This application is based upon and claims a priority to Chinese Patent Application Serial No. 201710492720.8, filed with the State Intellectual Property Office of P. R. China on Jun. 26, 2017, the entire contents of which are incorporated herein by reference.
The present disclosure relates to a field of new energy technologies in power systems, and more particularly, to a method for photovoltaic power forecast based on numerical weather prediction, an apparatus for photovoltaic power forecast based on numerical weather prediction and a computer-readable storage medium.
The photovoltaic power station is a typical intermittent power source because it can only generate electricity by daylight. The photovoltaic power of the photovoltaic power station has great volatility and randomness due to the influence of weather and environment on the photovoltaic power station. Therefore, it may cause adverse effects on power grids when large-scale photovoltaic power stations are merged into power grids.
Embodiments of the present disclosure provide a method for photovoltaic power forecast based on numerical weather prediction. The method includes: reading from a computer storage device, a plurality of first effective light intensities corresponding to a plurality of historical time periods and a plurality of first photovoltaic array temperatures corresponding to the plurality of historical time periods; reading from the computer storage device, a plurality of second effective light intensities corresponding to a plurality of prediction time periods and a plurality of second photovoltaic array temperatures corresponding to the plurality of prediction time periods; determining a historical key weather feature matrix based on the plurality of first effective light intensities and the plurality of first photovoltaic array temperatures; determining a prediction key weather feature matrix based on the plurality of second effective light intensities and the plurality of second photovoltaic array temperatures; reading from the computer storage device, a plurality of first groups of weather data corresponding to the plurality of historical time periods; determining a historical weather data matrix based on the plurality of first groups of weather data; reading from the computer storage device, a plurality of second groups of weather data corresponding to the plurality of prediction time periods; determining a prediction weather data matrix based on the plurality of second groups of weather data; determining a historical input matrix based on the historical key weather feature matrix and the historical weather data matrix; determining a prediction input matrix based on the prediction key weather feature matrix and the prediction weather data matrix; combining the historical input matrix and the prediction input matrix to determine an input matrix; performing singular value decomposition on the input matrix to obtain a principal component feature matrix; for a principal component feature corresponding to each prediction time period in the principal component feature matrix, calculating a Manhattan distance between the principal component feature corresponding to each prediction time period and a principal component feature corresponding to each historical time period; determining K principal component features corresponding to K historical time periods having the nearest Manhattan distances with the principal component feature corresponding to each prediction time period; acquiring a fitting relationship according to the K principal component features corresponding to the K historical time periods and K photovoltaic powers corresponding to the K historical time periods; inputting the principal component feature corresponding to each prediction time period to the fitting relationship to obtain a photovoltaic power corresponding to each prediction time period.
Embodiments of the present disclosure provide an apparatus for photovoltaic power forecast based on numerical weather prediction. The apparatus includes a computer storage device, a processor and computer programs stored in the computer storage device and executable by the processor. The computer storage device is configured to store a plurality of first effective light intensities corresponding to a plurality of historical time periods and a plurality of first photovoltaic array temperatures corresponding to the plurality of historical time periods, a plurality of second effective light intensities corresponding to a plurality of prediction time periods and a plurality of second photovoltaic array temperatures corresponding to the plurality of prediction time periods, a plurality of first groups of weather data corresponding to the plurality of historical time periods, and a plurality of second groups of weather data corresponding to the plurality of prediction time periods. The processor is configured to by reading the computer programs to: read from the computer storage device the plurality of first effective light intensities corresponding to the plurality of historical time periods and the plurality of first photovoltaic array temperatures corresponding to the plurality of historical time periods; read from the computer storage device the plurality of second effective light intensities corresponding to the plurality of prediction time periods and the plurality of second photovoltaic array temperatures corresponding to the plurality of prediction time periods; determine a historical key weather feature matrix based on the plurality of first effective light intensities and the plurality of first photovoltaic array temperatures; determine a prediction key weather feature matrix based on the plurality of second effective light intensities and the plurality of second photovoltaic array temperatures; read from the computer storage device the plurality of first groups