Power generation plan presentation device

Abstract

To provide a power generation plan presentation device capable of searching for an optimal solution satisfying a condition of being zero or equal to or more than a minium water-intake amount and equal to or less than a maximum water-intake amount by linear programming while reducing a calculation load.SOLUTION: A PU searches for an optimal solution by linear programming as a temporary constraint condition that a constraint condition of a water-intake amount is equal to or more than zero and equal to or less than a maximum water-intake amount. The PU makes the constraint condition a condition that the water-intake amount is zero when the water-intake amount indicated by a search result is less than a minimum water-intake amount. The PU makes the constraint condition a condition of being equal to or more than the minimum water-intake amount and equal to or less than the maximum water-intake amount when the water-intake amount indicated by the search result is equal to or more than the minimum water-intake amount.SELECTED DRAWING: Figure 10B

Description

Technical Field

[0001] The present invention relates to a power generation plan presentation device.

Background Art

[0002] The following Patent Document 1 describes a system for formulating power generation plans for a plurality of hydroelectric power plants. This system formulates plans for the power generation amounts at each of the plurality of hydroelectric power plants using linear programming. First, this system formulates a provisional power generation plan using a constraint condition that the amount of water used for power generation is equal to or greater than the lower limit value and equal to or less than the upper limit value. Then, when the output of the generator in the provisional power generation plan is below a predetermined threshold value, a constraint condition that the amount of water used for power generation is zero is added, and the final power generation plan is formulated using linear programming.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in the above case, in the second linear programming, there is a possibility that the output of the power plant subject to the constraint condition that the amount of water used for power generation is equal to or greater than the lower limit value and equal to or less than the upper limit value is below the above-mentioned predetermined threshold value.

Means for Solving the Problems

[0005] Hereinafter, means for solving the above problems and their operational effects will be described. 1. A power generation plan presentation device for planning the water intake amount by a power plant that generates power using a plurality of dams connected by a river, comprising an execution device configured to execute an optimization process, wherein the optimization process is a process of outputting a solution of the water intake amount that optimizes the value of a predetermined objective function and includes a search process and a constraint condition change process, the search process includes a process of searching for the water intake amount of at least a part of the plurality of dams by the linear programming method under a predetermined constraint condition, the constraint condition change process is a process of changing the predetermined constraint condition to the main constraint condition after the search process is executed under a temporary constraint condition as the predetermined constraint condition, the temporary constraint condition is a condition that the water intake amount is equal to or more than zero and equal to or less than the maximum water intake amount, the main constraint condition is a condition that the water intake amount is zero when the water intake amount searched by the search process under the temporary constraint condition is less than the minimum water intake amount, and the main constraint condition is a condition that the water intake amount is equal to or more than the minimum water intake amount and equal to or less than the maximum water intake amount when the water intake amount searched by the search process under the temporary constraint condition is equal to or more than the minimum water intake amount, and the minimum water intake amount is a value greater than zero.

[0006] The possible values of the water intake of the power plant may be zero or not less than the minimum water intake and not more than the maximum water intake. In that case, in order to make this condition a condition of linear programming, it may be considered to use mixed integer linear programming among linear programming methods. However, in that case, the computational load becomes extremely large. Therefore, in the above configuration, while adopting the condition that it is not less than zero and not more than the maximum water intake as a temporary constraint condition, the optimal solution is searched by linear programming. Since this is an ordinary constraint condition, it can be executed while suppressing the computational load. And when the obtained solution is less than the minimum water intake, this constraint condition is set to the condition that the water intake is zero. Also, when the obtained solution is not less than the minimum water intake, this constraint condition is set to the condition that the water intake is not less than the minimum water intake and not more than the maximum water intake. Thereby, the search for the optimal solution by linear programming using this constraint condition can also be executed while suppressing the computational load. Therefore, in the above configuration, it is possible to search for an optimal solution that satisfies the condition that the possible value of the water intake of the power plant is zero or not less than the minimum water intake and not more than the maximum water intake while reducing the computational load.

[0007] 2. At least a part of the plurality of dams includes two or more dams, and the constraint condition change process includes a process of selecting a part of the two or more dams and changing the constraint condition of the selected part of the dams to the present constraint condition. The search process is executed each time the predetermined constraint condition is changed by the constraint condition change process, and the constraint condition change process is repeatedly executed each time the search process is executed until there is no dam to which the temporary constraint condition is imposed among the two or more dams. The power generation plan presentation device according to item 1 above.

[0008] In the above configuration, the process of searching for the optimal solution in a state where the constraint conditions of a part of the two or more dams are changed to the present constraint conditions is executed until all the constraint conditions of the two or more dams are changed to the present constraint conditions. Thereby, compared with the case where all the constraint conditions of two or more dams are simultaneously changed from the temporary constraint conditions to the present constraint conditions, it is possible to suppress the occurrence of a situation where the optimal solution cannot be found.

[0009] In addition, when a change is newly made from the temporary constraint conditions to the main constraint conditions, the optimal solution is also searched for the water intake of the dams that had been changed to the main constraint conditions previously. Therefore, since the change in the water intake of the dams that had been changed to the main constraint conditions previously is allowed, the optimal solution can be searched for flexibly.

[0010] 3. The constraint condition change process of the power generation plan presentation device according to the above 2 includes a process of selecting some of the two or more dams that change the constraint conditions to the main constraint conditions according to a preset priority order.

[0011] Since the temporary constraint conditions are broader than the main constraint conditions, it is easier to find the optimal solution under the temporary constraint conditions. Therefore, depending on which of the two or more dams to shift to the main constraint conditions, the ease of obtaining the solution may vary. For example, when changing from a dam where it is difficult to obtain the optimal solution to the main constraint conditions, since the temporary constraint conditions are set for the remaining dams, it is easy to increase the degree of freedom in obtaining the optimal solution for the dam where it is difficult to obtain the optimal solution.

[0012] On the other hand, for example, which of the two or more dams is likely to obtain the optimal solution that satisfies the constraint conditions depends on the nature of the river upstream of the dam that affects the water inflow to the dam, the constraints of the power plant associated with the dam, etc. Therefore, it is possible to grasp to some extent in advance the dams among the two or more dams for which it is difficult to obtain the optimal solution that satisfies the constraint conditions. Therefore, to some extent in advance, it is possible to grasp which of the two or more dams is likely to obtain the optimal solution by shifting to the main constraint conditions.

[0013] Therefore, in the above configuration, by presetting the priority order of the dams that change the constraint conditions to the main constraint conditions, it is possible to increase the possibility of obtaining the optimal solution for all of the two or more dams.

[0014] 4. The water intake volume for which the search is performed by the search process is the volume at each time when a predetermined period is divided into a plurality of unit times. When there is no solution by the search process for the dam changed to the present constraint condition, the constraint condition change process includes a second change process of changing the constraint conditions of all the water intake volumes for the plurality of unit times related to the dam for which no solution is found to a condition that the volume is equal to or greater than the minimum water intake volume and equal to or less than the maximum water intake volume, without selecting a new dam to be changed to the present constraint condition. The power generation plan presentation device according to any one of 1 to 3 above.

[0015] According to the constraint condition change process, a condition that the water intake volume is zero may be set. In that case, since the water intake volume for which the condition of being zero is set cannot be adjusted, there is a concern that it becomes difficult to find an optimal solution for the water intake volume as a whole due to the constraint condition of being zero. Therefore, in the above configuration, when there is no solution in the search process, the constraint conditions of the water intake volumes for all time periods are set to a condition that the volume is equal to or greater than the minimum water intake volume and equal to or less than the maximum water intake volume. Thereby, the possibility of finding an optimal solution that satisfies the constraint conditions can be increased.

[0016] 5. The dam different from the two or more dams includes a discrete dam in which only a plurality of different values are allowed as the water intake volume. The search process includes a process of searching for a solution that optimizes the objective function together with the water intake volumes of the two or more dams while setting the water intake volume to each of the plurality of values for the discrete dam. The power generation plan presentation device according to any one of 1 to 4 above.

