Water volume prediction device, water volume prediction system, water volume prediction method, water volume prediction program, and recording medium

Abstract

To solve the problem that the prediction accuracy of a water amount is lowered with respect to external fluctuation.SOLUTION: The water quantity prediction device 100 includes an actual-value water quantity acquisition unit 1 that receives actual-value water quantity data indicating an actual water quantity acquired at an observation point and accumulates the actual-value water quantity data as actual-value time-series water quantity data, an external fluctuation removal unit 2 that determines water quantity fluctuation data indicating a temporary water quantity fluctuation in the actual-value time-series water quantity data by using the actual-value time-series water quantity data up to the current time by the actual-value water quantity acquisition unit 1, removes a temporary water quantity fluctuation element indicated by the determined water quantity fluctuation data from the actual-value time-series water quantity data, and obtains estimated value water quantity data up to the current time obtained by interpolating the removed actual-value water quantity data with the estimated value water quantity data, a water quantity prediction unit 3 that obtains future predicted value water quantity data by using the estimated value water quantity data up to the current time obtained by the external fluctuation removal unit 2, and a predicted water quantity output unit 4 that outputs the predicted value water quantity data obtained by the water quantity prediction unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a water quantity prediction device, a water quantity prediction system, a water quantity prediction method, a water quantity prediction program, and a recording medium that predict water quantities such as flow rates and water levels at observation points for low water management. [Background technology]

[0002] Low water management is being considered, including technology to predict the expected inflow (predicted flow rate) of dams in order to ensure the necessary discharge volume, and technology to predict the river flow rate required to maintain the normal function of rivers through dam discharges, etc. Non-Patent Document 1 shows a technique for predicting the predicted inflow of a dam.

[0003] The technology for predicting low water inflow at dams using AI, as described in Non-Patent Document 1, is a technology for constructing a model using a deep neural network (DNN). The data input to the DNN uses available input items that have a high correlation with dam discharge volume. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Furutajima, Itsuki, Miura, Shin, Kamise, Fumio, Kaneko, Takushi, and Yamawaki, Masashi, "Low Water Inflow Prediction at Dams Using AI," Proceedings of the Construction Consulting Business Research Conference, vol. 19th, pp. 13-16, 2019. Summary of the Invention [Problem to be solved by the invention]

[0005] The low water inflow prediction technology shown in Non-Patent Document 1 uses available input items as they are, so if there is a temporary flow rate fluctuation due to an external factor, such as the river water level, which is an input item, the inflow prediction may be higher than expected, and the actual inflow may not be predicted.

[0006] The present disclosure has been made in consideration of the above points, and aims to provide a water volume prediction device that can accurately predict water volumes such as flow rate and water level at an observation point even when there are temporary flow rate fluctuations due to external factors for which data is not available. [Means for solving the problem]

[0007] The water volume prediction device of the present disclosure comprises an actual water volume acquisition unit that receives actual water volume data indicating actual water volume acquired at an observation point and stores it as actual time-series water volume data; an external fluctuation removal unit that uses the actual time-series water volume data up to the current time obtained by the actual water volume acquisition unit to determine water volume fluctuation data indicating temporary water volume fluctuations in the actual time-series water volume data, removes the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from the actual time-series water volume data, and obtains estimated water volume data up to the current time by interpolating the removed actual water volume data with estimated water volume data; a water volume prediction unit that obtains future predicted water volume data using the estimated water volume data up to the current time obtained by the external fluctuation removal unit; and a predicted water volume output unit that outputs the predicted water volume data obtained by the water volume prediction unit. [Effects of the Invention]

[0008] According to the present disclosure, temporary flow rate fluctuations due to external factors for which data is unavailable can be removed, and the predicted values ​​of water volume such as flow rate and water level at the observation point can be predicted with high accuracy without any overshoot. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a block diagram showing an example of the configuration of a river flow rate prediction system including a river flow rate prediction device according to a first embodiment. [Figure 2] 3 is a diagram showing an example of a river flow rate observation value at a river flow rate measurement site measured by a flow rate observation device in the river flow rate prediction device according to the first embodiment. FIG. [Figure 3]4 is a diagram showing an example of a flow rate prediction result of a river's own flow at a river flow rate measurement site by an external fluctuation elimination unit in the river flow rate prediction device according to the first embodiment. FIG. [Figure 4] FIG. 3 is a diagram showing an example of a flow rate prediction result by the river flow rate prediction device in the river flow rate prediction device according to the first embodiment. [Figure 5] 1 is a configuration diagram showing a hardware configuration of a river flow rate prediction device according to a first embodiment. [Figure 6] 3 is a flowchart showing a river flow rate prediction method performed by the river flow rate prediction device according to the first embodiment. [Figure 7] FIG. 10 is a block diagram showing an example of the configuration of a river flow rate prediction system including a river flow rate prediction device according to a second embodiment. [Figure 8] FIG. 10 is a diagram conceptually showing the processing of each component in an external fluctuation elimination unit in the river flow rate prediction device according to the second embodiment. [Figure 9] 10 is a flowchart showing a river flow rate prediction method performed by the river flow rate prediction device according to the second embodiment. [Figure 10] FIG. 11 is a block diagram showing an example of the configuration of a river flow rate prediction system including a river flow rate prediction device according to a third embodiment. [Figure 11] FIG. 10 is a diagram showing an example of a river flow rate prediction result for a river flow rate measurement site, reflecting the rainfall detection processing result, by the external fluctuation elimination unit in the river flow rate prediction device of embodiment 3. [Figure 12] 10 is a flowchart showing a river flow rate prediction method performed by the river flow rate prediction device according to the third embodiment. [Figure 13] FIG. 10 is a block diagram showing an example of the configuration of a river flow rate prediction system including a river flow rate prediction device according to a fourth embodiment. [Figure 14] FIG. 10 is a diagram conceptually showing the processing of each component related to dam discharge in an external fluctuation elimination unit in a river flow rate prediction device according to a fourth embodiment. [Figure 15] 10 is a flowchart showing a river flow rate prediction method performed by a river flow rate prediction device according to a fourth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Embodiment 1 A water quantity prediction device according to a first embodiment will be described with reference to FIGS. 1 to 6. FIG. The water volume prediction device of embodiment 1 is a river flow rate prediction device that predicts the flow rate at an observation point on a river, a river water level prediction device that predicts the water level at an observation point on a river, or a dam inflow prediction device that predicts the inflow volume of a dam that is an observation point. In the first embodiment, the term "water volume" includes flow rate, water level, and inflow volume. The following description will be given taking a river flow rate prediction device as an example of the water volume prediction device. To avoid complication of explanation, the amount of water will be explained as a flow rate.

[0011] FIG. 1 shows a river flow rate prediction system (water volume prediction system) including a river flow rate prediction device 100, which is a water volume prediction device according to the first embodiment, a water volume (flow rate) observation device 200, and a predicted water volume (flow rate) display unit 300. The flow observation device 200 acquires the actual flow rate at an observation point in a river, that is, the actual flow rate, and outputs it as actual flow rate data indicating the actual water volume. The observation point here is a flow forecast point where the flow rate is predicted. The actual flow rate is periodically acquired by the flow rate observing device 200. The actual flow rate may be acquired at a regular interval or at an irregular interval.

[0012] The flow observation device 200 is composed of, for example, a water level meter that observes the water level at an observation point, and a water level-flow rate conversion device that converts the water level observation value observed by the water level meter into flow rate using a water level flow rate curve (HQ formula) that represents the relationship between the water level observation value and flow rate, and outputs it as actual value data. The flow rate observing device 200 may be the following device. This device directly observes the flow rate at the observation point using an ADCP (Acoustic Doppler Current Profiler) and outputs the actual flow rate data.

[0013] This device consists of a current meter that observes the flow velocity at the observation point, and a flow velocity-flow rate converter that converts the flow velocity observed by the current meter into flow rate using a hydraulic model that represents the relationship between flow velocity and flow rate, and outputs the data as actual flow rate data. In short, the flow observation device 200 may be any device that can obtain the actual flow rate at an observation point in a river.

[0014] The river flow rate prediction device 100 according to embodiment 1 removes temporary flow rate fluctuations (hereinafter referred to as external fluctuations) due to external factors for which data is unavailable, predicts natural flow rate fluctuations at observation points in the river under low water management without causing any upward fluctuations in the flow rate at observation points in the river, and outputs the predicted flow rate data.

[0015] Low water management defines the flow rate required to maintain the normal function of a river, i.e., normal flow rate, and when the river administrator predicts that the flow rate at an observation point on the river will decrease and fall below the normal flow rate, he or she can request the dam upstream of the observation point to release water so that the flow rate at the observation point does not fall below the normal flow rate.

[0016] Therefore, it is important for the river flow rate prediction device 100 to predict the flow rate at observation points in rivers with high accuracy. The external fluctuations referred to here are temporary fluctuations in flow rate at the observation point in a river due to artificial factors that cause flow rate fluctuations, such as temporary water discharges from a power plant upstream of the observation point in the river, where water for power generation is irregularly injected (returned) at intervals of several hours to half a day.

[0017] The river flow prediction device 100 of embodiment 1 uses actual flow data indicating the actual flow rate at the observation point obtained from the flow observation device 200 to remove external fluctuations to obtain estimated flow rate data, predicts the actual flow rate at the observation point (hereinafter, the actual flow rate in a river is referred to as the river flow) based on the estimated flow rate data, and outputs the predicted flow rate data (predicted flow rate (water volume) data) to the predicted flow rate display unit 300. The river flow rate prediction device 100 according to the first embodiment includes an actual flow rate acquisition unit 1, an external fluctuation elimination unit 2, a flow rate prediction unit 3, and a predicted flow rate output unit 4.

