Information processing apparatus, method for information processing, and program

Abstract

To accurately calculate an estimated probability distribution of the power generation from a renewable energy power source.SOLUTION: The information processing apparatus according to an embodiment includes a processing unit. The processing unit calculates a weighted average based on a plurality of estimation values related to power generation of a renewable energy power source, calculates a variance based on pairs of the plurality of estimation values and error values related to the power generation of the renewable energy power source, calculates a predetermined mean error based on the plurality of error values, calculates a predetermined probability distribution parameter based on past actual values related to the power generation of the renewable energy power source, and determines a distribution estimation value, which is an estimated probability distribution related to the power generation of the renewable energy power source, based on the weighted average, the variance, the mean error, and the probability distribution parameter.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] An embodiment of the present invention relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] Renewable energy (hereinafter also referred to as "renewable energy") power sources such as solar power generation systems and wind power generation systems vary in power output depending on weather conditions, so power output must be estimated (forecasted) when planning supply. Furthermore, because power supply plans are created, for example, one day or three to six hours before actual supply, the estimated power output contains errors. Representative conventional methods for quantifying errors include, for example, statistical quantile regression and physical weather ensemble forecasting. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2015-167439 [Patent Document 2] International Publication No. 2016 / 013089 [Patent Document 3] Japanese Patent Application Publication No. 2022-183981 Summary of the Invention [Problem to be solved by the invention]

[0004] While statistical quantile regression methods have the advantage of being able to output estimated values ​​for power generation at each quantile (e.g., 25%, 50%, and 75% in the case of quartiles), they have the disadvantage of being unable to adjust the estimated values ​​for each quantile based on weather conditions. Furthermore, physical weather ensemble forecasting methods have the advantage of being able to adjust the estimated values ​​for each quantile based on weather conditions, but they have the disadvantages of the confidence intervals being too narrow when the number of members is small, and the amount of calculation and data required is large. In other words, there is room for improvement in the calculation of estimated values ​​for the probability distribution of power generation from renewable energy sources using conventional methods.

[0005] Therefore, the problem to be solved by the present invention is to calculate with high accuracy an estimated value of the probability distribution regarding the amount of power generated by a renewable energy power source. [Means for solving the problem]

[0006] According to an embodiment, an information processing device includes a processing unit that calculates a weighted average based on a plurality of estimated values ​​for power generation of a renewable energy power source, calculates a variance based on a plurality of pairs of the estimated values ​​for power generation of the renewable energy power source and an error value, calculates a predetermined error average based on the plurality of error values, calculates a predetermined probability distribution parameter based on past actual values ​​for power generation of the renewable energy power source, and determines a distribution estimate that is an estimate of a probability distribution for power generation of the renewable energy power source based on the weighted average, the variance, the error average, and the probability distribution parameter. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram showing the configuration of an information processing apparatus according to an embodiment. [Figure 2] 10 is a flowchart showing a process of calculating an integrated predicted value by an information processing device. [Figure 3] 10 is a flowchart showing a process of calculating a predicted quantile by an information processing device. [Figure 4] 10 is a flowchart showing a process of optimizing parameters performed by an information processing device. [Figure 5] 10 is a graph showing an example of calculation results regarding a solar power generation device. [Figure 6] 10 is a graph showing an example of calculation results for a wind power generating device. DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, embodiments of an information processing device, an information processing method, and a program according to the present invention will be described with reference to the accompanying drawings. The target power generation facility in this embodiment is a renewable energy power source such as a solar power generation device or a wind power generation device. The unit of time period for predicting (estimating) the amount of power generation is, for example, 30 minutes (1 frame). The target time period for predicting the amount of power generation is, for example, 24 hours (48 frames), but is not limited to this and may be other periods (for example, 12 hours (48 frames), 1 week, etc.).

[0009] 1 is a diagram showing the configuration of an information processing device 1 according to an embodiment. The information processing device 1 is a computer device, and includes an input unit 2, a display unit 3, a storage unit 4, a communication unit 5, and a processing unit 6. The units 2 to 6 are connected to each other via a bus so that they can communicate with each other.

