Information processing device, information processing program, and information processing method

Abstract

To provide an information processor that can accurately predict an inflow amount to a dam as a prediction target.SOLUTION: The present invention includes: an acquisition unit for acquiring the amount of inflow to a dam as a prediction target in a predetermined period of time and the precipitation amount in a predetermined point in the predetermined period of time; and an output unit for outputting the inflow amount in a first period of time of the predetermined period of time and outputting a correction inflow amount with suppressed noise in a second period of time, which is the period of time other than the first period of time of the predetermined time. The precipitation amount in the predetermined point in the first period of time is larger than that in the second period of time.SELECTED DRAWING: Figure 13

Description

[Technical field]

[0001] The present invention relates to an information processing device, an information processing program, and an information processing method. [Background technology]

[0002] There is a known method for predicting the amount of water flowing into a dam from an upstream area of ​​the dam per unit time (inflow into the dam) using a prediction model trained by machine learning. For example, Patent Document 1 discloses a learning method of a neural network for predicting the inflow into a dam. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 5234085 Summary of the Invention [Problem to be solved by the invention]

[0004] By the way, it is necessary to obtain the past inflow amount to the dam as learning data for training the prediction model, but this cannot be measured directly. Therefore, the inflow amount to the dam is obtained by calculation taking into account the measurement value by the water level gauge installed in the dam and the opening degree of the outlet.

[0005] However, the lower the amount of precipitation, especially in the upstream area of ​​the dam, the more likely the water level of the dam is to fluctuate due to the effects of wind, and the more likely the water level readings on the water level gauge are to fluctuate accordingly. In such cases, the amount of inflow into the dam will include the effects of wind.

[0006] The effect of wind is an example of unnecessary information (noise) for predicting the inflow to a dam. If the inflow to a dam contains noise, the accuracy of the obtained prediction model will decrease due to the effect of the noise. However, the method disclosed in Patent Document 1 does not take into account the decrease in accuracy due to noise.

[0007] An object of the present invention is to provide an information processing device capable of predicting with high accuracy the inflow volume into a dam that is a prediction target. [Means for solving the problem]

[0008] One invention for achieving the above object is an information processing device including an acquisition unit that acquires an inflow to a dam to be predicted in a predetermined period and a precipitation amount at a predetermined point in the predetermined period, and an output unit that outputs the inflow amount in a first period of the predetermined period and outputs a corrected inflow amount in which noise is suppressed in a second period of the predetermined period excluding the first period, wherein the precipitation amount at the predetermined point in the first period is greater than the precipitation amount at the predetermined point in the second period. Other features of the present invention will become apparent from the description in this specification. Effect of the Invention

[0009] According to the present invention, it is possible to provide an information processing device capable of predicting with high accuracy the inflow amount into a dam that is the prediction target. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an information processing system. [Diagram 2] FIG. 2 illustrates an example of a hardware configuration of an information processing device. [Diagram 3] FIG. 2 is a diagram illustrating an example of functional blocks realized in the information processing device. [Figure 4] FIG. 11 is a diagram showing an example of inflow volume and precipitation volume. [Diagram 5] FIG. 13 is a diagram illustrating an example of a reference value. [Figure 6] FIG. 13 is a diagram illustrating an example of a fluctuation range. [Figure 7] FIG. 11 is a diagram illustrating an example of a method for generating a fluctuation range. [Figure 8] FIG. 11 is a diagram illustrating an example of a method for generating a fluctuation range. [Figure 9]FIG. 11 is a diagram illustrating an example of a method for generating a fluctuation range. [Figure 10] FIG. 11 is a diagram illustrating an example of a method for generating a fluctuation range. [Figure 11] FIG. 13 is a diagram showing a fluctuation range centered on a reference value. [Figure 12] 11 is a diagram for explaining a determination as to whether or not an inflow amount is within a fluctuation range. FIG. [Figure 13] FIG. 13 is a diagram showing an example of an inflow amount in which noise is suppressed. [Figure 14] FIG. 1 is a diagram illustrating an example of a method for generating a prediction model. [Figure 15] FIG. 1 is a diagram illustrating an example of a method for generating a prediction model. [Figure 16] 11 is a flowchart illustrating a process flow up to generating a prediction model. [Figure 17] FIG. 13 is a diagram showing another example of the fluctuation range. [Figure 18] 13A and 13B are diagrams illustrating another example of a method for generating a fluctuation range. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] == Implementation form == <<Information Processing System>> Fig. 1 is a diagram showing an overview of an information processing system 1 according to this embodiment. The information processing system 1 is a system for generating a prediction model for predicting the amount of water inflow into a prediction target dam (described later). The information processing system 1 includes a meteorological observation device 4 and an information processing device 5. Fig. 1 further shows a river including an upstream dam 2, a downstream dam 3, a main river R1, and a tributary R2 that merges with the main river R1. First, the upstream dam 2 and the downstream dam 3 will be described below.

