Power generation forecasting system, power system

Abstract

To provide a technique capable of accurately estimating a generated power of a solar battery by accurately estimating a solar radiation amount with respect to the solar battery.SOLUTION: A power generation prediction system learns a correlation between prediction data of solar radiation amounts based on meteorological information and estimated solar radiation amounts estimated from outputs of a solar battery, and predicts a solar radiation amount and a generated power on the basis thereof.SELECTED DRAWING: Figure 4A

Description

[Technical field]

[0001] The present invention relates to a technique for predicting the amount of power generated by a solar cell. [Background technology]

[0002] The amount of power generated by a solar cell varies depending on the amount of solar radiation it receives. When operating a power system that includes solar cells, it is sometimes necessary to predict the power output from the solar cells in advance and create an operation plan based on that prediction. For example, when it is necessary to declare the amount of power generated from a private solar cell system in advance, it is necessary to create such an operation plan, and it is necessary to predict the amount of power generated by the solar cells as part of the plan.

[0003] The amount of solar radiation on the solar cell can be known in advance, for example, according to a solar radiation forecast provided by a weather information provider. The output from the solar cell can be predicted according to this predicted amount of solar radiation. An operation plan for the private solar cell system can be created according to this power generation forecast.

[0004] Patent Document 1 aims to "provide a measurement device, a prediction method, and a computer program for calculating a predicted value of a prediction target with high accuracy," and describes the following technology (see abstract): "A prediction device 101 includes a similarity calculation unit 20 that calculates multiple similarities of the weather forecast data at the multiple first times to the weather forecast data at the second time based on weather forecast data at multiple first times at a point included in a geographical range having multiple points and weather forecast data at the point at a second time, and a prediction unit 21 that calculates a predicted value at the second time based on the multiple similarities and actual values ​​at the multiple first times in the geographical range."

[0005] Patent application 2020-127908 does not relate to prediction of power generation of a solar cell, but describes technology related to the present application regarding prediction of solar radiation and battery temperature. The application aims to "provide a battery diagnostic device capable of accurately detecting the occurrence of partial shading or partial disconnection in a solar cell." The application describes a technology in which "the battery diagnostic device according to the present invention converts the operating current and operating voltage of a solar cell into values ​​assumed when the solar cell is operating under standard solar radiation and standard temperature, and if the converted standard current is greater than the standard current when the solar cell is operating under the standard solar radiation and standard temperature by a threshold value or more, the solar cell is diagnosed as being in a partial shade state or a partial disconnection state" (see abstract). [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2020-149475 A Summary of the Invention [Problem to be solved by the invention]

[0007] The solar radiation forecast provided by the weather information provider is predicted using the measurement history of the pyranometers installed at each point. Since each pyranometer is placed at a certain distance from each other, it is considered that the solar radiation in the gap area is estimated, for example, by interpolating the measured values. In other words, the solar radiation forecast provided by the weather information provider is relatively coarse in accuracy, and therefore the power generation forecast using this is also assumed to be coarse in accuracy.

[0008] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a technology that can accurately estimate the amount of power generation of a solar cell by accurately estimating the amount of solar radiation on the solar cell. [Means for solving the problem]

[0009] The power generation prediction system according to the present invention learns the correlation between predicted data on solar radiation based on meteorological information and estimated solar radiation estimated from the output of solar cells, and predicts the amount of solar radiation and power generation based on this. Effect of the Invention

[0010] According to the power generation prediction system of the present invention, the amount of solar radiation on the solar cell can be accurately estimated, thereby making it possible to accurately estimate the amount of power generated by the solar cell. Problems, configurations, advantages, etc. other than those described above will become apparent from the description of the following embodiments. [Brief description of the drawings]

[0011] [Figure 1] This is an example of weather forecast data provided by a weather information provider. [Figure 2A] 1 is a graph comparing the amount of solar radiation described in solar radiation prediction data with the amount of solar radiation actually measured. [Figure 2B] 1 is a graph comparing the amount of solar radiation described in solar radiation prediction data with the amount of solar radiation actually measured. [Figure 2C] 1 is a graph comparing the amount of solar radiation described in solar radiation prediction data with the amount of solar radiation actually measured. [Diagram 3] 1 is a configuration diagram of a power generation prediction system 1 according to a first embodiment. [Figure 4A] 10 is a process flow diagram illustrating a learning process for estimating the actual amount of solar radiation on the solar cell 11 by the arithmetic device 15. FIG. [Figure 4B] FIG. 4B is a process flow diagram illustrating a procedure for predicting the amount of power generated by the solar cell 11 using the learning result in FIG. 4A. [Figure 5A] 13 is a process flow diagram illustrating another learning process for estimating the actual amount of solar radiation on the solar cell 11 by the arithmetic device 15. FIG. [Figure 5B] FIG. 5B is a process flow diagram illustrating a procedure for predicting the amount of power generated by the solar cell 11 using the learning result in FIG. 5A. [Figure 6A] This is an example of the installation inclination of solar cells. [Figure 6B]1 is a graph illustrating deviations of predicted solar radiation amount data from actual solar radiation amount due to the installation orientation of the solar cell. [Figure 7] 4 is a process flow diagram illustrating a procedure in which the power generation prediction system 1 estimates the installation orientation of the solar cell 11. FIG. [Figure 8A] 10 is a process flow diagram illustrating a learning process for estimating the actual amount of solar radiation on solar cell 11 and the temperature of solar cell 11 by a computing device 15. FIG. [Figure 8B] FIG. 8B is a process flow diagram illustrating a procedure for predicting the amount of power generated by the solar cell 11 using the learning result in FIG. 8A. [Figure 9A] Here is an example of further categorizing weather codes: [Figure 9B] FIG. 1 is a diagram comparing the actually measured amount of solar radiation and the predicted amount of solar radiation data on a certain day. [Figure 9C] FIG. 11 is a diagram comparing the actually measured amount of solar radiation and the predicted amount of solar radiation data on another day. [Figure 10A] FIG. 4 is a schematic diagram showing a procedure in which the calculation device 15 learns detailed weather classifications. [Figure 10B] FIG. 10B is a diagram illustrating a procedure for estimating a detailed weather classification using the results of learning according to the procedure in FIG. 10A. [Figure 11] 10 is a flowchart illustrating a procedure in which a calculation device 15 calculates the amount of solar radiation on a solar cell 11 and the temperature of the solar cell 11 from the output of the solar cell 11. [Figure 12] 2 is an example of a user interface presented by the computing device 15. [Figure 13] 13 is a graph illustrating a case where there is a large discrepancy between a previously created operation plan and the total output from the solar cell 11 and the storage battery 12. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] <Embodiment 1> Figure 1 is an example of weather forecast data provided by a weather information provider. In this data example, the combination of solar radiation and weather (weather code or weather code) at a specific location is described for each hour. In some cases, weather information other than solar radiation and weather is also described. In this data example, the temperature, probability of precipitation, and wind speed are described. The weather code represents one of five weather types.

