JP2016005347A - Equipment size estimation device, equipment size estimation method, and equipment size estimation program - Google Patents

Abstract

A processing load for estimating a facility scale is reduced. A peak cut facility size estimation device is provided. The apparatus has a plurality of N-dimensional coordinates expressed by a set of time series data of predicted values of demand electric power per unit period and time series data of predicted values of photovoltaic power generation per unit period for different periods. If present, a clustering unit that classifies a plurality of N-dimensional coordinates into one or a plurality of groups according to the proximity of the distance between the N-dimensional coordinates, and N that represents the group for each of the one or more groups It includes a representative time series extraction unit that specifies a dimension representative coordinate, and a problem solving unit that calculates the scale of the power supply facility based on the specified one or more N-dimensional representative coordinates. [Selection] Figure 3

Description

  The present invention relates to an equipment scale estimation device, an equipment scale estimation method, and an equipment scale estimation program.

  When power consumption exceeds the peak power consumption by combining solar power (PV) power generators and storage batteries in homes, factories, or offices, the power company can supply power through storage battery discharge or solar power generation. A technique for suppressing the normal power amount supplied from an external system such as a peak power amount or less is known. This technique is sometimes called peak cut.

  For example, a predetermined peak power amount is set for each time, and when the power demand exceeds the peak power amount, the power shortage is supplemented with the power from the solar power generation device and the power from the storage battery. As a result, power consumption is peak cut. As described above, the peak power consumption is cut, whereby the amount of normal power supplied from an external system such as an electric power company can be suppressed, and the occurrence of power shortage can be suppressed.

  For example, in connection with load control of power facilities, it is necessary to sequentially grasp the connection information and characteristics of power facilities (power consumption devices, distributed power supply devices, power storage devices, etc.) An electric power system that controls the total load amount is known (for example, Patent Document 1).

  In addition, an information supply system useful for a power consumer to determine the installation of a secondary battery for power storage is known (for example, Patent Document 2).

  Moreover, a system is known that controls the amount of electricity stored and the amount of discharge of the electricity storage means according to the amount of power generated and the amount of power used by the solar power generation means (for example, Patent Document 3).

  An apparatus that creates a representative load pattern used for estimating the amount of power used by a power consumer is known (for example, Patent Document 4).

A method for predicting the future of a time-series waveform is known (for example, Patent Document 5).
In addition, in energy management devices that control the operation of storage batteries and fuel cells according to the prediction of demand for power-using equipment by predicting the amount of load power, taking into account the risk of unforeseen risks, the risk of cost savings or increased costs There is known an apparatus that realizes an equipment operation plan with less (for example, Patent Document 6).

JP 2000-32658 A JP 2002-262456 A JP2013-27214A JP 2008-15921 A JP 2004-266927 A JP 2013-193360 A

  However, when estimating the number of storage batteries and the number of photovoltaic (PV) power generators for peak-cut equipment that can achieve the desired gain, if the equipment size estimate is formulated as a mixed integer programming problem, the number of unknown variables of the problem to be solved The number of simultaneous equations becomes enormous, and there is a problem that it is difficult to solve in practice.

  Therefore, as one aspect, an object of the present invention is to provide an equipment size estimation device, an equipment size estimation method, and an equipment size estimation program that can reduce the processing load for estimating the equipment size.

  A facility size estimation device is provided. The apparatus has different N-dimensional coordinates expressed by a set of time-series data of predicted values of power demand per unit period and time-series data of predicted values of photovoltaic power generation per unit period. When there are a plurality of N-dimensional coordinates, a clustering unit that classifies the plurality of N-dimensional coordinates into one or a plurality of groups according to the proximity of the distance between the N-dimensional coordinates, and represents the group for each of the one or more groups A representative time series extraction unit that identifies the N-dimensional representative coordinates to be performed, and a problem solving unit that calculates the scale of the power supply facility based on the identified one or more N-dimensional representative coordinates.

  The processing load for estimating the equipment scale can be reduced.