of weather data corresponding to the plurality of historical time periods; determine a historical weather data matrix based on the plurality of first groups of weather data; read from the computer storage device the plurality of second groups of weather data corresponding to the plurality of prediction time periods; determine a prediction weather data matrix based on the plurality of second groups of weather data; determine a historical input matrix based on the historical key weather feature matrix and the historical weather data matrix; determine a prediction input matrix based on the prediction key weather feature matrix and the prediction weather data matrix; combine the historical input matrix and the prediction input matrix to determine an input matrix; perform singular value decomposition on the input matrix to obtain a principal component feature matrix; for a principal component feature corresponding to each prediction time period in the principal component feature matrix, calculate a Manhattan distance between the principal component feature corresponding to each prediction time period and a principal component feature corresponding to each historical time period; determine K principal component features corresponding to K historical time periods having the nearest Manhattan distances with the principal component feature corresponding to each prediction time period; acquire a fitting relationship according to the K principal component features corresponding to the K historical time periods and K photovoltaic powers corresponding to the K historical time periods; input the principal component feature corresponding to each prediction time period to the fitting relationship to obtain a photovoltaic power corresponding to each prediction time period.
Embodiments of the present disclosure provide a non-transitory computer-readable storage medium having stored therein instructions. When the instructions are executed by a processor of a device, causes the device to perform acts of: reading from a computer storage device, a plurality of first effective light intensities corresponding to a plurality of historical time periods and a plurality of first photovoltaic array temperatures corresponding to the plurality of historical time periods; reading from the computer storage device, a plurality of second effective light intensities corresponding to a plurality of prediction time periods and a plurality of second photovoltaic array temperatures corresponding to the plurality of prediction time periods; determining a historical key weather feature matrix based on the plurality of first effective light intensities and the plurality of first photovoltaic array temperatures; determining a prediction key weather feature matrix based on the plurality of second effective light intensities and the plurality of second photovoltaic array temperatures; reading from the computer storage device, a plurality of first groups of weather data corresponding to the plurality of historical time periods; determining a historical weather data matrix based on the plurality of first groups of weather data; reading from the computer storage device, a plurality of second groups of weather data corresponding to the plurality of prediction time periods; determining a prediction weather data matrix based on the plurality of second groups of weather data; determining a historical input matrix based on the historical key weather feature matrix and the historical weather data matrix; determining a prediction input matrix based on the prediction key weather feature matrix and the prediction weather data matrix; combining the historical input matrix and the prediction input matrix to determine an input matrix; performing singular value decomposition on the input matrix to obtain a principal component feature matrix; for a principal component feature corresponding to each prediction time period in the principal component feature matrix, calculating a Manhattan distance between the principal component feature corresponding to each prediction time period and a principal component feature corresponding to each historical time period; determining K principal component features corresponding to K historical time periods having the nearest Manhattan distances with the principal component feature corresponding to each prediction time period; acquiring a fitting relationship according to the K principal component features corresponding to the K historical time periods and K photovoltaic powers corresponding to the K historical time periods; inputting the principal component feature corresponding to each prediction time period to the fitting relationship to obtain a photovoltaic power corresponding to each prediction time period.
Additional aspects and advantages of embodiments of the present disclosure will be given in part in the following descriptions, become apparent in part from the following descriptions, or be learned from the practice of the embodiments of the present disclosure.
The above and/or additional aspects and advantages of embodiments of the present disclosure will become apparent and more readily appreciated from the following descriptions made with reference to the drawings, in which:
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. The same or similar elements and elements having same or similar functions are denoted by like reference numerals throughout the descriptions. The embodiments described herein with reference to the accompanying drawings are explanatory and used to generally understand the present disclosure, which is not construed to limit the embodiments of the present disclosure.
The embodiments of the present disclosure disclose a method for photovoltaic power forecast based on numerical weather prediction. The method combines a physical model of the photovoltaic power with data-driven functionality.