[0013] A power generation prediction system that predicts the amount of power generated by solar cells can obtain forecast data such as that shown in Figure 1 from weather information providers and use it to estimate the amount of power generated by solar cells on an hourly basis. However, this solar radiation prediction may not necessarily be based on the measurement results of a pyranometer installed at the target location. This is because the pyranometers are installed at a certain distance apart. Therefore, this solar radiation prediction may deviate from the actual amount of solar radiation at the location where the solar cells are actually installed.

[0014] 2A to 2C are graphs comparing the amount of solar radiation described in the solar radiation prediction data with the actually measured amount of solar radiation. FIG. 2A shows an example when the weather code is 10 (sunny), FIG. 2B shows an example when the weather code is 20 (cloudy), and FIG. 2C shows an example when the weather code is 30 (rainy). The dotted lines show the results of the predicted amount of power generation based on the solar radiation prediction data. The solid lines show the actually measured amount of solar radiation. As these graphs show, solar radiation predictions provided by weather information providers and the like may deviate from the actually measured values. Therefore, if the predicted amount of solar radiation is used to estimate the amount of power generation from a solar cell, a large error may occur.

[0015] For example, in a power system that combines solar cells and storage batteries, in order to control the change over time in the power supplied to the outside of the system, it is useful to predict the change over time in the output from the solar cells in advance and create an operation plan for the power system based on the prediction result. However, as shown in Figures 2A to 2C, if the predicted data and the actual measured value are significantly different, the operation plan may be hindered. Therefore, in the first embodiment of the present invention, it is intended to estimate the actual amount of solar radiation on the solar cells without relying only on the solar radiation prediction data, and to estimate the amount of power generated by the solar cells based on the estimation result.

[0016] 3 is a configuration diagram of the power generation prediction system 1 according to the present embodiment 1. The power generation prediction system 1 is a system that predicts the power generated by a solar cell 11. The power generation prediction system 1 includes the solar cell 11, a storage battery 12, a power conditioner 13, an optical fiber 14, a computing device 15, a data server 16, and a control device 17.

[0017] The power conditioner 13 is a device that receives DC power output by the solar cell 11 or the storage battery 12, converts it into AC power, and supplies it to the power grid. The power conditioner 13 is appropriately equipped with conversion devices such as a chopper 131 and a DC / AC converter 132. The power conditioner 13 obtains state values ​​such as the output current and output voltage of each of the solar cell 11 and the storage battery 12, and outputs them to the optical fiber 14 via the communication interface 133.

[0018] The arithmetic device 15 is a computer device that predicts the amount of power generated by the solar cell 11. The arithmetic device 15 is connected to the power conditioner 13 via an optical fiber 14. The arithmetic device 15 is further connected to a data server 16 and a control device 17.

[0019] The data server 16 is configured with a storage device that stores solar radiation prediction data (as exemplified in FIG. 1) provided by a weather information provider. The control device 17 controls the power conditioner 13 according to a command value output to the power conditioner 13 by the arithmetic device 15. The power conditioner 13 suppresses the output from the solar cell 11 or controls the charging and discharging operation of the storage battery 12 according to the command value.

[0020] 4A is a process flow diagram illustrating a learning process for estimating the actual amount of solar radiation on the solar cell 11 by the arithmetic device 15. The arithmetic device 15 acquires an output voltage value from the solar cell 11 and an output current value from the solar cell 11 via the power conditioner 13. The arithmetic device 15 uses these values ​​to calculate the amount of solar radiation (estimated amount of solar radiation) that is assumed to be irradiated on the solar cell 11. The procedure for calculating the estimated amount of solar radiation will be described again in a later embodiment, and will not be described here.

[0021] The arithmetic device 15 acquires solar radiation prediction data provided by a weather information provider (for example, from the data server 16). This solar radiation prediction data may deviate from the actual amount of solar radiation as described above. Therefore, the arithmetic device 15 learns the correlation between the predicted value of the solar radiation prediction data and the calculated assumed amount of solar radiation, assuming that the assumed amount of solar radiation calculated from the output current value and the output voltage value is accurate.

[0022] Since the solar radiation forecast data describes a forecast value for each hour, for example, as described in Fig. 1, the correlation between the forecast value of the solar radiation forecast data and the calculated assumed solar radiation can be obtained for a plurality of forecast values. The calculation device 15 specifies a calculation formula expressing the correlation between the two by an appropriate machine learning method, an error minimization method, or the like. Although an example of expressing the correlation by a linear function is shown in Fig. 4A, the correlation is not limited to a linear function.