It is a figure for demonstrating the outline | summary of a peak cut. It is a figure for demonstrating the outline | summary of the clustering of time series data. It is a figure for demonstrating the outline | summary of the distance between the time series data for clustering. It is a figure for demonstrating the outline | summary of the estimation of the electric power peak change after a peak cut. It is a figure which shows the example of the functional block of the peak cut equipment scale estimation apparatus according to a 1st Example. It is a figure which shows the example of a structure of a peak cut equipment scale estimation apparatus. It is a figure which shows the example of the flow of a process of estimation of the peak cut equipment scale according to a 1st Example. It is a figure which shows the example of the functional block of the peak cut equipment scale estimation apparatus according to a 2nd Example. It is a figure which shows the example of the flow of a process of estimation of the peak cut equipment scale according to a 2nd Example. It is a figure which shows the example of the functional block of the peak cut equipment scale estimation apparatus according to a 3rd Example. It is a figure which shows the example of the flow of a process of estimation of the peak cut equipment scale according to a 3rd Example. It is a figure for demonstrating the outline | summary of the peak cut in a 4th Example. It is a figure which shows the example of the demand power after the peak cut in case storage battery capacity differs. It is a figure which shows the example of the functional block of the peak cut equipment scale estimation apparatus according to a 4th Example. It is a figure which shows the example of the flow of a process of estimation of the peak cut equipment scale according to a 4th Example.

  In the following, referring to the drawings, a storage battery for a peak cut facility that maximizes the gain due to peak cut in consideration of time series of demand power, time series of photovoltaic (PV) power generation, peak cut rate target value, etc. Embodiments of a peak cut facility size estimation device, a peak cut facility size estimation method, and a peak cut facility size estimation program for estimating the number and the number of photovoltaic (PV) power generation devices will be described.

  In the following, an embodiment of an equipment scale estimation device, equipment scale estimation method, and equipment scale estimation program that can reduce the scale of a mixed integer programming problem for estimating equipment scale will be described. It can also be applied to other mixed integer programming problems. In that case, a variable represented by an integer can be estimated using a time series of demand power, a time series of photovoltaic (PV) power generation, and another time series and gain as gains.

  Demand power time series, photovoltaic (PV) power generation time series, peak cut rate target value, etc. If the equipment size estimate is formulated as a mixed integer programming problem when estimating the number of devices, the number of unknown variables of the problem to be solved and the number of simultaneous equations may become enormous. Hereinafter, the number of unknown variables of the problem to be solved, the number of simultaneous equations, and the like may be simply referred to as “the scale of the problem”.

  In such a case, for example, it is conceivable to reduce the problem scale by clustering from a demand power time series for one year and a PV power generation time series and using a representative time series pattern.

  However, in general, when the number of representative time series patterns is small, the problem scale can be reduced, but the error of the peak cut result becomes large. When the number of representative time series patterns is large, the error can be reduced, but the problem scale cannot be reduced. In simple clustering, the trade-off between the number of clusters and peak cut error is unknown.

Therefore, in the following, a set of demand power time series and PV power generation time series for each day is regarded as one multidimensional vector. The distance between the time series data is defined using l ^∞ norm (l infinity norm), that is, the maximum absolute value of the difference in each time series time between these multidimensional vectors. By performing clustering at this distance, it is possible to evaluate the error of the system peak power as a result of performing each system power peak cut in the representative pattern time series. By this method, it is possible to extract a power demand amount and a PV power generation amount time series pattern in consideration of a trade-off between the number of representative pattern time series and the error of the peak cut result.

FIG. 1 is a diagram for explaining an outline of peak cut.
In FIG. 1, each point of D1, D2,..., D6 indicates power demand at times T1, T2,. Further, M1, M2,..., M6 are solar (PV) power generation amounts at times T1, T2,..., T6, and X1, X2, ..., X6 are storage battery discharge amounts at times T1, T2,. Show.

  In FIG. 1, it is shown that the electric power demands D1, D2,..., D6 before the peak cut decrease due to the photovoltaic power generation and the discharge from the storage battery. The maximum peak cut width is at time T4, and the peak cut amount at this time is indicated by Z.