[0012] <Upstream dam> In this example, the upstream dam 2 is provided upstream of the confluence of the main river R1 and the tributary R2. The upstream dam 2 is provided with a water level gauge for measuring the water level of the stored water, a gate for discharging the stored water, and an opening gauge for measuring the opening of the gate (not shown).

[0013] <Downstream dam> The downstream dam 3 is a dam to be predicted in the information processing system 1, and a prediction model for predicting the amount of water inflow into the downstream dam 3 is generated by the information processing system 1. In other words, the downstream dam 3 corresponds to the "dam to be predicted."

[0014] The downstream dam 3 is located downstream of the confluence of the main river R1 and the tributary R2. Water flowing down the main river R1 or the tributary R2 flows into the downstream dam 3. Here, the water flowing down the main river R1 includes water caused by meteorological factors such as precipitation, as well as water released from the upstream dam 2.

[0015] The downstream dam 3, like the upstream dam 2, is provided with a water level gauge, a gate, and an opening gauge (not shown).

[0016] <Weather Observation Equipment> The meteorological observation device 4 is a device for observing the weather at a point that affects the inflow amount to the downstream dam 3. The meteorological observation device 4 is provided at a predetermined point (hereinafter referred to as the "observation point") upstream of the downstream dam 3. The meteorological observation device 4 of this embodiment includes at least a rain gauge and can measure the amount of precipitation per unit time at the observation point. Note that multiple meteorological observation devices 4 may be provided. In this case, a meteorological observation device 4 is provided at each of the multiple observation points.

[0017] <Information processing device> The information processing device 5 is a device for generating a prediction model for predicting the amount of water inflow into the downstream dam 3 based on various actual value data acquired from the upstream dam 2, the downstream dam 3, and the meteorological observation device 4. Below, the hardware configuration of the information processing device 5 and the realized functional blocks will be described in that order.

[0018] Hardware Configuration 2 is a diagram showing an example of a hardware configuration of an information processing device 5. The illustrated information processing device 5 includes a processor 500, a main storage device 501, an auxiliary storage device 502, an input device 503, an output device 504, and a communication device 505. Note that the illustrated information processing device 5 may be realized, in whole or in part, by using virtual information processing resources provided by using a virtualization technique, a process space separation technique, or the like, such as a virtual server provided by a cloud system. In addition, all or in part of the functions provided by the information processing device 5 may be realized by a service provided by a cloud system via an API (Application Programming Interface), for example.

[0019] The processor 500 is configured using, for example, a CPU (Central Processing Unit) and the like.

[0020] The main memory device 501 is a device for storing programs and data, and is, for example, a read only memory (ROM) or a random access memory (RAM).

[0021] The auxiliary storage device 502 is, for example, a hard disk drive, a storage area of ​​a cloud server, etc. Programs and data can be read into the auxiliary storage device 502 via a recording medium reading device or a communication device 505. The programs and data stored (memorized) in the auxiliary storage device 502 are read into the main storage device 501 as needed.

[0022] The input device 503 is an interface that accepts input from the outside, and is, for example, a keyboard, a mouse, a touch panel, or the like.

[0023] The output device 504 is an interface that outputs various information such as the process progress, the process results, etc. The output device 504 is, for example, a display device (such as a liquid crystal monitor) that visualizes the various information described above, a device that converts the various information described above into voice (a voice output device (such as a speaker)), or a device that converts the various information described above into text (such as a printer).

[0024] The input device 503 and the output device 504 constitute a user interface that receives information from the user and presents information.

[0025] The communication device 505 is a device that realizes communication with other devices. The communication device 505 is a wired or wireless communication interface that realizes communication with other devices via a communication network NW such as the Internet.

[0026] The information processing device 5 may have installed therein, for example, an operating system, a file system, and the like.

[0027] Each function of the information processing system 1 is realized by a processor 500 of the information processing device 5 reading and executing a program stored in a main storage device 501, or by hardware constituting the information processing system 1. The information processing system 1 stores various types of information (data), for example, as a table of a database or a file managed by a file system.