[0023] The calculation formula expressing the correlation does not have to be one. For example, a year may be divided into a plurality of periods, and a correlation may be specified for each period. Although an example of specifying a calculation formula for each month is shown in FIG. 4A, this is not limiting. The same applies to the following embodiments.

[0024] Depending on the type of solar cell 11, for example, a bifacial light receiving panel may be adopted. Such solar cells receive light from the top surface and also receive reflected light from the ground from the bottom surface. In the case of such a configuration, the reflected light may vary significantly due to factors such as seasonal fluctuations. Therefore, it is desirable to learn the correlations individually for each period in which such fluctuations occur. Constructing a correlation calculation formula for each month is also useful from this perspective.

[0025] Fig. 4B is a process flow diagram explaining the procedure for predicting the amount of power generation from the solar cell 11 using the learning result in Fig. 4A. The calculation device 15 acquires solar radiation prediction data, and applies the predicted value to the learning result obtained in Fig. 4A (the correlation between the predicted value of the solar radiation prediction data and the calculated assumed solar radiation), thereby acquiring an estimate of the actual solar radiation on the solar cell 11 as an output from the learning model. When a calculation formula expressing the correlation is constructed for each period, the calculation formula corresponding to the current period is used. The same applies to the following embodiments.

[0026] The calculation device 15 calculates the amount of power generation E from the solar cell 11 as follows: PA may be acquired by a temperature sensor, may be estimated by a procedure described in an embodiment below, or a fixed value such as a standard temperature may be used if the temperature change is small. Alternatively, the air temperature may be replaced with the temperature of the solar cell 11 by adding or subtracting a fixed value that is predetermined for each installation mode of the solar cell 11 to or from the air temperature described in the solar radiation prediction data. Any other appropriate method may be used.

[0027] E [kW] = K × P AS ×I T,G ÷G S (1) P AS : Rated value I T,G : Estimated solar radiation obtained from the learning results G S :Standard solar radiation amount

[0028] K = K' x K PT ×e d K PT : Constant (array loss 0.97) e d : Degradation coefficient

[0029] K PT =1+α PMAX ×(T PA -25) / 100 α PMAX : Coefficient [% / ℃] T PA :Temperature of solar cell 11

[0030] FIG. 5A is a process flow diagram illustrating another learning process for the arithmetic device 15 to estimate the actual amount of solar radiation on the solar cell 11. The procedure described in FIG. 4A may be performed for each weather code described in the solar radiation prediction data. The arithmetic device 15 learns the correlation between the predicted value of the solar radiation prediction data and the calculated assumed amount of solar radiation for each weather code, as in FIG. 4A. In this example, five weather codes from weather codes 10 to 50 are described. Therefore, five correlations are learned.

[0031] Fig. 5B is a process flow diagram explaining the procedure for predicting the amount of power generation from solar cell 11 using the learning result in Fig. 5A. The calculation device 15 obtains an estimate of the actual amount of solar radiation for solar cell 11 by applying the predicted value of the prediction data to the learning result corresponding to the weather code described in the solar radiation prediction data (the correlation between the predicted value of the solar radiation prediction data and the calculated assumed amount of solar radiation). The rest is the same as Fig. 4B.

[0032] <First embodiment: Summary> The power generation prediction system 1 according to the first embodiment learns in advance the correlation between prediction data describing the results of predicting the amount of solar radiation based on meteorological information and the amount of solar radiation estimated from the output of the solar cell 11, and estimates the actual amount of solar radiation from the prediction data according to the learning results. This makes it possible to accurately estimate the actual amount of solar radiation for the solar cell 11 even if the prediction accuracy of the solar radiation prediction data is low.

[0033] <Embodiment 2> In principle, solar cells are installed with their light receiving surface facing due south. If the solar cells are installed facing due south, it is expected that they will receive the amount of solar radiation described in the solar radiation prediction data as is. However, in reality, solar cells may be installed tilted either east or west from due south. In this case, the amount of solar radiation that would have been received if the solar cells had been installed facing due south will be received slightly later (or slightly earlier) depending on the tilt. In the second embodiment of the present invention, a method for correcting the light receiving time difference caused by the installation orientation will be described. The configuration of the power generation prediction system 1 is the same as that of the first embodiment.

[0034] Figure 6A shows an example of the installation inclination of a solar cell. The installation angle when the light receiving surface of the solar cell faces due south is defined as 0°, and the installation angle when the light receiving surface is installed tilted toward either the east or west is defined as 0° east and 0° west, respectively. For example, 5° east means that the light receiving surface is tilted 5° east from due south.

[0035] Figure 6B is a graph that explains the deviation of the solar radiation prediction data from the actual solar radiation due to the installation orientation of the solar cell. If the installation orientation of the solar cell is tilted from due south, the time when the solar radiation is actually received will shift on the time axis corresponding to the tilt. For example, when comparing a case where the light receiving surface is tilted toward the east with a case where it faces due south, the time when the solar radiation is actually received will be earlier in the former case, even if the predicted amount of solar radiation is the same.

[0036] In consideration of the above, in the present embodiment 2, the installation orientation of the solar cell 11 is estimated, and the solar radiation prediction data is shifted on the time axis by an amount corresponding to the estimated orientation. Even if the solar cell 11 does not face due south, it is considered that the accuracy of the prediction data itself can be improved by time-shifting the solar radiation prediction data in accordance with the installation orientation.

[0037] If the installation direction of the solar cell 11 is significantly deviated from due south, not only the light reception time but also the amount of light received will vary. However, if the range of the orientation deviation is small, the variation in the amount of light received can be ignored. Therefore, in this embodiment 2, only the shift on the time axis due to the deviation of the installation direction is considered, and the variation in the amount of light received is ignored.

