CN114511127A - Neural network-based regional photovoltaic output characteristic long-term prediction method and system - Google Patents
Neural network-based regional photovoltaic output characteristic long-term prediction method and system Download PDFInfo
- Publication number
- CN114511127A CN114511127A CN202011285927.6A CN202011285927A CN114511127A CN 114511127 A CN114511127 A CN 114511127A CN 202011285927 A CN202011285927 A CN 202011285927A CN 114511127 A CN114511127 A CN 114511127A
- Authority
- CN
- China
- Prior art keywords
- photovoltaic output
- weather type
- probability
- climate
- 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
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000007774 longterm Effects 0.000 title claims abstract description 35
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 22
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000008859 change Effects 0.000 claims description 9
- 238000003066 decision tree Methods 0.000 claims description 7
- 238000012417 linear regression Methods 0.000 claims description 7
- 230000001932 seasonal effect Effects 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 238000010248 power generation Methods 0.000 description 4
- 238000012950 reanalysis Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 2
- 239000000443 aerosol Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012353 t test Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention provides a neural network-based regional photovoltaic output characteristic long-term prediction method and system, which comprises the following steps: performing interpolation processing on the acquired cross-season climate application data to obtain a predicted value time sequence, obtaining the probability of each weather type of photovoltaic output of a region to be measured based on the predicted value time sequence and a pre-trained multiple regression prediction model of each weather type, and estimating the cross-season region photovoltaic output characteristic for a long time based on the probability of each weather type of the photovoltaic output of the region to be measured, wherein the multiple regression prediction model is formed by training a time sequence of historical contemporaneous climate factors and weather type probabilities of historical contemporaneous set times; according to the method, the regional photovoltaic output characteristic multivariate regression prediction model is trained by using the climate factor of the cross-season climate data and the regional photovoltaic output probability, and the long-term photovoltaic output characteristic conjecture is effectively solved through the regional photovoltaic output characteristic multivariate regression prediction model and the cross-season climate data.
Description
Technical Field
The invention relates to long-term prediction of photovoltaic output characteristics, in particular to a neural network-based regional photovoltaic output characteristic long-term prediction method and system.
Background
The 0-48 hour change of large-scale photovoltaic power generation is mainly influenced by the geographical conditions of the area, the altitude angle of the sun, and weather phenomena such as cloud systems and precipitation in the daytime. The regular part of the regional photovoltaic output characteristics is mainly determined by a headroom radiation hypothesis, namely, smooth unimodal distribution is presented by removing significant factors such as weather phenomenon and aerosol particle concentration in the atmosphere. In actual operation, the phenomenon of coexistence of regularity and randomness is presented in regional photovoltaic power generation on the basis of clearance radiation influence, and the randomness characteristic of the phenomenon is mainly determined by various weather phenomena and weather processes with different life cycles within 1 hour to tens of hours.
Image recognition research on weather phenomena, weather processes and the like is a relatively new topic in computer vision. The artificial intelligence technology generally shows remarkable differences in recognition methods for different targets and scenes, and the implementation scheme and link design are correspondingly adjusted according to the target and scene requirements, so that the artificial intelligence technology has higher technical difficulty in general. The atmospheric three-dimensional structural features related to the photovoltaic power generation enrichment area have regional differences and category diversity, so that the conventional technical means are difficult to find out the general rule. Therefore, considering the cost and efficiency advantages of the artificial intelligence technology, the innovative application of the weather influence factor related to the power grid operation control under specific premise has important practical significance, but the main characteristics of the monthly photovoltaic output characteristic cannot be represented according to the probability of different regions, different seasons and different weather in the prior art.
Disclosure of Invention
Aiming at the problem that the main characteristics of the moonlight photovoltaic output characteristics can not be represented according to the probabilities of different regions, different seasons and different weather in the prior art, the invention provides a regional photovoltaic output characteristic long-term prediction method based on a neural network, which comprises the following steps:
performing interpolation processing on the acquired cross-season climate application data to obtain a predicted value time sequence;
obtaining the probability of each weather type of photovoltaic output of the area to be tested based on the predicted value time sequence and a pre-trained multiple regression prediction model of each weather type;
estimating the photovoltaic output characteristics of the cross-season area for a long time based on the probability of each weather type of the photovoltaic output of the area to be detected;
the multiple regression prediction model is formed by training a time sequence of historical contemporaneous climate factors and weather type probabilities of historical contemporaneous set times.