[0010] The input unit 2 is an input device such as a mouse, a keyboard, a touch panel, etc. The input unit 2 receives information according to an operation by a user as an instruction signal, and outputs the instruction signal to the processing unit 6.

[0011] The display unit 3 is a display device such as an LCD (Liquid Crystal Display), etc. The display unit 3 displays various information based on a display signal from the processing unit 6.

[0012] The storage unit 4 is a semiconductor storage medium such as a flash memory, or a magnetically or optically recordable storage medium.

[0013] The communication unit 5 communicates with an external device via a network under the control of the processing unit 6.

[0014] The processing unit 6 is configured with, for example, a CPU (Central Processing Unit), RAM (Random Access Memory), and ROM (Read Only Memory). The CPU uses a predetermined area of ​​the RAM as a working area and executes various processes based on programs stored in the ROM, the storage unit 4, etc. The processing unit 6 includes, as functional units, a parameter adjustment unit 61, a prediction mean calculation unit 62, a prediction variance calculation unit 63, an error mean calculation unit 64, a prediction distribution determination unit 65, and a quantile calculation unit 66.

[0015] Note that some or all of the units 61 to 66 may be realized by hardware such as a dedicated integrated circuit (IC). In the following, processing other than that performed by the units 61 to 66 or processing performed by more than one of the units 61 to 66 may be described as being performed by the processing unit 6.

[0016] The processing unit 6 receives input of a predicted value (estimated value), an error value, and an actual value regarding the amount of power generated by the renewable energy power source from an external device or the like.

[0017] 1 simply refers to a "forecasted value," but the "forecasted value" here means multiple forecasted values ​​for the same future forecast period. The multiple forecasted values ​​may be, for example, forecasted values ​​calculated using different forecasting methods, or may be forecasted values ​​calculated using the same forecasting method under different initial conditions.

[0018] The error value indicates the error value for each predicted value. In other words, the predicted value and the error value are paired. The error value may be calculated, for example, from the predicted value and the actual value (for example, the actual value when the weather conditions, etc. are the same as when the predicted value was calculated), or may be calculated from an empirical formula.

[0019] The parameter adjustment unit 61 calculates predetermined probability distribution parameters based on the actual values. The probability distribution parameters are parameters that define probability distributions such as normal distribution, beta distribution, and gamma distribution. For example, parameters that define normal distribution are the mean and variance. In this case, for example, the mean may be a constant (a constant close to 1) multiplied by the weighted mean of the predicted values, and the variance may be a weighted sum of the harmonic mean of the errors (details will be described later) and the prediction variance (details will be described later). Note that multiplying the weighted mean of the predicted values ​​by a constant when calculating the mean occurs when there is a so-called bias error and the bias error is corrected.

[0020] Furthermore, the probability distribution parameters may be calculated by optimizing the actual values ​​of the probability distribution parameters, or may be calculated from an empirical formula.

[0021] The predicted average calculation unit 62 calculates a weighted average (hereinafter also referred to as a predicted average or an integrated predicted value) based on a plurality of predicted values. The weighting coefficient used when calculating the weighted average may be, for example, a value proportional to the inverse square of the corresponding error value. The weighted average may also be a simple average.

[0022] The prediction variance calculation unit 63 calculates the variance (prediction variance) based on a plurality of "pairs of predicted values ​​and error values."

[0023] The error average calculation unit 64 calculates a predetermined error average (for example, a harmonic mean of the errors) based on the plurality of error values.

[0024] The predicted distribution determination unit 65 determines a distribution predicted value, which is a predicted value of the probability distribution regarding the amount of power generated by the renewable energy power source, based on the weighted mean, variance (prediction variance), error mean, and probability distribution parameters.