[0038] 7 is a process flow diagram explaining the procedure for the power generation prediction system 1 to estimate the installation orientation of the solar cell 11. The step of calculating the amount of solar radiation (estimated amount of solar radiation) for the solar cell 11 and the step of acquiring solar radiation prediction data provided by a weather information provider are the same as those in the first embodiment.

[0039] The calculation device 15 temporarily sets an estimated value of the installation orientation of the solar cell 11. The initial estimated value is, for example, 30° east. The calculation device 15 shifts the solar radiation prediction data on the time axis by an amount corresponding to the estimated orientation. For example, if the estimated orientation is 30° east, the solar cell 11 will receive the amount of solar radiation described in the solar radiation prediction data earlier than expected by 30° east. A time shift corresponding to this time difference is applied to the solar radiation prediction data.

[0040] The calculation device 15 calculates the square sum error between the expected amount of solar radiation (p1 to p12) at each time calculated from the output of the solar cell 11 and the amount of solar radiation (p'1 to p'12) at each time described in the solar radiation prediction data. Past history may be used as the amount of solar radiation at each time. In other words, this flowchart can be executed at any time.

[0041] The arithmetic device 15 scans the estimated value of the installation orientation of the solar cell 11 within a predetermined range. In this flowchart, the scanning is performed within a range of 30° east to 30° west. The arithmetic device 15 changes the estimated orientation, for example, by 1° and performs the same procedure as above. The scanning range is not limited to this example, and may be set appropriately within a range where fluctuations in the amount of received light due to orientation deviations can be ignored.

[0042] The computing device 15 regards the estimated orientation in which the square sum error between the assumed solar radiation and the solar radiation forecast data is the smallest as the actual installation orientation of the solar cell 11. The computing device 15 determines a time shift corresponding to the determined installation orientation. When referring to the solar radiation forecast data provided by the weather information provider thereafter, the computing device 15 applies this time shift to the forecast data before using it.

[0043] <Embodiment 2: Summary> The power generation prediction system 1 according to the second embodiment determines the installation orientation of the solar cell 11 that minimizes the difference between the assumed solar radiation calculated from the output of the solar cell 11 and the solar radiation prediction data, and regards this as the actual installation orientation of the solar cell 11. The power generation prediction system 1 subsequently uses the solar radiation prediction data after shifting it on the time axis by an amount corresponding to the estimated installation orientation. This makes it possible to improve the accuracy of the solar radiation prediction data or the power generation prediction based on the solar radiation prediction data, even if the solar cell 11 does not face due south.

[0044] <Embodiment 3> In the first embodiment, an example is described in which the amount of solar radiation (estimated amount of solar radiation) for the solar cell 11 is calculated from the output of the solar cell 11. Furthermore, an example is described in which the temperature of the solar cell 11 is defined based on, for example, air temperature when calculating the amount of power generated by the solar cell 11. However, the actual temperature of the solar cell 11 is not necessarily uniformly derived from the air temperature, and may vary greatly depending on the installation environment, etc. As shown in Equation 1, the amount of power generated by the solar cell 11 is greatly influenced by the temperature. Therefore, in the third embodiment of the present invention, an example is described in which the temperature of the solar cell 11 is estimated from the output of the solar cell 11, and the amount of power generated by the solar cell 11 is calculated using the temperature. The configuration of the power generation prediction system 1 is the same as that of the first embodiment.

[0045] 8A is a process flow diagram illustrating a learning process for arithmetic device 15 to estimate the actual amount of solar radiation on solar cell 11 and the temperature of solar cell 11. The step of calculating the amount of solar radiation on solar cell 11 (estimated amount of solar radiation), the step of acquiring solar radiation prediction data provided by a meteorological information provider, and the step of learning the correlation between the estimated amount of solar radiation and the predicted amount of solar radiation are the same as those in the first embodiment.

[0046] The arithmetic device 15 acquires an output voltage value from the solar cell 11 and an output current value from the solar cell 11 via the power conditioner 13 (already acquired values ​​may be reused). The arithmetic device 15 uses these values ​​to calculate the temperature (estimated temperature) of the solar cell 11. The procedure for calculating the estimated temperature will be described again in the embodiment described later, and therefore will not be described here.

[0047] The arithmetic device 15 learns the correlation between the expected solar radiation amount and the expected temperature. The arithmetic device 15 identifies a calculation formula expressing the correlation between the two, for example, by an appropriate machine learning method or an error minimization method. Although an example of expressing the correlation by a linear function is shown in FIG. 8A, the correlation is not limited to a linear function. As in the first embodiment, the correlation may be learned for each period (for example, each month).

[0048] FIG. 8B is a process flow diagram for explaining the procedure for predicting the amount of power generation from the solar cell 11 using the learning result in FIG. 8A. The steps for acquiring an estimated value of the actual amount of solar radiation on the solar cell 11 are the same as those in the first embodiment. The calculation device 15 acquires the temperature of the solar cell 11 as an output from the learning model by applying the estimated amount of solar radiation to the correlation between the expected amount of solar radiation and the battery temperature. The calculation device 15 calculates the acquired battery temperature as T PA By substituting for, the power generation amount E of the solar cell 11 is calculated.

[0049] <Embodiment 3: Summary> The power generation prediction system 1 according to the third embodiment estimates the temperature of the solar cell 11 using the output from the solar cell 11, and calculates the amount of power generation of the solar cell 11 using the estimated temperature. This makes it possible to accurately predict the amount of power generation of the solar cell 11 even if there is a large discrepancy between the air temperature described in the solar radiation prediction data or the like and the temperature of the solar cell 11.

