public meteorological data.
Background
With the advance of national 'double-carbon' targets and whole county photovoltaic test points, more and more distributed roof small photovoltaic power stations are built and connected into a power distribution network, so that the problems of bus voltage rise, three-phase imbalance and the like of the power distribution network are caused, and equipment faults of the power distribution network and even the interruption of the power distribution network are further caused. One of the most effective solutions to the problems brought by large-scale distributed photovoltaic power distribution network access is to accurately predict the ultra-short-term power generation power of the distributed photovoltaic power station, so that more precise and economic optimal scheduling of the power distribution network is supported, and the challenges brought by the distributed photovoltaic power station access to the power distribution network are solved. Along with the upgrading construction of a power system, particularly the updating and upgrading of a distributed photovoltaic grid-connected inverter, more and more distributed photovoltaic power station data are accessed to a cloud for storage and management, so that basic data support is provided for ultra-short-term prediction of distributed photovoltaic power generation power, and ultra-short-term accurate prediction of the distributed photovoltaic power generation power becomes possible.
Most of the existing photovoltaic power generation power prediction algorithms are directed at medium and large photovoltaic power stations of MW and above, and the default meteorological data of the algorithms mainly comprise professional numerical meteorological forecast data which mainly comprise illumination intensity, irradiance and the like of the area where the photovoltaic power station is located. However, for a large number of small-scale distributed photovoltaic power stations, due to the problems of cost, property ownership and the like, specific positions of areas where distributed photovoltaics are located and professional numerical weather forecast data cannot be obtained, so that the operation performance of a photovoltaic power generation power prediction algorithm is seriously affected. Therefore, research aims at the generated power prediction algorithm of a large number of small-sized distributed photovoltaic power stations, public meteorological data which can be obtained freely are used for replacing professional numerical weather forecast data to serve as input data, the construction requirements of a high-quality novel power distribution network under the background of 'double carbon' can be better met, and the method has strong social requirements and great economic value.
Disclosure of Invention
In order to solve the technical problems, the invention provides a photovoltaic power generation power prediction method based on public meteorological data, which predicts the photovoltaic power generation power based on free public meteorological data which is conveniently obtained, and effectively avoids dependence on professional meteorological data.
The invention relates to a photovoltaic power generation power prediction method based on public meteorological data, which comprises the following steps:
step 1, acquiring historical meteorological data of a photovoltaic power station in a district and county with historical time not less than 100 days, and the temperature, humidity, wind speed and wind direction of an hour level 24 hours in the future, and meteorological data of the highest temperature, the lowest temperature, the average temperature, sunrise time and sunset time of each day (the daily level) from a public meteorological data provider by using an API (application program interface) calling interface;
step 2, obtaining historical generating power data of the photovoltaic power station, which is not less than 100 days at the level of 5 minutes, from a grid-connected inverter of the photovoltaic power station or management software of the photovoltaic power station;
step 3, processing the meteorological data and the historical generated power data to align the meteorological data and the historical generated power data;
step 4, selecting data from 45 days to 15 days ago as a training set, data from 15 days ago as a verification set, data from 24 hours in the future as a test set, and respectively extracting photovoltaic power generation power data characteristics as input data of a next prediction algorithm;
step 5, training a regression prediction model by using the characteristic data extracted from the historical meteorological data and the historical generated power data in the step 4, and screening out optimal parameters and an optimal model by using a verification set;
and 6, predicting a photovoltaic power generation power predicted value 24 hours in the future by using the regression prediction model trained in the step 5 to obtain a final prediction result.
Further, in step 3, the step of aligning the meteorological data and the generated power data comprises:
step 3-1, dividing historical generated power data of the photovoltaic power station into intervals of 15 minutes, averaging a plurality of generated powers falling in the same interval, and taking the averaged generated powers as aligned generated powers;
step 3-2, intersecting the aligned generated power in the step 3-1 with the meteorological data of the hour level and the daily level in a left full-connection mode to obtain the intersection of the historical generated power of the 15-minute level and the meteorological data of the hour level;
step 3-3, the missing meteorological data in the intersection set obtained in the step 3-2 comprise continuous meteorological data and discrete meteorological data, and the continuous meteorological data is filled with the first effective value before the missing value; and filling the discrete type meteorological data with the previous effective value of the missing value, and if the previous effective value does not exist, filling the discrete type meteorological data with the first subsequent effective value.