[0025] Furthermore, a plurality of calculation methods may be stored for the process of calculating the variance by the prediction variance calculation unit 63, the process of calculating the error average by the error average calculation unit 64, and the process of calculating the probability distribution parameters by the parameter adjustment unit 61, and the prediction distribution determination unit 65 may determine the distribution prediction value by switching the calculation method to be used depending on the weighted average. Here, for example, when the renewable energy power source is a solar power generation device, research has shown that when the weather is neither sunny nor rainy but cloudy (i.e., the amount of solar radiation is medium), it may be more accurate to calculate the distribution prediction value without using the error value. Therefore, when the weighted average is a value corresponding to cloudy weather, it is sufficient to use one of the plurality of calculation methods that is suitable for cloudy weather.

[0026] Furthermore, the forecast distribution determination unit 65 may determine the distribution forecast value using quantile forecast values ​​related to the power generation amount of the renewable energy power source, instead of a plurality of "pairs of forecast value and error value."

[0027] The quantile calculation unit 66 calculates and determines the predicted value (estimated value) of the quantile specified in the probability distribution parameter (for example, 25%, 50%, 75% in the case of the quartiles) based on the distribution predicted value calculated by the predicted distribution determination unit 65. The processing unit 6 may also cause the display unit 3 to display the predicted value of the quantile and the weighted average.

[0028] Next, a description will be given of an example of calculation by the processing unit 6. Also, a description will be given of CRPS (Continuous Ranked Probability Score), an evaluation index of the calculation results. In the following, the predicted value may be referred to as predicted data, the error value as error data, the actual value as actual data, the weighted average as weighted average value, the error average as predicted error, and the parameters of the probability distribution as predictive distribution parameters.

[0029] The CRPS is defined as follows using the cumulative distribution function F(y) of the probabilistic forecast and the realized value x: x(y)≡θ(yx) can be considered as the cumulative distribution function for the delta function probability distribution δ(yx). In other words, CRPS is an index that measures the distance between cumulative distribution functions, and becomes zero when both cumulative distributions are the same.

number

[0030] As shown in the following equation (2), the CRPS has a zero prediction width, i.e., the cumulative distribution function is F μ If (y)=θ(y-μ), it is the absolute value of the error.

number

[0031] CRPS is a condition where the predicted probability distribution is a normal distribution N(μ,σ 2 ), it can be expressed by the Gaussian function φ(ω) and the error function Φ(ω) as shown in the following equation (3). However, the symbol ω≡(x-μ) / σ is used. This expression is expressed as the inverse σ in the curly brackets {…} when the prediction width σ approaches zero. -1 These numerical limiting processes are difficult to handle in numerical calculations, so when the prediction range is small, the calculation is switched to the theoretical limiting value.

number

[0032] CRPS is an index that measures whether the width of the probability distribution of a prediction is set appropriately. The smaller the CRPS, the better.

[0033] (Adjustment of the predictive distribution)

number

[0034] The above formulas (4) to (6) will now be explained. A predicted value of renewable energy output (power output from renewable energy sources) is generally uniquely identified by giving a triplet of the target time t of the prediction, type p, and version number v. For example, the prediction type p is identification information that identifies the underlying weather forecast, and the version number v is identification information that identifies the age of the update time.

[0035] Weather forecasts have a forecast length (how many hours into the future the forecast is) and an update interval (how often it is updated), and since the former is longer than the latter, multiple forecast values ​​with different update times will overlap at the target time t of the forecast, so they are distinguished by a version number v.

[0036] The integrated prediction value Y(t,f) is subscripted with the target time t and the version number f. The version number f can be interpreted as the age of the update time when the target time t is used as the base, and as the length of the prediction when the update time is used as the base. The set N of predictions to be input to the integrated prediction value f Since the length of the integrated prediction, f, increases (as the overlap of predictions decreases), the subscript f is used. Note that the integrated prediction value here is the same as the weighted average with the bias error (μ) set to zero.