Preferably, the training of the multiple linear regression model includes:
acquiring cross-season climate application data information and regional photovoltaic output historical data;
classifying and counting according to weather features based on regional photovoltaic output historical data to obtain the occurrence probability of each weather type within set time, and determining regional photovoltaic output characteristics;
extracting the cross-season climate application data information to obtain a time sequence of historical contemporaneous climate factors;
calculating a correlation coefficient of each weather type through an exponential correlation formula based on the time series of the historical contemporaneous climate factors and the regional photovoltaic output characteristics of each weather type respectively,
performing data combination based on the relevant coefficients of each weather type to obtain the probability of each weather type at each set time in the historical period;
forming a training sample by the time sequence of the historical contemporaneous climate factors and the probability of each weather type at each set time of the historical contemporaneous climate factors;
taking the time sequence of the historical contemporaneous climate factors in the training samples as the input of a multivariate regression prediction model, and taking the probability of each weather type at each set time in the historical contemporaneous as the output of the multivariate regression prediction model for training;
and estimating parameters by using a least square method to obtain an estimation model corresponding to each weather type.
Preferably, the classifying and counting based on the regional photovoltaic output historical data according to the weather features to obtain the occurrence probability of each weather type within a set time, and determining the regional photovoltaic output characteristics includes:
selecting all historical contemporary photovoltaic power stations in the area to form a set, and obtaining regional photovoltaic output data of one-by-one set time through addition calculation according to the obtained one-by-one set time output data of each photovoltaic power station;
processing the regional photovoltaic output data according to set time one by adopting a table look-up method to obtain regional photovoltaic output data at a certain moment, and calculating to obtain photovoltaic output data of a clear sky region by adopting an ideal photoelectric conversion model;
and training a decision tree calculation program based on the regional photovoltaic output characteristics and the clear sky region photovoltaic output data, and dividing each weather type characteristic of the regional photovoltaic output data by taking the set number of days as a time unit to obtain the regional photovoltaic output probability of each weather type.
Preferably, the extracting the cross-season climate application data information to obtain the time series of the historical contemporaneous climate factors includes:
based on the cross-season climate application data, obtaining historical contemporaneous climate factors of appointed time one by one in corresponding time periods according to the photovoltaic output data time length of the region to be calculated;
obtaining a climate factor value through a climate factor calculation formula based on the historical contemporaneous climate factor;
based on the climate factor values, forming a time series of historical contemporaneous climate factors for the values;
the climate factors include: ocean according to the principles of the invention, ocean-related PNA is selected from ocean AO, ocean NAO, North American and ocean related PNA.
Preferably, the obtaining of the probability of each weather type at each setting time in the historical synchronization by performing data combination based on the correlation coefficient of each weather type includes:
according to the change curve of the weather types at each set time in a certain period and the probability of each weather type in the month, the change curve is used as a quantitative index for describing the photovoltaic output characteristics of the region;
based on the quantitative indexes describing the regional photovoltaic output characteristics, calculating correlation coefficients of the climate factor indexes and each weather type set one by one;
and based on the correlation coefficient of each weather type, performing data combination on the quantitative index of the regional photovoltaic output characteristic and the time sequence of the historical contemporaneous climate factor to obtain the probability of each weather type at each set time in the historical contemporaneous period.
Preferably, the interpolating the acquired seasonal climate application data to obtain a predicted value time series includes:
obtaining cross-season climate factor prediction information issued by an authoritative meteorological institution based on the cross-season climate application data;
and obtaining a time sequence of the predicted value of the cross-season climate factor through interpolation based on the cross-season climate factor prediction information.
Preferably, the long-term estimation of the photovoltaic output characteristics of the cross-season area based on the probability of each weather type of the photovoltaic output of the area to be measured includes:
inputting the time sequence of the predicted value of the cross-season climate factor into the multiple regression prediction model to obtain the weather type probability of the photovoltaic output of the region to be tested;
and obtaining comparison information of each weather type at set time and each weather type at historical set time based on the historical weather type probability of the photovoltaic output of the area to be measured and the weather type probability of the photovoltaic output of the area to be measured, and performing long-term estimation on the regional photovoltaic output.
Based on the same invention concept, the invention provides a regional photovoltaic output characteristic long-term prediction system based on a neural network, which comprises an acquisition module, an estimation module and a long-term estimation module;
the acquisition module: the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring seasonal weather application data;
the obtaining probability module: the multivariate regression prediction model is used for obtaining the probability of each weather type of the photovoltaic output of the area to be tested for the predicted value time sequence and each pre-trained weather type;
the long-term estimation module: probability for each weather type for photovoltaic contribution to the region under test versus long-term estimation of cross-season regional photovoltaic contribution characteristics
Preferably, the obtaining module comprises a prediction information sub-module and a time sequence sub-module;
the prediction information sub-module: the system is used for applying data to the cross-season climate to obtain cross-season climate factor prediction information issued by an authoritative meteorological institution;
the time series submodule: and the time sequence of the predicted value of the cross-season climate factor is obtained by interpolation according to the cross-season climate factor prediction information.