Further, the continuous weather comprises temperature, humidity, wind speed, highest temperature, lowest temperature and average temperature; the discrete type weather comprises wind direction, weather type, sunrise time and sunset time.
Further, in step 4, the step of extracting the photovoltaic power generation power data features is as follows:
4-1, selecting data from 45 days ago to 15 days ago as a training set, data from 15 days ago to the present as a verification set, and data for 24 hours in the future as a test set;
step 4-2, respectively extracting the maximum value, the minimum value, the average value and the median value of the interval between the first 3 days, the first 7 days and the first 30 days of the lowest temperature, the maximum temperature, the humidity, the wind speed and the maximum temperature corresponding to each record in the training set, the verification set and the test set as characteristics;
4-3, extracting whether the current time corresponding to each record in the training set, the verification set and the test set is between sunrise and sunset time intervals or not, and taking the time distance relative to sunrise as a characteristic;
4-4, taking the current time corresponding to each record in the training set, the verification set and the test set, and the weather type, the wind power level and the wind direction of the current day as characteristics;
and 4-5, combining all the characteristics to serve as input data of a next prediction algorithm.
Further, the step of training the regression prediction model and selecting the optimal parameters and the optimal model in the step 5 is as follows:
step 5-1, the regression prediction model is one of XGboost, LightGBM, neural Network, Gradient Boosting Decision Tree, LSTM and GRU;
step 5-2, training a regression model by respectively using different parameters by using the training set characteristics extracted in the step 4-5 as input and RMSE as an evaluation index;
and 5-3, selecting the model with the minimum RMSE as the optimal model by using the scores of the regression models with different parameters trained in the step 5-2 on the verification set, and taking the corresponding parameters as the optimal parameters.
The invention has the beneficial effects that: the method of the invention predicts the photovoltaic power generation power based on the free public meteorological data of hour-level and day-level temperature, humidity, wind speed, wind direction, sunrise time and sunset time on the Internet, and avoids the dependence of the existing photovoltaic power generation power algorithm on professional meteorological data containing irradiance, illumination intensity and the like in 15 minutes; compared with the existing ultra-short term prediction method for the photovoltaic power generation power, the method provided by the invention uses the temperature, humidity, wind speed, wind direction, highest temperature of the day level, lowest temperature, average temperature, sunrise time and sunset time of the hour level to replace professional meteorological data, and can further improve the accuracy of the prediction of the photovoltaic power generation power by constructing the characteristics of the photovoltaic power generation power, the temperature, humidity, wind speed, maximum temperature, 3 days before the lowest temperature, 7 days before the lowest temperature, 1 month before the highest temperature and the maximum, minimum, average, median and sunrise time and sunset time of the same time interval.
Drawings
In order that the present invention may be more readily and clearly understood, a more particular description of the invention briefly described above will be rendered by reference to specific embodiments that are illustrated in the appended drawings.
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a diagram of the predicted effect of the embodiment of the present invention.
Detailed Description
As shown in fig. 1, the photovoltaic power generation power prediction method based on public meteorological data according to the present invention includes the following steps:
step 1, calling an interface or a crawler by using an API (application programming interface), and acquiring historical meteorological data of the photovoltaic power station in the county at the historical time of not less than 100 days and the temperature, humidity, wind speed and wind direction at the hour level of 24 hours in the future from a public meteorological data provider, and the meteorological data of the highest temperature, the lowest temperature, the average temperature, the sunrise time and the sunset time of each day (the daily level);
step 2, obtaining historical generating power data of the photovoltaic power station, which is not less than 100 days at the level of 5 minutes, from a grid-connected inverter of the photovoltaic power station or management software of the photovoltaic power station;
step 3, processing the meteorological data and the historical generated power data to align the meteorological data and the historical generated power data;
step 3-1, dividing historical generated power data of the photovoltaic power station into intervals of 15 minutes, averaging a plurality of generated powers falling in the same interval, and taking the averaged generated powers as aligned generated powers;
step 3-2, intersecting the aligned generated power in the step 3-1 with the meteorological data of the hour level and the daily level in a left full-connection mode to obtain the intersection of the historical generated power of the 15-minute level and the meteorological data of the hour level;
3-3, filling continuous meteorological data such as temperature, humidity, wind speed, highest temperature, lowest temperature and average temperature which are obtained in the intersection set in the step 3-2 with a first effective value before the missing value, filling discrete type meteorological data such as wind direction, weather type, sunrise time and sunset time with a first effective value before the missing value, and if the former effective value does not exist, filling the former effective value with the latter first effective value;