[0037]

number

[0038]

number

[0039] In this embodiment, the probability prediction algorithm first assumes that the probability distribution parameter of the predicted value is the weighted sum of the prediction error and the prediction distribution. The weighting coefficients of this weighted sum are determined by minimizing the CRPS. Note that the weighting coefficients depend on the input prediction data set (the combination of the prediction method and the predicted time), so they are determined for each predicted time.

[0040] In addition, when switching between the above-mentioned multiple calculation methods, for example, multiple relational expressions with different β1 and β2 may be prepared as the relational expression of the above equation (8), and different relational expressions may be used depending on the weighted average.

[0041] Next, a process of calculating an integrated predicted value by the information processing device 1 will be described with reference to Fig. 2. Fig. 2 is a flowchart showing a process of calculating an integrated predicted value by the information processing device 1.

[0042] First, in step S21, the processing unit 6 receives information on a plurality of "pairs of predicted values ​​and error values" from an external device or the like.

[0043] Next, in step S22, the prediction average calculation unit 62 calculates weighting coefficients (weighting coefficients for each prediction value) used when calculating the weighted average.

[0044] Next, in step S23, the prediction average calculation unit 62 calculates a weighted average of the prediction values ​​using the weighting coefficients calculated in step S22.

[0045] Next, in step S24, the prediction average calculation unit 62 outputs the weighted average calculated in step S23 as an integrated prediction value.

[0046] Next, a process of calculating a predicted quantile point by the information processing device 1 will be described with reference to Fig. 3. Fig. 3 is a flowchart showing a process of calculating a predicted quantile point by the information processing device 1.

[0047] First, in step S31, the processing unit 6 receives input of information on a plurality of "pairs of predicted values ​​and error values" from an external device or the like.

[0048] Next, in step S32, the prediction average calculation unit 62 calculates a prediction average based on a plurality of prediction values.

[0049] Next, in step S33, the error average calculation unit 64 calculates the error average (harmonic mean of the errors) based on the plurality of error values.

[0050] Next, in step S34, the prediction variance calculation unit 63 calculates the prediction variance based on a plurality of "pairs of predicted values ​​and error values."

[0051] Next, in step S35, the predictive distribution determination unit 65 determines a distribution prediction value (predictive distribution), which is a predicted value of the probability distribution regarding the power generation amount of the renewable energy power source, based on the weighted average, predictive variance, error mean, and probability distribution parameters (calculated using the flowchart in Figure 4).

[0052] Next, in step S36, the quantile calculation unit 66 calculates and determines the predicted quantile, which is the predicted value of the quantile specified in the probability distribution parameter (for example, 25%, 50%, 75% in the case of the quartiles), based on the distribution predicted value calculated in step S35, and outputs it.

[0053] Next, a process of optimizing parameters by the information processing device 1 will be described with reference to Fig. 4. Fig. 4 is a flowchart showing a process of optimizing parameters by the information processing device 1.

[0054] First, in step S41, the parameter adjusting unit 61 inputs a performance value from an external device or the like.

[0055] Next, in steps S42 and S43, the parameter-adjusting unit 61 selects parameters (probability distribution parameters) based on the performance values ​​input in step S41 and the calculation results in step S47.

[0056] Next, in step S44, the parameter adjustment unit 61 inputs probability prediction data such as an integrated predicted value used in calculating the objective function (the above formula (9)).

[0057] Next, in step S45, the parameter adjusting unit 61 calculates the objective function (the above formula (9)).

[0058] Next, in step S46, the parameter-adjusting unit 61 determines whether or not the CRPS is minimized in the objective function (the above equation (9)). If Yes, the process proceeds to step S47, and if No, the process returns to step S46.

[0059] In step S47, the parameter adjustment unit 61 outputs parameters (probability distribution parameters) (for example, the mean and variance that define the normal distribution).

[0060] Next, an example of calculation results will be described with reference to Fig. 5 and Fig. 6. Fig. 5 is a graph showing an example of calculation results related to a solar power generation system. In (a) ((a1) to (a3)), the vertical axis represents the amount of power generated, and the horizontal axis represents time. In (b), the vertical axis represents CRPS per amount of power generated, and the horizontal axis represents time. In the following, the static width method is a conventional method that uses only prediction error. The dynamic width method is a conventional method that uses only prediction variance. The mixed width method is a method of this embodiment that uses both prediction error and prediction variance.