Preferably, the estimation module comprises a weather type sub-module and an estimation result sub-module;
the weather type sub-module: the multi-regression prediction model is used for inputting the time sequence of the cross-season climate factor predicted value into the multi-regression prediction model to obtain the weather type probability of the photovoltaic output of the region to be tested;
the estimation result sub-module: and the method is used for obtaining comparison information of each weather type at set time and each weather type at historical set time for the historical weather type probability of the photovoltaic output of the area to be measured and the weather type probability of the photovoltaic output of the area to be measured, and performing long-term estimation result of the photovoltaic output of the area.
Compared with the prior art, the invention has the beneficial effects that:
1. the regional photovoltaic output characteristic long-term prediction method based on the neural network comprises the following steps: performing interpolation processing on the acquired cross-season climate application data to obtain a predicted value time sequence, obtaining the probability of each weather type of photovoltaic output of a region to be measured based on the predicted value time sequence and a pre-trained multiple regression prediction model of each weather type, and estimating the cross-season region photovoltaic output characteristic for a long time based on the probability of each weather type of the photovoltaic output of the region to be measured, wherein the multiple regression prediction model is formed by training a time sequence of historical contemporaneous climate factors and weather type probabilities of historical contemporaneous set times; the method comprises the steps of inputting application information of cross-season weather data to a multiple regression prediction model corresponding to each type of pre-trained weather to obtain cross-season estimation of each region and each type of weather, and conjecturing the monthly photovoltaic output characteristic according to the probability of each type of weather;
2. according to the method, the regional photovoltaic output characteristic multiple regression prediction model is trained by using the climate factor of the cross-season climate data and the regional photovoltaic output probability, and the prediction of the monthly photovoltaic output characteristic is effectively solved through the regional photovoltaic output characteristic multiple regression prediction model and the cross-season climate data.
Drawings
FIG. 1 is a schematic diagram of a neural network-based regional photovoltaic output characteristic long-term prediction method of the present invention;
fig. 2 is a flow chart of the regional photovoltaic output characteristic long-term prediction method based on the neural network.
Detailed Description
The embodiments of the present invention will be further explained with reference to the drawings.
Example 1
With reference to fig. 1, the present invention provides a neural network-based regional photovoltaic output characteristic long-term prediction method, including:
the method comprises the following steps: performing interpolation processing on the acquired cross-season climate application data to obtain a predicted value time sequence;
step two: obtaining the probability of each weather type of photovoltaic output of the area to be tested based on the predicted value time sequence and a pre-trained multiple regression prediction model of each weather type;
step three: estimating the photovoltaic output characteristics of the cross-season area for a long time based on the probability of each weather type of the photovoltaic output of the area to be measured;
the multivariate regression prediction model is formed by training a time sequence of historical contemporaneous climate factors and weather type probabilities at each set time in the historical contemporaneous period.
Wherein, the step two: training of a multiple linear regression model, comprising:
acquiring cross-season climate application data information and regional photovoltaic output historical data;
classifying and counting according to weather features based on regional photovoltaic output historical data to obtain the occurrence probability of each weather type within set time, and determining regional photovoltaic output characteristics;
extracting the cross-season climate application data information to obtain a time sequence of historical contemporaneous climate factors;
calculating the correlation coefficient of each weather type through an exponential correlation formula based on the time sequence of the historical contemporaneous climate factor and the regional photovoltaic output characteristics of each weather type,
performing data combination based on the relevant coefficients of each weather type to obtain the probability of each weather type at each set time in the historical period;
forming a training sample by the time sequence of the historical contemporaneous climate factors and the probability of each weather type at each set time of the historical contemporaneous climate factors;
taking the time sequence of historical contemporaneous climate factors in the training samples as the input of a multivariate regression prediction model, and taking the probability of each weather type at each set time in the historical contemporaneous period as the output of the multivariate regression prediction model for training;
and estimating parameters by using a least square method to obtain an estimation model corresponding to each weather type.
Based on regional photovoltaic output historical data, classifying and counting according to weather features, obtaining the occurrence probability of each weather type in set time, and determining regional photovoltaic output characteristics, wherein the method comprises the following steps:
selecting all historical contemporary photovoltaic power stations in the area to form a set, and obtaining regional photovoltaic output data of one-by-one set time through addition calculation according to the obtained one-by-one set time output data of each photovoltaic power station;
processing the regional photovoltaic output data one by one according to set time by adopting a table look-up method to obtain regional photovoltaic output data at a certain moment, and calculating to obtain photovoltaic output data of a clear sky region by adopting an ideal photoelectric conversion model;
training a decision tree calculation program based on the regional photovoltaic output characteristics and the photovoltaic output data of the clear sky region, and dividing each weather type characteristic of the regional photovoltaic output data by taking the set number of days as a time unit to obtain the regional photovoltaic output probability of each weather type.