[0017] In the above configuration, it is possible to simultaneously search for an optimal solution for the water intake volume of a discrete dam that can only take discrete values. 6. The plurality of dams includes an ALR dam, and the execution device is configured to execute a water storage amount initial value acquisition process and a stream flow rate acquisition process. The water storage amount initial value acquisition process is a process of acquiring the initial value of the water storage amount in the ALR dam, and the stream flow rate acquisition process is a process of acquiring the stream flow rate flowing into the ALR dam. The optimization process includes a process of substituting the sum of the discharge amount of the upstream dam adjacent to the ALR dam and the water intake amount and the stream flow rate flowing into the ALR dam into the inflow amount of the ALR dam, and a process of updating the initial value of the water storage amount according to the inflow amount of the ALR dam. The search process is a process of searching for the water intake amount further including a condition that the water storage amount of the ALR dam is equal to or less than the maximum water storage amount. The maximum water storage amount is set to the smaller of the values calculated by two linear functions with the inflow amount into the ALR dam as an independent variable. The power generation plan presentation device according to any one of 1 to 5 above.

[0018] The upper limit water level of the ALR dam varies depending on the inflow amount into the ALR dam. And the relationship between the upper limit water level and the inflow amount has non-linearity. On the other hand, a non-linear relationship is difficult to handle with linear programming. Therefore, in the above configuration, attention is paid to the fact that the relationship between the upper limit water level and the inflow amount is a convex upward curve. And in the above configuration, the curve is approximated by the smaller of two linear functions. Thereby, a linear relationship that is easy to handle with linear programming can be introduced.

[0019] 7. In the power generation plan presentation device according to 6 above, the linear function is such that the relationship between the water level of the dam and the water storage amount of the dam is approximated by a straight line connecting the upper limit water level of the dam and the corresponding water storage amount and the lower limit water level and the corresponding water storage amount.

[0020] When the relationship between the water level of the dam and the water storage amount of the dam has non-linearity and the upper limit water level varies, it is difficult to replace the upper limit water level with the maximum water storage amount. Therefore, in the above configuration, in the region of the upper limit water level and the lower limit water level of the dam, the relationship between the water level and the water storage amount is linearly approximated. Thereby, the relationship between the upper limit water level and the maximum water storage amount can be approximated by a linear relationship.

[0021] 8. Among some of the plurality of dams, there are rising-restricted dams, which are configured to execute correction processing. The correction processing is a process of correcting to a predetermined rise when the rise of the water intake amount by the search processing regarding the rising-restricted dam is large. The power generation plan presentation device according to any one of 1 to 7 above.

[0022] When there is a restriction on the rise of the water intake amount in the dam, it is difficult to make this a constraint condition in the linear programming method. Therefore, in the above configuration, when the rise of the optimal solution of the water intake amount searched by the linear programming method is large, by executing the process of correcting the rise, a solution reflecting the rise restriction can be obtained.

[0023] 9. The objective function is configured using at least one of three functions: a function in which the smaller the total amount of the discharge water of the plurality of dams, the higher the evaluation; a function in which the larger the total amount of the power generation of the plurality of dams, the higher the evaluation; and a function in which the larger the selling electricity amount of the plurality of dams, the higher the evaluation. The power generation plan presentation device according to any one of 1 to 8 above.

[0024] According to the above objective function, an optimal solution corresponding to at least one of the three degrees of desire to minimize the discharge water amount, maximize the power generation amount, and maximize the selling electricity price can be obtained.

[0025] 10. The execution device is configured to execute a process for acquiring an initial value of the water storage volume and a process for acquiring the stream flow rate. The process for acquiring the initial value of the water storage volume is a process for acquiring the initial value of the water storage volume in each of the plurality of dams. The process for acquiring the stream flow rate is a process for acquiring the stream flow rate flowing into each of the plurality of dams. The optimization process includes a process of substituting, into the inflow amount to the dam, the sum of the water discharge amount of the dam upstream of the dam and the water intake amount and the stream flow rate flowing into the dam, and a process of updating the initial value of the water storage volume of the dam according to the inflow amount of the dam. The search process is a process of searching for the water intake amount, further including a condition that the water storage volume is equal to or greater than the minimum water storage volume and equal to or less than the maximum water storage volume. The power generation plan presentation device according to any one of 1 to 9 above.

[0026] In the above configuration, an optimal solution of the water intake amount that satisfies the condition that the water storage volume of the dam is equal to or greater than the minimum water storage volume and equal to or less than the maximum water storage volume can be obtained. 11. A storage device is provided, and model definition data is stored in the storage device. The model definition data is data for defining a regression model that outputs the stream flow rate when values of input variables are input. The input variables include at least one of a soil rainfall index and a soil runoff amount. The power generation plan presentation device according to 10 above is configured to execute a process for acquiring a predicted value of precipitation and a process for calculating input variables. The process for acquiring a predicted value of precipitation is a process for acquiring a predicted value of precipitation in a predetermined area near the dam. The process for calculating input variables is a process for calculating at least one of them using a tank model with the predicted value of precipitation as input. The process for acquiring the stream flow rate is a process for calculating the stream flow rate by inputting the values of the input variables calculated by the process for calculating input variables into the regression model.

[0027] In the above configuration, the stream flow rate can be calculated with high accuracy by using the soil rainfall index or the soil runoff amount. 12. The power generation plan presentation device according to 11 above, wherein the input variables include at least one time series data.

[0028] In the above configuration, compared with the case where a single value of the at least one variable is used as an input variable, more detailed information can be included in the input variable. Therefore, the prediction accuracy of the stream flow rate can be improved.

[0029] 13. The power generation plan presentation device according to claim 11 or 12, wherein the input variable includes the at least one value for each of a plurality of divided regions in a predetermined region near the dam. The influence on the stream flow rate may be different between different regions in the region near the dam. Therefore, in the above configuration, the input variable of the regression model is set as a variable for each of the regions obtained by dividing the predetermined region into a plurality of regions. As a result, compared with the case where the predetermined region is not subdivided, more detailed information can be included in the input variable of the regression model. Therefore, in the above configuration, the prediction accuracy of the stream flow rate can be improved compared with the case where the predetermined region is not subdivided.

[0030] 14. The power generation plan presentation device according to any one of claims 11 to 13, wherein the input variable includes a weighted average value of the at least one value for each of the regions obtained by dividing the predetermined region near the dam into a plurality of regions.

[0031] The influence on the stream flow rate may be different between different regions in the region near the dam. Therefore, in the above configuration, the input variable of the regression model is set as the weighted average value of the at least one value for each of the regions obtained by dividing the predetermined region into a plurality of regions. As a result, compared with the case where the predetermined region is not subdivided, information adapted to the actual situation of the predetermined region can be included in the input variable of the regression model. In other words, compared with the case where it is a simple average value of the values in a plurality of regions, information adapted to the actual situation of the predetermined region can be included in the input variable of the regression model. Therefore, in the above configuration, the prediction accuracy of the stream flow rate can be improved compared with the case where the predetermined region is not subdivided.

[0032] 15. A memory device is provided, and the memory device stores past plan data for each of the plurality of dams. The past plan data is data in which, for each of the plurality of dams, the transition of the water intake amount planned in the past is associated with the value of a situation variable that is a variable for specifying the situation on the target date for the planning of the water intake amount. The execution device is configured to execute a selection process, and the selection process is a process of selecting, from the past plan data, data in which the value of the situation variable is close to the value of the situation variable regarding the target date. The power generation plan presentation device according to any one of 1 to 14 above.

[0033] In the above configuration, by using the past plan data, it is possible to obtain information on the transition of the water intake amount actually planned on a past day in a situation similar to the target date. 16. The plurality of dams are divided into a plurality of groups, and the dams included in each of the plurality of groups are connected by the river. The selection process includes a process of selecting the transition of the water intake amount on the same past day for the water intake amounts of the grouped dams. The power generation plan presentation device according to 15 above.