[0018] When the water volume prediction device 100 is used as a river water level prediction device, the actual flow rate acquisition unit 1, the external fluctuation removal unit 2, the flow rate prediction unit 3, and the predicted flow rate output unit 4 function as an actual water level acquisition unit, an external fluctuation removal unit, a water level prediction unit, and a predicted water level output unit, respectively. Furthermore, when the water volume prediction device 100 is used as a dam inflow prediction device, the actual flow rate acquisition unit 1, the external fluctuation removal unit 2, the flow rate prediction unit 3, and the predicted flow rate output unit 4 function as an actual inflow rate acquisition unit, an external fluctuation removal unit, an inflow rate prediction unit, and a predicted inflow rate output unit, respectively.

[0019] That is, the actual flow rate acquisition unit 1, the actual water level acquisition unit, and the actual inflow volume acquisition unit are actual water volume acquisition units. The flow rate prediction unit 3, the water level prediction unit, and the inflow prediction unit are water volume prediction units. The predicted flow rate output unit 4, the predicted water level output unit, and the predicted inflow volume output unit are predicted water volume output units.

[0020] The actual flow rate acquiring unit 1 receives actual flow rate data indicating actual flow rates acquired at observation points from a flow rate observing device 200, and accumulates past actual flow rate data up to the present time as actual flow rate time-series data. FIG. 2 shows an example of the actual flow rate data at the observation point accumulated by the actual flow rate acquisition unit 1 in a time series. In FIG. 2, the horizontal axis represents time, the vertical axis represents flow rate, t represents the current time, and the dashed curve A represents the actual flow rate data.

[0021] In the actual flow rate data A, a1, a2, and a3 are external fluctuations, that is, flow rate fluctuation data indicating temporary flow rate fluctuations due to external factors. That is, the actual flow data A represents flow fluctuation data that includes the natural flow fluctuation of the river and the natural flow fluctuation plus external fluctuations. The actual flow rate data A from the flow rate observation device 200 is data in which fluctuations in the river's own flow rate and external fluctuations are not distinguished.

[0022] The external fluctuation elimination unit 2 uses the actual value time-series flow rate data A up to the current time obtained by the actual flow rate acquisition unit 1 to identify flow rate fluctuation data a1, a2, a3 that indicate temporary flow rate fluctuations in the actual value time-series flow rate data A, removes the temporary flow rate fluctuation elements indicated by the identified flow rate fluctuation data a1, a2, a3 from the actual value time-series flow rate data A, and obtains interpolated estimated value flow rate data at the observation point up to the current time by linearly interpolating the removed actual value flow rate data.

[0023] That is, the external fluctuation elimination unit 2 obtains estimated time-series flow rate data up to the present time by interpolating the actual time-series flow rate data A from which the external fluctuation has been eliminated using the estimated water volume data. The estimated water volume data obtained by the external fluctuation elimination unit 2 is shown in FIG. In FIG. 3, the horizontal axis represents time, the vertical axis represents flow rate, t represents the current time, and the solid curve B represents the estimated flow rate data.

[0024] As is clear from Figure 3, the estimated flow rate data B shown by the solid line represents flow rate data obtained by removing the temporary flow rate fluctuation elements shown by the flow rate fluctuation data a1, a2, and a3 shown by the dashed lines from the actual time-series flow rate data and interpolating the estimated water volume data. The estimated flow rate data B is the past flow rate data up to the current time t that estimates the flow rate of the river's natural flow at the observation point.

[0025] For example, the external fluctuation removal unit 2 detects a rising point of temporary flow rate fluctuation when the flow rate difference value in the actual flow rate data for the Δt period is equal to or greater than a first threshold value indicating a positive value, and after detecting the rising point and detecting a negative flow rate difference value, detects a falling point of temporary flow rate fluctuation when the flow rate difference value in the actual flow rate data for the Δt period is equal to or less than a second threshold value indicating a negative value.

[0026] The external fluctuation removal unit 2 determines the interval between the detected rising and falling points as flow rate fluctuation data a1, a2, a3 indicating temporary flow rate fluctuation, and linearly interpolates the interval between the detected rising and falling points of the actual value time-series flow rate data A to obtain estimated value flow rate data (estimated value time-series flow rate data) B. The external fluctuation elimination unit 2 uses the actual flow rate data as it is as estimated flow rate data except for the period from the detected rising point to the falling point.

[0027] The external fluctuation elimination unit 2 may use a low-pass filter or a Kalman filter to remove instantaneous flow rate fluctuations, to identify and remove flow rate fluctuation data a1, a2, a3 indicating temporary flow rate fluctuations from the actual value time-series flow rate data A, thereby obtaining estimated value time-series flow rate data B. In the first embodiment, the low-pass filter or Kalman filter is a commonly known filter implemented by software.

[0028] The flow rate prediction unit 3 obtains future predicted flow rate data at the observation point using the estimated flow rate data B up to the current time obtained by the external fluctuation removal unit 2. The predicted flow rate data obtained by the flow rate prediction unit 3 is shown in FIG. In FIG. 4, the horizontal axis represents time, the vertical axis represents flow rate, t represents the current time, and the thick solid curve C shown after the current time t represents the predicted flow rate data.

[0029] The predicted flow rate data C is shown as continuous flow rate data from the estimated flow rate data B indicated by a thin solid line at the current time t. The predicted flow rate data C is the future flow rate data from the current time t that predicts the flow rate of the river's natural flow at the observation point.

[0030] As is clear from the predicted value flow rate data C shown in Figure 4, the predicted value flow rate data C is calculated based on the estimated value flow rate data B obtained by removing the flow rate fluctuation data a1, a2, a3, which indicate temporary flow rate fluctuations (external fluctuations), using the external fluctuation removal unit 2.Therefore, there is no upward fluctuation in the flow rate data at the observation point due to the flow rate fluctuation data a1, a2, a3, and the predicted value flow rate data C indicates flow rate data that can accurately predict the natural flow of the river at the observation point.

[0031] The flow rate prediction unit 3 obtains predicted flow rate data C by, for example, making a prediction by time extrapolation using a polynomial function. The flow rate prediction unit 3 may obtain the predicted flow rate data C by making a prediction using a state space model such as a trend model or an AR (AutoRegressive model) model. The flow rate prediction unit 3 may obtain the predicted flow rate data C by making predictions using a combination of a state space model and a Kalman filter. The flow rate prediction unit 3 may obtain the predicted flow rate data C by making a prediction using a machine learning method such as a neural network.

[0032] The predicted flow rate output unit 4 outputs the predicted flow rate data C obtained by the flow rate prediction unit 3 to at least the predicted flow rate display unit 300 . Furthermore, the predicted flow rate output unit 4 may output the estimated flow rate data B and the flow rate fluctuation data a1, a2, and a3 together with the predicted flow rate data C to the predicted flow rate display unit 300.

[0033] The predicted flow rate display unit 300 displays the predicted flow rate data C from the predicted flow rate output unit 4, and is, for example, a liquid crystal display. The predicted flow rate display unit 300 displays the graph shown in FIG. 4 as an example.

[0034] The river flow prediction device 100 is realized by a computer hardware configuration, and as shown in Figure 5, includes a CPU (Central Processing Unit) 10A, a large-capacity semiconductor memory (RAM: Random Access Memory) 10B, a storage device (ROM: Read only memory) 10C such as a hard disk device or a non-volatile storage device such as an SSD device, an input interface unit 10D, an output interface unit 10E, and a signal path (bus) 10F.

[0035] The CPU 10A controls and manages the RAM 10B, the ROM 10C, the input interface unit 10D, and the output interface unit 10E. The CPU 10A loads the programs stored in the ROM 10C into the RAM 10B, and executes various processes based on the programs loaded into the RAM.

[0036] The actual flow rate acquisition unit 1 is composed of a CPU 10A, a RAM 10B, and an input interface unit 10D, and based on the program loaded into RAM 10B, the CPU 10A acquires actual flow rate data from the flow observation device 200 via the input interface unit 10D, and stores the actual time-series flow rate data in RAM 10B.

[0037] The external fluctuation elimination unit 2 is made up of a CPU 10A and a RAM 10B, and based on the program loaded into RAM 10B, the CPU 10A uses the actual flow rate data for the current time and past actual flow rate data in the actual flow rate time-series data stored in RAM 10B to identify flow rate fluctuation data a1, a2, a3 that indicate temporary flow rate fluctuations, removes the temporary flow rate fluctuation elements indicated by the flow rate fluctuation data a1, a2, a3 from the actual flow rate time-series data A, and obtains estimated flow rate data B at the observation point up to the current time that has been interpolated by linearly interpolating the removed actual flow rate data, and stores the estimated flow rate data B in RAM 10B.

[0038] The flow rate prediction unit 3 is composed of a CPU 10A and a RAM 10B, and based on the program loaded into RAM 10B, the CPU 10A obtains future predicted flow rate data C using the estimated flow rate data B stored in RAM 10B, and stores the predicted flow rate data C in RAM 10B. The predicted flow rate output unit 4 is composed of a CPU 10A, a RAM 10B, and an output interface unit 10E, and based on the program loaded into RAM 10B, the CPU 10A reads out at least the predicted value flow rate data C stored in RAM 10B; in embodiment 1, the predicted value flow rate data C, estimated value flow rate data B, and flow rate fluctuation data a1, a2, a3, and outputs them to the predicted flow rate display unit 300 via the output interface unit 10E.