[0028] Function blocks 3 is a diagram showing an example of functional blocks realized in the information processing device 5. When the processor 500 of the information processing device 5 executes a predetermined program, a calculation unit 510, an acquisition unit 511, an output unit 512, and a model generation unit 513 are realized in the information processing device 5. Each of these will be described below.

[0029] [Calculation section] The calculation unit 510 calculates the inflow (F1(t)) into the downstream dam 3, where t is time and F1(t) is the calculated value of the inflow into the downstream dam 3 at time t. The inflow into the downstream dam 3 includes the contributions of the inflow from the main river R1 and the inflow from the tributary R2. The inflow into the downstream dam 3 is calculated based on the output from a measuring instrument such as a water level gauge that measures the water level of the downstream dam 3.

[0030] In this embodiment, an upstream dam 2 is provided. Therefore, the calculation unit 510 calculates the inflow amount (F1(t)) to the downstream dam 3 by subtracting contributions from artificial controls, such as the discharge amount (Fa(t)) from the upstream dam 2 and the amount of water used for power generation at the upstream dam 2 (Fb(t)), from the total inflow amount (Ft(t)) to the downstream dam 3.

[0031] The contribution of the discharge from the upstream dam 2 is calculated based on the output from measuring instruments such as an opening meter of the upstream dam 2 and the flow time (Δt) from the upstream dam 2 to the downstream dam 3.

[0032] In other words, the inflow (F1(t)) to the downstream dam 3 at time t is expressed by the following equation. F1(t)=Ft(t)-(Fa(t-Δt)+Fb(t-Δt))

[0033] [Acquisition Department] The acquisition unit 511 acquires the inflow amount to the dam to be predicted during a predetermined observation period (corresponding to a predetermined period) and the precipitation amount at the observation point (corresponding to a predetermined point) during the observation period. The acquisition unit 511 acquires the inflow amount from the calculation unit 510 and the precipitation amount from the meteorological observation device 4.

[0034] 4 is a diagram showing an example of the inflow amount F1 and the precipitation amount W acquired by the acquisition unit 511. The observation period in this example is 29 hours from 2:00 AM on the observation start date to 7:00 AM on the following day. In addition, in this example, data on the inflow amount F1 and the precipitation amount W is acquired every hour.

[0035] Here, the smaller the amount of precipitation W in the upstream area of ​​the downstream dam 3, the more likely the inflow F1 is to contain noise. Here, "noise" refers to information that is not necessary for predicting the inflow to the dam. One of the factors that causes the inflow F1 to contain noise is the influence of wind near the water surface of the downstream dam 3.

[0036] The smaller the amount of precipitation W in the upstream area of ​​the downstream dam 3, the more likely the dam's water level will fluctuate due to the influence of wind, and the more likely the water level gauge measurement value will fluctuate accordingly. In Figure 4, the smaller the amount of precipitation W, the more vibrations with relatively low amplitude and relatively high frequency appear in the inflow F1. Such vibrations are presumed to be noise.

[0037] If such noise is included in the inflow amount F1, the inflow amount F1 will deviate from the actual inflow amount to the downstream dam 3, and the accuracy will decrease. Therefore, it is preferable to output a value in which noise has been suppressed from the inflow amount F1.

[0038] [Output section] The output unit 512 outputs the corrected inflow amount (corresponding to the corrected inflow amount) (details will be described later). Note that in order to clearly distinguish between the inflow amount F1 calculated by the calculation unit 510 and acquired by the acquisition unit 511 and the corrected inflow amount, the inflow amount F1 may be referred to as the "inflow amount F1 before correction." Also, when simply referred to as the "inflow amount F1," it means the "inflow amount F1 before correction."

[0039] As will be described in detail later, the corrected inflow volume will be the inflow volume before correction during a portion of the observation period (corresponding to the first period), and will be a value with noise suppressed during the period excluding the portion of the observation period during which the inflow volume before correction is obtained (corresponding to the second period).

[0040] Here, the amount of precipitation W at the observation point in the above-mentioned part of the period is greater than the amount of precipitation W at the observation point in the period excluding the part of the period.

[0041] Hereinafter, a detailed description will be given of the output unit 512. The output unit 512 includes a generation unit 512a, a setting unit 512b, a determination unit 512c, and a correction value output unit 512d.

[0042] <Generation unit 512a> The generating unit 512a generates a reference value in which noise is suppressed for the inflow amount F1 into the downstream dam 3. At this time, the generating unit 512a generates a reference value in which noise is suppressed over the entire period of the inflow amount F1.

[0043] 5 is a diagram showing an example of the reference value S generated by the generating unit 512a. In FIG 5, the reference value S (broken line) is further shown in addition to FIG 4.