In detail, the physical model of the photovoltaic power is denoted by a formula of:
where, Pmp denotes the photovoltaic power; Ee denotes an effective light intensity; Tc denotes the temperature of a photovoltaic array; E0 denotes a reference light intensity (E0 is a constant value, for example, 1000 W/m2); T0 denotes a reference temperature (T0 is a constant value, for example, 25° C.); Pmp0 denotes a photovoltaic rating power; and γ denotes a temperature coefficient of the photovoltaic array (γ is a constant value, depending on factors such as material of the photovoltaic array).
It can be known from the physical model of the photovoltaic power that weather features directly related to the photovoltaic power are the effective light intensity and the photovoltaic array temperature.
Based on the physical model of the photovoltaic power, a linear expression of the photovoltaic power with respect to the effective light intensity and the photovoltaic array temperature is denoted by a formula of:
where, the formula (2) is the linear expression of Pmp with respect to Ee and Ee·Tc. Therefore, Ee and Ee·Tc are marked as key weather features that affects the photovoltaic power.
The computer can calculate the above effective light intensities and the above photovoltaic array temperatures and record them into the computer storage device.
In an embodiment of the present disclosure, calculating the effective light intensity includes the following acts (1-1), (1-2), (1-3) and (1-4).
(1-1) A direct sunlight component, a ground reflection component and a sky diffuse component are acquired by a light intensity meter, or through a weather database.
(1-2) A photovoltaic array light intensity is acquired based on the direct sunlight component, the ground reflection component and the sky diffuse component by a formula of:
EPOA=Eb+Eg+Ed (3),
where, EPOA denotes the photovoltaic array light intensity; Eb denotes the direct sunlight component; Eg denotes the ground reflection component; and Ed denotes the sky diffuse component.
(1-3) A photovoltaic array cleanliness is measured.
The photovoltaic array cleanliness may be determined by field experiments.
(1-4) The effective light intensity is acquired based on the photovoltaic array cleanliness and the photovoltaic array light intensity by a formula of:
Ee=EPOA·SF (4),
where, SF denotes the photovoltaic array cleanliness and SF∈[0,1]. When SF=1, it indicates that the photovoltaic array is completely clean.
In an embodiment of the present disclosure, calculating the photovoltaic array temperature includes the following acts (2-1), (2-2), (2-3) and (2-4).
(2-1) The direct sunlight component, the ground reflection component and the sky diffuse component are acquired based on the act (1-1).
(2-2) The photovoltaic array light intensity is acquired based on the act (1-2).
(2-3) A wind speed and an ambient temperature are measured.
(2-4) The photovoltaic array temperature is acquired based on the photovoltaic array light intensity, the wind speed and the ambient temperature by a formula of
where, Ta denotes the ambient temperature; WS denotes the wind speed; ΔT denotes a temperature difference coefficient of the photovoltaic array and indicates a temperature difference between the photovoltaic array and the photovoltaic element, which is related to factors such as material of the photovoltaic array, in which ΔT>0; U0 denotes a thermal conductivity constant of the photovoltaic array, for example, 25 W/m2K; and U1 denotes a thermal convection constant of the photovoltaic array, for example, 6.84 W/m3sK.
At block 101, a plurality of first effective light intensities corresponding to a plurality of historical time periods and a plurality of first photovoltaic array temperatures corresponding to the plurality of historical time periods are read from a computer storage device.
At block 102, a plurality of second effective light intensities corresponding to a plurality of prediction time periods and a plurality of second photovoltaic array temperatures corresponding to the plurality of prediction time periods are read from the computer storage device.
In the embodiments of the present disclosure, acts in block 101 and block 102 are executed in a non-sequential order.
At block 103, a historical key weather feature matrix is determined based on the plurality of first effective light intensities and the plurality of first photovoltaic array temperatures.