[0039] Next, the operation of the river flow rate prediction device 100 according to the first embodiment, that is, the river flow rate prediction method, will be described with reference to FIG. In step ST1, the actual flow rate acquisition unit 1 acquires actual flow rate data indicating the actual flow rate acquired at the observation point from the flow rate observation device 200, and accumulates past actual flow rate data up to the current time t as actual flow rate time series data A. The actual flow rate acquiring unit 1 periodically acquires the actual flow rate data. Step ST1 is a data acquisition step for the actual flow rate data.

[0040] Once the actual flow rate acquisition unit 1 acquires and stores the actual flow rate data at the current time, the process proceeds to step ST2. In step ST2, the external fluctuation elimination unit 2 compares the actual flow rate data at the current time t with the actual flow rate data at the time immediately preceding the current time t (t-Δt1) to obtain a flow rate difference value, and compares the flow rate difference value with a threshold value to detect the rising and falling points of temporary flow rate fluctuations, and determines the interval between the detected rising and falling points as flow rate fluctuation data a1, a2, a3 indicating temporary flow rate fluctuations.

[0041] In step ST2, linear interpolation is performed between the rising and falling points detected by the external fluctuation elimination unit 2, and the flow rate data obtained by the interpolation is used as estimated flow rate data between the rising and falling points. In step ST2, the external fluctuation elimination unit 2 uses the actual flow rate data as it is as estimated flow rate data except for the period from the rising point to the falling point. The external fluctuation elimination unit 2 accumulates past estimated flow rate data B up to the current time t.

[0042] In short, in step ST2, the external fluctuation removal unit 2 uses the actual value time-series flow rate data A up to the current time t from the actual flow rate acquisition unit 1 to identify flow rate fluctuation data a1, a2, a3 that indicate temporary flow rate fluctuations in the actual value time-series flow rate data A, removes the temporary flow rate fluctuation elements indicated by the identified flow rate fluctuation data a1, a2, a3 from the actual value time-series flow rate data A, and obtains estimated value flow rate data B at the observation point up to the current time. Step ST2 is an estimation step in which external fluctuations are removed and the past river flow rate up to the current time t is estimated.

[0043] When the external fluctuation elimination unit 2 obtains and stores the estimated flow rate data for the current time, the process proceeds to step ST3. In step ST3, the flow rate prediction unit 3 uses the estimated flow rate data B to obtain future predicted flow rate data C at the observation point. Step ST3 is a prediction step in which the future river flow rate is predicted from the current time t.

[0044] When the flow rate prediction unit 3 obtains the predicted flow rate data C, the process proceeds to step ST4. In step ST4, the predicted flow rate output unit 4 outputs the predicted value flow rate data C obtained by the flow rate prediction unit 3, the estimated value flow rate data B obtained by the external fluctuation removal unit 2, and the flow rate fluctuation data a1, a2, and a3 to the predicted flow rate display unit 300. Step ST4 is a data output step that outputs predicted flow rate data C that predicts the future flow rate of the river's natural flow.

[0045] The river flow rate prediction method from step ST1 to step ST4 is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C). That is, the program stored in ROM10C includes a data acquisition procedure for acquiring actual water volume data indicating actual water volumes obtained at observation points and storing it as actual time-series water volume data; an estimation procedure for using the actual time-series water volume data to determine water volume fluctuation data indicating temporary water volume fluctuations in the actual time-series water volume data, removing the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from the actual time-series water volume data, and interpolating the removed actual flow rate data with estimated flow rate data to obtain estimated water volume data up to the current time; a prediction procedure for using the estimated water volume data to obtain predicted water volume data for the future; and a data output procedure for outputting the predicted water volume data.

[0046] The river flow rate prediction device 100, which is a water volume prediction device according to the first embodiment, comprises an actual flow rate acquisition unit 1 (actual water volume acquisition unit), an external fluctuation elimination unit 2, a flow rate prediction unit 3 (water volume prediction unit), and a predicted flow rate output unit 4 (predicted water volume output unit). The external fluctuation elimination unit 2 uses actual value time-series flow rate data A up to the current time from the actual flow rate acquisition unit 1 to determine flow rate fluctuation data a1, a2, a3 that indicate temporary flow rate fluctuations in the actual value time-series flow rate data A, removes the temporary flow rate fluctuation elements indicated by the determined flow rate fluctuation data a1, a2, a3 from the actual value time-series flow rate data A, and interpolates the removed actual value flow rate data using estimated value flow rate data to obtain estimated value flow rate data B up to the current time. As a result, there is no over-excursion of the flow rate data at the observation point due to the flow rate fluctuation data a1, a2, a3, and predicted value flow rate data C is obtained as flow rate data that can accurately predict the river flow at the observation point.

[0047] Embodiment 2 A water quantity prediction device according to a second embodiment will be described with reference to FIGS. The water volume prediction device according to the second embodiment will also be described using a river flow rate prediction device as an example, similar to the first embodiment. The river flow rate prediction device 101 of embodiment 2 differs from the river flow rate prediction device 100 of embodiment 1 in that the external fluctuation elimination unit 2 has been replaced with an external fluctuation elimination unit 21, but is otherwise the same.

[0048] That is, the external fluctuation elimination unit 21 has the same function as the external fluctuation elimination unit 2 in the first embodiment, but has different components. Therefore, the following description will focus on the components that make up the external fluctuation elimination unit 21. 7 to 9, the same reference numerals as those in FIGS. 1 to 6 designate the same or corresponding parts.

[0049] FIG. 7 shows a river flow rate prediction system (water volume prediction system) including a river flow rate prediction device 101, a flow rate observation device 200, and a predicted flow rate display unit 300, which is a water volume prediction device according to the second embodiment. The river flow prediction device 101 of embodiment 2 uses actual flow data A indicating the actual flow rate at the observation point obtained from the flow observation device 200, removes external fluctuations a1, a2, and a3, and linearly interpolates the removed actual flow data to obtain interpolated estimated flow data B, predicts the river flow at the observation point based on the estimated flow data B, and outputs the predicted flow data (predicted flow data C) to the predicted flow display unit 300.

[0050] The river flow rate prediction device 101 according to the second embodiment includes an actual flow rate acquisition unit 1, an external fluctuation elimination unit 21, a flow rate prediction unit 3, and a predicted flow rate output unit 4. The actual flow rate acquisition unit 1, flow rate prediction unit 3, and predicted flow rate output unit 4 are essentially the same as the actual flow rate acquisition unit 1, flow rate prediction unit 3, and predicted flow rate output unit 4 in the river flow rate prediction device 100 of embodiment 1.

[0051] The external fluctuation elimination unit 21 uses the actual value time-series flow rate data A up to the current time obtained by the actual flow rate acquisition unit 1 to identify flow rate fluctuation data a1, a2, a3 that indicate temporary flow rate fluctuations in the actual value time-series flow rate data A, removes the temporary flow rate fluctuation elements indicated by the identified flow rate fluctuation data a1, a2, a3 from the actual value time-series flow rate data A, and obtains estimated value flow rate data at the observation point up to the current time by interpolating the removed actual value flow rate data.

[0052] The actual flow rate data A at the observation point accumulated by the actual flow rate acquisition unit 1 and the estimated water volume data B obtained by the external fluctuation elimination unit 21 are shown in the upper diagram of FIG. In the upper diagram of Figure 8, the horizontal axis represents time, the vertical axis represents flow rate, and t represents the current time. Since the actual flow rate data A and the estimated flow rate data B are the same except for the periods indicated by the flow rate fluctuation data a1, a2, and a3, both are shown as thin solid curves up to the current time t.

[0053] The external fluctuation removal unit 21 includes a difference value calculation unit 21a, a peak detection unit 21b, an external fluctuation determination unit 21c, and an interpolation unit 21d. The difference value calculation unit 21a calculates a primary flow rate difference value (primary water flow rate difference value), which is the difference between the actual flow rate data (actual water rate data) at the current time and the actual flow rate data (actual water rate data) at the time immediately before the current time in the actual value time series water rate data A of the observation point obtained by the actual flow rate acquisition unit 1, and a secondary flow rate difference value (secondary water flow rate difference value), which is the difference between the primary flow rate difference value at the current time and the primary flow rate difference value at the time immediately before the current time, and stores these as secondary flow rate difference value time series data.

[0054] The primary flow differential value q1(t) at the observation point at the current time t is expressed by the following equation (1). q1(t)=q(t)-q(t-Δt1) (1) In the above equation (1), q(t) is the actual flow rate value based on the actual flow rate data at the observation point at the current time t, q(t-Δt1) is the actual flow rate value based on the actual flow rate data at the observation point at the time (t-Δt1) immediately before the current time t, and Δt1 is the time difference between the current time t and the time immediately before (t-Δt1).

[0055] When the primary flow rate difference value q1(t) is positive, it indicates a rise in the actual flow rate value in the actual value time-series water volume data A, and when it is negative, it indicates a fall in the actual flow rate value in the actual value time-series water volume data A.

[0056] The secondary flow difference value q2(t) at the observation point at the current time t is expressed by the following equation (2). q2(t)=q1(t)-q1(t-Δt2) (2) In the above equation (2), q1(t) is the primary flow differential value calculated by the above equation (1) at the current time t at the observation point, q1(t-Δt2) is the primary flow differential value calculated by the above equation (1) at the time (t-Δt2) immediately before the current time t at the observation point, and Δt2 is the time difference between the current time t and the time immediately before the current time t (t-Δt2).

[0057] The peak detection unit 21b detects a maximum value in the secondary water flow difference value q2(t) using the secondary flow difference value time series data from the difference value calculation unit 21a, and sets the time corresponding to the detected maximum value as the inflection point time. The peak detection unit 21b searches for peaks in the secondary flow rate difference value time series data, detects a maximum value, and regards the detected maximum value as an inflection point. The time when the inflection point is detected is taken as the inflection point time.