Extracting the cross-season climate application data information to obtain a time sequence of historical contemporaneous climate factors, wherein the time sequence comprises the following steps:
based on the seasonal climate application data, obtaining historical contemporaneous climate factors of designated time one by one in a corresponding time period according to the photovoltaic output data time length of the to-be-calculated region;
obtaining a climate factor value through a climate factor calculation formula based on historical contemporaneous climate factors;
forming a time series of historical contemporaneous climate factors of the values based on the climate factor values;
the climate factors include: ocean according to the principles of the invention, ocean-related PNA is selected from ocean AO, ocean NAO, North American and ocean related PNA.
The method for obtaining the probability of each weather type at each set time in the historical synchronization by carrying out data combination based on the relevant coefficient of each weather type comprises the following steps:
according to the change curve of the weather types at each set time in a certain period and the probability of each weather type in the month, the change curve is used as a quantitative index for describing the photovoltaic output characteristics of the region;
calculating correlation coefficients of the climate factor indexes and each weather type set one by one based on quantitative indexes describing regional photovoltaic output characteristics;
and based on the correlation coefficient of each weather type, performing data combination on the quantitative index of the regional photovoltaic output characteristic and the time sequence of the historical contemporaneous climate factor to obtain the probability of each weather type at each set time in the historical contemporaneous period.
Wherein, the first step: carrying out interpolation processing on the acquired cross-season climate application data to obtain a predicted value time sequence, wherein the method comprises the following steps:
obtaining cross-season climate factor prediction information issued by an authoritative meteorological institution based on the cross-season climate application data;
and obtaining a time sequence of the predicted value of the cross-season climate factor through interpolation based on the cross-season climate factor prediction information.
Wherein, the third step: the long-term estimation of the photovoltaic output characteristics of the cross-season area based on the probability of each weather type of the photovoltaic output of the area to be measured comprises the following steps:
inputting the time sequence of the predicted value of the cross-season climate factor into a multiple regression prediction model to obtain the weather type probability of the photovoltaic output of the region to be measured;
and obtaining comparison information of each weather type at set time and each weather type at historical set time based on the historical weather type probability of the photovoltaic output of the area to be measured and the weather type probability of the photovoltaic output of the area to be measured, and performing long-term estimation on the regional photovoltaic output.
Example 2
In order to achieve the purpose, the invention provides a regional photovoltaic output characteristic long-term prediction method based on a neural network. With reference to fig. 2, the method is performed as follows:
1. and (3) performing time scale calibration on the output data of each photovoltaic power station in the region, and calculating to obtain the total photovoltaic output (hereinafter referred to as regional photovoltaic output) of the region by 15 minutes.
2. And (3) according to the regional photovoltaic output time sequence, applying a decision tree classification algorithm to classify and identify weather types of the day-by-day time sequence, and extracting 3 main categories of sunny sky days, cloudy days and cloudy and rainy days to form the historical day-by-day regional photovoltaic output weather type time sequence of each month.
3. A weather and climate index analysis product is obtained from climate data, and 3 indexes, such as Arctic billows, Aratlantic billows and Pacific North American type remote correlation, used by the patent are monthly average indexes. The calculation method of the weather and climate index is drawn from a professional meteorological research institution and related authoritative documents, and the content of the invention only comprises the application skill. The indices used in this patent are specifically The Arctic throw (AO), North Atlantic throw (NAO), and Pacific North American circulation pad lines (PNA).
4. And obtaining the probability of a clear sky day, the probability of a cloudy day and the probability of a rainy day each month based on the time sequence of the photovoltaic output weather type of the day-by-day region of each month.
5. And establishing a multiple linear regression estimation model by respectively combining the monthly clear sky day probability time sequence, the cloudy day probability time sequence and the rainy day probability time sequence with AO, NAO and PNA.
6. And extracting cross-season time scale forecast information in future prediction of AO, NAO and PNA released by an authoritative meteorological research institution, and substituting the cross-season time scale forecast information into a multiple linear regression estimation model to obtain a special cross-level estimation result of the photovoltaic power generation output in the cross-season region.
To illustrate the actual flow and details of a regional photovoltaic output seasonal variation characteristic estimation method, the following example is given.