In an embodiment of the present disclosure, each historical time period is one sampling unit time. That is, the unit time for acquiring photovoltaic power data and weather data. The length of each historical time period is not limited, generally from 15 minutes to 1 hour. For example, this embodiment selects 1 hour.
For any historical time period t∈{1, 2 . . . Th}, Th denotes the number of the plurality of historical time periods and depends on a size of historical data samples. For example, the number of historical time periods in this embodiment is 8760 hours per year. Assuming that the first effective light intensity in the historical time period t is denoted as Eeh[t] and the first photovoltaic array temperature in the historical time period t is denoted as Tch[t], the historical key weather feature matrix XCh in the plurality of historical time periods is denoted as follows:
At block 104, a prediction key weather feature matrix is determined based on the plurality of second effective light intensities and the plurality of second photovoltaic array temperatures.
In an embodiment of the present disclosure, each prediction time period is one sampling unit time. The length of each prediction time period is the same with that of each historical time period.
For any prediction time period t ∈{1, 2 . . . Tf}, Tf denotes the number of the plurality of prediction time periods and depends on a size of prediction data samples. For example, the number of prediction time periods in this embodiment is 168 hours per week. Assuming that the second effective light intensity in the prediction time period t is denoted as Eef [t] and the second photovoltaic array temperature in the prediction time period t is denoted as Tef [t], the prediction key weather feature matrix XCf in the plurality of prediction time periods is denoted as follows:
In the embodiments of the present disclosure, acts in block 103 and block 104 are executed in a non-sequential order.
At block 105, a plurality of first groups of weather data corresponding to the plurality of historical time periods are read from the computer storage device and a historical weather data matrix is determined based on the plurality of first groups of weather data.
The historical database of numerical weather forecasts can be stored in the computer storage device. From the historical database, the plurality of first groups of weather data corresponding to the plurality of historical time periods can be acquired. Each first group of weather data may include one or more of an atmospheric pressure, an air sediment, an air humidity, a sky cloud volume, a wind speed, an air temperature and a surface heat radiation. The air sediment refers to a particulate matter deposited in the air.
The historical weather data matrix XWh∈RT
At block 106, a plurality of second groups of weather data corresponding to the plurality of prediction time periods are read from the computer storage device and a prediction weather data matrix is determined based on the plurality of second groups of weather data.
The prediction database of numerical weather forecasts can be stored in the computer storage device. From the prediction database, the plurality of second groups of weather data corresponding to the plurality of prediction time periods can be acquired. Each second group of weather data may include one or more of an atmospheric pressure, an air sediment, an air humidity, a sky cloud volume, a wind speed, an air temperature and a surface heat radiation.
The prediction weather data matrix XWf∈RT
In the embodiments of the present disclosure, acts in block 105 and block 106 are executed in a non-sequential order.
At block 107, a historical input matrix is determined based on the historical key weather feature matrix and the historical weather data matrix.
At block 108, a prediction input matrix is determined based on the prediction key weather feature matrix and the prediction weather data matrix.
In the embodiments of the present disclosure, acts in block 107 and block 108 are executed in a non-sequential order.
The historical input matrix is denoted by a formula of:
XIf=[XCf,XWf] (8).
The prediction input matrix is denoted by a formula of:
XIf=[XCf,XWf] (8).
At block 109, the historical input matrix is combined with the prediction input matrix to obtain an input matrix.
The input matrix is denoted by
At block 110, singular value decomposition is performed on the input matrix to obtain a principal component feature matrix.
The principal component analysis is applied to the combined input matrix to extract features of the historical input matrix and the prediction input matrix.
The principal component analysis converts a group of variables that may have correlation with each other into a group of linearly uncorrelated variables, thereby reducing redundancy of the input matrix and extracting the principal component features.
The singular value decomposition is performed on the input matrix
to obtain a principal component feature matrix
where, XPh∈RT
At block 111, for a principal component feature corresponding to each prediction time period in the principal component feature matrix, a Manhattan distance between the principal component feature corresponding to each prediction time period and a principal component feature corresponding to each historical time period is calculated, and K principal component features corresponding to K historical time periods having the nearest Manhattan distances with the principal component feature corresponding to each prediction time period are determined.