[0044] In this embodiment, the generating unit 512a performs a moving average process on the inflow F1 to generate the reference value S. Specifically, the reference value S is set to the average value of the inflow F1 for the previous two hours at each time. The reference value S shown in FIG. 5 is a value obtained by smoothing the above-mentioned vibrations of relatively low amplitude and relatively high frequency that appeared in the inflow F1.

[0045] The method for generating the reference value S is not limited to this. The period for which the moving average process is performed and the number of data items for which the moving average process is performed may be set appropriately. In addition, instead of simply taking the average value during the period for which the moving average process is performed, a weighted average value using a predetermined weight may be used.

[0046] <Setting section 512b> The setting unit 512b sets the fluctuation range based on the inflow F1 and the precipitation W. Here, the "fluctuation range" is an index for determining the range within which the absolute value of the difference between the inflow F1 and the reference value S (hereinafter referred to as "variation") is within which the inflow F1 is estimated to include noise. The fluctuation range is a function that depends on the precipitation W, and is set to decrease as the precipitation W increases.

[0047] 6 is a diagram showing an example of the fluctuation range V1 set by the setting unit 512b. In this example, the fluctuation range V1 is a linear function that decreases as the amount of precipitation W increases in a region where the amount of precipitation W ranges from 0 (zero) to a predetermined threshold value T1, and is 0 in a region where the amount of precipitation W is greater than the predetermined threshold value T1. Here, the fluctuation range V1 is at a maximum value M1 when the amount of precipitation W is 0, and is 0 when the amount of precipitation W is the threshold value T1.

[0048] In other words, setting the fluctuation band V1 in this embodiment corresponds to setting the maximum value M1 and the threshold value T1 for the amount of precipitation W.

[0049] As described above, the inflow F1 is likely to be affected by noise, particularly when the amount of precipitation W is small. Therefore, it is preferable to set the maximum value M1 so that it is easy to determine that the inflow F1 contains noise when the amount of precipitation W is smaller than a predetermined value W1 preset by the user, for example.

[0050] On the other hand, as can be seen from Fig. 5, there is variation even during periods when the amount of precipitation W is relatively large, such as from 8:00 to 15:00. However, it is estimated that such variation is not caused by noise. Therefore, it is preferable to set the threshold value T1 of the amount of precipitation W as a threshold value between the amount of precipitation W when the inflow amount F1 is estimated to include noise and the amount of precipitation W when the inflow amount F1 is estimated to include no noise.

[0051] 7 to 10, an example of a method for setting the maximum value M1 of the fluctuation range V1 and the threshold value T1 of the amount of precipitation W will be described. First, a method for setting the maximum value M1 of the fluctuation range V1 will be described with reference to FIGS.

[0052] FIG. 7 is a diagram showing a set of data on the amount of precipitation W and the amount of inflow F1 at each time shown in FIG. 4, expressed in a two-dimensional space with the amount of precipitation W on the horizontal axis and the amount of inflow F1 on the vertical axis.

[0053] In Fig. 7, a boundary with a predetermined precipitation amount W1 as a boundary value is indicated by a dashed line. The precipitation amount W1 can be arbitrarily set by a user of the information processing system 1. As will be described later, in setting the maximum value M1, data with a precipitation amount smaller than W1 is used from among the data pairs of the precipitation amount W and the inflow amount F1 at each time shown in Fig. 4. Therefore, W1 should be set so that data that is relatively free of noise, that is, data with a relatively large precipitation amount W, is excluded.

[0054] Figure 8 shows a histogram of the variance of data sets of precipitation W and inflow F1 in Figure 7, where the precipitation W is smaller than W1. Figure 8 also shows the standard deviation σ1 of the variance.

[0055] The maximum value M1 can be set based on the standard deviation σ1. For example, the maximum value M1 can be set to any one of σ1, 2×σ1, 3×σ1, etc. When the maximum value M1 is set to σ1, it is estimated that 68.3% of the data, particularly those in which the amount of precipitation W is smaller than W1, contains noise.

[0056] As another example, the maximum value M1 may be set to a value that includes all the variations shown in Fig. 8. In other words, the maximum value M1 may be set to the greater of the difference between the maximum and average values ​​of the variations, or the difference between the minimum and average values ​​of the variations. In this case, all data, particularly those with a precipitation amount W smaller than W1, are estimated to include noise.

[0057] Next, a method for setting the threshold T1 for the amount of precipitation W will be described with reference to FIGS.