First, regional photovoltaic output data processing
1. The photovoltaic output data is sorted, all historical contemporary photovoltaic power stations in the area are selected to form a set, and 15min output data of each photovoltaic power station obtained in an Energy Management System (EMS for short) is utilized.
2. And performing quality control on EMS output data of each photovoltaic power station, removing power limiting data caused by active power control, and adding and calculating to obtain 15 min-by-15 min regional photovoltaic output data.
3. The time length of the regional photovoltaic output data required to be obtained through calculation is not less than 3 years.
Second, climate analysis data application
1. Based on weather reanalysis data made by the American weather environment forecasting center (NCEP) and the American national atmospheric research center (NCAR), weather element data in the reanalysis data is extracted according to the photovoltaic output data time length of the region to be calculated.
2. Processing meteorological elements in the meteorological reanalysis data, referring to important climate factors AO, NAO and PNA calculation formulas in authoritative documents to obtain monthly AO, NAO and PNA values in corresponding time intervals, and respectively forming a historical time sequence of the index.
3. Based on AO, NAO and PNA forecast information issued by an authoritative meteorological institution, a monthly forecast value time sequence of cross-season AO, NAO and PNA is obtained through interpolation.
Thirdly, classifying regional photovoltaic output characteristic decision trees
1. And processing the data of the regional photovoltaic output for 15min into a time sequence with days, months and years as time scales by adopting a table look-up method according to the sunrise and sunset time, wherein the representations of the days, the months and the years are i, j and k respectively. And recording the regional photovoltaic output data at a specific moment as P (i, j, k).
2. And calculating to obtain photovoltaic output data of the clear sky region and recording the photovoltaic output data as PC (i, j, k) by using the total clearance solar radiation data in the reanalysis data and adopting an ideal photoelectric conversion model.
3. Selecting a CART (classification And Regression Tree) decision tree calculation program, wherein a training set is composed of P (i, j, k) And PC (i, j, k).
4. And classifying the weather type characteristics of P (i, j, k) by taking the day as a time unit to obtain the weather type categories of any day in the period to be analyzed, wherein the weather type categories respectively comprise 3 main categories of sunny sky days, cloudy days and rainy days, and are respectively marked as cle, clo and ra.
5. The treated regional photovoltaic output daily characteristics are respectively recorded as P (cle, j, k), P (clo, j, k) and P (ra, j, k).
6. And (5) counting to obtain the probability of P (cle, j, k), P (clo, j, k) and P (ra, j, k) in any one month.
Fourth, index association degree analysis of regional photovoltaic output characteristics and historical contemporaneous climate factors
1. And setting the variation curves of the weather types cle, clo and ra of each month and the probabilities of P (cle, j, k), P (clo, j, k) and P (ra, j, k) of the current month in a certain period as quantitative indexes for describing the regional photovoltaic output characteristics.
2. Calculating the correlation coefficients of the month-by-month AO, NAO and PNA indexes and P (cle, j, k), P (clo, j, k) and P (ra, j, k), and determining the correlation significance by adopting a t test method.
3. And respectively determining a dependent variable combination method of the multiple linear regression equation to be established, namely obtaining the climate factor index combination related to the lunar output characteristics P (cle, j, k), P (clo, j, k) and P (ra, j, k).
Area photovoltaic output characteristic multiple regression prediction model based on multiple climate factor indexes
1. Setting month-by-month AO, NAO and PNA as multiple regression independent variables, and respectively establishing multiple linear regression equations with dependent variables as historical same-period weather type probabilities P (cle, j, k), P (clo, j, k) and P (ra, j, k).
2. And estimating parameters by using a least square method to obtain respective estimation models of P (cle, j, k), P (clo, j, k) and P (ra, j, k).
Sixth, long-term estimation of photovoltaic output characteristics in cross-season region
1. And accessing and analyzing the cross-season prediction results of AO, NAO and PNA issued by an authoritative meteorological institution.
2. By utilizing the regional photovoltaic output characteristic multiple regression prediction model provided by the patent, the regional photovoltaic output probability characteristic of 3-6 months in the future is obtained
3. Comparing the weather type probability condition of photovoltaic output of each month area within 3-6 months in the future, obtaining the same-ratio and ring-ratio conditions of sunny sky days, cloudy days and cloudy-rainy days of the month according to the historical weather type probability statistics in the step (2), and giving the estimation result of the photovoltaic output of the month.