[0058] Inflection points d1 to d7 in the secondary flow rate difference value time series data D obtained by the peak detection unit 21b are shown as an example in the lower diagram of FIG. In the lower diagram of FIG. 8, the horizontal axis represents time and the vertical axis represents the secondary flow rate difference value q2(t).

[0059] The external fluctuation determination unit 21c determines a flow rate fluctuation period that indicates a temporary flow rate fluctuation in the performance value time-series flow rate data A using the inflection point times at which the inflection points d1 to d7 obtained by the peak detection unit 21b are detected. The external fluctuation determination unit 21c excludes the maximum value obtained when the actual flow rate value indicated by the actual flow rate data A is equal to or greater than the peak removal determination curve from the inflection points for determining the flow rate fluctuation period.

[0060] For example, among the inflection points d1 to d7 shown as examples in the lower diagram of Figure 8, the external fluctuation determination unit 21c excludes the inflection point d6, which is the maximum value obtained when the actual flow rate value is equal to or greater than the peak removal determination curve H, from the inflection points used to determine the flow rate fluctuation period. The peak removal judgment curve H is, for example, a curve obtained using a moving average or a curve obtained using percentiles.

[0061] Furthermore, the external fluctuation determination unit 21c may exclude from the inflection points used to determine the flow rate fluctuation period in such a manner that the actual flow rate value at the time indicating the target inflection point, for example, inflection point d6 shown in the lower diagram of Figure 8, is greater than the actual flow rate values ​​at the times indicating the inflection points before and after it, for example, inflection points d5 and d7 shown in the lower diagram of Figure 8, by a set threshold value or more.

[0062] As can be seen from the secondary flow rate difference value time series data D shown in the lower diagram of Figure 8, maximum values ​​are detected at the times when the primary flow rate difference value q1(t) takes a positive value, indicating a rise in the actual flow rate value, and at the times when the primary flow rate difference value q1(t) takes a negative value, indicating a fall in the actual flow rate value, and inflection points d1 to d5 and d7 in the secondary flow rate difference value time series data D are obtained.

[0063] Under low water management, the external fluctuation determination unit 21c determines the time at which the primary flow rate difference value q1(t) shows a rising edge of the actual flow rate value among the inflection point times at which inflection points d1 to d5 and d7 (excluding the excluded inflection point d6) are detected from among the inflection points d1 to d7 obtained by the peak detection unit 21b as the external fluctuation start time, determines the time at which the primary flow rate difference value q1(t) shows a falling edge of the actual flow rate value as the external fluctuation end time, and determines the period from the external fluctuation start time to the external fluctuation end time as the flow rate fluctuation period in the actual value time-series water volume data A.

[0064] In FIG. 8, the flow rate fluctuation periods determined by the external fluctuation determination unit 21c are from the inflection point time when inflection point d1 is detected to the inflection point time when inflection point d2 is detected, from the inflection point time when inflection point d3 is detected to the inflection point time when inflection point d4 is detected, and from the inflection point time when inflection point d5 is detected to the inflection point time when inflection point d7 is detected, and each of these periods corresponds to the periods indicated by the flow rate fluctuation data a1, a2, and a3, respectively.

[0065] The interpolation unit 21d removes the actual flow rate data during the flow rate fluctuation period determined by the external fluctuation determination unit 21c from the actual flow rate time-series flow rate data A obtained by the actual flow rate acquisition unit 1, and obtains interpolated estimated flow rate data B up to the current time by linearly interpolating the estimated flow rate data during the flow rate fluctuation period determined by the external fluctuation determination unit 21c. The estimated flow rate data linearly interpolated by the interpolation unit 21d is an estimated flow rate value represented by a line segment connecting the actual flow rate value at the external fluctuation start time and the actual flow rate value at the external fluctuation end time. The estimated flow rate data linearly interpolated by the interpolation unit 21d may be obtained using a polynomial function or a neural network.

[0066] The estimated flow rate data linearly interpolated by the interpolation unit 21d are shown as b1, b2, and b3 in the upper diagram of FIG. The estimated flow rate data B obtained by the interpolation unit 21d is past flow rate data up to the present time t that estimates the flow rate of the river's own flow at the observation point. The flow rate prediction unit 3 functions in the same manner as the flow rate prediction unit 3 in the river flow rate prediction device 100 according to embodiment 1, and obtains future predicted flow rate data C at the observation point using estimated flow rate data B up to the current time obtained by the interpolation unit 21d in the external fluctuation removal unit 2. In the upper diagram of FIG. 8, the predicted flow rate data C obtained by the flow rate prediction unit 3 is shown by a thick solid curve C.

[0067] The river flow rate prediction device 101 according to the second embodiment is realized by a computer hardware configuration shown in FIG. 5, similar to the river flow rate prediction device 100 according to the first embodiment. Next, the operation of the river flow rate prediction device 101 according to the second embodiment, that is, the river flow rate prediction method, will be described with reference to FIG. Step ST1 is the same as step ST1 in the river flow rate prediction device 100 according to the first embodiment, which is the data acquisition step of the actual flow rate data.

[0068] Step ST21 is an estimation step in which, like step ST2 in the river flow rate prediction device 100 according to the first embodiment, external fluctuations are removed and the past river flow rate up to the present time t is estimated. That is, in step ST21a, the differential value calculation unit 21a uses the actual value flow rate data for the current time t and the actual value flow rate data for the time just before the current time t (t-Δt1) in the actual value time series flow rate data A from the actual water volume acquisition unit 1 to calculate the primary flow rate differential value q1(t) using the above equation (1), and further calculates the secondary flow rate differential value q2(t) using the above equation (2), stores them as secondary flow rate differential value time series data, and proceeds to step ST21b. Step ST21a is a step for calculating the secondary flow rate difference value q2(t).

[0069] In step ST21b, the peak detection unit 21b detects a maximum value in the secondary flow differential value q2(t) using the secondary flow differential value time series data, sets the time corresponding to the detected maximum value as the inflection point time, and stores it as secondary flow differential value time series data D, which is inflection point time series data, before proceeding to step ST21c. Step ST21b is a peak detection step in which a maximum value is detected from the secondary flow rate difference value time series data to obtain an inflection point time.

[0070] In step ST21c, the external fluctuation determination unit 21c first excludes the inflection point time obtained from the secondary flow differential value time series data D when the actual flow value at the inflection point time is equal to or greater than the peak removal determination curve H from the inflection point times for determining the flow fluctuation period. Next, the external fluctuation determination unit 21c determines the external fluctuation start time and external fluctuation end time from the secondary flow rate difference value time series data D excluding the inflection point times excluded, and determines the period from the external fluctuation start time to the external fluctuation end time as a water volume fluctuation period indicating temporary water volume fluctuation, and proceeds to step ST21d. Step ST21c is a determination step for determining a flow rate fluctuation period that indicates a temporary flow rate fluctuation using the inflection point time obtained by the peak detection unit.

[0071] In step ST21d, the interpolation unit 21d removes the actual flow rate data during the flow rate fluctuation period from the actual flow rate time-series flow rate data A, and obtains estimated flow rate data B up to the current time by linearly interpolating the estimated flow rate data b1, b2, and b3 during the flow rate fluctuation period, and then proceeds to step ST3. Step ST21d is a flow rate estimation step in which the flow rate due to the river's natural flow is estimated to obtain estimated flow rate data B.

[0072] Steps ST3 and ST4 are the same as step ST3 of obtaining future predicted value flow rate data C and step ST4 of outputting predicted value flow rate data C in the river flow rate prediction device 100 according to the first embodiment.

[0073] In the river flow rate prediction device 101 according to the second embodiment, the river flow rate prediction method from step ST1 to step ST4 is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C). That is, the program stored in ROM 10C includes a data acquisition procedure for acquiring actual water volume data indicating actual water volumes acquired at observation points and storing the data as actual time-series water volume data; a calculation procedure for using the actual time-series water volume data to calculate a first water volume difference value, which is the difference between the actual water volume data at the current time and the actual water volume data at the time immediately preceding the current time, and a second water volume difference value, which is the difference between the first water volume difference value at the current time and the first water volume difference value at the time immediately preceding the current time, and storing the data as second water volume difference value time-series data; and a calculation procedure for using the second water volume difference value time-series data to calculate a second water volume difference value time-series data. The system includes a peak detection procedure that detects a maximum value in the secondary water volume difference value and sets the time corresponding to the detected maximum value as the inflection point time; a discrimination procedure that uses the inflection point time to discriminate a water volume fluctuation period that indicates temporary water volume fluctuations; an estimation procedure that removes actual water volume data during the water volume fluctuation period from the actual value time-series water volume data and obtains estimated water volume data up to the current time by linearly interpolating the estimated water volume data during the water volume fluctuation period; a prediction procedure that uses the estimated water volume data to obtain predicted water volume data for the future; and a data output procedure that outputs the predicted water volume data.

[0074] The river flow rate prediction device 101 of embodiment 2 has the same effects as the river flow rate prediction device 100 of embodiment 1, and in addition, the peak detection unit 21b calculates the secondary flow rate difference value of the flow fluctuation, and the external fluctuation determination unit 21c determines the period of flow rate fluctuation due to external fluctuation based on the secondary flow rate difference value, so even when the flow rate of the river's own flow is on an upward or downward trend, the predicted flow rate data C can be obtained as flow rate data that can accurately predict the river's own flow at the observation point.