[0034] For example, when the transition of the water intake amount on different past days is selected for each dam, there is a possibility that the amount of water flowing from upstream to downstream through the river is significantly different from the amount of water on the selected past day. And in that case, the possibility that an appropriate transition of the water intake amount, such as the water storage amount of the dam not satisfying the constraints, increases. Therefore, in the above configuration, for the grouped dams, by selecting the transition of the water intake amount on the same past day, it is possible to increase the possibility that the selected water intake amount is achievable.

[0035] 17. The selection process is a process of selecting, from the past plan data, data in which the value of the situation variable is close to the value of the situation variable on the target date for some of the dams within the group. The power generation plan presentation device according to 16 above.

[0036] In the above configuration, the values of the situation variables for some of the dams within the group are matched with the values of the situation variables in the past planned data. As a result, compared with the case of matching the values of the situation variables for all the dams within the group with the values of the situation variables in the past planned data, the computational load can be reduced.

[0037] 18. The power generation plan presentation device according to any one of the above 15 to 17, wherein the situation variable includes the water intake amount on the day before the target day and the amount of available water on the target day. In the above configuration, past plans can be retrieved using the factors that affect the planning of the transition of the water intake amount.

[0038] 19. The power generation plan presentation device according to any one of the above 1 to 18, wherein the execution device is configured to execute a presentation process, and the presentation process is a process of presenting information regarding the water intake amount output by the optimization process to a person.

[0039] In the above configuration, the water intake amount obtained by the optimization process can be presented to a person. 20. The power generation plan presentation device according to any one of claims 15 to 18, wherein the execution device is configured to execute a presentation process, and the presentation process is a process of presenting, to a person, the information regarding the water intake amount output by the optimization process and the information regarding the water intake amount output by the selection process.

[0040] In the above configuration, the transition of the water intake amount planned in the past and the optimal solution of the water intake amount explored from the perspective of optimizing the objective function independently of the past plan can be presented to a person. As a result, the planner of the water intake amount plan can formulate the final plan in consideration of both the information on what plans were made in the past and the information on what transition of the water intake amount is good for optimizing the objective function.

Brief Description of the Drawings

[0041]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10A

Figure 10B

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0042] Hereinafter, an embodiment will be described with reference to the drawings. 「Premise Configuration」 FIG. 1 shows the configuration of the power generation plan presentation system according to this embodiment. The power generation plan presentation device 10 shown in FIG. 1 includes a PU 12, a storage device 14, and a communication device 16. The PU 12 is a software processing device such as a CPU and a GPU. The storage device 14 may be a non-volatile memory that is electrically non-rewritable. The storage device 14 may also be a non-volatile memory that is electrically rewritable and a storage medium such as a disk medium.

[0043] 「Water System Targeted for Planning」 Fig. 2 exemplifies the water system that is the target of power generation plan formulation by the power generation plan presentation device 10 shown in Fig. 1.

[0044] The first dam 40(1), the second dam 40(2),... shown in Fig. 2 are provided in order from upstream to downstream. That is, the second dam 40(2) is provided downstream of the first dam 40(1). Assuming "i = 1, 2, 3,...", the water discharged from the i-th dam 40(i) is discharged to the downstream (i + 1)-th dam 40(i + 1) via the dam discharge line 42(i) and the power generation discharge line 44(i).

[0045] The dam discharge line 42(i) is a line that directly discharges water to the downstream dam. The power generation discharge line 44(i) is a line that discharges water to the downstream dam via the power plant 46(i).

[0046] Fig. 2 shows an example in which two dam discharge lines 42(m), 42(m + 1) and power generation discharge lines 44(m), 44(m + 1) are connected to the (m + 2)-th dam 40(m + 2).

[0047] "Formulation of power generation plan" The power generation plan presentation device 10 executes a process for assisting a person in formulating a power generation plan. Fig. 3 shows the procedure of the process executed by the power generation plan presentation device 10. The series of processes shown in Fig. 3 is realized by the PU12 repeatedly executing the program stored in the storage device 14 at a predetermined cycle, for example. Hereinafter, the step numbers of each process are represented by numbers preceded by "S".

[0048] In the series of processes shown in Fig. 3, the PU12 first executes a past plan search process (S10). The past plan search process is a process of searching, using the past plan data 14a stored in the storage device 14 shown in Fig. 1, what plans were set in the past in a situation close to the target date of plan formulation.

[0049] "Past plan search process" Fig. 4 shows the details of the process of S10. In the series of processes shown in FIG. 4, PU12 first searches for past plans regarding the first group obtained by classifying the first dam 40(1) to the Nth dam 40(N) shown in FIG. 2 into two groups (S100). The first group is composed of the first dam 40(1) to the mth dam 40(m). On the other hand, the second group is composed of the mth dam 40(m) to the Nth dam.

[0050] In this way, the first group and the second group are groups divided at a point in the water flow proceeding from the upstream side to the downstream side of the river. Therefore, the first group does not include dams between the uppermost dam and the lowermost dam of the second loop. Also, the second group does not include dams between the uppermost dam and the lowermost dam of the first group. Further, the mth dam 40(m) is shared between the first group and the second group. The shared mth dam 40(m) is the lowermost dam of the first group and the uppermost dam of the second group.

[0051] FIG. 5 shows the detailed procedure of the process of S100. In the series of processes shown in FIG. 5, PU12 first acquires the daily amount and the initial value of the water intake amount for the target date of the first dam 40(1) (S102). Here, the daily amount is the amount of water that can be used in one day. The daily amount is set, for example, when a person formulates a power generation plan for one week. The initial value of the water intake amount is the water intake amount on the day before the target day. When the execution time of the process of S102 is before just before 12:00 am of the target date of the plan formulation, the initial value of the water intake amount does not have to be the actual value. Specifically, for example, it may be a planned value that has been formulated. Note that the process of S102 may be a process of acquiring the daily amount and the initial value of the water intake amount input via the input unit 22 shown in FIG. 1.

[0052] Next, PU12 selects a plurality of values from the past daily amounts and the initial values of the water intake amounts indicated by the past plan data 14a that approximate the values obtained by the process of S102 (S104). Fig. 6 illustrates the past planning data 14a. As shown in Fig. 6, the past planning data 14a is data showing the past daily volume and the initial water intake value, and the waveform of the actually formulated water intake. The waveform of the water intake shows the change in the water intake according to time.

[0053] Note that the past planning data 14a does not necessarily have to be data regarding all the plans formulated in the past. For example, data regarding irregular plans that are not suitable as the reference target may be excluded. Examples of cases where irregular plans are made include, for example, cases where the power plant has been continuously shut down during a specific period, or cases where the water consumption significantly deviates between the plan and the actual performance.

[0054] Returning to Fig. 5, the PU12 acquires the daily volume and the initial water intake value of the target date for planning for the third dam 40(3) (S106). Next, the PU12 selects several pieces of the past planning data selected by the process of S104 that are close to the initial water intake value and the daily volume acquired by the process of S106 (S108). The process of S108 is a process of selecting a part from among the plurality of plan performances selected by the process of S104.

[0055] Next, the PU12 acquires the initial water intake value and the daily volume of the m-th dam 40(m), which is the most downstream dam of the first group (S110). Then, the PU12 selects one piece that is closest to the daily volume and the initial water intake value acquired by the process of S110 from among the plurality of plan performances selected by the process of S108 (S112). Note that when the process of S112 is completed, the PU12 completes the process of S100 in Fig. 4.

[0056] Returning to FIG. 4, PU12 searches for past plans regarding the second group (S120). The process of S120 is the same as that of S100. That is, a plan formulated on a specific past day is selected using the daily volume and initial water intake values of some of the dams from the m-th dam 40(m) to the N-th dam 40(N). Here, the specific day may be different from the day on which the past plan finally selected in the process of S112 was made in the first group.

[0057] Note that in the processes of S100 and S120, the selection of dams excluded from the search for similarity between the daily volume and initial water intake values on the target day of planning and the past daily volume and initial water intake values is arbitrary. However, for example, in the case of a dam of the inflow type that uses the water flowing in a river directly for power generation without storing it, that dam may be excluded from the target.