[0058] Fig. 9 shows the same data as in Fig. 7. In Fig. 9, a predetermined inflow F3 is set as a boundary value, and the boundary is indicated by a dashed line. In this example, F3 is the sum of the average value A of the reference value S during the period when the amount of precipitation W is smaller than W1 and the maximum value M1 of the fluctuation range V1.

[0059] Figure 10 shows a histogram of precipitation W, extracted from the data pairs of precipitation W and inflow F1 in Figure 9, where inflow F1 is smaller than F3. Figure 10 also shows the standard deviation equivalent σ2 for precipitation W (the range of precipitation W included in the 68.3% of the total data in Figure 9 on the side with the smallest precipitation W).

[0060] The threshold T1 of the precipitation amount W can be set based on the standard deviation equivalent σ2. For example, the threshold T1 of the precipitation amount W can be set to σ2, 2×σ2, 3×σ2, and the like.

[0061] The determination unit 512c determines whether the inflow amount F1 is within a fluctuation range V1 with the reference value S as the center.

[0062] Fig. 11 is a diagram showing a fluctuation range V1 centered on a reference value S. Fig. 11 shows an upper limit U (dash-dotted line) and a lower limit L (dash-two-dot line). The upper limit U is a value obtained by adding a fluctuation range V1 corresponding to the amount of precipitation W to the reference value S (dashed line). The lower limit L is a value obtained by subtracting the fluctuation range V1 corresponding to the amount of precipitation W from the reference value S.

[0063] 11, in periods when the amount of precipitation W is relatively high, the region between the upper limit value U and the lower limit value L (the range of the fluctuation range V1) is relatively narrow, or is 0. On the other hand, in periods when the amount of precipitation W is relatively low, the range of the fluctuation range V1 is relatively wide.

[0064] 12 is a diagram for explaining the determination of whether the inflow F1 is within the range of the fluctuation band V1. In this example, the inflow F1 is determined to be within the range of the fluctuation band V1 during the periods from 2:00 to 7:00 on the observation start date and from 16:00 to 7:00 the following day. On the other hand, the inflow F1 is determined to be not within the range of the fluctuation band V1 during the period from 8:00 to 15:00.

[0065] The correction value output unit 512d outputs the corrected inflow amount (corresponding to the corrected inflow amount). The corrected inflow amount F2 is data that becomes the inflow amount F1 before correction during a period (corresponding to a first period) when the inflow amount F1 before correction is outside the range of the fluctuation width V1 centered on the reference value S, and becomes the reference value S during a period (corresponding to a second period) when the inflow amount F1 before correction is within the range of the fluctuation width V1 centered on the reference value S.

[0066] Fig. 13 is a diagram showing an example of the inflow volume F2 after correction. In Fig. 13, the inflow volume F2 after correction (bold line) and the inflow volume F1 before correction (dotted line) are shown. In this example, the inflow volume F2 after correction is the pre-correction F1 during the period from 8:00 to 15:00 on the observation start date, and is the reference value S during the periods from 2:00 to 7:00 and from 16:00 to 7:00 the next day on the observation start date.

[0067] That is, the corrected inflow F2 becomes the pre-correction inflow F1 during the period when precipitation W is relatively high from 8:00 to 15:00 on the observation start date. Also, the corrected inflow F2 becomes the noise-suppressed reference value S during the periods when precipitation W is relatively low from 2:00 to 7:00 on the observation start date and from 16:00 to 7:00 the following day.

[0068] Moreover, the corrected inflow amount F2 is not corrected over the entire period of the pre-correction F1, but is corrected during a portion of the period in which the amount of precipitation W is smaller than a predetermined value.

[0069] From the above, the correction value output unit 512d can remove noise and output a corrected inflow F2 with improved accuracy that is close to the original inflow to the downstream dam 3. The corrected inflow F2 has improved accuracy, especially in periods when the amount of precipitation W is small.

[0070] [Model generation part] The model generation unit 513 generates a prediction model for predicting the inflow to the downstream dam 3. The prediction model is generated based on the amount of precipitation W, the inflow F1 before correction, and the inflow F2 after correction.

[0071] In this embodiment, the model generation unit 513 generates a prediction model using a regression equation. Hereinafter, a procedure in which the model generation unit 513 generates a prediction model will be described with reference to Figs.

[0072] Fig. 14 is a diagram explaining the procedure for expressing in two-dimensional space the data set of the amount of precipitation W at each time and the corrected inflow amount F2 shown in Fig. 13. The lower diagram in Fig. 14 is a diagram showing the data set of Fig. 13 in two-dimensional space with the amount of precipitation W on the horizontal axis and the corrected inflow amount F2 on the vertical axis.