step 4, extracting photovoltaic power generation power data characteristics as input data of a next prediction algorithm;
4-1, selecting data from 45 days ago to 15 days ago as a training set, data from 15 days ago to the present as a verification set, and data for 24 hours in the future as a test set;
step 4-2, respectively extracting the maximum value, the minimum value, the average value and the median value of the interval between the first 3 days, the first 7 days and the first 30 days of the lowest temperature, the maximum temperature, the humidity, the wind speed and the maximum temperature corresponding to each record in the training set, the verification set and the test set as characteristics;
4-3, extracting whether the current time corresponding to each record in the training set, the verification set and the test set is between sunrise and sunset time intervals or not, and taking the time distance relative to sunrise as a characteristic;
4-4, taking the current time corresponding to each record in the training set, the verification set and the test set, and the weather type, the wind power level and the wind direction of the current day as characteristics;
4-5, combining all the characteristics to be used as input data of a next prediction algorithm;
step 5, training a regression prediction model by using the characteristic data extracted from the historical meteorological data and the historical generated power data in the step 4, and screening out optimal parameters and an optimal model by using a verification set;
step 5-1, the regression prediction model is one of XGboost, LightGBM, neural Network, Gradient Boosting Decision Tree, LSTM and GRU;
step 5-2, training a regression model by respectively using different parameters by using the training set characteristics extracted in the step 4-5 as input and RMSE as an evaluation index;
step 5-3, selecting the model with the minimum RMSE as the optimal model by using the scores of the regression models with different parameters trained in the step 5-2 on the verification set, and taking the corresponding parameters as the optimal parameters;
and 6, predicting a predicted value of the photovoltaic power generation power at the level of 15 minutes in 24 hours in the future by using the regression prediction model trained in the step 5, and obtaining a final prediction result.
The process of the present invention is illustrated below with reference to examples.
The photovoltaic power generation power prediction method based on the public meteorological data comprises the following steps:
step 1, acquiring public meteorological data WT of CT (computed tomography) in county of photovoltaic power station from darksky by using API (application programming interface) interface function, wherein the public meteorological data WT contains temperature, humidity, wind speed, wind direction at hour level, maximum temperature, minimum temperature, average temperature, sunrise time and sunset time at each day, and data are shown in tables 1 and 2:
TABLE 1 Meteorological data example on the hourly scale
| Time | Temperature of | Humidity | Wind speed | Wind direction | Air pressure |
| 2021-01-01 00:00:00 | 8.5 | 0.86 | 2.77 | 90 | 1016.76 |
| 2021-01-01 01:00:00 | 9.2 | 0.89 | 2.95 | 90 | 1016.5 |
| … | … | … | … | … | … |
TABLE 2 weather data on a daily scale
| Time | Maximum temperature | Minimum temperature | Mean temperature | Sunrise time | Time of sunset |
| 2021-01-01 | 15 | 6 | 9 | 06:35 | 17:52 |
| 2021-01-02 | 14 | 2 | 7 | 06:37 | 17:54 |
| … | … | … | … | … | … |
Step 2, obtaining historical generating power data SOHD of the photovoltaic power station from a photovoltaic power station management platform, as shown in table 3 below:
TABLE 3 example data for historical generated power of photovoltaic power plant
| Time | Power station numbering | Generated power |
| 2021-01-01 07:00:00 | 380021 | 27.5 |
| 2021-01-01 07:05:00 | 380021 | 28.2 |
| 2021-01-01 07:10:00 | 380021 | 29.1 |
| … | … | … |
Step 3, aligning the data of the WT and the SOHD and then splicing the WT and the SOHD together:
and 3-1, dropping the SOHD data into time windows which are divided into one window every 15 minutes according to time fields in the SOHD data, grouping each time window, and taking an average value of the generated power data in the same group as the generated power so that only one generated power exists in each time window finally, thereby obtaining the photovoltaic generated power data SHD15 in the 15-minute level. Examples after conversion to data on the 15 minute scale are given in table 4 below:
TABLE 415 minute-scale photovoltaic Power Generation data
| Time | Power station numbering | Generated power |
| 2021-01-01 07:00:00 | 380021 | 29.2 |
| 2021-01-01 07:15:00 | 380021 | 30.5 |
| 2021-01-01 07:30:00 | 380021 | 33.4 |
| … | … | … |
Step 3-2, adding a time axis with a value every 15 minutes to the meteorological data WT, then adopting a missing value filling method of filling the previous effective value, and filling the vacant weather data on the time axis by using the data of the WT to obtain new meteorological data WT15 in the level of 15 minutes; performing full left connection on the SHD15 obtained in the step 3-1 and the WT to obtain aligned photovoltaic power generation power and meteorological data;
and 3-3, filling continuous meteorological data such as the temperature, the humidity, the wind speed, the highest temperature and the lowest temperature which are missing in the intersection set obtained in the step 3-2 with a first effective value before the missing value, filling discrete meteorological data such as the wind direction, the weather type, the sunrise time and the sunset time with a first effective value before the missing value, and if the former effective value does not exist, filling the former effective value with the latter first effective value, and finally obtaining aligned and filled data DS 1.