Example 3
Based on the same invention concept, the invention provides a regional photovoltaic output characteristic long-term prediction system of a neural network, which comprises an acquisition module, an estimation module and a long-term estimation module;
an acquisition module: the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring seasonal weather application data;
an obtaining probability module: the multivariate regression prediction model is used for obtaining the probability of each weather type of the photovoltaic output of the area to be tested for the predicted value time sequence and the pre-trained multivariate regression prediction model of each weather type;
a long-term estimation module: long-term estimation of cross-season regional photovoltaic output characteristics by probability for each weather type for photovoltaic output of region under test
The acquisition module comprises a prediction information submodule and a time sequence submodule;
the prediction information submodule: the system is used for applying data to the cross-season climate to obtain cross-season climate factor prediction information issued by an authoritative meteorological institution;
time series submodule: and the method is used for obtaining the time sequence of the predicted value of the cross-season climate factor through interpolation on the prediction information of the cross-season climate factor.
The probability obtaining module comprises a regional photovoltaic output probability obtaining submodule and a weather type probability obtaining submodule;
obtaining a regional photovoltaic output probability submodule: the system is used for carrying out classification statistics on the regional photovoltaic output historical data according to weather features to obtain the occurrence probability of each weather type within set time and determine the regional photovoltaic output characteristics;
obtaining the probability of each weather type submodule: and the method is used for carrying out data combination on the related coefficients of each weather type to obtain the probability of each weather type at each set time in the historical period.
The sub-module for acquiring the regional photovoltaic output probability comprises a regional photovoltaic output data acquisition system, a clear sky region photovoltaic output data acquisition system and a regional photovoltaic output probability acquisition system for each weather type;
acquiring a regional photovoltaic output data system: selecting all historical contemporary photovoltaic power stations in the area to form a set, and obtaining regional photovoltaic output data of one-by-one set time through addition calculation according to the obtained one-by-one set time output data of each photovoltaic power station;
acquiring a photovoltaic output data system of a clear sky area: processing the regional photovoltaic output data according to set time one by adopting a table look-up method to obtain regional photovoltaic output data at a certain moment, and calculating to obtain photovoltaic output data of a clear sky region by adopting an ideal photoelectric conversion model;
the system for acquiring the photovoltaic output probability of each weather type region comprises the following steps: and training a decision tree calculation program based on the regional photovoltaic output characteristics and the clear sky region photovoltaic output data, and dividing each weather type characteristic of the regional photovoltaic output data by taking the set number of days as a time unit to obtain the regional photovoltaic output probability of each weather type.
The obtain probability of each weather type sub-module includes: acquiring a quantitative index system, a related system and a probability system of each weather type at set time;
obtaining a quantization index system: according to the change curve of the weather types at each set time in a certain period and the probability of each weather type in the month, the change curve is used as a quantitative index for describing the photovoltaic output characteristics of the region;
acquiring a correlation coefficient system: based on the quantitative indexes describing the regional photovoltaic output characteristics, calculating correlation coefficients of the climate factor indexes and each weather type set one by one;
the system for acquiring the probability of each weather type at set time comprises: and based on the correlation coefficient of each weather type, performing data combination on the quantitative index of the regional photovoltaic output characteristic and the time sequence of the historical contemporaneous climate factor to obtain the probability of each weather type at each set time in the historical contemporaneous period.
The estimation module comprises a weather type submodule and an estimation result submodule;
weather type submodule: the weather type probability of the photovoltaic output of the region to be tested is obtained by inputting the time sequence of the predicted value of the cross-season weather factor into the multiple regression prediction model;
an estimation result sub-module: and the method is used for obtaining comparison information of each weather type at set time and each weather type at historical set time for the historical weather type probability of the photovoltaic output of the area to be measured and the weather type probability of the photovoltaic output of the area to be measured, and performing long-term estimation result of the photovoltaic output of the area.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The present invention is not limited to the above embodiments, and any modifications, equivalent substitutions, improvements, etc. within the spirit and principle of the present invention are included in the scope of the claims of the present invention as filed.
Claims (10)
1. The regional photovoltaic output characteristic long-term prediction method based on the neural network is characterized by comprising the following steps of:
performing interpolation processing on the acquired cross-season climate application data to obtain a predicted value time sequence;
obtaining the probability of each weather type of photovoltaic output of the area to be tested based on the predicted value time sequence and a pre-trained multiple regression prediction model of each weather type;
estimating the photovoltaic output characteristics of the cross-season area for a long time based on the probability of each weather type of the photovoltaic output of the area to be detected;
the multiple regression prediction model is formed by training a time sequence of historical contemporaneous climate factors and weather type probabilities of historical contemporaneous set times.