[0073] Fig. 15 is a diagram illustrating a prediction model generated by a regression equation. In this embodiment, a quadratic regression equation is used. In Fig. 15, a prediction model M in which the predicted value of the inflow is a quadratic function of the precipitation amount W is shown by a solid line.

[0074] Here, we will explain the advantage of generating a prediction model using a data set of precipitation W and corrected inflow F2. The lower diagram in Fig. 14 is a two-dimensional space with the horizontal axis representing precipitation W and the vertical axis representing corrected inflow F2, while Fig. 7 is a two-dimensional space with the horizontal axis representing precipitation W and the vertical axis representing pre-correction inflow F1.

[0075] As can be seen from these comparisons, the data in Figure 7 is more widely dispersed in the area of ​​low precipitation W than in Figure 14. Such dispersion is due to the fact that each of the data in the area of ​​low precipitation W contains noise.

[0076] In contrast, in Figure 14, the variance of the data is suppressed in areas with low precipitation W. This is because the noise-suppressed corrected inflow F2 is used. As a result, the accuracy of the prediction model using the regression equation is improved, especially in areas with low precipitation W.

[0077] In this embodiment, a prediction model is generated using a regression equation, but this is not limited to this, and other examples such as a neural network, a recurrent neural network, deep learning, and a reservoir function method may be used.

[0078] By using the prediction model M generated by the information processing device 5 described above, it becomes possible to accurately predict the inflow amount into the prediction target dam.

[0079] <Processing up to generating a predictive model> 16 is a flowchart illustrating the flow of processing up to generation of a prediction model by the information processing device 5. The processing up to generation of a prediction model includes steps S101 to S107.

[0080] First, in step 101 , the calculation unit 510 calculates the inflow amount to the downstream dam 3 .

[0081] Next, in step 102, the acquisition unit 511 acquires the inflow amount to the dam to be predicted and the precipitation amount at the observation point in the observation period. The inflow amount is acquired as calculated by the calculation unit 510 in step 101. The precipitation amount is acquired from the meteorological observation device 4. Fig. 4 shows an example of the inflow amount and the precipitation amount acquired by the acquisition unit 511.

[0082] Next, in step 103, the generating unit 512a generates a reference value in which noise of the inflow amount to the downstream dam 3 is suppressed. The reference value S shown in FIG. 5 is an example of the reference value generated by the generating unit 512a in step 103.

[0083] Next, in step 104, the setting unit 512b sets a fluctuation range based on the inflow amount and the precipitation amount. The fluctuation range V1 shown in FIG.

[0084] Next, in step 105, the determination unit 512c determines whether or not the inflow amount is within a range of a fluctuation width centered around a reference value.

[0085] Next, in step 106, the correction value output unit 512d outputs the corrected inflow amount. The corrected inflow amount is a value that becomes the inflow amount before correction during a period in which the inflow amount is outside a fluctuation range centered on the reference value, and becomes the reference value during a period in which the inflow amount before correction is within a fluctuation range centered on the reference value.

[0086] Next, in step 107, the model generation unit 513 generates a prediction model for predicting the inflow amount to the downstream dam 3. The prediction model M shown in Fig. 15 is an example of the prediction model generated by the model generation unit 513 in step 107 using a regression equation.

[0087] By using the prediction model generated by the method described above, it is possible to accurately predict the inflow volume into the dam to be predicted.

[0088] == Variation 1 == The fluctuation range set by the setting unit 512b is not limited to the fluctuation range V1 exemplified in Fig. 6 in the above embodiment. Fig. 17 is a diagram showing another example of the fluctuation range. In Fig. 17, in addition to the example in Fig. 6 (fluctuation range V1), two other examples (fluctuation ranges V2 and V3) are shown.

[0089] Both of the fluctuation ranges V2 and V3 have a maximum value M1 when the amount of precipitation W is 0, and are 0 in the region where the amount of precipitation is greater than a threshold value T1.

[0090] The fluctuation range V2 is a step function. Specifically, the fluctuation range V2 has a maximum value M1 in a region where the amount of precipitation W is smaller than the threshold T1, and is a constant value.

[0091] The fluctuation range V3 is a downward convex function in the region where the amount of precipitation W is smaller than the threshold T1. Specifically, the fluctuation range V3 is a downward convex quadratic function in the region where the amount of precipitation W is smaller than the threshold T1.