[0058] As a method for searching for similarity between the daily volume and initial water intake values on the target day of planning and the past daily volume and initial water intake values, any pattern matching method can be selected. However, when there are characteristics specific to a dam in the selection of water intake, a selection method utilizing such characteristics may be adopted.

[0059] For example, as shown in FIG. 7(a), in the case of a dam where most of the data has a waveform with little change in water intake, the computational load of the narrowing-down process may be reduced as follows. That is, even when most of the waveforms are like those in FIG. 7(a) and there are only a few waveforms where the water intake changes according to time as in FIG. 7(b), for example, if narrowing down is done only based on the initial water intake value, the waveform shown by the solid line and the waveform shown by the one-dot chain line in FIG. 7(c) cannot be distinguished. On the other hand, if narrowing down is done only based on the daily volume, the waveform shown by the one-dot chain line and the waveform shown by the two-dot chain line in FIG. 7(c) cannot be distinguished.

[0060] Here, the value obtained by dividing the daily water intake by the initial value of the water intake is close to "1" in the case of the waveform in Fig. 7(a) or a waveform similar thereto, while it is a smaller value in the case of the waveform in Fig. 7(b). Therefore, PU12 may first narrow down which of the waveforms illustrated in Fig. 7(a) and the waveform illustrated in Fig. 7(b) it corresponds to by setting a threshold value for drawing a line between them and comparing the magnitude of the divided value and the threshold value. Then, thereafter, PU12 may further narrow down, for example, by selecting past data close to the obtained daily water intake.

[0061] Returning to Fig. 4, PU12 determines either one of the water intake plans for the m-th dam 40(m) selected by the process of S100 and the water intake plan for the m-th dam 40(m) selected by the process of S120 as the final plan for the m-th dam 40(m) (S140). PU12 may, for example, select the one closer to the daily water intake of the m-th dam 40(m) among the two selected water intake plans.

[0062] Note that when PU12 completes the process of S140, it completes the process of S10 in Fig. 3. Returning to Fig. 3, PU12 executes a process of predicting the stream flow rate to each of the first dam 40(1) to the N-th dam 40(N) (S12).

[0063] "Stream flow rate prediction process" Fig. 8 shows the detailed procedure of the stream flow rate prediction process. In the series of processes shown in Fig. 8, PU12 first substitutes "1" into a variable i for designating a dam (S30). Next, PU12 acquires, via the communication device 16, the predicted value of the precipitation amount in each mesh into which a predetermined area in the basin of the i-th dam 40(i) is divided (S32). The predicted value obtained here is the predicted value for each time zone within a predetermined period in each mesh. The predetermined period may be, for example, 24 hours. Also, each time zone may have a length of, for example, 1 hour. Note that in the process of S32, PU12 may acquire, in addition to the predicted value, the actual situation value (the actually measured value) of the precipitation amount.

[0064] Next, PU12 substitutes "1" into variable t that designates the time period for calculating the soil rainfall index ISR and the soil runoff amount SRA (S34). As an example, variable t that designates time designates the time in 30 - minute units.

[0065] Then, PU12 calculates the soil rainfall index ISR and the soil runoff amount SRA in the time periods designated by each of variables t - t0, t - t0 - 6, …, t - t0 - 12, …, t - t0 - 42 (S36). Here, the time period designated by variable t - t0 - 6 is a past time period compared to the time period designated by variable t - t0. Here, among the time periods for calculating the soil rainfall index ISR and the soil runoff amount SRA, there may be a past time period compared to the time period corresponding to the predicted value acquired in S32 in the current control cycle. In that case, PU12 uses the predicted value acquired by the process of S32 in the control cycle before the previous one.

[0066] As shown in FIG. 8, the soil rainfall index ISR and the soil runoff amount SRA to be calculated are specified by the three - dimensional components within the parentheses attached to them. Here, the first component within the parentheses is variable i that identifies the dam. The second component within the parentheses is the component that identifies the mesh. That is, for example, the soil rainfall index ISR(i, 1, t - t0) and the soil rainfall index ISR(i, 2, t - t0) have different meshes. In other words, the soil rainfall index ISR(i, 1, t - t0) and the soil rainfall index ISR(i, 2, t - t0) are for different regions where the soil rainfall index ISR is being obtained. The third component within the parentheses is variables t - t0, t - t0 - 6, …, t - t0 - 12, …, t - t0 - 42 that indicate the time period. Note that the above variable t indicates the time for which the stream flow is to be predicted. And the leading time t0 is the time that represents how far in the past it is with respect to the time to be predicted.

[0067] PU12 calculates the soil rainfall index ISR and the soil runoff amount SRA using a well - known tank model based on the value acquired by the process of S32 as an input variable. The tank model models the state of rain flowing through the soil.

[0068] Figure 9 shows a tank model. The tank model shown in Figure 9 consists of three tanks: the uppermost first tank 60, the middle second tank 62, and the lowermost third tank 64. Water is stored in the first tank 60 by rainfall. The water stored in the first tank 60 does not flow out from the side holes 60a until the water level reaches the height of the side holes 60a. The water stored in the first tank 60 penetrates into the second tank 62. The water stored in the second tank 62 does not flow out from the side holes 62a until the water level reaches the height of the side holes 62a. The water stored in the second tank 62 penetrates into the third tank 64. The water stored in the third tank 64 does not flow out from the side holes 64a until the water level reaches the height of the side holes 64a.

[0069] The total outflow rates q1, q2, and q3 of water from the side holes 60a, 62a, and 64a respectively become the outflow rate to the river. PU12 calculates the outflow rates q1, q2, and q3, and the water levels S1 of the first tank 60, S2 of the second tank 62, and S3 of the third tank 64 using the tank model. The tank model has a permeability coefficient β1 from the first tank 60 to the second tank, a permeability coefficient β2 from the second tank 62 to the third tank 64, and a permeability coefficient β3 from the third tank 64.

[0070] The soil rainfall index ISR is the sum of the water levels S1 of the first tank 60, S2 of the second tank 62, and S3 of the third tank 64. The soil outflow rate SRA is the total of the outflow rates q1, q2, and q3.

[0071] Returning to Figure 8, PU12 calculates the stream flow MF using a regression model defined by the model definition data 14b stored in the storage device 14 shown in Figure 1 (S38). The input variables of the regression model are the time series data of the soil rainfall index ISR and the soil outflow rate SRA respectively.

[0072] The time-series data of the soil rainfall index ISR as an input variable includes eight temporally adjacent data for each mesh. That is, the input variable includes the soil rainfall index ISR(i,j,t-t0) to ISR(i,j,t-t0-42) for all values of the variable j that defines the mesh. Further, the time-series data of the soil rainfall index ISR as an input variable includes the time-series data of the weighted average value of the soil rainfall index ISR in each mesh. Here, the weighting coefficient may have a larger value the closer it is to the area for which the stream flow is to be determined, for example.

[0073] The time-series data of the soil runoff amount SRA as an input variable includes eight temporally adjacent data for each mesh. That is, the input variable includes the soil runoff amount SRA(i,j,t-t0) to SRA(i,j,t-t0-42) for all values of the variable j that defines the mesh. Further, the time-series data of the soil runoff amount SRA as an input variable includes the time-series data of the weighted average value of the soil runoff amount SRA in each mesh. Here, the weighting coefficient may have a larger value the closer it is to the area for which the stream flow is to be determined, for example.

[0074] The output variable of the regression model is the stream flow MF. PU12 calculates the stream flow MF by inputting the time-series data of the soil rainfall index ISR and the soil runoff amount SRA calculated by the process of S36 into the regression model.

[0075] The first component among the two-dimensional components within the parentheses following the stream flow MF shown in FIG. 8 is the variable i that identifies the dam. The second component among the two-dimensional components within the parentheses following the stream flow MF shown in FIG. 8 is the variable indicating the time zone.