The meteorological data after the processing of steps 3-2 and 3-3 are shown in the following table 5:
TABLE 5 weather data after 15 min conversion
And 4, respectively extracting the characteristics of the photovoltaic power generation power and the characteristics of the meteorological data, and combining into a final characteristic set:
step 4-1, selecting data from the data set DS1 before 45 days to 15 days as a training set, data from 15 days to the present as a verification set, and data of 24 hours in the future as a test set;
step 4-2, respectively extracting the maximum value, the minimum value, the average value and the median value of the same time interval of the first 3 days, the first 7 days and the first 1 month of the average temperature corresponding to the photovoltaic power generation power, the temperature, the humidity, the wind speed, the maximum temperature, the minimum temperature and the average temperature of each record in the training set, the verification set and the test set to form a feature set SF 1;
4-3, extracting whether the current time corresponding to each record in the training set, the verification set and the test set is between sunrise and sunset time intervals, and taking the time distance relative to sunrise as a characteristic to obtain a sunrise and sunset characteristic set WF 1;
4-4, extracting current time corresponding to each record in the training set, the verification set and the test set, and taking the weather type, the wind power level and the wind direction of the current day as characteristics to obtain a characteristic set SF 2;
step 4-5, merging the obtained feature sets to obtain a final complete feature set FF;
step 5, training a regression prediction model M by using the acquired feature data set FF and historical time data in the feature data set FF as a training set and using the conventional open-source LightGBM regression algorithm; screening out optimal parameters and optimal models by using a verification set;
step 5-1, the regression prediction model is one of XGboost, LightGBM, neural Network, Gradient Boosting Decision Tree, LSTM and GRU;
step 5-2, training a regression model by using the extracted training set characteristics as input and RMSE (reduced form-factor analysis) carried by a regression algorithm as an evaluation index and respectively using different parameters;
step 5-3, selecting the model with the minimum RMSE as the optimal model by using the scores of the regression models with different parameters trained in the step 5-2 on the verification set, and taking the corresponding parameters as the optimal parameters;
and 6, predicting the photovoltaic power generation power of the time period to be predicted by using the acquired feature data of the time period to be predicted in the feature data set FF as the input of the model M trained in the step 5, and taking the predicted photovoltaic power generation power as the final predicted value of the photovoltaic power generation power of the future 24 hours to be predicted. The predicted effect is shown in fig. 2.
The prediction method comprises the steps of firstly, carrying out time-axis expansion on public meteorological data to align the public meteorological data with generated power data, so that an association relation can be established between the generated power and the meteorological data; secondly, by using the principle that the photovoltaic power generation power has strong autocorrelation and periodicity, the change characteristic of the photovoltaic power generation power is described by constructing the statistical characteristics of maximum, minimum, average, median and the like of the photovoltaic power generation power in the same period of nearly 3 days, 7 days and nearly 1 month, so that the prediction of the photovoltaic power generation power in the ultra-short period is realized. The method can realize the photovoltaic power generation power prediction by using the least open free meteorological data, has lower cost and lower requirements on computing resources, meteorological data and the like, and has better practical engineering application value.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and all equivalent variations made by using the contents of the present specification and the drawings are within the protection scope of the present invention.