[0092] ==Variation 2== In the fluctuation range V1 in the above embodiment, the method of setting the threshold T1 of the amount of precipitation W is not limited to the example shown in Figures 9 and 10. Figure 18 is a diagram showing another example of the method of setting the threshold T1 of the amount of precipitation W.

[0093] In Fig. 18, the data set in Fig. 4 is shown in a two-dimensional space with the horizontal axis representing the amount of precipitation W and the vertical axis representing the inflow amount F1 before correction. Furthermore, in Fig. 18, the regression equation R of the inflow amount generated based on these data sets is shown by a solid line.

[0094] In a modified example, the threshold T2 for precipitation W may be set to a precipitation amount corresponding to the average value A during the period in which the precipitation W of the reference value S is smaller than W1 in the regression equation R plus the maximum value M1 of the fluctuation range V1.

[0095] Even in the case of such modified examples, by using a prediction model, it is possible to accurately predict the inflow volume into the dam being predicted.

[0096] == Variation 3 == In the above embodiment, a reference value was generated for the entire period of the inflow amount F1. Then, a period to be corrected and a period not to be corrected were specified according to the amount of precipitation W, and the corrected inflow amount F2 was output as a reference value for the period to be corrected.

[0097] However, the present invention is not limited to this example, and a period to be corrected and a period not to be corrected may be specified first according to the amount of precipitation W. Then, a reference value may be generated only for the period to be corrected among the entire period of the inflow amount F1, and the corrected inflow amount F2 may be output as the reference value for the period to be corrected. ==Summary== As described above, the information processing device 5 of the embodiment includes an acquisition unit 511 that acquires the inflow amount F1 to the dam to be predicted in a specified period and the precipitation amount W at a specified point in the specified period, and an output unit 512 that outputs the inflow amount F1 in a first period of the specified period and outputs a corrected inflow amount with noise suppressed in a second period of the specified period excluding the first period, and the precipitation amount W at the specified point in the first period is greater than the precipitation amount W at the specified point in the second period.

[0098] According to this configuration, it is possible to generate a corrected inflow F2 by suppressing noise, which is likely to appear when the precipitation W is small, from the inflow F1. As a result, the prediction model generated based on the precipitation W and the corrected inflow F2 has reduced influence of noise and has improved prediction accuracy.

[0099] In the information processing device 5, the output unit 512 includes a generation unit 512a that generates a reference value S in which noise in the inflow amount F1 is suppressed, a determination unit 512c that determines whether the inflow amount F1 is within a range of a fluctuation width V1 that is centered on the reference value S and that decreases as the amount of precipitation W increases, and a correction value output unit 512d that outputs a corrected inflow amount that becomes the inflow amount F1 in a first period in which the inflow amount F1 is outside the range of the fluctuation width V1 centered on the reference value S and becomes the reference value S in a second period in which the inflow amount F1 is within the range of the fluctuation width V1 centered on the reference value S. With this configuration, it is possible to easily estimate the period in which the inflow amount F1 includes noise and the period in which it does not include noise.

[0100] In the information processing device 5, the generating unit 512a performs moving average processing on the inflow amount F1 to generate the reference value S. With this configuration, noise can be suppressed by a simple procedure.

[0101] In the information processing device 5, the output unit 512 includes a setting unit 512b that sets the fluctuation range V1 based on the inflow amount F1 and the precipitation amount W. With this configuration, it is possible to accurately estimate the period in which the inflow amount F1 includes noise and the period in which the inflow amount F1 does not include noise.

[0102] The information processing device 5 further includes a model generating unit 513 that generates a prediction model for predicting the inflow to the dam based on the precipitation W, the inflow, and the corrected inflow. With this configuration, it is possible to generate a prediction model with improved prediction accuracy by suppressing the influence of noise.

[0103] A prediction model using a regression equation or a neural network is generated in the information processing device 5. With this configuration, the prediction accuracy of the prediction model is further improved.

[0104] The information processing device 5 further includes a calculation unit 510 that calculates the inflow amount F1 to the dam based on the output from a measuring instrument that measures the water level of the dam. With this configuration, it is possible to calculate the inflow amount excluding the contribution of artificial control such as the discharge amount from the upstream dam 2.

[0105] The information processing program of the embodiment causes a computer to implement an acquisition unit 511 that acquires the inflow amount to a dam to be predicted during a specified period and the precipitation amount W at a specified point during the specified period, and an output unit 512 that outputs the inflow amount F1 during a first period of the specified period and outputs a corrected inflow amount with noise suppressed during a second period of the specified period excluding the first period, and the precipitation amount W at the specified point during the first period is greater than the precipitation amount W at the specified point during the second period.