2. The method of claim 1, wherein the training of the multiple linear regression model comprises:
acquiring cross-season climate application data information and regional photovoltaic output historical data;
classifying and counting according to weather features based on regional photovoltaic output historical data to obtain the occurrence probability of each weather type within set time, and determining regional photovoltaic output characteristics;
extracting the cross-season climate application data information to obtain a time sequence of historical contemporaneous climate factors;
calculating a correlation coefficient of each weather type through an exponential correlation formula based on the time series of the historical contemporaneous climate factors and the regional photovoltaic output characteristics of each weather type respectively,
performing data combination based on the relevant coefficients of each weather type to obtain the probability of each weather type at each set time in the historical period;
forming a training sample by the time sequence of the historical contemporaneous climate factors and the probability of each weather type at each set time of the historical contemporaneous climate factors;
taking the time sequence of the historical contemporaneous climate factors in the training samples as the input of a multivariate regression prediction model, and taking the probability of each weather type at each set time in the historical contemporaneous period as the output of the multivariate regression prediction model for training;
and estimating parameters by using a least square method to obtain an estimation model corresponding to each weather type.
3. The method of claim 2, wherein the step of determining the regional photovoltaic output characteristics based on the statistical classification of the regional photovoltaic output historical data according to the weather features to obtain the occurrence probability of each weather type within a set time comprises the steps of:
selecting all historical contemporary photovoltaic power stations in the area to form a set, and obtaining regional photovoltaic output data of one-by-one set time through addition calculation according to the obtained one-by-one set time output data of each photovoltaic power station;
processing the regional photovoltaic output data according to set time one by adopting a table look-up method to obtain regional photovoltaic output data at a certain moment, and calculating to obtain photovoltaic output data of a clear sky region by adopting an ideal photoelectric conversion model;
and training a decision tree calculation program based on the regional photovoltaic output characteristics and the clear sky region photovoltaic output data, and dividing each weather type characteristic of the regional photovoltaic output data by taking the set number of days as a time unit to obtain the regional photovoltaic output probability of each weather type.
4. The method of claim 3, wherein extracting the cross-season climate application data information to obtain a time series of historical contemporaneous climate factors comprises:
based on the cross-season climate application data, obtaining historical contemporaneous climate factors of appointed time one by one in corresponding time periods according to the photovoltaic output data time length of the region to be calculated;
obtaining a climate factor value through a climate factor calculation formula based on the historical contemporaneous climate factor;
forming a time series of historical contemporaneous climate factors for the values based on the climate factor values;
the climate factors include: ocean according to the principles of the invention, ocean-related PNA is selected from ocean AO, ocean NAO, North American and ocean related PNA.
5. The method as claimed in claim 4, wherein the step of combining data based on the correlation coefficient of each weather type to obtain the probability of each weather type at each set time in the historical synchronization comprises:
according to the change curve of the weather types at each set time in a certain period and the probability of each weather type in the month, the change curve is used as a quantitative index for describing the photovoltaic output characteristics of the region;
based on the quantitative indexes describing the regional photovoltaic output characteristics, calculating correlation coefficients of the climate factor indexes and each weather type set one by one;
and based on the correlation coefficient of each weather type, performing data combination on the quantitative index of the regional photovoltaic output characteristic and the time sequence of the historical contemporaneous climate factor to obtain the probability of each weather type at each set time in the historical contemporaneous.
6. The method of claim 5, wherein interpolating the acquired cross-season climate application data to obtain a predicted value time series comprises:
obtaining cross-season climate factor prediction information issued by an authoritative meteorological institution based on the cross-season climate application data;
and obtaining a time sequence of the predicted value of the cross-season climate factor through interpolation based on the cross-season climate factor prediction information.
7. The method of claim 6, wherein the long-term evaluation of the photovoltaic output characteristics across seasonal areas based on the probability for each weather type of photovoltaic output for the area under test comprises:
inputting the time sequence of the predicted value of the cross-season climate factor into the multiple regression prediction model to obtain the weather type probability of the photovoltaic output of the region to be tested;
and obtaining comparison information of each weather type at set time and each weather type at historical set time based on the historical weather type probability of the photovoltaic output of the area to be measured and the weather type probability of the photovoltaic output of the area to be measured, and performing long-term estimation on the regional photovoltaic output.
8. The regional photovoltaic output characteristic long-term prediction system based on the neural network is characterized by comprising an acquisition module, an estimation module and a long-term estimation module;
the acquisition module: the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring seasonal weather application data;
the obtaining probability module: the multivariate regression prediction model is used for obtaining the probability of each weather type of the photovoltaic output of the area to be tested for the predicted value time sequence and each pre-trained weather type;
the long-term estimation module: and the probability of each weather type for the photovoltaic output of the region to be measured is used for long-term estimation of the photovoltaic output characteristics of the cross-season region.