In detail, the K historical time periods closest to the prediction time period are determined based on a K-nearest neighbor clustering method. That is, for any prediction time period t∈{1, 2 . . . Tf}, the Manhattan distance between the principal component feature corresponding to this prediction time period and the principal component feature corresponding to each historical time period is calculated, and the K historical time periods having the nearest Manhattan distances are selected. The value of K may affect prediction effect. The best value of K can be determined by parameter sensitivity analysis during cross-checking. For example, K=300 is used in this embodiment.
At block 112, a fitting relationship is acquired according to the K principal component features corresponding to the K historical time periods and K photovoltaic powers corresponding to the K historical time periods.
Based on support vector machine, the relationship between the principal component features of the K historical time periods closest to the prediction time period and the photovoltaic powers corresponding to the K historical time periods closest to the prediction time period may be fitted. The fitting relationship of the prediction time period t is obtained as: gt: R1×L→R1×1, where, gt represents a mapping function.
At block 113, the principal component feature corresponding to each prediction time period is inputted to the fitting relationship to obtain a photovoltaic power corresponding to each prediction time period.
In detail, the principal component feature corresponding to the prediction time period t is inputted to the fitting relationship gt to obtain the photovoltaic power corresponding to the prediction time period t.
In the related art, the method for predicting the photovoltaic power of the photovoltaic power station mainly includes the followings.
(1) Collect weather data of the historical time periods, such as the atmospheric pressure, the air sediment, the air humidity, the sky cloud volume, the wind speed, the air temperature and the surface heat radiation.
(2) Use the weather data of the historical time periods as input directly. Through data-driven algorithms, such as neural network or support vector machine, the relationship between historical weather data and historical photovoltaic power is statistically learned.
(3) Collect weather data of the prediction time periods. Input this weather data directly to the relationship statistically learned between historical weather data and historical photovoltaic power, so as to obtain the photovoltaic powers of the prediction time periods.
However, the method in the related art strongly rely on statistical learning abilities of the data-driven algorithms such as neural networks and support vector machines and ignore the role of the physical model of the photovoltaic power on the data-driven algorithms. The physical model of the photovoltaic power represents an analytical formula among the photovoltaic power and weather features such as ambient temperature, wind speed and surface light intensity, and implies the physical law among the photovoltaic power and weather features.
The calculation results of the method for predicting the photovoltaic power of the photovoltaic power station according to the embodiments of the present disclosure are compared with that of the method for predicting the photovoltaic power of the photovoltaic power station according to the related art, which are illustrated in
The embodiments of the present disclosure provide the method for predicting the photovoltaic power of the photovoltaic power station by aiming at practical demands in industry and academia and considering deficiencies of data-driven methods in the related art. The method combines the physical model of the photovoltaic power with the data-driven functionality. The method determines the key weather features affecting the photovoltaic power through the physical model of the photovoltaic power, further uses the data-driven method to fit the relationship among the key features and the photovoltaic power, and accurately predicts the photovoltaic power. The present disclosure has significant improvement in the prediction accuracy.
To achieve the above embodiments, the present disclosure further provides an apparatus for photovoltaic power forecast based on numerical weather prediction.
To achieve the above embodiments, the present disclosure further provides non-transitory computer-readable storage medium having stored therein computer programs. When the computer programs are executed by a processor, the above method according to any one of the above embodiments of the present disclosure is implemented.
Reference throughout this specification to “an embodiment,” “some embodiments,” “one embodiment”, “another example,” “an example,” “a specific example,” or “some examples,” means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present disclosure. Thus, the appearances of the above phrases in various places throughout this specification are not necessarily referring to the same embodiment or example of the present disclosure. Furthermore, the particular features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art may combine the different embodiments or examples described in this specification and features of different embodiments or examples without conflicting with each other.