需要のピークＺが正という条件は、各時刻Ｔ１、Ｔ２、…、Ｔｎに対して１つずつ存在するので、合計でｎ個の不等式で表される。また、容量Ａが正であるという条件は、各時刻Ｔ１、Ｔ２、…、Ｔｎに対して１つずつ存在する。もし、３６５日分の条件の下で最適な設備規模を見積るためには、３６５×２ｎ個の条件の下で、ｍｉｎＺを求める問題に帰着する。 The discharge amounts X1, X2,..., Xn at times T1, T2,..., Tn are determined so that the demand peak Z is minimized when the PV power generation amount is Mi with the capacity (accumulated charge amount) A and the initial charge amount B. . If the value of Xi is positive, it is discharging, and if it is negative, it is charging. That is, estimating the optimum equipment scale results in the problem of determining minZ under the following conditions.

The condition that the demand peak Z is positive exists for each time T1, T2,..., Tn, and therefore is represented by a total of n inequalities. Further, there is one condition that the capacity A is positive for each time T1, T2,..., Tn. In order to estimate the optimum equipment scale under the conditions for 365 days, it results in the problem of obtaining minZ under 365 × 2n conditions.

  When designing a peak cut facility, it is necessary to determine the scale of the facility (the number of storage batteries and solar power generation) in consideration of the cost required for the facility and the system power saving effect due to the peak cut.

  The method of estimating the number of storage batteries and photovoltaic power generation that optimizes the total cost including the cost of electricity and the electricity cost saving effect due to the system power peak cut is useful when designing the equipment for peak cut. In the following, considering the daily power demand and PV power generation amount, the device that estimates the number of storage batteries and solar power generation equipment that optimizes the total cost including the equipment cost and the electricity cost saving effect due to peak cut Will be described.

  The following assumptions may be made when formulating the problem. First, the electricity charge system may be a peak charge per day (Critical Peak Pricing). Further, the peak cut target time may be 15 hours from 8:00 to 23:00. Further, the storage battery charge / discharge amount and the PV power generation amount may change every hour. Moreover, it can be assumed that the power demand and the PV power generation amount change in a cycle of one year, and the life of the facility (storage battery) is about 10 years.

Here, symbols used in the following will be described.
The number of days is T, and the number of storage battery charge / discharge amounts and PV power generation amount per day is N. In the above example, T = 365 and N = 16. Let D _{t, n be the} amount of power demand (estimated) when the date is t and the time is n. Here, t = 1,..., T, n = 1,.

によって定義する。 Also, the maximum power demand (estimated) amount Dt for date t

Defined by.

With regard to constants and variables related to costs, the amount of PV power generation (estimated) at date t and time n is _{Pt, n} . The PV power generation (estimated) amount can be assumed to be a positive number _{Pt and n} . Here, t = 1,..., T, n = 1,. Let the target peak cut rate be pcr. The storage battery capacity is A, the number of storage batteries is nb, the price per storage battery is Mb, the number of solar power generation devices (hereinafter sometimes referred to simply as solar cells) is np, and the price per solar power generation device Is Mp. Let Mdt be the electricity charge per unit of date t (1 kWh). It is assumed that the storage battery charge / discharge amount at time t and time n is X _{t, n} , and the maximum power value after peak cut of date t is Zt. Here, t = 1,..., T, n = 1,. The variables are the number of storage batteries nb, the number of solar power generation devices np, the storage battery charge / discharge amount X _{t, n} , and the maximum power value after peak cut is Zt.

ただし、

である。また、［数３］内の定数「１０」は、１０年分を仮定していることに起因する定数である。上の定式化では、変数が（Ｔ＋１）×Ｔ＋２個、制約条件が（２Ｎ＋１）×Ｔ個の混合整数計画問題となる。 The facility scale that optimizes the total cost (equipment cost-peak cut gain), for example, the number nb of storage batteries and the number np of photovoltaic power generation devices, minimizes (reduces) the following amount of mathematical programming problems Can be formulated as:

However,

It is. In addition, the constant “10” in [Equation 3] is a constant caused by assuming 10 years. The above formulation results in a mixed integer programming problem with (T + 1) × T + 2 variables and (2N + 1) × T constraints.