Claims (5)
1. A photovoltaic power generation power prediction method based on public meteorological data is characterized by comprising the following steps:
step 1, acquiring historical meteorological data of the historical time of a district or county where a photovoltaic power station is located, the temperature, the humidity, the wind speed and the wind direction of the historical meteorological data of the hour level of 24 hours in the future, the highest temperature, the lowest temperature, the average temperature, the sunrise time and the sunset time of the day, namely the daily level, from a public meteorological data provider by utilizing an API (application program interface) calling interface;
step 2, obtaining historical generating power data of the photovoltaic power station, which is not less than 100 days at the level of 5 minutes, from a grid-connected inverter of the photovoltaic power station or management software of the photovoltaic power station;
step 3, processing the meteorological data and the historical generated power data to align the meteorological data and the historical generated power data;
step 4, selecting data from 45 days to 15 days ago as a training set, data from 15 days ago as a verification set, data from 24 hours in the future as a test set, and respectively extracting photovoltaic power generation power data characteristics as input data of a next prediction algorithm;
step 5, training a regression prediction model by using the characteristic data extracted from the historical meteorological data and the historical generated power data in the step 4, and screening out optimal parameters and an optimal model by using a verification set;
and 6, predicting a photovoltaic power generation power predicted value 24 hours in the future by using the regression prediction model trained in the step 5 to obtain a final prediction result.
2. The method for predicting the photovoltaic power generation based on the public meteorological data as claimed in claim 1, wherein in the step 3, the step of aligning the meteorological data and the power generation data comprises the following steps:
step 3-1, dividing historical generated power data of the photovoltaic power station into intervals of 15 minutes, averaging a plurality of generated powers falling in the same interval, and taking the averaged generated powers as aligned generated powers;
step 3-2, intersecting the aligned generated power in the step 3-1 with the meteorological data of the hour level and the daily level in a left full-connection mode to obtain the intersection of the historical generated power of the 15-minute level and the meteorological data of the hour level;
step 3-3, the missing meteorological data in the intersection set obtained in the step 3-2 comprise continuous meteorological data and discrete meteorological data, and the continuous meteorological data is filled with the first effective value before the missing value; and filling the discrete type meteorological data with the previous effective value of the missing value, and if the previous effective value does not exist, filling the discrete type meteorological data with the first subsequent effective value.
3. The method of claim 1, wherein the continuous weather comprises temperature, humidity, wind speed, maximum temperature, minimum temperature, average temperature; the discrete type weather comprises wind direction, weather type, sunrise time and sunset time.
4. The method for predicting the photovoltaic power generation power based on the public meteorological data as claimed in claim 1, wherein in the step 4, the step of extracting the photovoltaic power generation power data features comprises the following steps:
4-1, selecting data from 45 days ago to 15 days ago as a training set, data from 15 days ago to the present as a verification set, and data for 24 hours in the future as a test set;
step 4-2, respectively extracting the maximum value, the minimum value, the average value and the median value of the interval between the first 3 days, the first 7 days and the first 30 days of the lowest temperature, the maximum temperature, the humidity, the wind speed and the maximum temperature corresponding to each record in the training set, the verification set and the test set as characteristics;
4-3, extracting whether the current time corresponding to each record in the training set, the verification set and the test set is between sunrise and sunset time intervals or not, and taking the time distance relative to sunrise as a characteristic;
4-4, taking the current time corresponding to each record in the training set, the verification set and the test set, and the weather type, the wind power level and the wind direction of the current day as characteristics;
and 4-5, combining all the characteristics to serve as input data of a next prediction algorithm.
5. The photovoltaic power generation power prediction method based on the public meteorological data as claimed in claim 1, wherein the step of training the regression prediction model and selecting the optimal parameters and the optimal model in the step 5 comprises the following steps:
step 5-1, the regression prediction model is one of XGboost, LightGBM, neural Network, Gradient Boosting Decision Tree, LSTM and GRU;
step 5-2, training a regression model by respectively using different parameters by using the training set characteristics extracted in the step 4-5 as input and RMSE as an evaluation index;