[0076] The regression model is a trained model. The regression model is a model learned by supervised learning using the time series data of the soil rainfall index ISR and the soil runoff amount SRA respectively, and the measured values of the corresponding stream flow as training data. When the regression model is a parametric model, the model specification data 14b includes the parameters representing the regression model. When the regression model is a non-parametric model, the model specification data 14b includes a part of the training data, etc. That is, for example, when the regression model is a support vector regression, the model specification data 14b includes support vectors.

[0077] Next, PU12 determines whether the variable t indicating the time zone is 48 (S40). This is a process of determining whether the prediction of the stream flow MF for one day is completed. When PU12 determines that the variable t is smaller than 48 (S40: NO), after incrementing the variable t (S42), it returns to the process of S36.

[0078] On the other hand, when PU12 determines that the variable t is 48 (S40: YES), it determines whether the value of the variable i designating the dam is N (S44). This process is a process of determining whether the prediction of the stream flow MF from the first dam 40(1) to the Nth dam 40(N) is completed. When PU12 determines that the variable i is smaller than N (S44: NO), after incrementing the variable i (S46), it returns to the process of S32.

[0079] Note that when making an affirmative determination in the process of S44, PU12 completes the process of S12 shown in FIG. 3. Returning to FIG. 3, PU12 searches for the optimal solution of the power generation plan based on the linear programming method (S14).

[0080] "Optimal solution search process" FIGS. 10A and 10B show the procedure of the optimal solution search process. The processes shown in FIGS. 10A and 10B are realized by PU12 repeatedly executing the program stored in the storage device 14, for example, at a predetermined period.

[0081] In the series of processes shown in FIGS. 10A and 10B, PU12 first acquires the stream flow rate MF in each time zone of each dam and the water storage amount WSA of each dam (S50). That is, PU12 acquires the stream flow rates MF(1,1), MF(1,2), … MF(2,1), MF(2,2), … MF(N,47), MF(N,48). Further, PU12 acquires the water storage amounts WSA(1,0), WSA(2,0), …, WSA(N,0). Among the two-dimensional variables in the parentheses after the water storage amount WSA, the first component is a variable that designates a dam. Among the two-dimensional variables in the parentheses after the water storage amount WSA, the second component is a component that designates a time zone. In the process of S50, the value of the variable that designates the time zone is zero. The fact that the value of the variable that designates the time zone is zero means that it is the last time zone of the previous day. That is, the water storage amounts WSA(1,0), WSA(2,0), …, WSA(N,0) are the initial values of the water storage amount WSA. Note that if the process of FIG. 10 is not executed at 0:00 am on the target day for planning, the water storage amounts WSA(1,0), WSA(2,0), …, WSA(N,0) are values calculated based on the water intake plan established on the previous day.

[0082] Next, PU12 sets the constraint conditions in the linear programming method (S52). The constraint conditions include the condition that the water storage amount WSA(i,t) of each dam is equal to or greater than the lower limit water storage amount WSAL(i) and equal to or less than the upper limit water storage amount WSAH(i). The lower limit water storage amount WSAL(i) and the upper limit water storage amount WSAH(i) are set for each dam. That is, for each dam, since the upper limit water level and the lower limit water level are determined, the upper limit water storage amount WSAH(i) is set according to the upper limit water level and the lower limit water storage amount WSAL(i) is set according to the lower limit water level. The lower limit water storage amount WSAL(i) and the upper limit water storage amount WSAH(i) are fixed values as an example, except for dams of the ALR (Automatic Load Regulator) type. For dams of the ALR type, the upper limit water storage amount WSAH(i) is calculated according to the water inflow amount IA into the dam. The inflow amount IA is the sum of the amount of water flowing in through the dam discharge line 42 and the power generation discharge line 44 and the stream flow rate MF. FIG. 10 illustrates that the k-th dam 40(k) is an ALR dam.

[0083] PU12 sets the upper limit water storage capacity WSAH(k) of the k-th dam 40(k) to the smaller value of the values of the following two equations (c1) and (c2). -a1·IA(k,t)+b1 …(c1) -a2·IA(k,t)+b2 …(c2) This is a device for easily handling the upper limit water storage capacity WSAH(k) in the dam of the ALR formula by linear programming. This will be described below.

[0084] FIG. 11 illustrates the relationship between the water level and the water storage capacity of the dam. As shown in FIG. 11, the relationship between the water level and the water storage capacity of the dam has non-linearity. Therefore, when the upper limit water level of the dam changes, the upper limit water level and the upper limit water storage capacity have a non-linear relationship. On the other hand, in linear programming, the water intake for each time period is obtained. The water intake and the water storage capacity have a linear relationship. Therefore, it is difficult to handle the changing upper limit water level in optimizing the water intake based on linear programming. Therefore, in this embodiment, the relationship between the water level and the water storage capacity is linearly approximated. That is, as shown in FIG. 11, as long as the water level is in the region above the lower limit water level, the relationship between the water level and the water storage capacity can be regarded as a linear relationship with high accuracy. Therefore, the relationship between the water level and the water storage capacity is linearly approximated by a linear equation with the slope of the change in the water storage capacity with respect to the change in the water level in the region where the water level is above the lower limit water level as the proportionality coefficient.

[0085] FIG. 12 shows the relationship between the water inflow rate into the dam and the upper limit water level. As shown in FIG. 12, the relationship between the inflow rate and the upper limit water level has non-linearity. Since this is difficult to handle by linear programming, in this embodiment, it is linearly approximated by the straight line f1 and the straight line f2. That is, when the water inflow rate into the dam is taken as the horizontal axis and the upper limit water level is taken as the vertical axis, their relationship is a convex-up relationship. Therefore, the relationship is approximated by a broken line of two line segments.

[0086] The above equation (c1) is the equation obtained by converting the vertical axis to the water storage capacity for the straight line f1 in FIG. 12. The above equation (c2) is the equation obtained by converting the vertical axis to the water storage capacity for the straight line f2 in FIG. 12.

[0087] Returning to FIG. 10A, PU12 sets a constraint condition for the water intake WI(q) of the q-th dam where the possible values of the water intake are only discrete values. FIG. 10 shows, as an example, an example where the water intake WI(q) can only take one of the discrete values W1, W2, W3,....

[0088] PU12 sets the following temporary constraint condition for the water intake (S54). The temporary constraint condition is a condition that the water intake WI(i,t) is equal to or greater than zero and equal to or less than the upper limit water intake WIM(i). Here, the upper limit water intake WIM(i) is the actual upper limit water intake for the i-th dam 40(i). However, the possible water intake of the i-th dam 40(i) is not all values equal to or greater than zero and equal to or less than the upper limit water intake WIM(i). The possible water intake of the i-th dam 40(i) is zero or a value equal to or greater than the lower limit water intake WIL(i) and equal to or less than the upper limit water intake WIM(i). However, when such a condition is used as a constraint condition, although it can be addressed by using the mixed-integer linear programming method, the computational load becomes extremely large. Therefore, here, in order to search for the optimal solution without using the mixed-integer linear programming method, a temporary constraint condition is provided.

[0089] PU12 optimizes the water intake WI and the water discharge DA by linear programming under the given constraint conditions (S56). As an example, PU12 searches for the optimal solution of the water intake WI and the water discharge DA using the simplex method. However, for the water intake WI(q) of the q-th dam, PU12 searches all of the values W1, W2, W3,....

[0090] The water discharge DA is the amount of water flowing out of the dam without being used for power generation. In other words, the water discharge DA is the amount of water flowing out through the dam discharge line 42. The process of S56 includes a process where PU12 calculates the water storage amount WSA by the following formula using the inflow amount IA and the dam outflow amount OA.

[0091] WSA(i,t)=WSA(i,t - 1)+IA(i,t - 1)+OA(i,t - 1) The process of S56 includes the process in which PU12 calculates the inflow rate IA according to the following formula.