[0075] Embodiment 3 A water quantity prediction device according to a third embodiment will be described with reference to FIGS. The water volume prediction device according to the third embodiment will also be described using a river flow rate prediction device as an example, similar to the first and second embodiments. The river flow prediction device 102 of embodiment 3 differs from the river flow prediction device 101 of embodiment 2 in that information due to rainfall is added to the prediction of flow fluctuations in the river's own flow at the observation point, but is otherwise the same.

[0076] That is, the river flow rate prediction device 102 according to embodiment 3 differs from the external fluctuation removal unit 21 in the river flow rate prediction device 101 according to embodiment 2 in that it has an external fluctuation removal unit 22 which adds a rainfall information acquisition unit 22a and a rainfall detection unit 22b. Therefore, the following description will be focused on the rainfall information acquisition unit 22a and the rainfall detection unit 22b in the external fluctuation removal unit 22. 10 to 12, the same reference numerals as those in FIGS. 1 to 9 indicate the same or corresponding parts.

[0077] FIG. 10 shows a river flow rate prediction system (water volume prediction system) including a river flow rate prediction device 102, a flow observation device 200, and a predicted flow rate display unit 300, which is a water volume prediction device according to the third embodiment. The river flow prediction device 102 of embodiment 3 uses actual flow data A indicating the actual flow rate at the observation point obtained from the flow observation device 200, removes external fluctuations a1, a2, and a3, and interpolates the removed actual flow data by linear interpolation to detect the rainfall period, obtains estimated flow data B by converting the flow data during the rainfall period into flow data based on rainfall, predicts the river flow at the observation point based on the estimated flow data B, and outputs the predicted flow data (predicted flow data C) to the predicted flow display unit 300.

[0078] The river flow rate prediction device 102 according to the third embodiment includes an actual flow rate acquisition unit 1, an external fluctuation elimination unit 22, a flow rate prediction unit 3, and a predicted flow rate output unit 4. The actual flow rate acquisition unit 1, flow rate prediction unit 3, and predicted flow rate output unit 4 are essentially the same as the actual flow rate acquisition unit 1, flow rate prediction unit 3, and predicted flow rate output unit 4 in the river flow rate prediction device 101 of embodiment 2.

[0079] The external fluctuation elimination unit 22 uses the actual value time-series flow data A up to the current time from the actual flow acquisition unit 1 to identify flow fluctuation data a1, a2, a3 that indicate temporary flow fluctuations in the actual value time-series flow data A, removes the temporary flow fluctuation elements indicated by the identified flow fluctuation data a1, a2, a3 from the actual value time-series flow data A, linearly interpolates the removed actual value flow data, and uses rainfall time-series data from the basin rainfall data to detect the start time of the flow increase and the end time of the flow increase, and treats the flow data for the rainfall period from the start time of the flow increase to the end time of the flow increase as flow data based on rainfall, thereby removing the flow fluctuation data a1, a2, a3 and interpolating using linear interpolation to obtain estimated value flow data at the observation point up to the current time as flow data based on rainfall during the rainfall period.

[0080] The external fluctuation removal unit 22 has a difference value calculation unit 21a, a peak detection unit 21b, an external fluctuation determination unit 21c, an interpolation unit 21d, a rainfall information acquisition unit 22a, and a rainfall detection unit 22b. The difference value calculation unit 21a, peak detection unit 21b, and external fluctuation determination unit 21c are the same as the difference value calculation unit 21a, peak detection unit 21b, and external fluctuation determination unit 21c of the external fluctuation removal unit 21 in the river flow prediction device 101 of embodiment 2.

[0081] The rainfall information acquisition unit 22a acquires regional rainfall data in a rainfall observation region that affects flow rate fluctuations due to rainfall at observation points in a river, and accumulates rainfall time-series data that represents basin rainfall data in the rainfall observation region in a time series. The basin rainfall data in the rainfall observation area includes measured rainfall data from rain gauges in the rainfall observation area, regional rainfall data from radar rainfall in the rainfall observation area using weather radar, regional rainfall data in the rainfall observation area from a weather information service via the Internet or a dedicated line, or regional rainfall data in the rainfall observation area from a river information center. The regional rainfall data for the rainfall observation region may be actual rainfall up to the present time or predicted rainfall in the future.

[0082] The rainfall detection unit 22b detects the start time and end time of the flow rate increase due to rainfall using the rainfall time series data from the rainfall information acquisition unit 22a, and obtains estimated flow rate data B by converting the flow rate data for the rainfall period from the start time of the flow rate increase to the end time of the flow rate increase in the estimated flow rate data B up to the current time obtained by the interpolation unit 21d into flow rate data based on rainfall, and provides the obtained estimated flow rate data B to the flow rate prediction unit 3. The estimated flow rate data B obtained by the rainfall detection unit 22b is past flow rate data up to the present time t that estimates the flow rate of the river's own flow taking into account rainfall at the observation point, excluding flow changes due to external fluctuations.

[0083] The start time of the increase in flow rate due to rainfall detected by the rainfall detection unit 22b is the time a set time after rainfall based on the regional rainfall data acquired by the rainfall information acquisition unit 22a, the time at which the actual value indicated by the actual flow rate data from the actual flow rate acquisition unit 1 becomes equal to or exceeds a set threshold within the set time after rainfall, or the time at which the primary flow rate difference value q1(t) calculated by the difference value calculation unit 21a rises to or exceeds a set threshold within the set time after rainfall.

[0084] The end time of the flow rate increase due to rainfall detected by the rainfall detection unit 22b is the time after the flow rate increase start time when the actual value indicated by the actual flow rate data from the actual flow rate acquisition unit 1 becomes equal to or less than the flow rate estimated value before the flow rate increase start time, the time after the flow rate increase start time when the actual value indicated by the actual flow rate data from the actual flow rate acquisition unit 1 becomes equal to or less than a set threshold, or the time after the flow rate increase start time when the secondary flow rate difference value q2(t) calculated by the difference value calculation unit 21a reaches its maximum value and the falling point when the primary flow rate difference value q1(t) becomes equal to or less than a set threshold.

[0085] The flow rate data based on rainfall obtained by the rainfall detection unit 22b is actual flow rate data obtained by the actual flow rate acquisition unit 1. Furthermore, the flow rate data based on rainfall obtained by the rainfall detection unit 22b may be flow rate data obtained by adding a flow rate increase value obtained by using rainfall time-series data obtained by the rainfall information acquisition unit 22a to the estimated flow rate data in the estimated value flow rate data B.

[0086] In short, the rainfall detection unit 22b detects a rainfall period and sets estimated rainfall data during the rainfall period based on the regional rainfall data acquired by the rainfall information acquisition unit 22a, or detects a rainfall period and sets estimated rainfall data during the rainfall period as actual flow rate data based on the regional rainfall data acquired by the rainfall information acquisition unit 22a and the actual flow rate data from the actual flow rate acquisition unit 1.

[0087] The estimated water volume data B obtained by the rainfall detection unit 22b is shown in FIG. In FIG. 11, the horizontal axis represents time and the vertical axis represents flow rate. Since the actual flow rate data A and the estimated flow rate data B are the same except for the periods indicated by the flow rate fluctuation data a1 and a2, both are shown by solid curves. In addition, in Figure 11, b1 and b2 represent estimated flow rate data linearly interpolated by the interpolation unit 21d, e1 represents the start time of the increase in flow rate due to rainfall, e2 represents the end time of the increase in flow rate due to rainfall, e represents actual flow rate data A and estimated flow rate data B during the rainfall period, and E represents the basin rainfall data, which is the amount of rainfall.

[0088] As can be seen from the estimated value flow rate data B shown in Figure 11, there is a clear distinction between the actual flow rate values ​​indicated by the flow rate fluctuation data a1, a2 due to external factors in the actual value flow rate data A and the actual flow rate values ​​due to rainfall in the actual value flow rate data A. The actual flow rate values ​​due to external factors are excluded from the actual value flow rate data A, and estimated value flow rate data linearly interpolated to the actual value flow rate data A is used, and the actual flow rate values ​​in the actual value flow rate data A are used as they are during the rainfall period to obtain estimated value flow rate data B. Therefore, the estimated flow rate data B is past flow rate data up to the current time t that estimates the flow rate of the river's natural flow taking into account rainfall at the observation point, excluding flow changes due to external fluctuations.

[0089] The river flow rate prediction device 102 according to the third embodiment is realized by a computer hardware configuration shown in FIG. 5, similar to the river flow rate prediction device 101 according to the second embodiment. Next, the operation of the river flow rate prediction device 102 according to the third embodiment, that is, the river flow rate prediction method, will be described with reference to FIG.

[0090] Step ST1 and steps ST21a to ST21d are respectively the same as step ST1 in the river flow prediction device 101 of embodiment 2, which is the step of acquiring actual flow data A, step ST21a, which is the step of calculating the secondary flow difference value q2(t), step ST21b, which is the peak detection step of obtaining the inflection point time, step ST21c, which is the determination step of determining the flow fluctuation period, and step ST21d, which is the flow estimation step of obtaining estimated flow data B.

[0091] In step ST22a, the rainfall information acquisition unit 22a acquires basin rainfall data in the rainfall observation area, accumulates rainfall time-series data e that represents the basin rainfall data in the rainfall observation area in a time series, and then proceeds to step ST22b. Step ST22a is a step for acquiring the amount of rainfall from the rainfall time-series data e.

[0092] In step ST22b, the rainfall detection unit 22b detects the start time e1 of the water volume increase due to rainfall and the end time e2 of the water volume increase, and obtains the flow data for the rainfall period from the start time e1 of the water volume increase to the end time e2 of the water volume increase in the estimated flow data B as flow data based on rainfall, which in embodiment 3 is the actual flow data A, and proceeds to step ST3. Step ST22b is a flow rate estimation step in which the flow rate due to the natural flow of the river is estimated taking into account rainfall at the observation point, and estimated flow rate data B is obtained.