[0050] <Fourth embodiment> In the first embodiment, it has been described that five weather types are described as weather codes in the solar radiation prediction data provided by the weather information provider. However, even if the weather code is the same, the actual solar radiation amount may vary greatly. For example, weather code 10 (sunny) indicates that the cloud ratio is 0 to 80%, and even if the weather code is the same, the solar radiation amount varies greatly depending on the individual situation. Therefore, in the fourth embodiment of the present invention, an example will be described in which a more detailed weather type classification is estimated using the solar radiation prediction value described in the solar radiation prediction data. The configuration of the power generation prediction system 1 is the same as that of the first embodiment.

[0051] 9A shows an example of classifying weather codes into more detailed categories. Here, weather code 10 is further classified into two categories, and weather code 20 is further classified into two categories. In the following, detailed weather classification into which weather code 10 is classified will be described as an example.

[0052] Figure 9B is a graph comparing the measured solar radiation amount and the predicted solar radiation amount data on a certain day. This graph shows an example of a day with a weather code of 10 (sunny). In this example, it can be seen that the measured solar radiation amount and the predicted solar radiation amount data match relatively well.

[0053] FIG. 9C is a diagram comparing the measured solar radiation amount and the predicted solar radiation amount data on another day. This graph shows an example of another day where the weather code is 10 (clear). In this example, the weather code is the same as in FIG. 9B, 10 (clear), but it can be seen that there is a large deviation between the measured solar radiation amount and the predicted solar radiation amount data. In this way, even with the same weather code, the solar radiation amount may deviate significantly from the predicted data.

[0054] 10A is a schematic diagram showing a procedure for the calculation device 15 to learn detailed weather classification. The calculation device 15 can learn the correlation between the solar radiation forecast data and the detailed weather classification using a learning device constructed by, for example, a neural network. Here, a neural network is exemplified in which the input layer receives solar radiation forecast values ​​for 12 hours, the intermediate layer (hidden layer) is composed of three neurons, and the output layer is composed of two neurons.

[0055] This neural network is constructed to further classify weather code 10 into two detailed weather categories 11 and 12. If the outputs from the two neurons in the output layer are "1" and "0", respectively, it means that the input solar radiation forecast value is for weather code 11. If the outputs from the two neurons in the output layer are "0" and "1", respectively, it means that the input solar radiation forecast value is for weather code 12.

[0056] The predicted solar radiation value for 12 hours on the first day (DAY1) is input to the first neuron in the hidden layer. In this example, the detailed weather classification for DAY1 is assumed to be weather code 11 (output layer "1" "0"), which is the correct answer. The first neuron in the hidden layer has a weighting matrix composed of 12 elements that are multiplied by each of the 12 elements of the matrix describing the predicted solar radiation value. The result of multiplying both matrices is substituted into the sigmoid function represented by the first calculation formula in the lower part of Figure 10A. The threshold of the first neuron in the hidden layer is specified as 15.02. The output y of the first neuron in the hidden layer is calculated by the above procedure.

[0057] The predicted solar radiation amount for 12 hours on the second day (DAY2) is input to the second neuron in the hidden layer. The same is true for the third and subsequent days, which are input to the corresponding neurons in the hidden layer. The output y of each hidden layer is calculated in the same way. For ease of description, solar radiation amount prediction data up to the third day is used in this example, but the same is true for the fourth and subsequent days.

[0058] The output y of the first neuron in the hidden layer is input to the first neuron in the output layer. The first neuron in the output layer has a weight matrix composed of three elements that are multiplied by the outputs y of the first to third neurons in the hidden layer. The result of multiplying the two matrices is substituted into the sigmoid function represented by the second calculation formula in the lower part of FIG. 10A. The threshold of the first neuron in the output layer is specified as 19.84. The output z of the first neuron in the output layer is obtained by the above procedure. The same is true for the second neuron in the output layer.

[0059] From the above calculations, the output of the first neuron in the output layer is 1.00, and the output of the second neuron in the output layer is 0.33. On the other hand, since the correct answers for the weather code on DAY1 are "1" and "0", the error Q between the output from the output layer and the correct answer can be calculated using the third formula in the lower part of Figure 10A. The same calculations are performed for DAY2 to DAY3. Learning is completed by optimizing the matrix elements and thresholds of each layer so that the error Q is minimized. Methods for optimizing neural networks are well known, such as backpropagation, so a description of these will be omitted.

[0060] Fig. 10B is a diagram explaining the procedure for estimating a detailed weather classification using the results of learning according to the procedure in Fig. 10A. The calculation device 15 inputs solar radiation forecast data to the neural network constructed according to Fig. 10A. The output from each neuron in the output layer is obtained according to the same procedure as during learning. For example, if the output from the first neuron in the output layer is equal to or greater than the threshold and the output from the second neuron is less than the threshold, the estimation result by the neural network is "1" or "0", i.e., weather code 11. This results in an estimation result of the detailed weather classification.

[0061] The calculation device 15 learns in advance the correlation between the expected amount of solar radiation and the solar radiation prediction data for each detailed weather classification in the same manner as described in Fig. 5A. The calculation device 15 estimates the amount of solar radiation for the solar cell 11 using the correlation corresponding to the estimated detailed weather classification. The subsequent steps are the same as those in the first embodiment.

[0062] <Embodiment 4: Summary> The power generation prediction system 1 according to the fourth embodiment learns in advance the correlation between the solar radiation prediction data and the detailed weather classification, and estimates the detailed weather classification from the solar radiation prediction data according to the learning result. The calculation device 15 learns in advance the correlation between the assumed solar radiation and the solar radiation prediction data for each detailed weather classification. This makes it possible to accurately estimate the solar radiation on the solar cell 11 and the power generation amount of the solar cell 11, even if the correlation between the solar radiation prediction data and the actual solar radiation greatly varies depending on the weather.

[0063] <Fifth Preferred Embodiment> In the fifth embodiment of the present invention, a specific procedure for calculating the amount of solar radiation on solar cell 11 and the temperature of solar cell 11 using the output current from solar cell 11 and the output voltage from solar cell 11 will be described.