[0092] IA(i,t)=WI(i-1,t-J)+DA(i-1,t-K)+MF(i,t) Here, the constant J represents the delay time until the water intake of the upstream dam affects the inflow rate IA. The constant J may have different values for each dam. The constant K represents the delay time until the water discharge from the upstream dam affects the inflow rate IA. The constant K may have different values for each dam.

[0093] The process of S56 includes the process in which PU12 calculates the dam outflow rate OA according to the following formula. OA(i,t)=WI(i,t+L)+DA(i,t) Here, the constant L represents the delay time until the water intake of the accompanying power plant is affected. The constant L may have different values for each dam.

[0094] The process of S56 is a process of searching for the water intake WI and the water discharge DA that maximize the selling electricity price. Specifically, PU12 searches for the water intake WI and the water discharge DA that maximize the objective function. The objective function is the value obtained by adding, for all times and all dams, the product of the electricity-water ratio of the i-th dam 40(i) and the water intake WI(i,t) and the selling electricity price of the i-th dam 40(i) at the time indicated by the variable t. Here, the electricity-water ratio is a coefficient that converts the water intake of the dam into the power generation amount. The selling electricity price is the selling electricity price per unit power generation amount. The selling electricity price depends on the power plant and the time.

[0095] In this way, the optimal solutions of the 48 water intakes WI and water discharges DA from the first dam 40(1) to the N-th dam 40(N) are searched. Note that after searching for the optimal solution, if the rising gradient of the k-th dam 40(k) does not satisfy the constraint defined by the rising waveform data 14d stored in the storage device 14, PU12 corrects the rising gradient.

[0096] When the process of S56 is completed, PU12 selects the p-th dam 40(p) which is the dam with the maximum priority among the dams for which the present constraint conditions are not set (Fig. 10B: S58). As an example, the priority is set to a higher value for a dam for which it is more difficult to obtain a solution that satisfies the constraint conditions. A dam for which it is more difficult to obtain a solution that satisfies the constraint conditions is, for example, a dam where a plurality of dam discharge lines 42 and power generation discharge lines merge. In the example of Fig. 2, the (m + 2)-th dam 40(m + 2) corresponds. PU12 selects the p-th dam 40(p) by referring to the priority definition data 14c stored in the storage device 14.

[0097] Then, PU12 changes the constraint conditions of the p-th dam 40(p) from the temporary constraint conditions to the present constraint conditions (S60). Here, when the optimal water intake WI is less than the lower limit water intake WIL, PU12 sets the present constraint conditions to the condition that the water intake WI is zero. Also, when the optimal water intake WI is greater than or equal to the lower limit water intake WIL, PU12 changes the present constraint conditions to the effect that it is greater than or equal to the lower limit water intake WIL and less than or equal to the upper limit water intake WIH. PU12 sets the present constraint conditions separately for each value of the variable t indicating the time.

[0098] Then, while newly adopting the present constraint conditions as the constraint conditions of the p-th dam 40(p), PU12 executes a search process by the linear programming method (S62). The process of S62 is a process of searching for the optimal solutions of the water intake WI and the discharge amount DA of all the dams from the first dam 40(1) to the N-th dam 40(N). Note that the process of S62 may include a correction process for the rising gradient of the searched k-th dam 40(k).

[0099] Next, PU12 determines whether there is a solution by the linear programming method (S64). If there is no solution among the water intake WI and the discharge amount DA for each of the variables t = 1 to 48 in the first dam 40(1) to the N-th dam 40(N) (S64: NO), the process proceeds to the process of S66.

[0100] In the process S66, PU12 changes the constraint condition of the p-th dam 40(p) to the condition that it is not less than the lower limit water intake WIL and not more than the upper limit water intake WIH at all times. Then, PU12 returns to the process S62 to execute the search process for the optimal solution based on the linear programming method.

[0101] On the other hand, when PU12 determines that there is a solution to the linear programming method (S64: YES), it determines whether the search with the present constraint condition has been completed for all dams from the first dam 40(1) to the N-th dam 40(N) (S68). When PU12 determines that there is still a dam for which the search based on the present constraint condition has not been completed (S68: NO), it returns to the process S58. In the process S58, PU12 sets the dam with the highest priority among the dams for which the present constraint condition has not yet been set as the p-th dam and shifts to the process S60.

[0102] Note that when making an affirmative determination in the process S68, PU12 completes the process S14 in FIG. 3. Returning to FIG. 3, PU12 operates the display unit 20 to display the results of the processes S10 and S14 (S16). Note that when completing the process S16, PU12 temporarily ends the series of processes shown in FIG. 3.

[0103] "Actions and Effects of the Present Embodiment" PU12 selects values close to the daily amount and the initial water intake value for the target day for power generation planning among the daily amount and the initial water intake value included in the past plan data 14a. Then, it displays this as the past established power generation plan results on the display unit 20. Also, PU12 obtains the water intake that maximizes the selling electricity price by the linear programming method. Then, it displays the obtained water intake WI on the display unit 20.

[0104] Thereby, the person formulating the power generation plan can formulate the plan while referring to the information on what plans were made in the past in a situation similar to the target day and the information on the water intake for maximizing the selling electricity price.

[0105] <Corresponding Relationship> The correspondence between the matters in the above-described embodiment and the matters described in the column of "Means for Solving the Problems" is as follows. Below, the correspondence is shown for each number of the solution means described in the column of "Means for Solving the Problems". [1] The plurality of dams correspond to the first dam 40(1) to the Nth dam 40(N). The execution device corresponds to PU12. The search process corresponds to the processes of S56 and S62. The constraint condition change process corresponds to the processes of S58 and S60. [2] Two or more dams correspond to the dams other than the qth dam 40(q) among the first dam 40(1) to the Nth dam 40(N). [3] The constraint condition change process corresponds to the process of S58. [4] The constraint condition change process corresponds to the process of S66. [5] The discrete dam corresponds to the qth dam 40(q). [6] The ALR dam corresponds to the kth dam 40(k). The initial water storage amount acquisition process and the stream flow rate acquisition process correspond to the process of S50. [7] It corresponds to the setting shown in FIG. 11. [8] The rising regulation dam corresponds to the kth dam 40(k). [9] The function in which the higher the power selling amount, the higher the evaluation corresponds to the objective function in which the power selling price is the dependent variable.

[10] The initial water storage amount acquisition process and the stream flow rate acquisition process correspond to the process of S50. [11 to 14] The storage device corresponds to the storage device 14. The precipitation prediction value acquisition process corresponds to the process of S32. The input variable calculation process corresponds to the process of S36. [15, 18] The storage device corresponds to the storage device 14. The selection process corresponds to the process of S10.

[16] The fact that the first dam 40(1) to the Nth dam 40(N) are grouped into the first dam 40(1) to the mth dam 40(m) and the mth dam 40(m) to the Nth dam 40(N) corresponds.

[17] It corresponds to the process illustrated in FIG. 5. [19, 20] The presentation process corresponds to the process of S16.

[0106] <Other Embodiments> Note that this embodiment can be implemented with the following modifications. This embodiment and the following modification examples can be implemented in combination with each other within a technically non-conflicting range.

[0107] "Regarding the Constraint Condition Change Process" ·The priority definition data 14c is not limited to data that defines the priorities of the "N-1" dams from the first dam 40(1) to the Nth dam 40(N) excluding the qth dam, from the first priority to the "N-1"th priority. For example, the priority definition data may be data that defines the same priorities for some dams. In that case, the PU12 may simultaneously change the constraint conditions from the temporary constraint conditions to the actual constraint conditions for the dams with the same priority.

[0108] ·It is not essential that the constraint condition change process is a process of determining the dams to be changed from the temporary constraint conditions to the actual constraint conditions according to the priority definition data 14c. For example, it may be a process of determining the dam with the largest degree of deviation from the condition that the water intake is zero or within the range from the minimum water intake to the maximum water intake as the dam to be changed based on the search result of the optimal solution under the temporary constraint conditions. At this time, the number of dams changed from the temporary constraint conditions to the actual constraint conditions at one time is not limited to one.