Terms such as “first” and “second” are used herein for purposes of description and are not intended to indicate or imply relative importance or significance or imply the number of technical features. Furthermore, the feature defined by “first” or “second” may indicate or imply including at least one feature. In the description of the present disclosure, “a plurality or refers to two or more unless otherwise specified.
Any process or method described in a flow chart or described herein in other ways may be understood to include one or more modules, segments or portions of” codes of executable instructions for achieving specific logical functions or steps in the process, and the scope of a preferred embodiment of the present disclosure includes other implementations, in which the functions may be executed in other orders instead of the order illustrated or discussed, including in a basically simultaneous manner or in a reverse order, which should be understood by those skilled in the art.
Any process or method described in a flow chart or described herein in other ways may be understood to be a sequence table of executable instructions for achieving logical functions, which may be realized in any computer-readable medium for being used by the instruction execution system, device or apparatus (for example, the system based on the computer, the system including the processor or other systems capable of reading instructions from the instruction execution system, device or apparatus and executing the instructions) or being used in combination with the instruction execution system, device or apparatus. In the specification, “computer-readable medium” may be any device including, storing, communicating, broadcasting or transmitting programs for being used by the instruction execution system, device or apparatus or being used in combination with the instruction execution system, device or apparatus. Specific examples of the computer-readable medium (non-exhaustiveness list) include: electrical connection (electronic device) having one or one wires, portable computer disk box (magnetic device), random access memory (RAM), read only memory (ROM), electrically programmable read-only-memory (EPROM or flash memory), fiber device, and portable CD-ROM. In addition, the computer-readable medium may even to paper on which programs can be printed or other appropriate medium, this is because optical scanning may be performed on the paper or the other medium, and then edit, interpretation or any other appropriate way if necessary are performed to electrically obtain the programs, and then the programs are stored in the computer storage.
It should be understood that each part of the present disclosure may be realized by the hardware, software, firmware or their combination. In the above embodiments, a plurality of steps or methods may be realized by the software or firmware stored in the memory and executed by the appropriate instruction execution system. For example, if it is realized by the hardware, likewise in another embodiment, the steps or methods may be realized by one or a combination of the following techniques known in the art: a discrete logic circuit having a logic gate circuit for realizing a logic function of a data signal, an application-specific integrated circuit having an appropriate combination logic gate circuit, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
Those skilled in the art shall understand that all or parts of the steps in the above exemplifying method of the present disclosure may be achieved by commanding the related hardware with programs. The programs may be stored in a computer readable storage medium, and the programs include one or a combination of the steps in the method embodiments of the present disclosure when run on a computer.
In addition, each function cell of the embodiments of the present disclosure may be integrated in a processing module, or these cells may be separate physical existence, or two or more cells are integrated in a processing module. The integrated module may be realized in a form of hardware or in a form of software function modules. When the integrated module is realized in a form of software function module and is sold or used as a standalone product, the integrated module may be stored in a computer readable storage medium.
The storage medium mentioned above may be read-only memories, magnetic disks or CD, etc. Although explanatory embodiments have been shown and described, it would be appreciated by those skilled in the art that the above embodiments cannot be construed to limit the present disclosure, and changes, alternatives, and modifications can be made in the embodiments without departing from scope of the present disclosure.
Number | Date | Country | Kind |
---|---|---|---|
2017 1 0492720 | Jun 2017 | CN | national |
Number | Name | Date | Kind |
---|---|---|---|
20160104084 | Philip | Apr 2016 | A1 |
20180046924 | Huang | Feb 2018 | A1 |
20180175790 | Sanfilippo | Jun 2018 | A1 |
20180225416 | Wong | Aug 2018 | A1 |
Entry |
---|
Wen, “An Investigation of the Effect of Wind Cooling on Photovoltaic Arrays”, Mar. 1982, Jet Propulsion Laboratory (Year: 1982). |
Number | Date | Country | |
---|---|---|---|
20180375334 A1 | Dec 2018 | US |