[0064] Fig. 11 is a flow chart for explaining the procedure by which the arithmetic unit 15 calculates the amount of solar radiation on the solar cell 11 and the temperature of the solar cell 11 from the output of the solar cell 11. Each step in Fig. 11 will be explained below. The process of deriving the calculation formula for each step will be described later.

[0065] (FIG. 11: Step S1101) The computing device 15 calculates the operating voltage V of the solar cell 11. PE and the operating current I PE The calculation device 15 further obtains specification data describing the specification values ​​of the solar cell 11. The specification data describes the following specification values: (a) short-circuit current I SC_ST , (b) Open circuit voltage V OC_ST , (c) Optimum operating current I OP_ST , (d) Optimum operating voltage V OP_ST (e) Coefficients α and β in the calculation formula described later. α is the temperature characteristic of the short circuit current [% / °C], and β is the temperature characteristic of the operating voltage [mV / °C]. The specification data may be obtained, for example, from the power conditioner 13, or may be stored in advance in a storage device provided in the calculation device 15. The specification data may also be obtained by other appropriate means.

[0066] (FIG. 11: Step S1102) The computing device 15 calculates the short-circuit current I SC The specification value I SC_ST Initialize it by:

[0067] (FIG. 11: Step S1103) The calculation device 15 calculates the estimated solar radiation amount p for the solar cell 11. The estimated solar radiation amount p is calculated based on the short circuit current I SC_ST The current operating current at the current solar radiation level is I PE Therefore, the short circuit current at the present solar radiation level is I PE Therefore, the expected solar radiation p can be defined as (short circuit current at the current solar radiation) / I SC_ST That is, it is expressed by the following formula 2. Using this formula is useful when there is a situation where the maximum power operating point for the current amount of solar radiation cannot be specified.

[0068] (Figure 11: Step S1103: Calculation formula) p=(I PE / j) / I SC_ST (2) j = (operating current at the current operating point at the current solar radiation) / (short circuit current at the current solar radiation)

[0069] (FIG. 11: Step S1104) The calculation device 15 calculates the operating voltage V at the assumed solar radiation p and the battery temperature T=298K using kT / q=0.026 at 298K. P is calculated according to the following formula:

[0070] (Figure 11: Step S1104: Calculation formula) V p =n f 0.026 N cell ln(p)+V op_ST

[0071] (FIG. 11: Step S1105) The computing device 15 calculates the assumed temperature T of the solar cell 11 according to the following formula 7. β is generally a temperature characteristic of silicon, and is approximately −2 mV / K.

[0072] (Figure 11: Step S1105: Calculation formula) T=298+(V PE -V p ) / (N cell β) (7)

[0073] (FIG. 11: Step S1106) The arithmetic device 15 ends this flowchart if S1103 to S1105 have been repeated a predetermined number of times or more (for example, three times), and if not, returns to S1103 after performing S1107.

[0074] (FIG. 11: Step S1107) The computing device 15 calculates the short circuit current I SC is updated according to the following formula, and the process returns to step S1103.

[0075] (Figure 11: Step S1107: Calculation formula) I SC =I SC_ST ·{1+(α·(T-298)) / 100}

[0076] <Embodiment 5: Process of deriving the calculation formula in S1104> I SC =I SC_ST , kT / q=0.026 at 298K, I=I SC_ST Using p j, V p can be expressed by the following Equation 3.

[0077] V p =n f 0.026 ln(((1-j) I SC_ST p) / Is) (3)

[0078] Standard operating voltage V when p=1.0 OP_ST is expressed by the following formula 4.

[0079] V OP_ST =n f 0.026 ln(((1-j) I SC_ST ) / Is)···(4)

[0080] By subtracting equation 4 from equation 3 to calculate the voltage difference, we obtain the following equation 5.

[0081] V p -V OP_ST =n f 0.026 ln(p) (5)

[0082] Therefore, the following formula 6 can be derived. This formula 6 is used in S1104.

[0083] V p =(V p -V OP_ST )+V OP_ST =n f 0.026 ln(p)+V OP_ST (6)

[0084] Through the above steps, the estimated amount of solar radiation p for the solar cell 11 and the estimated temperature T of the solar cell 11 can be calculated by repeated convergence calculation. When the calculation device 15 estimates the estimated amount of solar radiation and the estimated temperature of the solar cell 11 in each embodiment, the method described in the fifth embodiment can be used.

[0085] <Sixth embodiment> Fig. 12 is an example of a user interface presented by the arithmetic device 15. The arithmetic device 15 may present the results of calculations in the above embodiments on a user interface as shown in Fig. 12. The user interface can be presented as a screen interface on a display provided in the arithmetic device 15, for example, but is not limited to this and any type of user interface may be presented.

[0086] Possible items to be presented include, for example, the following: (a) predicted solar radiation value for each time provided by the weather information provider; (b) expected solar radiation at each time calculated by the computing device 15; (c) weather code for each time described in the solar radiation prediction data; (d) detailed weather classification estimated by the computing device 15; (e) temperature of the solar cell 11 calculated by the computing device 15; (f) formula expressing the correlation between the expected solar radiation and the solar radiation prediction data (for each weather classification); (g) formula expressing the overall relationship between the expected solar radiation and the battery temperature (for each weather classification).

[0087] <Embodiment 7> In the above embodiment, the amount of power generated by the solar cell 11 estimated by the power generation prediction system 1 may be used to create an operation plan for the solar cell 11 and the storage battery 12. For example, the computing device 15 can plan in advance the total power output by the solar cell 11 and the storage battery 12 at each time of day according to the estimated amount of power generated by the solar cell 11. This allows the power generation prediction system 1 to operate as a power system that creates an operation plan for the solar cell 11 and the storage battery 12.