[0109] ·In the constraint condition change process, the dams changed from the temporary constraint conditions to the actual constraint conditions at one time are not limited to a part of the plurality of dams for which it is desired to obtain the final optimal solution under the actual constraint conditions. In other words, the constraint condition change process may be a process of changing all of the plurality of dams for which it is desired to obtain the final optimal solution under the actual constraint conditions from the temporary constraint conditions to the actual constraint conditions at one time.

[0110] "Regarding the search process" ·It is not essential that the search process is a process of searching for the optimal solution of the water intake for each time period obtained by dividing the predetermined period of 24 hours at 30-minute intervals. For example, the predetermined time may be 12 hours. Also, for example, the predetermined time may be 48 hours. Also, for example, the time interval for dividing the predetermined time may be 1 hour.

[0111] ·When a negative determination is made in the process of S64 in FIG. 10B, it may be notified to people that there is no solution. In other words, it is not essential to execute the process of S66. ·It is not essential that the search process is a process of searching for an optimal solution by a single method.

[0112] "Regarding the use of the search process" ·The use of the search process is not limited to providing information for the planner to formulate the final plan. For example, the search result may be set as the final plan.

[0113] "Regarding a predetermined objective function" ·The objective function is not limited to a function in which the value of the dependent variable is the selling electricity price. For example, the dependent variable of the objective function may be the power generation amount. In that case, PU12 may search for a solution that maximizes the power generation amount by the linear programming method. Also, for example, the dependent variable of the objective function may be the total amount of dam overflow water that is the flow rate by the dam discharge line 42. In other words, it may be the total amount for a predetermined time of the respective discharge amounts DA of the first dam 40(1) to the Nth dam 40(N). In that case, PU12 may search for a solution that minimizes the overflow water amount by the linear programming method. Also, the dependent variable of the objective function is not limited to any one of the three variables of the selling electricity price, the power generation amount, and the overflow water amount. For example, it may be a weighted average value of two or three of the three variables. However, in that case, when PU12 searches for a solution that maximizes the weighted average value and the overflow water amount is included in the weighted average value, the weight coefficient of the overflow water amount is negative. Furthermore, it is not essential that the variable to which the weighted average process is applied is any one of the selling electricity price, the power generation amount, and the overflow water amount. Also, when the dependent variable of the objective function is defined by a combination of a plurality of variables, it is not essential that the sum of the weight coefficients multiplied by each variable becomes "1".

[0114] "Regarding the regression model" ·The time-series data of the soil rainfall index as an input variable of the regression model in which the stream flow rate is the output variable is not limited to the data for 24 hours every 3 hours. The time-series data of the soil rainfall index may be, for example, the data for 24 hours every 1 hour. Also, for example, the time-series data of the soil rainfall index may be the data for 48 hours every 3 hours.

[0115] · The time-series data of soil erosion amount as an input variable of a regression model with stream flow as an output variable is not limited to 24-hour data every 3 hours. The time-series data of soil erosion amount may be, for example, 24-hour data every 1 hour. Also, for example, the time-series data of soil erosion amount may be 48-hour data every 3 hours.

[0116] · It is not essential that the soil rainfall index as an input variable of a regression model with stream flow as an output variable is the soil rainfall index in each mesh where a predetermined area is divided into a plurality. For example, the soil rainfall index as an input variable of a regression model with stream flow as an output variable may be the soil rainfall index over the entire predetermined area. Here, the soil rainfall index over the entire predetermined area may be the weighted average value of the soil rainfall indices in the respective meshes.

[0117] · It is not essential that the soil erosion amount as an input variable of a regression model with stream flow as an output variable is the soil erosion amount in each mesh where a predetermined area is divided into a plurality. For example, the soil erosion amount as an input variable of a regression model with stream flow as an output variable may be the soil erosion amount over the entire predetermined area. Here, the soil erosion amount over the entire predetermined area may be the weighted average value of the soil erosion amounts in the respective meshes.

[0118] · The input variables of a regression model with stream flow as an output variable are not limited to the time-series data of soil rainfall index and the time-series data of soil erosion amount. For example, the input variable of a regression model with stream flow as an output variable may include only one of the time-series data of soil rainfall index and the time-series data of soil erosion amount.

[0119] · The input variables of a regression model with stream flow as an output variable are not limited to the time-series data of a predetermined variable. For example, the input of a regression model with stream flow as an output variable may be the value at one point in time of each of the soil rainfall index and the soil erosion amount.

[0120] · It is not essential that the soil rainfall index as an input variable of the regression model with stream flow as the output variable be a value calculated at 30 - minute intervals. The soil rainfall index may be, for example, a value calculated at 1 - hour intervals.

[0121] · It is not essential that the soil runoff as an input variable of the regression model with stream flow as the output variable be a value calculated at 30 - minute intervals. The soil runoff may be, for example, a value calculated at 1 - hour intervals.

[0122] · The input variables of the regression model with stream flow as the output variable are not limited to only at least one of the two, i.e., the soil rainfall index and the soil runoff. For example, precipitation may be included in the input variables. Also, for example, snowfall, snow accumulation, and air temperature may be included in the input variables. Thereby, stream flow considering an increase in stream flow due to snowmelt can be calculated.

[0123] · The stream flow MF is not limited to a variable predicted at 30 - minute intervals. For example, it may be a variable predicted in 1 - hour units. In that case, if the stream flow MF used in FIG. 10A is in 30 - minute units, the same predicted value may be used for two 30 - minute - unit stream flow MFs.

[0124] "Regarding the tank model" · The tank model is not limited to a model consisting of three - stage tanks, i.e., the first tank 60, the second tank 62, and the third tank 64 as illustrated in FIG. 9. The tank model may be, for example, a model having four or more stages of tanks.

[0125] "Regarding the selection process" · The selection process is not limited to a process of selecting, only for some dams within the group, days with similar daily amounts and initial water intake amounts between the target day and the past planned data 14a. For example, it may be a process of selecting, for all dams within the group, days with similar daily amounts and initial water intake amounts between the target day and the past planned data 14a.

[0126] · The group that sets the constraint for selecting data with the same date among the past plan data 14a is not limited to two groups. For example, there may be three or more groups. · It is not essential that the selection process is a process of selecting data of the same date among the past plan data 14a for each group.

[0127] · In the narrowing-down process, the situation variable, which is a variable for comparing the situation between the target date and the past date, is not limited to the variables described above. For example, the amount of water available during a predetermined period is not limited to the amount of water available in 24 hours. The predetermined period may be, for example, 12 hours. Also, for example, when the power plant has a plurality of generators, a variable indicating the number of generators to be used may be included in the situation variable. For example, when there are two generators, the variable indicating the number of generators to be used can take values of 0, 1, or 2.

[0128] · It is not essential to execute the selection process. For example, PU12 may execute only the search process. "Regarding the execution device" · The execution device is not limited to those that execute software processing. For example, it may include a dedicated hardware circuit such as an ASIC that executes at least a part of the processing executed in the above embodiment. That is, the execution device may include a processing circuit having any of the following configurations (a) to (c). (a) A processing circuit including a processing device that executes all of the above processing according to a program and a program storage device such as a storage device that stores the program. (b) A processing circuit including a processing device and a program storage device that execute a part of the above processing according to a program, and a dedicated hardware circuit that executes the remaining processing. (c) A processing circuit including a dedicated hardware circuit that executes all of the above processing. Here, there may be a plurality of software execution devices including a processing device and a program storage device. Also, there may be a plurality of dedicated hardware circuits.

[0129] "Regarding the water system" ·The water systems targeted for power generation plan formulation are not limited to those exemplified in FIG. 2. For example, there may be a plurality of dams into which water flows from a plurality of upstream dams.

[0130] ·The water systems targeted for power generation plan formulation are not limited to those assumed in FIGS. 10A and 10B. For example, there may be a plurality of dams whose water intake can only take discrete values. Also, for example, the number of dams whose water intake can only take discrete values may be zero. Also, for example, there may be a plurality of ALR dams. Also, for example, the number of ALR dams may be zero.