and 5-3, selecting the model with the minimum RMSE as the optimal model by using the scores of the regression models with different parameters trained in the step 5-2 on the verification set, and taking the corresponding parameters as the optimal parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210011626.7A CN114330923A (en) | 2022-01-06 | 2022-01-06 | Photovoltaic power generation power prediction method based on public meteorological data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210011626.7A CN114330923A (en) | 2022-01-06 | 2022-01-06 | Photovoltaic power generation power prediction method based on public meteorological data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114330923A true CN114330923A (en) | 2022-04-12 |
Family
ID=81024740
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210011626.7A Pending CN114330923A (en) | 2022-01-06 | 2022-01-06 | Photovoltaic power generation power prediction method based on public meteorological data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114330923A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114919433A (en) * | 2022-05-27 | 2022-08-19 | 深圳先进技术研究院 | Electric vehicle cluster charging and discharging control method, system and related equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657881A (en) * | 2019-01-14 | 2019-04-19 | 南京国电南自电网自动化有限公司 | A kind of neural network photovoltaic power generation prediction technique and system suitable for small sample |
| CN109685257A (en) * | 2018-12-13 | 2019-04-26 | 国网青海省电力公司 | A kind of photovoltaic power generation power prediction method based on Support vector regression |
| CN112257941A (en) * | 2020-10-28 | 2021-01-22 | 福州大学 | Photovoltaic power station short-term power prediction method based on improved Bi-LSTM |
| CN112906757A (en) * | 2021-01-29 | 2021-06-04 | 广东晨兴智能科技有限公司 | Model training method, optical power prediction method, device, equipment and storage medium |
-
2022
- 2022-01-06 CN CN202210011626.7A patent/CN114330923A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109685257A (en) * | 2018-12-13 | 2019-04-26 | 国网青海省电力公司 | A kind of photovoltaic power generation power prediction method based on Support vector regression |
| CN109657881A (en) * | 2019-01-14 | 2019-04-19 | 南京国电南自电网自动化有限公司 | A kind of neural network photovoltaic power generation prediction technique and system suitable for small sample |
| CN112257941A (en) * | 2020-10-28 | 2021-01-22 | 福州大学 | Photovoltaic power station short-term power prediction method based on improved Bi-LSTM |
| CN112906757A (en) * | 2021-01-29 | 2021-06-04 | 广东晨兴智能科技有限公司 | Model training method, optical power prediction method, device, equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 杨荣新;孙朝云;徐磊;: "基于Stacking模型融合的光伏发电功率预测", 计算机系统应用, no. 05, 15 May 2020 (2020-05-15) * |
| 甄皓;: "有限信息下基于深度学习模型的小型分布式光伏电站功率预测", 上海节能, no. 04, 30 April 2020 (2020-04-30) * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114919433A (en) * | 2022-05-27 | 2022-08-19 | 深圳先进技术研究院 | Electric vehicle cluster charging and discharging control method, system and related equipment |
| WO2023226368A1 (en) * | 2022-05-27 | 2023-11-30 | 深圳先进技术研究院 | Electric vehicle cluster charging/discharging control method and system, and related equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113496311A (en) | Photovoltaic power station generated power prediction method and system | |
| CN109840633B (en) | Photovoltaic output power prediction method, system and storage medium | |
| CN106022538B (en) | Photovoltaic power generation prediction method based on K-means clustering improved generalized weather | |
| CN110334870B (en) | Photovoltaic power station short-term power prediction method based on gated cyclic unit network | |
| CN107832881B (en) | Wind power prediction error evaluation method considering load level and wind speed segmentation | |
| CN109978242A (en) | The photovoltaic power generation cluster power forecasting method and device of scale are risen based on statistics | |
| Cui et al. | Algorithm for identifying wind power ramp events via novel improved dynamic swinging door | |
| CN113128786B (en) | Wind, light and water complementary capacity and site selection optimization method based on space grid search | |
| CN104638672A (en) | Determining method of photovoltaic transmission power limit considering variable correlation | |
| CN114792156A (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
| CN114021420A (en) | Distributed photovoltaic ultra-short-term power generation prediction method and system | |
| Liao et al. | Ultra-short-term interval prediction of wind power based on graph neural network and improved bootstrap technique | |
| WO2023216576A1 (en) | Photovoltaic power generation short-term power prediction method and system | |
| CN115829126A (en) | Photovoltaic power generation power prediction method based on multi-view self-adaptive feature fusion | |
| CN113205210A (en) | Wind speed and power prediction method, system, equipment and storage medium for wind power plant with complex terrain | |
| Xue et al. | Adaptive ultra-short-term wind power prediction based on risk assessment | |
| CN114819374A (en) | Regional new energy ultra-short term power prediction method and system | |
| CN114330923A (en) | Photovoltaic power generation power prediction method based on public meteorological data | |
| CN114065634A (en) | Data-driven power quality monitoring and stationing optimization method and device | |
| CN108694475B (en) | Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model | |
| CN105512763A (en) | Method and system for predicting photovoltaic power station middle-short term power generation | |
| CN112653194A (en) | New energy source limit consumption capacity evaluation method | |
| CN107748943A (en) | A kind of grid power load management Forecasting Methodology based on cloud computing | |
| CN114676931B (en) | Electric quantity prediction system based on data center technology | |
| CN115329899A (en) | Clustering equivalent model construction method, system, equipment and storage medium |
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 |