[0061] In (a1) to (a3), the graph GR shows actual values. Graph G1_P25 in (a1) is the predicted value of the 25th quantile of the static width method. Graph G1_P75 in (a1) is the predicted value of the 75th percentile of the static width method.

[0062] Graph G2_P25 in (a2) is the predicted value of the 25th quantile of the dynamic width method. Graph G2_P75 in (a2) is the predicted value of the 75th percentile of the dynamic width method.

[0063] Graph G3_P25 in (a3) ​​is the predicted value of the 25% quantile of the mixed width method. Graph G3_P75 in (a3) ​​is the predicted value of the 75% quantile of the mixed width method.

[0064] The width of the static width method (a1) (the width between the predicted value of the 25th quantile and the predicted value of the 75th quantile; the same applies below) is almost the same on days 1 to 3. The width of the dynamic width method (a2) decreases from day 1 to day 3. And the width of the mixed width method (a3) ​​is intermediate between the two.

[0065] Also, in (b), graph G1 is a CRPS of the static width method. Graph G2 is a dynamic width method CRPS. Graph G3 is a mixed-width CRPS. The CRPS (graph G3) of the mixed width method is generally the smallest, and it is clear that the method of this embodiment is effective.

[0066] Figure 6 is a graph showing an example of calculation results for a wind power generation device. In (a) ((a1) to (a3)), the vertical axis is the amount of power generated, and the horizontal axis is time. In (b), the vertical axis is CRPS per amount of power generated, and the horizontal axis is time. The meanings of the static width method, dynamic width method, and mixed width method are the same as in Figure 5.

[0067] In (a1) to (a3), the graph GR shows actual values. Graph G1_P25 in (a1) is the predicted value of the 25th quantile of the static width method. Graph G1_P75 in (a1) is the predicted value of the 75th percentile of the static width method.

[0068] Graph G2_P25 in (a2) is the predicted value of the 25th quantile of the dynamic width method. Graph G2_P75 in (a2) is the predicted value of the 75th percentile of the dynamic width method.

[0069] Graph G3_P25 in (a3) ​​is the predicted value of the 25% quantile of the mixed width method. Graph G3_P75 in (a3) ​​is the predicted value of the 75% quantile of the mixed width method.

[0070] The width of the static width method (a1) does not change much over time. The width of the dynamic width method (a2) changes significantly over time. And the width of the mixed width method (a3) ​​is somewhere between the two.

[0071] Also, in (b), graph G1 is a CRPS of the static width method. Graph G2 is a dynamic width method CRPS. Graph G3 is a mixed-width CRPS. The CRPS (graph G3) of the mixed width method is generally the smallest, and it is clear that the method of this embodiment is effective.

[0072] As described above, according to the information processing device 1 of this embodiment, it is possible to calculate with high accuracy a distribution estimate (distribution forecast value), which is an estimate of the probability distribution regarding the amount of power generated by a renewable energy power source, based on the weighted mean, variance, error mean, and probability distribution parameters.

[0073] In other words, it is possible to output a probabilistic forecast taking into account both statistical and physical forecast ranges from a relatively small amount of input data. Specifically, this overcomes the drawback of conventional techniques, in that the statistical forecast range is not dynamically adjusted according to weather conditions, and has the effect of appropriately quantifying the uncertainty in power generation forecasts.

[0074] As shown in Figures 5 and 6 above, using the probability prediction evaluation index CRPS, it has been confirmed that the method of this embodiment (mixed width method) is superior to the conventional methods, the static width method and the dynamic width method.

[0075] Furthermore, since it is possible to obtain not only distribution forecast values ​​but also quantile forecast values ​​as forecast results, it is possible to directly enjoy the benefits of using a statistical quantile regression method.