[0093] Steps ST3 and ST4 are the same as step ST3 in the river flow rate prediction device 101 according to the second embodiment, which is the prediction step of obtaining future predicted value flow rate data C, and step ST4, which is the data output step of outputting predicted value flow rate data C.

[0094] In the river flow rate prediction device 102 according to the third embodiment, the river flow rate prediction method from step ST1 to step ST4 is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C).

[0095] That is, the program stored in ROM 10C includes a data acquisition procedure for acquiring actual water volume data indicating actual water volumes acquired at observation points and storing the data as actual time-series water volume data; a calculation procedure for using the actual time-series water volume data to calculate a first water volume difference value, which is the difference between the actual water volume data at the current time and the actual water volume data at the time immediately preceding the current time, and a second water volume difference value, which is the difference between the first water volume difference value at the current time and the first water volume difference value at the time immediately preceding the current time, and storing the data as second water volume difference value time-series data; a peak detection procedure for using the second water volume difference value time-series data to detect a maximum value in the second water volume difference value and setting the time corresponding to the detected maximum value as an inflection point time; a discrimination procedure for using the inflection point time to discriminate a water volume fluctuation period indicating a temporary water volume fluctuation; The system comprises a first estimation procedure that removes actual water volume data during the water volume fluctuation period from the time-series water volume data and obtains first estimated water volume data up to the current time by linearly interpolating estimated water volume data during the water volume fluctuation period; a rainfall acquisition procedure that acquires basin rainfall data in the rainfall observation area and accumulates rainfall time-series data that represents the basin rainfall data in the rainfall observation area in a time series; a second estimation procedure that detects the start time and end time of the water volume increase due to rainfall and obtains second estimated water volume data by using the water volume data for the rainfall period from the start time of the water volume increase to the end time of the water volume increase in the first estimated water volume data as water volume data based on rainfall; a prediction procedure that obtains future predicted water volume data using the second estimated water volume data; and a data output procedure that outputs the predicted water volume data.

[0096] The river flow rate prediction device 102 according to the third embodiment has the same effect as the river flow rate prediction device 101 according to the second embodiment, and in addition, the rainfall detection unit 22b detects the rainfall period and obtains estimated water volume data using the flow rate data during the rainfall period as flow rate data based on rainfall, so that the flow rate data can be obtained that can accurately predict the river flow at the observation point during rainfall.

[0097] The river flow rate prediction device 102 of embodiment 3 is a river flow rate prediction device equipped with an external fluctuation removal unit 22 in which a rainfall detection unit 22b and a rainfall information acquisition unit 22a are added to the external fluctuation removal unit 21 in the river flow rate prediction device 101 of embodiment 2, but it may also be a river flow rate prediction device equipped with an external fluctuation removal unit 22 in which a rainfall detection unit 22b and a rainfall information acquisition unit 22a are added to the external fluctuation removal unit 2 in the river flow rate prediction device 100 of embodiment 1.

[0098] In this case, the river flow rate prediction method executed in the river flow rate prediction device 102 according to the third embodiment is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C). That is, the program stored in ROM 10C includes a data acquisition procedure for acquiring actual water volume data indicating actual water volumes acquired at observation points and storing the data as actual time-series water volume data, and a first estimation procedure for using the actual time-series water volume data to determine water volume fluctuation data indicating temporary water volume fluctuations in the actual time-series water volume data, removing the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from the actual time-series water volume data, and interpolating the removed actual flow rate data with estimated flow rate data to obtain estimated water volume data up to the current time. The system includes a rainfall acquisition procedure for acquiring basin rainfall data in a rainfall observation area and accumulating rainfall time series data that represents the basin rainfall data in the rainfall observation area in a time series; a second estimation procedure for detecting the start time and end time of the water volume increase due to rainfall and obtaining estimated water volume data by using the flow rate data for the rainfall period from the start time of the water volume increase to the end time of the water volume increase in the estimated water volume data as water volume data based on rainfall; a prediction procedure for obtaining future predicted water volume data using the estimated water volume data; and a data output procedure for outputting the predicted water volume data.

[0099] Embodiment 4 A water quantity prediction device according to a fourth embodiment will be described with reference to FIGS. The water volume prediction device according to the fourth embodiment will also be described using a river flow rate prediction device as an example, similar to the first embodiment. The river flow prediction device 103 of embodiment 4 differs from the river flow prediction device 100 of embodiment 1 in that information from the release of water from a dam upstream of the observation point is added to the prediction of flow fluctuations in the river's natural flow at the observation point, but is otherwise the same.

[0100] That is, the river flow prediction device 103 of embodiment 4 differs from the river flow prediction device 100 of embodiment 1 in that it adds a dam discharge volume acquisition unit 51, a dam discharge volume removal unit 52, and a dam discharge volume addition unit 53. Therefore, the following description will focus on the dam discharge amount acquisition unit 51, the dam discharge amount removal unit 52, and the dam discharge amount addition unit 53. 13 to 15, the same reference numerals as those in FIGS. 1 to 6 designate the same or corresponding parts.

[0101] FIG. 13 shows a river flow rate prediction system (water volume prediction system) including a river flow rate prediction device 103, a flow observation device 200, and a predicted flow rate display unit 300, which is a water volume prediction device according to the fourth embodiment. The river flow rate prediction device 103 according to the fourth embodiment uses time-series flow rate data obtained by subtracting time-series discharge data indicating the discharge rate of a dam located upstream of an observation point from actual value flow rate data A indicating the actual flow rate at the observation point acquired from the flow observation device 200 as virtual actual value time-series flow rate data, and calculates a virtual external fluctuation a 11 , a 21 , a 31 The interpolated estimated flow rate data B is obtained by linearly interpolating the virtual actual flow rate data that has been removed, and first predicted flow rate data is obtained based on the estimated flow rate data B. The time-series discharge flow rate data is added to the first predicted flow rate data, and second predicted flow rate data that predicts the river's natural flow at the observation point is output to the predicted flow rate display unit 300.

[0102] The river flow prediction device 103 of embodiment 4 includes an actual flow acquisition unit 1, an external fluctuation removal unit 2, a flow prediction unit 3, a predicted flow output unit 4, a dam discharge acquisition unit 51, a dam discharge removal unit 52, and a dam discharge addition unit 53. The actual flow rate acquisition unit 1, external fluctuation elimination unit 2, flow rate prediction unit 3, and predicted flow rate output unit 4 are essentially the same as the actual flow rate acquisition unit 1, external fluctuation elimination unit 2, flow rate prediction unit 3, and predicted flow rate output unit 4 in the river flow rate prediction device 100 of embodiment 1.

[0103] The dam discharge amount acquisition unit 51 receives discharge amount data indicating the discharge amount of a dam located upstream of the observation point and accumulates the data as time-series discharge amount data. The dam discharge volume is the current discharge volume communicated by the dam manager or the future dam discharge plan. The time-series discharge data F is shown in the lower diagram of Figure 14. The time-series discharge data F indicates the time-series discharge data released from the dam up to time T. In the lower diagram of Figure 14, the horizontal axis represents time and the vertical axis represents the amount of water discharged from the dam.

[0104] The dam discharge volume removal unit 52 provides the time series flow data obtained by subtracting the time series discharge volume data F from the dam discharge volume acquisition unit 51 from the actual time series flow volume data A from the actual flow volume acquisition unit 1 to the external fluctuation removal unit 2 as virtual actual time series water volume data A1 up to the current time from the actual flow volume acquisition unit 1. The time of the time-series discharge data F is delayed by the propagation delay time of the dam's discharge water from the dam to the observation point.

[0105] That is, the time of the time-series discharge data F and the time of the actual time-series discharge data A coincide at the observation point. The propagation delay time of the dam's discharge water is set in advance as a parameter. The propagation delay time of the dam's discharge water may be estimated by pairing the rising and falling times of fluctuations in the dam's discharge volume with the rising and falling times of the actual volume of discharge.

[0106] The hypothetical actual time-series water volume data A1 obtained by the dam discharge volume removal unit 52 is shown in the upper diagram of FIG. In the upper diagram of Figure 14, the horizontal axis represents time and the vertical axis represents flow rate, and it shows that there is a difference equivalent to the amount of water discharged from the dam between the actual time-series flow rate data A and the virtual actual time-series water volume data A1. External fluctuations in the hypothetical actual time-series water volume data A1 11 , a 21 , a 31 correspond to external fluctuations a1, a2, and a3 in the actual time-series flow data A.

[0107] The external fluctuation elimination unit 2 functions in the same way as the external fluctuation elimination unit 2 in the river flow rate prediction device 100 according to the first embodiment, and uses the virtual actual value time-series water volume data A1 obtained by the dam discharge volume elimination unit 52 to generate virtual flow rate fluctuation data a 11 , a 21 , a 31 The determined virtual flow rate fluctuation data a 11 , a 21 , a31 The temporary flow rate fluctuation element indicated by is removed from the virtual actual value time-series flow rate data A1, and the interpolated estimated value flow rate data B at the observation point up to the current time is obtained by linearly interpolating the removed virtual actual value flow rate data.

[0108] Estimated flow data B is an estimate of the river's natural flow at the observation point, with the discharge from the dam subtracted. In the upper diagram of FIG. 14, the estimated flow rate data is shown by curve B, and the virtual actual flow rate data A1 and the estimated flow rate data B are plotted against the virtual flow rate fluctuation data a 11 , a 21 , a 31 Since the periods other than the period indicated by are the same, both are shown by solid curves.