[0088] 13 is a graph illustrating a case where there is a large deviation between a previously created operation plan and the total output from the solar cell 11 and the storage battery 12. When there is a large deviation between the two, the control device 17 controls the solar cell 11 or the storage battery 12 as follows, thereby making it possible to bring the total power output by the solar cell 11 and the storage battery 12 closer to the plan.

[0089] When the total output exceeds the plan, the control device 17 controls the solar cell 11 and the storage battery 12 so as to suppress the output from the solar cell 11 (i.e., the output power is not connected to the power grid and is discarded as thermal energy) or to use the surplus output to charge the storage battery 12. This suppresses the surplus and makes it possible to bring the total output closer to the plan.

[0090] When the total output falls short of the plan, the control device 17 controls the storage battery 12 so that the storage battery 12 discharges the shortfall. This makes up for the shortfall and allows the total output to approach the plan.

[0091] <Modifications of the present invention> The present invention is not limited to the above-described embodiment, and includes various modified examples. For example, the above-described embodiment has been described in detail to clearly explain the present invention, and is not necessarily limited to those having all of the configurations described. In addition, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. In addition, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

[0092] For example, in the second embodiment, the installation direction of the solar cell 11 is estimated, and the estimated direction is used to shift the solar radiation prediction data on the time axis, and then the solar radiation prediction method and the power generation prediction method described in the other embodiments can be used. The other embodiments can also be used in appropriate combinations. [Explanation of symbols]

[0093] 1: Power generation forecasting system 11: Solar cell 12: Battery 13: Power conditioner 15: Arithmetic device 16: Data Server 17: Control device

Claims

1. A power generation prediction system for predicting the amount of power generation of a solar cell, comprising: a calculation device that predicts the amount of power generation using prediction data describing the results of predicting an amount of solar radiation based on meteorological information; the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; The calculation device estimates the amount of solar radiation for each orientation in which the solar cell is installed, The computing device estimates the direction in which an error between the amount of solar radiation estimated for each direction and the amount of solar radiation described in the prediction data is minimum as the direction in which the solar cell is installed. A power generation prediction system comprising:

2. A power generation prediction system for predicting the amount of power generation of a solar cell, comprising: a calculation device that predicts the amount of power generation using prediction data describing the results of predicting an amount of solar radiation based on meteorological information; the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; the computing device calculates a third predicted value of the temperature of the solar cell using an output current from the solar cell, an output voltage from the solar cell, and a specification value of the solar cell; The arithmetic device learns in advance a second correlation between the first predicted value and the third predicted value; The arithmetic device estimates the temperature by applying the amount of solar radiation estimated using the first correlation to the second correlation obtained as a result of the learning; and The calculation device predicts the power generation amount by using the temperature estimated using the second correlation. A power generation prediction system comprising:

3. A power generation prediction system for predicting the amount of power generation of a solar cell, comprising: a calculation device that predicts the amount of power generation using prediction data describing the results of predicting an amount of solar radiation based on meteorological information; the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; the computing device acquires an operating current at an operating point of the solar cell and an operating voltage at the operating point; The arithmetic device converts the operating current and the operating voltage into a converted standard current and a converted standard voltage, which are values ​​assumed to be obtained when the solar cell is operating under a standard amount of solar radiation and a standard temperature; The computing device acquires specification data describing specification values ​​of the solar cell, The computing device estimates the amount of solar radiation on the solar cell by repeatedly performing an estimation process that calculates a virtual amount of solar radiation on the solar cell using the specification values ​​and calculates a virtual temperature of the solar cell using the specification values. A power generation prediction system comprising:

4. The forecast data describes a forecast value of the amount of solar radiation together with a weather classification, The computing device learns the first correlation for each weather classification, The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation corresponding to the weather classification described in the prediction data.

4. The power generation prediction system according to claim 1,

5. The arithmetic device learns the first correlation for each period obtained by dividing a year into a plurality of periods, The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation corresponding to the period at the time point when the amount of solar radiation is estimated.

5. The power generation prediction system according to claim 4.

6. the forecast data describes forecast values ​​of the amount of solar radiation for a plurality of times; The calculation device estimates the amount of solar radiation for each time, The computing device estimates the orientation in which an error between the amount of solar radiation estimated for each time and the amount of solar radiation described in the prediction data is minimized as the orientation in which the solar cell is installed.

2. The power generation prediction system according to claim 1.

7. the calculation device calculates a correction amount for converting the amount of solar radiation for each time period described in the prediction data into an amount of solar radiation actually received by the solar cell according to the estimated direction; The calculation device corrects the prediction data according to the correction amount.

7. The power generation prediction system according to claim 6.

8. The calculation device converts the amount of solar radiation for each time period described in the prediction data into the amount of solar radiation actually received by the solar cell by shifting the amount of solar radiation for each time period described in the prediction data on a time axis. The power generation prediction system according to claim 7.

9. The arithmetic device learns the second correlation for each period obtained by dividing a year into a plurality of periods, The calculation device estimates the temperature by applying the amount of solar radiation estimated using the first correlation to the second correlation corresponding to the period at the time point when the temperature is estimated.

3. The power generation prediction system according to claim 2.

10. The arithmetic device learns the first correlation for each detailed weather classification obtained by further classifying the weather classification, the calculation device estimates the detailed weather classification by applying the predicted value of the amount of solar radiation described in the prediction data to the first correlation learned for each detailed weather classification; The power generation prediction system according to claim 4, characterized in that the calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation corresponding to the estimated detailed weather classification.

11. The power generation prediction system further includes an interface for presenting at least one of the estimated solar radiation amount or the predicted power generation amount.