[0076] Furthermore, by displaying the quantile predicted values ​​and the weighted averages (integrated predicted values) in those cases on the display unit 3, the user can recognize and respond to them.

[0077] Furthermore, multiple calculation methods are stored for the process of calculating the variance, the process of calculating the error mean, and the process of calculating the probability distribution parameters, and the calculation method to be used can be switched depending on the weighted average to determine the distribution forecast value and the quantile forecast value. This allows the appropriate method to be used depending on the value of the weighted average.

[0078] Furthermore, instead of multiple "pairs of forecast values ​​and error values," it is possible to determine distribution forecast values ​​and quantile forecast values ​​using quantile estimates (quantile forecast values) for the amount of power generated by renewable energy power sources. As a result, even if multiple "pairs of forecast values ​​and error values" are not available, it is possible to determine highly accurate distribution forecast values ​​and quantile forecast values ​​as long as a quantile estimate for the amount of power generated by renewable energy power sources is available.

[0079] The program to be executed by the computer is provided as a file in a format that can be installed on the computer or in a format that can be executed by the computer, and is recorded on a computer-readable recording medium such as a CD-ROM, a flexible disk, a CD-R, or a DVD (Digital Versatile Disk).

[0080] This program may also be configured to be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. This program may also be configured to be provided or distributed via a network such as the Internet. The program executed by the information processing device 1 may also be configured to be provided by being pre-installed in a ROM or the like.

[0081] Although an embodiment of the present invention has been described, this embodiment is presented as an example and is not intended to limit the scope of the invention. This novel embodiment can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. This embodiment and its modifications are included within the scope and spirit of the invention, and are also included in the invention described in the claims and their equivalents.

[0082] For example, the information processing device 1 may be realized by a server device or a cloud computer on a network.

[0083] (Addendum) A configuration example of the embodiment will be described below.

[0084] (Configuration example 1) Calculating a weighted average based on multiple estimates of renewable energy generation capacity; calculating a variance based on a plurality of pairs of the estimated values ​​and error values ​​related to the amounts of power generated by the renewable energy power sources; calculating a predetermined error average based on a plurality of said error values; calculating predetermined probability distribution parameters based on past performance values ​​regarding the amount of power generated by the renewable energy power source; a processing unit that determines a distribution estimate that is an estimate of a probability distribution regarding the amount of power generation of the renewable energy power source based on the weighted average, the variance, the error average, and the probability distribution parameters, Information processing device.

[0085] (Configuration example 2) When determining the distribution estimate based on the weighted average, the variance, the error average, and the probability distribution parameters, the processing unit calculates and determines an estimate of a quantile specified in the probability distribution parameters. The information processing device according to configuration example 1.

[0086] (Configuration example 3) the processing unit causes a display unit to display the estimated quantile value and the weighted average. The information processing device according to configuration example 2.

[0087] (Configuration example 4) The processing unit a plurality of calculation methods are stored for the process of calculating the variance, the process of calculating the error mean, and the process of calculating the probability distribution parameters; determining the distribution estimate by switching the calculation method to be used depending on the weighted average; The information processing device according to any one of configuration examples 1 to 3.

[0088] (Configuration Example 5) the processing unit determines the distribution estimate using a quantile estimate for the amount of power generation of the renewable energy power source instead of a plurality of pairs of the estimate value and the error value. The information processing device according to configuration example 2 or 3.

[0089] (Configuration Example 6) When calculating the predetermined error average based on the plurality of error values, the processing unit calculates a harmonic mean. 6. The information processing device according to any one of configuration examples 1 to 5.