[0109] The flow rate prediction unit 3 functions in the same manner as the flow rate prediction unit 3 in the river flow rate prediction device 100 according to embodiment 1, and obtains future predicted flow rate data at the observation point using the estimated flow rate data B up to the current time obtained by the external fluctuation removal unit 2. 14 shows estimated flow rate data B1 obtained by adding the time-series discharge amount data F from the dam discharge amount acquisition unit 51 to the estimated flow rate data B.

[0110] The actual value time-series flow rate data A and the estimated value flow rate data B1 are the same except for the periods indicated by the flow rate fluctuation data a1, a2, and a3, and therefore both are shown by solid curves. The estimated flow rate data B1 is an estimate of the river's natural flow at the observation point to which the discharge from the dam has been added.

[0111] The dam discharge amount addition unit 53 adds the time-series discharge amount data from the dam discharge amount removal unit 52 to the future predicted flow amount data from the flow amount prediction unit 3 and provides the resulting flow amount data to the predicted flow amount output unit 4 as future predicted water volume data. The predicted flow rate output unit 4 functions in the same manner as the predicted flow rate output unit 4 in the river flow rate prediction device 100 of embodiment 1, and outputs the predicted flow rate data obtained by the dam discharge rate addition unit 53 to at least the predicted flow rate display unit 300.

[0112] The river flow rate prediction device 101 according to the fourth embodiment is realized by a computer hardware configuration shown in FIG. 5, similar to the river flow rate prediction device 100 according to the first embodiment. Next, the operation of the river flow rate prediction device 101 according to the fourth embodiment, that is, the river flow rate prediction method, will be described with reference to FIG.

[0113] Step ST1 is the same as step ST1 in the river flow rate prediction device 100 according to the first embodiment, which is the data acquisition step of the actual flow rate data. In step ST11, the dam discharge amount acquisition unit 51 acquires dam discharge amount data indicating the discharge amount of a dam located upstream of the observation point, accumulates the data as time-series discharge amount data, and proceeds to step ST12. Step ST11 is a step for acquiring the discharge amount from the time-series discharge amount data.

[0114] In step ST12, the dam discharge removal unit 52 obtains virtual actual value time series flow data A1 by subtracting the time series discharge data F from the dam discharge acquisition unit 51 from the actual value time series flow data A from the actual flow acquisition unit 1, and proceeds to step ST2. In step ST2, the external fluctuation elimination unit 2 uses the virtual actual value time-series flow rate data A1 to generate virtual flow rate fluctuation data a 11 , a 21 , a 31 and calculate the virtual flow rate fluctuation data a from the virtual actual time-series water volume data A1. 11 , a 21 , a 31 The virtual actual flow rate data thus removed is interpolated with the estimated flow rate data to obtain estimated flow rate data B up to the present time, and the process proceeds to step ST3. Step ST2 is an estimation step in which external fluctuations are removed and the past river flow rate up to the current time t is estimated.

[0115] In step ST3, the flow rate prediction unit 3 obtains future predicted flow rate data at the observation point using the estimated flow rate data B, and the process proceeds to step ST31. Step ST3 is the first prediction step for predicting the future river flow rate from the current time t.

[0116] In step ST31, the dam discharge amount addition unit 53 obtains future predicted flow rate data by adding the time-series discharge amount data F from the dam discharge amount acquisition unit 51 to the future predicted flow rate data from the flow rate prediction unit 3, and then proceeds to step ST4. Step ST31 is the second prediction step for predicting the future river flow rate from the current time t. In step ST 4 , the predicted flow rate output unit 4 outputs the predicted flow rate data obtained by the dam discharge flow rate addition unit 53 to the predicted flow rate display unit 300 .

[0117] The river flow rate prediction method from step ST1 to step ST4 is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C). That is, the program stored in ROM 10C includes an actual data acquisition procedure for acquiring actual water volume data indicating actual water volumes acquired at observation points and storing the data as actual time-series water volume data; an outflow data acquisition procedure for acquiring dam outflow data indicating outflow volumes of dams located upstream of the observation points and storing the data as time-series outflow volume data; a virtual actual data calculation procedure for obtaining virtual actual time-series water volume data by subtracting the time-series outflow volume data from the actual time-series water volume data; and a virtual actual data calculation procedure for calculating a value in the virtual actual time-series water volume data by using the virtual actual time-series water volume data. The system comprises an estimation procedure for determining water volume fluctuation data that indicates temporal water volume fluctuations, removing the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from virtual actual value time-series water volume data, and obtaining estimated value water volume data up to the current time by interpolating the removed virtual actual value water volume data using estimated value water volume data; a first prediction procedure for obtaining first predicted value water volume data in the future using the estimated value water volume data; a first prediction procedure for obtaining second predicted value water volume data in the future by adding time-series discharge amount data to the first predicted value water volume data; and a data output procedure for outputting the second predicted value water volume data.

[0118] The river flow rate prediction device 103 of embodiment 4 has the same effects as the river flow rate prediction device 100 of embodiment 1, and in addition, the dam discharge volume removal unit 52 obtains virtual actual value time series water volume data A1 by subtracting the time series discharge volume data F from the dam discharge volume acquisition unit 51 from the actual value time series water volume data A from the actual water volume acquisition unit 1, and the external fluctuation removal unit 2 obtains estimated value flow rate data B using the virtual actual value time series water volume data A1, so that flow rate data can be obtained that can accurately predict the river flow at the observation point when a dam located upstream of the observation point discharges water.

[0119] The river flow rate prediction device 103 of embodiment 4 is a river flow rate prediction device in which a dam discharge amount acquisition unit 51, a dam discharge amount removal unit 52, and a dam discharge amount addition unit 53 are added to the river flow rate prediction device 100 of embodiment 1, but it may also be a river flow rate prediction device in which a dam discharge amount acquisition unit 51, a dam discharge amount removal unit 52, and a dam discharge amount addition unit 53 are added to the river flow rate prediction device 101 of embodiment 2.

[0120] In this case, the river flow rate prediction method executed in the river flow rate prediction device 103 according to the fourth embodiment is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C).

[0121] That is, the program stored in ROM 10C includes an actual data acquisition procedure for acquiring actual water volume data indicating actual water volumes acquired at an observation point and storing it as actual time-series water volume data; a discharge data acquisition procedure for acquiring dam discharge volume data indicating the discharge volume of a dam upstream of the observation point and storing it as time-series discharge volume data; a virtual actual data calculation procedure for obtaining virtual actual time-series water volume data by subtracting the time-series discharge volume data from the actual time-series water volume data; and a virtual primary water volume difference value for using the virtual actual time-series water volume data to calculate a virtual primary water volume difference value, which is the difference between the virtual actual water volume data at the current time and the virtual actual water volume data at the time immediately preceding the current time, and a virtual secondary water volume difference value, which is the difference between the virtual primary water volume difference value at the current time and the virtual primary water volume difference value at the time immediately preceding the current time, and calculating the virtual secondary water volume difference value time. The system comprises a calculation procedure for accumulating the virtual secondary water volume difference data as series data, a peak detection procedure for detecting a maximum value in the virtual secondary water volume difference value using the virtual secondary water volume difference value time series data and setting the time corresponding to the detected maximum value as the inflection point time, a discrimination procedure for discriminating a water volume fluctuation period indicating temporary water volume fluctuations using the inflection point time, an estimation procedure for removing virtual actual value water volume data during the water volume fluctuation period from the virtual actual value time series water volume data and obtaining estimated value water volume data up to the present time by linearly interpolating the estimated value water volume data during the water volume fluctuation period, a first prediction procedure for obtaining first predicted value water volume data in the future using the estimated value water volume data, a first prediction procedure for obtaining second predicted value water volume data in the future by adding time series discharge amount data to the first predicted value water volume data, and a data output procedure for outputting the second predicted value water volume data.

[0122] In addition, the river flow rate prediction device 103 of embodiment 4 may be a river flow rate prediction device in which a dam discharge amount acquisition unit 51, a dam discharge amount removal unit 52, and a dam discharge amount addition unit 53 are added to the river flow rate prediction device 102 of embodiment 3. In this case, the river flow rate prediction method executed in the river flow rate prediction device 103 according to the fourth embodiment is performed by the CPU 10A executing processes in accordance with a program stored in the RAM 10B (ROM 10C).

[0123] That is, the program stored in ROM 10C includes a performance data acquisition procedure for acquiring performance value water volume data indicating the performance water volume acquired at the observation point and storing it as performance value time-series water volume data, a discharge data acquisition procedure for acquiring dam discharge volume data indicating the discharge volume of a dam located upstream of the observation point and storing it as time-series discharge volume data, a virtual performance data calculation procedure for obtaining virtual performance value time-series water volume data by subtracting the time-series discharge volume data from the performance value time-series water volume data, and a calculation procedure for calculating virtual performance value time-series water volume data by using the virtual performance value time-series water volume data. a calculation step of calculating a virtual first water quantity difference value, which is the difference between virtual actual water quantity data at the current time and virtual actual water quantity data at the time immediately preceding the current time, and a virtual second water quantity difference value, which is the difference between the virtual first water quantity difference value at the current time and the virtual first water quantity difference value at the time immediately preceding the current time, and storing these as virtual second water quantity difference value time series data; and a peak detection step of detecting a maximum value in the virtual second water quantity difference value using the virtual second water quantity difference value time series data, and setting the time corresponding to the detected maximum value as an inflection point time. a determination procedure for determining a water volume fluctuation period that indicates temporary water volume fluctuations using an inflection point time; a first estimation procedure for removing virtual actual value water volume data in the water volume fluctuation period from virtual actual value time-series water volume data and linearly interpolating estimated value water volume data in the water volume fluctuation period to obtain first estimated value water volume data up to the current time; a rainfall acquisition procedure for acquiring basin rainfall data in a rainfall observation area and accumulating rainfall time-series data that represents the basin rainfall data in the rainfall observation area in a time series; and a start time of water volume increase due to rainfall. The system comprises a second estimation procedure for detecting the end time of the water volume increase and obtaining second estimated water volume data by using water volume data for the rainfall period from the start time of the water volume increase to the end time of the water volume increase in the first estimated water volume data as water volume data based on rainfall; a first prediction procedure for obtaining first predicted water volume data for the future using the second estimated water volume data; a first prediction procedure for obtaining second predicted water volume data for the future by adding time-series discharge volume data to the first predicted water volume data; and a data output procedure for outputting the second predicted water volume data.