  Formulating 365 days a year for every day becomes a mixed integer programming problem with 6207 variables and 12045 constraints, which is difficult to solve in practice.

  However, experience shows that the annual power demand and PV power generation time series are represented by several typical patterns.

  However, the peak cut result by the representative time series pattern of power demand and PV power generation has an error from the peak cut result of the original time series. Generally, when considering the peak cut of one year, a small number of representative time series patterns. Then, the error of the peak cut result is large, and the error becomes small when a large number of representative patterns are used. For this reason, it is necessary to extract representative patterns in consideration of the trade-off between the number of patterns (number of patterns) and peak cut error, instead of just clustering the time series.

  In the following, considering the trade-off between the number of representative patterns and the peak cut result error, a time series pattern of annual power demand and PV power generation is extracted and used to reduce the problem scale.

FIG. 2 is a diagram for explaining an overview of time-series data clustering.
As described above, time series clustering is performed to extract representative patterns of power demand and PV power generation time series.

  For example, as shown in FIG. 2, clustering refers to the amount of power demand at times T1, T2,..., T6 as sunlight (PV, D2, D2,..., D6, times T1, T2,. ) Let the power generation amounts be M1, M2,..., M6, respectively. It may mean that these sets are regarded as points in a 12-dimensional space. The points in the 12-dimensional space may be called representative time series patterns.

  In general, in clustering, when the number of representative time series patterns is small, the problem size can be reduced, but the error of the peak cut result becomes large. If the number of representative time series patterns is large, the error can be reduced but the problem cannot be reduced. Furthermore, in simple clustering, the relationship between the representative time series pattern and the resulting peak cut error is unknown. Therefore, the distance between time series for clustering that can take into account the trade-off between the number of representative patterns and the peak cut result error is introduced.

  FIG. 3 is a diagram for explaining the outline of the distance between time-series data for clustering.

  As shown in FIG. 3, the distance between time series is defined as follows for clustering that can reflect the error of the peak cut result.

と定義する。上記の距離ｄは、ｌ^∞ノルムに他ならない。 The distance d between the time series data {Di} (i = 1,..., N) and {D ′ i} (i = 1,.

It is defined as The distance d is none other than l ^∞ norm.

  Further, the distance d between the PV power generation time series {Mi}, (i = 1,..., N) and {M ′ i} (i = 1,..., N) can be defined similarly.

によって定義する。この距離ｄは、ピークカット結果の誤差を反映できるようなものである。 Further, between the time series data sets of power demand and PV power generation ({Di}, {Mi}) (i = 1, ..., n) and ({D'i}, {M'i}) The distance d is

Defined by. This distance d is such that the error of the peak cut result can be reflected.

FIG. 4 is a diagram for explaining an outline of estimation of power peak change after peak cut.
With reference to FIG. 4, it will be described that it is possible to estimate the power peak change after the peak cut when the time series data of the demand power amount and the PV power generation amount is changed by the time series distance d defined above.

  Let Z be the power peak after the peak cut when the time series data of the demand power amount is {Di} and the time series data of the PV power generation amount is {Pi}. The power peak after the peak cut when the time series data of the demand power amount is {D′ i} and the time series data of the PV power generation amount is {P′i} is Z ′.

距離ｄ１とｄ２の定義より各時刻で、

が成立する。つまり、需要電力量の時系列データ｛Ｄｉ｝とＰＶ発電量の時系列データ｛Ｐｉ｝が、需要電力量の時系列データ｛Ｄ’ｉ｝とＰＶ発電量の時系列データ｛Ｐ’ｉ｝に変わっても、需要電力量は高々ｄ１しか上がらず、ＰＶ発電量は高々ｄ２しか下がらない。よって、蓄電池の充放電計画が同じなら、図４に示されているように、変更後のピーク電力の差は高々ｄ１＋ｄ２である。実際にはピーク電力は最適化されるので差は更に縮まる。以上によりピーク電力の差は、

のように評価される。この式は、時系列データ間の距離により、ピークカット後の電力ピークの誤差が見積可能であることを意味している。 At this time, the maximum difference between Z and Z ′ is considered.
For this purpose, the distances d1 and d2 are defined as follows.