9. The system of claim 8, wherein the acquisition module includes a prediction information sub-module and a time series sub-module;
the prediction information sub-module: the system is used for applying data to the cross-season climate to obtain cross-season climate factor prediction information issued by an authoritative meteorological institution;
the time series submodule: and the method is used for obtaining the time sequence of the cross-season climate factor predicted value through interpolation on the cross-season climate factor predicted information.
10. The system of claim 9, wherein the prognostics module includes a weather type sub-module and a prognostics sub-module;
the weather type sub-module: the multi-regression prediction model is used for inputting the time sequence of the cross-season climate factor predicted value into the multi-regression prediction model to obtain the weather type probability of the photovoltaic output of the region to be tested;
the estimation result sub-module: and the method is used for obtaining comparison information of each weather type at set time and each weather type at historical set time for the historical weather type probability of the photovoltaic output of the area to be measured and the weather type probability of the photovoltaic output of the area to be measured, and performing long-term estimation result of the photovoltaic output of the area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011285927.6A CN114511127A (en) | 2020-11-17 | 2020-11-17 | Neural network-based regional photovoltaic output characteristic long-term prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011285927.6A CN114511127A (en) | 2020-11-17 | 2020-11-17 | Neural network-based regional photovoltaic output characteristic long-term prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114511127A true CN114511127A (en) | 2022-05-17 |
Family
ID=81547167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011285927.6A Pending CN114511127A (en) | 2020-11-17 | 2020-11-17 | Neural network-based regional photovoltaic output characteristic long-term prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114511127A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116742624A (en) * | 2023-08-10 | 2023-09-12 | 华能新能源股份有限公司山西分公司 | Photovoltaic power generation amount prediction method and system |
-
2020
- 2020-11-17 CN CN202011285927.6A patent/CN114511127A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116742624A (en) * | 2023-08-10 | 2023-09-12 | 华能新能源股份有限公司山西分公司 | Photovoltaic power generation amount prediction method and system |
| CN116742624B (en) * | 2023-08-10 | 2023-11-03 | 华能新能源股份有限公司山西分公司 | Photovoltaic power generation amount prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109002915B (en) | Photovoltaic power station short-term power prediction method based on Kmeans-GRA-Elman model | |
| CN109919353B (en) | Distributed photovoltaic prediction method of ARIMA model based on spatial correlation | |
| CN113496311A (en) | Photovoltaic power station generated power prediction method and system | |
| CN111079980A (en) | Optical power prediction method based on self-adaptive classification strategy and hybrid optimization SVR | |
| CN111598337B (en) | Method for predicting short-term output of distributed photovoltaic system | |
| CN113657662B (en) | Downscaling wind power prediction method based on data fusion | |
| Aliberti et al. | Forecasting Short-term Solar Radiation for Photovoltaic Energy Predictions. | |
| CN115965132A (en) | Distributed photovoltaic digital twin system power prediction method based on GA-BP neural network | |
| CN114399081A (en) | Photovoltaic power generation power prediction method based on weather classification | |
| CN115935809A (en) | Photovoltaic power generation power interval prediction method based on cloud cover characteristic of foundation cloud cover | |
| CN115936177A (en) | Photovoltaic output power prediction method and system based on neural network | |
| CN112352523A (en) | Tea garden water and fertilizer irrigation control method and system based on intelligent decision | |
| CN115759389A (en) | Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy | |
| CN116799796A (en) | Photovoltaic power generation power prediction method, device, equipment and medium | |
| CN114882373A (en) | Multi-feature fusion sandstorm prediction method based on deep neural network | |
| Meng et al. | A new PV generation power prediction model based on GA-BP neural network with artificial classification of history day | |
| CN114511127A (en) | Neural network-based regional photovoltaic output characteristic long-term prediction method and system | |
| CN115860797B (en) | Electric quantity demand prediction method suitable for new electricity price reform situation | |
| CN116050666A (en) | Photovoltaic power generation power prediction method for irradiation characteristic clustering | |
| CN113052386B (en) | Random forest algorithm-based distributed photovoltaic daily power generation amount prediction method and device | |
| CN116470486A (en) | Ultra-short-term photovoltaic power generation power prediction method, device and storage medium | |
| CN115238967A (en) | Photovoltaic power prediction method and device combining cloud picture and adjacent power station cluster | |
| CN115563848A (en) | Distributed photovoltaic total radiation prediction method and system based on deep learning | |
| CN112132344A (en) | Short-term wind power prediction method based on similar day and FRS-SVM | |
| CN113487080B (en) | Wind speed dynamic scene generation method, system and terminal based on wind speed classification |
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 |