[0124] It should be noted that the embodiments may be freely combined, any of the components of the embodiments may be modified, or any of the components of the embodiments may be omitted. [Industrial Applicability]

[0125] The water volume prediction device of the present disclosure can be applied to a river flow rate prediction device that predicts the flow rate at an observation point on a river, a river water level prediction device that predicts the water level at an observation point on a river, or a dam inflow prediction device that predicts the inflow volume of a dam that is an observation point, and is particularly suitable for a river flow rate prediction device. [Explanation of symbols]

[0126] 100 River flow prediction device, 1 Actual flow acquisition unit, 2, 21 External fluctuation removal unit, 21a Difference value calculation unit, 21b Peak detection unit, 21c External fluctuation determination unit, 21d Interpolation unit, 22a Rainfall information acquisition unit, 22b Rainfall detection unit, 3 Flow prediction unit, 4 Predicted flow output unit, 51 Dam discharge acquisition unit, 52 Dam discharge removal unit, 53 Dam discharge addition unit, 200 Flow observation device.

Claims

1. an actual water volume acquisition unit that receives actual water volume data indicating actual water volumes acquired at observation points and accumulates the data as actual water volume time-series data; an external fluctuation removal unit that uses the actual water volume time-series water volume data up to the current time obtained by the actual water volume acquisition unit to determine water volume fluctuation data that indicates temporary water volume fluctuations in the actual water volume time-series water volume data, removes the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from the actual water volume time-series water volume data, and obtains estimated water volume data up to the current time by interpolating the removed actual water volume data with estimated water volume data; a water volume prediction unit that obtains future predicted water volume data using the estimated water volume data up to the current time obtained by the external fluctuation removal unit; a predicted water volume output unit that outputs the predicted water volume data obtained by the water volume prediction unit; A water volume prediction device comprising:

2. The determination of the water volume fluctuation data in the external fluctuation elimination unit is carried out by detecting the rising and falling points of temporary water volume fluctuations using the water volume difference value between the actual water volume data at the current time and the actual water volume data at the time immediately before the current time, The estimated water volume data between the rising point and the falling point obtained by the external fluctuation elimination unit is water volume data obtained by interpolating the rising point and the falling point. The water volume prediction device according to claim 1 .

3. The water volume fluctuation data is discriminated in the external fluctuation elimination unit, and estimated water volume data at the time of discrimination is obtained using a low-pass filter or a Kalman filter. The water volume prediction device according to claim 1 .

4. an actual water volume acquisition unit that receives actual water volume data indicating actual water volumes acquired at observation points and accumulates the data as actual water volume time-series data; a difference value calculation unit that calculates a first water volume difference value, which is the difference between the actual water volume data for the current time and the actual water volume data for the time immediately preceding the current time, and a second water volume difference value, which is the difference between the first water volume difference value for the current time and the first water volume difference value for the time immediately preceding the current time, in the actual value time-series water volume data obtained by the actual water volume acquisition unit, and stores these as second water volume difference value time-series data; a peak detection unit that detects maximum values ​​in the second water volume difference value using the second water volume difference value time-series data obtained by the difference value calculation unit, and sets the time corresponding to the detected maximum value as an inflection point time; an external fluctuation determination unit that determines a water volume fluctuation period indicating temporary water volume fluctuation using the inflection point time obtained by the peak detection unit; and an external fluctuation removal unit that has an interpolation unit that removes actual value water volume data during the water volume fluctuation period determined by the external fluctuation determination unit from the actual value time-series water volume data obtained by the actual water volume acquisition unit, and obtains interpolated estimated value water volume data up to the current time by interpolating estimated value water volume data into the water volume fluctuation period determined by the external fluctuation determination unit; a water volume prediction unit that obtains future predicted water volume data using the estimated water volume data up to the current time obtained by the external fluctuation removal unit; a predicted water volume output unit that outputs the predicted water volume data obtained by the water volume prediction unit; A water volume prediction device comprising:

5. a rainfall information acquisition unit that acquires basin rainfall data in a rainfall observation area and accumulates rainfall time series data that represent the basin rainfall data in the rainfall observation area in a time series; a rainfall detection unit that detects the start time and end time of a water volume increase due to rainfall, obtains estimated water volume data by converting the water volume data for the rainfall period from the start time of the water volume increase to the end time of the water volume increase in the estimated water volume data into water volume data based on rainfall, and provides the obtained estimated water volume data to the water volume prediction unit; The water volume prediction device according to claim 1 , further comprising:

6. The water volume increase start time detected by the rainfall detection unit is the time at which a first-order water volume difference value, which is the difference between the actual water volume data at the current time and the actual water volume data at the time immediately before the current time in the actual value time-series water volume data obtained by the actual water volume acquisition unit, rises to or exceeds a set threshold value within a set time from rainfall based on the regional rainfall data obtained by the rainfall information acquisition unit, The water volume increase end time detected by the rainfall detection unit is the time of the falling point after the water volume increase start time at which the secondary water volume difference value, which is the difference between the primary water volume difference value at the current time and the primary water volume difference value at the time immediately before the current time, reaches a maximum value. The water volume prediction device according to claim 5.

7. a dam discharge volume acquisition unit that receives discharge volume data indicating the discharge volume of a dam located upstream of the observation point and accumulates the data as time-series discharge volume data; a dam discharge amount removal unit that provides time-series water volume data obtained by subtracting the time-series discharge amount data obtained by the dam discharge amount removal unit from the actual time-series water volume data obtained by the actual water volume removal unit to the external fluctuation removal unit as the actual time-series water volume data obtained by the actual water volume removal unit up to the current time; a dam discharge amount addition unit that adds the time-series discharge amount data obtained by the dam discharge amount acquisition unit to the future predicted value water volume data obtained by the water volume prediction unit and provides the water volume data to the predicted water volume output unit as future predicted value water volume data; The water volume prediction device according to claim 1 , further comprising:

8. a dam discharge volume acquisition unit that receives dam discharge volume data indicating the discharge volume of a dam located upstream of the observation point and accumulates the data as time-series discharge volume data; a dam discharge amount removal unit that provides time-series water volume data obtained by subtracting the time-series discharge amount data obtained by the dam discharge amount removal unit from the actual time-series water volume data obtained by the actual water volume removal unit to the external fluctuation removal unit as the actual time-series water volume data obtained by the actual water volume removal unit up to the current time; a dam discharge amount addition unit that adds the time-series discharge amount data obtained by the dam discharge amount acquisition unit to the future predicted value water volume data obtained by the water volume prediction unit and provides the water volume data to the predicted water volume output unit as future predicted value water volume data; The water volume prediction device according to claim 5 , further comprising:

9. The water volume prediction device according to any one of claims 1 to 4, a water quantity observation device that acquires the actual water quantity at the observation point and outputs the acquired actual water quantity as actual value water quantity data to the water quantity prediction device; a predicted water volume display unit that displays predicted water volume data from the water volume prediction device; A water volume prediction system equipped with:

10. A water volume prediction method in a water volume prediction device including an actual water volume acquisition unit, an external fluctuation removal unit, a water volume prediction unit, and a predicted water volume output unit, The actual water volume acquisition unit acquires actual water volume data indicating actual water volumes acquired at observation points and stores the data as actual time-series water volume data; the external fluctuation removal unit uses the actual value time-series water volume data to determine water volume fluctuation data that indicates temporary water volume fluctuations in the actual value time-series water volume data, removes the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from the actual value time-series water volume data, and obtains estimated water volume data up to the current time by interpolating the removed actual value water volume data with estimated value water volume data; a step in which the water volume prediction unit obtains future predicted water volume data using the estimated water volume data; a step in which the predicted water volume output unit outputs the predicted water volume data; A water volume prediction method comprising:

11. A procedure for acquiring actual water volume data indicating actual water volumes acquired at observation points and accumulating the data as actual time-series water volume data; a step of using the actual value time-series water volume data to determine water volume fluctuation data that indicates temporary water volume fluctuations in the actual value time-series water volume data, removing the temporary water volume fluctuation elements indicated by the determined water volume fluctuation data from the actual value time-series water volume data, and interpolating the removed actual value water volume data with estimated value water volume data to obtain estimated value water volume data up to the current time; A procedure for obtaining future predicted water volume data using the estimated water volume data; a step of outputting the predicted water volume data; A water volume prediction program that runs on a computer.

12. A procedure for acquiring actual water volume data indicating actual water volumes acquired at observation points and accumulating the data as actual time-series water volume data; a step of using the actual value time-series water quantity data to determine water quantity fluctuation data that indicates temporary water quantity fluctuations in the actual value time-series water quantity data, and removing the temporary water quantity fluctuation elements indicated by the determined water quantity fluctuation data from the actual value time-series water quantity data to obtain estimated value water quantity data up to the current time; A procedure for obtaining future predicted water volume data using the estimated water volume data; a step of outputting the predicted water volume data; A recording medium storing a program that causes a computer to execute the above.

Images