[0090] (Configuration Example 7) An information processing method executed by a processing unit in an information processing device, calculating a weighted average based on a plurality of estimates of power generation from renewable energy sources; calculating a variance based on a plurality of pairs of the estimated values ​​and error values ​​related to the power generation amounts of the renewable energy power sources; calculating a predetermined error average based on a plurality of said error values; calculating predetermined probability distribution parameters based on past performance values ​​of the power generation amount of the renewable energy power source; determining a distribution estimate that is an estimate of a probability distribution regarding the amount of power generated by the renewable energy power source based on the weighted mean, the variance, the error mean, and the probability distribution parameters; An information processing method including:

[0091] (Configuration Example 8) Computer, Calculating a weighted average based on multiple estimates of renewable energy generation capacity; calculating a variance based on a plurality of pairs of the estimated values ​​and error values ​​related to the amounts of power generated by the renewable energy power sources; calculating a predetermined error average based on a plurality of said error values; calculating predetermined probability distribution parameters based on past performance values ​​regarding the amount of power generated by the renewable energy power source; A program for causing the program to function as a processing unit that determines a distribution estimate, which is an estimate of the probability distribution regarding the power generation amount of the renewable energy power source, based on the weighted average, the variance, the error average, and the probability distribution parameters. [Explanation of symbols]

[0092] 1...information processing device, 2...input unit, 3...display unit, 4...storage unit, 5...communication unit, 6...processing unit, 61...parameter adjustment unit, 62...prediction mean calculation unit, 63...prediction variance calculation unit, 64...error mean calculation unit, 65...prediction distribution determination unit, 66...quantile calculation unit

Claims

1. Calculating a weighted average based on multiple estimates of renewable energy generation capacity; calculating a variance based on a plurality of pairs of the estimated values ​​and error values ​​related to the amounts of power generated by the renewable energy power sources; calculating a predetermined error average based on a plurality of said error values; calculating predetermined probability distribution parameters based on past performance values ​​regarding the amount of power generated by the renewable energy power source; a processing unit that determines a distribution estimate that is an estimate of a probability distribution regarding the amount of power generation of the renewable energy power source based on the weighted average, the variance, the error average, and the probability distribution parameters, Information processing device.

2. When determining the distribution estimate based on the weighted average, the variance, the error average, and the probability distribution parameters, the processing unit calculates and determines an estimate of a quantile specified in the probability distribution parameters. The information processing device according to claim 1 .

3. the processing unit causes a display unit to display the estimated quantile value and the weighted average. The information processing device according to claim 2 .

4. The processing unit a plurality of calculation methods are stored for the process of calculating the variance, the process of calculating the error mean, and the process of calculating the probability distribution parameters; determining the distribution estimate by switching the calculation method to be used depending on the weighted average; The information processing device according to claim 1 .

5. the processing unit determines the distribution estimate using a quantile estimate for the amount of power generation of the renewable energy power source instead of a plurality of pairs of the estimate value and the error value. The information processing device according to claim 2 .

6. When calculating the predetermined error average based on the plurality of error values, the processing unit calculates a harmonic mean. The information processing device according to claim 1 .

7. An information processing method executed by a processing unit in an information processing device, calculating a weighted average based on a plurality of estimates of power generation from renewable energy sources; calculating a variance based on a plurality of pairs of the estimated values ​​and error values ​​related to the power generation amounts of the renewable energy power sources; calculating a predetermined error average based on a plurality of said error values; calculating predetermined probability distribution parameters based on past performance values ​​of the power generation amount of the renewable energy power source; determining a distribution estimate that is an estimate of a probability distribution regarding the amount of power generated by the renewable energy power source based on the weighted mean, the variance, the error mean, and the probability distribution parameters; An information processing method including:

8. Computer, Calculating a weighted average based on multiple estimates of renewable energy generation capacity; calculating a variance based on a plurality of pairs of the estimated values ​​and error values ​​related to the amounts of power generated by the renewable energy power sources; calculating a predetermined error average based on a plurality of said error values; calculating predetermined probability distribution parameters based on past performance values ​​regarding the amount of power generated by the renewable energy power source; A program for causing the program to function as a processing unit that determines a distribution estimate, which is an estimate of the probability distribution regarding the power generation amount of the renewable energy power source, based on the weighted average, the variance, the error average, and the probability distribution parameters.

Images