Explanation of Signs

[0131] 10…Power generation plan presentation device 14…Storage device 14a…Past plan data 14b…Model specification data 14c…Priority definition data 14d…Startup waveform data 16…Communication device 20…Display unit 42…Dam discharge line 44…Power generation discharge line 46…Power plant 60…First tank 60a…Side hole 62…Second tank 62a…Side hole 64…Third tank 64a…Side hole

Claims

1. A power generation plan presentation device for planning the water intake amount by a power plant that generates power using a plurality of dams connected by a river, comprising: an execution device configured to execute an optimization process; The optimization process is a process of outputting a solution of the water intake amount that optimizes the value of a predetermined objective function, and includes a search process and a constraint condition change process. The search process includes a process of searching for the water intake amount of at least a part of the plurality of dams by linear programming under a predetermined constraint condition. The constraint condition change process is a process of changing the predetermined constraint condition to the actual constraint condition after the search process is executed under a temporary constraint condition as the predetermined constraint condition. The temporary constraint condition is a condition that the water intake amount is equal to or greater than zero and equal to or less than the maximum water intake amount. The actual constraint condition is a condition that the water intake amount is zero if the water intake amount searched by the search process under the temporary constraint condition is less than the minimum water intake amount, and the water intake amount is equal to or greater than the minimum water intake amount and equal to or less than the maximum water intake amount if the water intake amount searched by the search process under the temporary constraint condition is equal to or greater than the minimum water intake amount. A power generation plan presentation device, wherein the minimum water intake amount is a value greater than zero.

2. At least a part of the plurality of dams includes two or more dams. The constraint condition change process includes a process of selecting a part of the two or more dams and changing the constraint condition of the selected part of the dams to the actual constraint condition. The search process is executed each time the predetermined constraint condition is changed by the constraint condition change process. The power generation plan presentation device according to claim 1, wherein the constraint condition change process is repeatedly executed each time the search process is executed until there is no dam to which the temporary constraint condition is imposed among the two or more dams.

3. The power generation plan presentation device according to claim 2, wherein the constraint condition change process includes a process of selecting a part of the two or more dams for changing the constraint condition to the actual constraint condition according to a preset priority order.

4. The water intake amount searched by the search process is the amount at each time when a predetermined period is divided into a plurality of unit times. When there is no solution obtained by the search process for the dam whose constraint condition has been changed to the main constraint condition, the constraint condition change process includes a second change process of changing the constraint condition of all the water intake amounts for the plurality of unit times related to the dam for which no solution has been obtained to a condition that the amount is not less than the minimum water intake amount and not more than the maximum water intake amount, without selecting a new dam to be changed to the main constraint condition. The power generation plan presentation device according to claim 1.

5. Among the dams different from the two or more dams, there are discrete dams that allow only a plurality of different values as the water intake amounts. The search process includes, for the discrete dam, a process of searching for a solution that optimizes the objective function together with the water intake amounts of the two or more dams while setting the water intake amount to each of the plurality of values. The power generation plan presentation device according to claim 1.

6. Among the plurality of dams, there is an ALR dam. The execution device is configured to execute a storage water amount initial value acquisition process and a stream flow rate acquisition process. The storage water amount initial value acquisition process is a process of acquiring the initial value of the storage water amount in the ALR dam. The stream flow rate acquisition process is a process of acquiring the stream flow rate flowing into the ALR dam. The optimization process includes a process of substituting, into the inflow amount of the ALR dam, the sum of the discharge amount of the upstream dam adjacent to the ALR dam and the water intake amount and the stream flow rate flowing into the ALR dam, and a process of updating the initial value of the storage water amount according to the inflow amount of the ALR dam. The search process is a process of searching for the water intake amount while further including a condition that the storage water amount of the ALR dam is not more than the maximum storage water amount. The maximum storage water amount is set to the smaller one of the values calculated by two linear functions having the inflow amount into the ALR dam as an independent variable. The power generation plan presentation device according to claim 1.

7. The linear function is such that the relationship between the water level of the dam and the storage water amount of the dam is approximated by a straight line connecting the upper limit water level of the dam and the corresponding storage water amount and the lower limit water level and the corresponding storage water amount. The power generation plan presentation device according to claim 6.

8. Among the plurality of dams, some dams include rising regulation dams. It is configured to execute a correction process. The correction process is a process of correcting to a predetermined rise when the rise of the water intake amount obtained by the search process for the rising regulation dam is large. The power generation plan presentation device according to claim 1.

9. The objective function is configured using at least one of the following three functions: a function that gives a higher evaluation as the total amount of water release from a plurality of dams is smaller; a function that gives a higher evaluation as the total amount of power generation from a plurality of dams is larger; and a function that gives a higher evaluation as the total amount of electricity sales from a plurality of dams is larger. The power generation plan presentation device according to claim 1.

10. The execution device is configured to execute a water storage amount initial value acquisition process and a stream flow rate acquisition process. The water storage amount initial value acquisition process is a process of acquiring the initial value of the water storage amount in each of the plurality of dams. The stream flow rate acquisition process is a process of acquiring the stream flow rate flowing into each of the plurality of dams. The optimization process includes a process of substituting, into the inflow amount to a dam, the sum of the water release amount of the dam upstream of the dam and the water intake amount and the stream flow rate flowing into the dam, and a process of updating the initial value of the water storage amount of the dam according to the inflow amount to the dam. The search process is a process of searching for the water intake amount, further including a condition that the water storage amount is equal to or greater than the minimum water storage amount and equal to or less than the maximum water storage amount. The power generation plan presentation device according to claim 1.

11. It includes a storage device. Model specification data is stored in the storage device. The model specification data is data for specifying a regression model that outputs the stream flow rate when the values of input variables are input. The input variables include at least one of two variables: a soil rainfall index and a soil runoff amount. It is configured to execute a precipitation prediction value acquisition process and an input variable calculation process. The precipitation prediction value acquisition process is a process of acquiring the predicted value of the precipitation in a predetermined area near the dam. The input variable calculation process is a process of calculating the at least one using a tank model with the predicted value of the precipitation as input. The stream flow rate acquisition process is a process of calculating the stream flow rate by inputting the values of the input variables calculated by the input variable calculation process into the regression model. The power generation plan presentation device according to claim 10.

12. The input variables include the at least one time series data. The power generation plan presentation device according to claim 11.

13. The input variables include the at least one value for each of a plurality of divisions of a predetermined area near the dam. The power generation plan presentation device according to claim 11.

14. The power generation plan presentation device according to claim 11, wherein the input variable includes a weighted average value of the at least one value for each of a plurality of divided predetermined regions near the dam.

15. Comprising a storage device, The storage device stores past plan data for each of the plurality of dams, The past plan data is data in which the transition of the water intake planned in the past for each of the plurality of dams is associated with the value of a situation variable that is a variable specifying the situation on the target date for the planning of the water intake, The execution device is configured to execute a selection process, The selection process is a process of selecting, from the past plan data, data in which the value of the situation variable is close to the value of the situation variable regarding the target date. The power generation plan presentation device according to claim 1.

16. The plurality of dams are divided into a plurality of groups, The dams included in each of the plurality of groups are connected by the river, The selection process includes a process of selecting the transition of the water intake on the same day in the past for the water intake of the grouped dams. The power generation plan presentation device according to claim 15.

17. The selection process is a process of selecting, from the past plan data, data in which the value of the situation variable is close to the value of the situation variable on the target date for some of the dams within the group. The power generation plan presentation device according to claim 16.

18. The situation variable includes the water intake on the day before the target date and the amount of available water on the target date. The power generation plan presentation device according to claim 15.

19. The execution device is configured to execute a presentation process, The presentation process is a process of presenting to a person information regarding the water intake output by the optimization process. The power generation plan presentation device according to claim 1.

20. The execution device is configured to execute a presentation process, The presentation process is a process of presenting to a person information regarding the water intake output by the optimization process and information regarding the water intake output by the selection process. The power generation plan presentation device according to claim 15.

Images