[0106] According to such a program, it is possible to generate a corrected inflow F2 that suppresses noise from the inflow F1, which is likely to occur when the precipitation W is small. As a result, the prediction model generated based on the precipitation W and the corrected inflow F2 has reduced effects of noise and has improved prediction accuracy.

[0107] An information processing method according to an embodiment includes the steps of acquiring an inflow F1 to a dam to be predicted for a specified period and a precipitation W at a specified point for the specified period, outputting the inflow F1 for a first period within the specified period, and outputting a corrected inflow with noise suppressed for a second period within the specified period excluding the first period, wherein the precipitation W at the specified point for the first period is greater than the precipitation W at the specified point for the second period.

[0108] According to this method, it is possible to generate a corrected inflow F2 that suppresses noise from the inflow F1, which is likely to occur when the precipitation W is small. As a result, the prediction model generated based on the precipitation W and the corrected inflow F2 has reduced noise effects and has improved prediction accuracy.

[0109] The above-mentioned embodiment is for the purpose of facilitating understanding of the present invention, and is not intended to limit the present invention. Furthermore, the present invention can be modified or improved without departing from the spirit of the present invention, and it goes without saying that the present invention includes equivalents thereof.

[0110] For example, in the above embodiment, the information processing device 5 is shown not to have a means for predicting the inflow volume at a future time point using the generated prediction model, but may further have a prediction unit for performing this, an output unit for outputting the prediction result, etc. [Explanation of symbols]

[0111] 1: Information processing system 2: Upstream dam 3: Downstream dam 4: Weather observation equipment 5: Information processing device 500: Processor 501: Main memory 502:Auxiliary storage device 503: Input device 504: Output device 505: Communication equipment 510: Calculation Department 511: Acquisition Department 512: Output section 512a: Generation part 512b: Setting section 512c: Judgment section 512d: Correction value output unit 513: Model generation unit

Claims

1. An acquisition unit that acquires an inflow amount into a dam to be predicted for a predetermined period and a precipitation amount at a predetermined point for the predetermined period; an output unit that outputs the inflow amount during a first period of the predetermined period, and outputs a corrected inflow amount in which noise is suppressed during a second period of the predetermined period excluding the first period; Equipped with The amount of precipitation at the specific point in the first period is greater than the amount of precipitation at the specific point in the second period. Information processing device.

2. 2. The information processing device according to claim 1, The output unit is A generation unit that generates a reference value in which noise of the inflow amount is suppressed; a determination unit that determines whether the inflow amount is within a range of a fluctuation width that is centered around the reference value and that decreases as the amount of precipitation increases; a correction value output unit that outputs the inflow amount during the first period when the inflow amount is outside the range of the fluctuation range centered on the reference value, and outputs the corrected inflow amount that becomes the reference value during the second period when the inflow amount is within the range of the fluctuation range centered on the reference value; An information processing device comprising:

3. 3. The information processing device according to claim 2, The generating unit performs a moving average process on the inflow amount to generate the reference value. Information processing device.

4. 4. The information processing device according to claim 2, The output unit is A setting unit that sets the fluctuation range based on the inflow amount and the precipitation amount. Information processing device.

5. The information processing device according to any one of claims 1 to 4, A model generating unit generates a prediction model for predicting an inflow into the dam based on the precipitation, the inflow, and the corrected inflow. Information processing device.

6. 6. The information processing device according to claim 5, generating the prediction model using a regression equation or a neural network; Information processing device.

7. The information processing device according to any one of claims 1 to 6, A calculation unit is further provided for calculating an inflow amount into the dam based on an output from a measuring instrument that measures a water level of the dam. Information processing device.

8. On the computer, An acquisition unit that acquires an inflow amount into a dam to be predicted for a predetermined period and a precipitation amount at a predetermined point for the predetermined period; an output unit that outputs the inflow amount during a first period of the predetermined period, and outputs a corrected inflow amount in which noise is suppressed during a second period of the predetermined period excluding the first period; Realize this, The amount of precipitation at the specific point in the first period is greater than the amount of precipitation at the specific point in the second period. Information processing program.

9. A step of acquiring an inflow amount into a dam to be predicted for a predetermined period of time and a precipitation amount at a predetermined point for the predetermined period of time; outputting the inflow amount during a first period of the predetermined period, and outputting a corrected inflow amount in which noise is suppressed during a second period of the predetermined period excluding the first period; Including, The amount of precipitation at the specific point in the first period is greater than the amount of precipitation at the specific point in the second period. Information processing methods.

Images