4. The power generation prediction system according to claim 1,

12. A power system that creates a power generation amount of a solar cell and an operation plan for a storage battery, a calculation device that predicts the amount of power generation using prediction data describing the results of predicting an amount of solar radiation based on meteorological information; the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; the computing device creates an operation plan instructing a total power output from the solar cell and the storage battery according to the predicted power generation amount; The calculation device estimates the amount of solar radiation for each orientation in which the solar cell is installed, The computing device estimates the direction in which an error between the amount of solar radiation estimated for each direction and the amount of solar radiation described in the prediction data is minimum as the direction in which the solar cell is installed.

1. A power system comprising:

13. A power system that creates a power generation amount of a solar cell and an operation plan for a storage battery, a calculation device that predicts the amount of power generation using prediction data describing the results of predicting an amount of solar radiation based on meteorological information; the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; the computing device creates an operation plan instructing a total power output from the solar cell and the storage battery according to the predicted power generation amount; the computing device calculates a third predicted value of the temperature of the solar cell using an output current from the solar cell, an output voltage from the solar cell, and a specification value of the solar cell; The arithmetic device learns in advance a second correlation between the first predicted value and the third predicted value; The arithmetic device estimates the temperature by applying the amount of solar radiation estimated using the first correlation to the second correlation obtained as a result of the learning; and The calculation device predicts the power generation amount by using the temperature estimated using the second correlation.

1. A power system comprising:

14. A power system that creates a power generation amount of a solar cell and an operation plan for a storage battery, a calculation device that predicts the amount of power generation using prediction data describing the results of predicting an amount of solar radiation based on meteorological information; the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; the computing device creates an operation plan instructing a total power output from the solar cell and the storage battery according to the predicted power generation amount; the computing device acquires an operating current at an operating point of the solar cell and an operating voltage at the operating point; The arithmetic device converts the operating current and the operating voltage into a converted standard current and a converted standard voltage, which are values ​​assumed to be obtained when the solar cell is operating under a standard amount of solar radiation and a standard temperature; The computing device acquires specification data describing specification values ​​of the solar cell, The computing device estimates the amount of solar radiation on the solar cell by repeatedly performing an estimation process that calculates a virtual amount of solar radiation on the solar cell using the specification values ​​and calculates a virtual temperature of the solar cell using the specification values.

1. A power system comprising:

15. A power system that creates a power generation amount of a solar cell and an operation plan for a storage battery, a calculation device for predicting the amount of power generation using prediction data describing the results of predicting the amount of solar radiation based on meteorological information; a control device for controlling the operation of the solar cell and the operation of the storage battery; Equipped with the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; The control device controls the total power output from the solar cell and the storage battery according to the power generation amount predicted by the arithmetic device so as to conform to a pre-created operation plan; The calculation device estimates the amount of solar radiation for each orientation in which the solar cell is installed, The computing device estimates the direction in which an error between the amount of solar radiation estimated for each direction and the amount of solar radiation described in the prediction data is minimum as the direction in which the solar cell is installed.

1. A power system comprising:

16. A power system that creates a power generation amount of a solar cell and an operation plan for a storage battery, a calculation device for predicting the amount of power generation using prediction data describing the results of predicting the amount of solar radiation based on meteorological information; a control device for controlling the operation of the solar cell and the operation of the storage battery; Equipped with the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; The control device controls the total power output from the solar cell and the storage battery according to the power generation amount predicted by the arithmetic device so as to conform to a pre-created operation plan; the computing device calculates a third predicted value of the temperature of the solar cell using an output current from the solar cell, an output voltage from the solar cell, and a specification value of the solar cell; The arithmetic device learns in advance a second correlation between the first predicted value and the third predicted value; The arithmetic device estimates the temperature by applying the amount of solar radiation estimated using the first correlation to the second correlation obtained as a result of the learning; and The calculation device predicts the power generation amount by using the temperature estimated using the second correlation.

1. A power system comprising:

17. A power system that creates a power generation amount of a solar cell and an operation plan for a storage battery, a calculation device for predicting the amount of power generation using prediction data describing the results of predicting the amount of solar radiation based on meteorological information; a control device for controlling the operation of the solar cell and the operation of the storage battery; Equipped with the calculation device calculates a first predicted value of the amount of solar radiation using an output current from the solar cell and an output voltage from the solar cell; The computing device obtains a second predicted value of the amount of solar radiation described in the prediction data, The arithmetic device learns in advance a first correlation between the first predicted value and the second predicted value; The calculation device estimates the amount of solar radiation by applying a predicted value described in the prediction data to the first correlation obtained as a result of the learning, The calculation device predicts the power generation amount using the amount of solar radiation estimated using the first correlation; The control device controls the total power output from the solar cell and the storage battery according to the power generation amount predicted by the arithmetic device so as to conform to a pre-created operation plan; the computing device acquires an operating current at an operating point of the solar cell and an operating voltage at the operating point; The arithmetic device converts the operating current and the operating voltage into a converted standard current and a converted standard voltage, which are values ​​assumed to be obtained when the solar cell is operating under a standard amount of solar radiation and a standard temperature; The computing device acquires specification data describing specification values ​​of the solar cell, The computing device estimates the amount of solar radiation on the solar cell by repeatedly performing an estimation process that calculates a virtual amount of solar radiation on the solar cell using the specification values ​​and calculates a virtual temperature of the solar cell using the specification values.

1. A power system comprising:

18. When the power generation amount predicted by the arithmetic device exceeds a planned power generation amount determined by the operation plan, the control device suppresses an output of the solar cell or charges the storage battery using the output of the solar cell, thereby adjusting the total power to conform to the operation plan; When the power generation amount predicted by the arithmetic device is less than the planned power generation amount defined by the operation plan, the control device discharges the storage battery to make the total power conform to the operation plan.

18. The power system according to any one of claims 15 to 17.

Images