At each time from the definition of distances d1 and d2,

Is established. That is, the time series data {Di} of demand electric energy and the time series data {Pi} of PV power generation amount are time series data {D'i} of demand electric energy and time series data {P'i} of PV power generation amount. Even if it changes to, the amount of demand electric power will only increase at most d1, and the amount of PV power generation will only decrease at most d2. Therefore, if the charging / discharging plan of the storage battery is the same, as shown in FIG. 4, the difference in peak power after the change is at most d1 + d2. In practice, the peak power is optimized, so the difference is further reduced. Thus, the difference in peak power is

It is evaluated as follows. This equation means that the power peak error after the peak cut can be estimated based on the distance between the time series data.

ただし、ｔ＝０、…、Ｓ−１に対して、

なる拘束条件が課される。上記問題の変数は（Ｎ＋１）×Ｓ＋２、拘束条件の数は(２×Ｎ＋１)×Ｓである。たとえば、Ｓ＝１０、２０、５０であっても良い。しかしながら、Ｓの値は、装置の処理能力によって適切な値を選んでも良い。このように、問題規模を縮小することにより最適化問題を解くのに要する処理時間を短縮することができる。 Gives a peak cut error e, and S representative patterns by clustering whose cluster diameter is equal to or smaller than e, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} , ..., P _{i, N} } (i = 0, ..., the use of S-1), the power consumption for one year, PV power generation pattern _{is {D t, 1, ...,} D t, N; P t, 1, ..., P t, N} with The optimization problem for estimating the optimum facility scale (np, nb), where ND (t) (t = 0,..., S-1) is the sum of electricity charges on a certain day is as follows.

However, for t = 0,..., S−1,

A constraint is imposed. The variable in question is (N + 1) × S + 2, and the number of constraint conditions is (2 × N + 1) × S. For example, S may be 10, 20, or 50. However, the value of S may be selected appropriately depending on the processing capability of the apparatus. In this way, the processing time required to solve the optimization problem can be shortened by reducing the problem scale.

となる。 The cost error due to the extraction of the pattern of the demand power amount and the PV power generation amount is defined as follows. When et is the distance between the demand power amount of the date t, the time series data of the PV power generation amount and the time series data of the representative pattern,

It becomes.

In this way, a set of time-series data of demand electric energy for each day and time-series data of PV power generation is regarded as one multidimensional vector, and l ^∞ norm (time series time at each time of the time series) is considered as one multidimensional vector. Define the distance using the maximum of the absolute difference. By performing clustering at this distance, it is possible to evaluate the error of the system peak power as a result of performing the system power peak cut with the time series data of the representative pattern. With this method, it is possible to extract the power demand and PV power generation time series data patterns in consideration of the trade-off between the number of representative pattern time series data and the peak cut result error. As a result, it is possible to reduce the scale of the mixed integer programming problem for estimating the optimum equipment scale while taking into account the error of the peak cut result and shorten the processing time.

  And the peak cut facility scale estimation apparatus which implement | achieves the above processes is each one of the time series data of the some demand electric power regarding electric power load, and the several photovoltaic power generation amount generated with a photovoltaic power generation apparatus. A plurality of sets composed of one of the time series data of the two, the distance between two of the plurality of sets is externally supplied by supplying a part of the demand power of the power load from the photovoltaic power generation device and the storage battery Clustering is performed to form a cluster in the clustering section so that the peak cut error when the power supply from the peak is cut is less than or equal to the time series data of demand power representing the cluster and the amount of photovoltaic power generation The representative time series including the series data is extracted by the representative time series extraction unit, and the representative time series is used to install the storage battery and the photovoltaic power generation device. Define the problem to reduce the total cost including the cost of preparative optimization problem creating unit, solve the problem in optimization problem solver may be obtained the number and the number of photovoltaic device of the storage battery.

At this time, as in [Equation 6], the first set composed of one of the time series data of the demand power and one of the time series data of the amount of photovoltaic power generation, and the time series data of the demand power The distance between one of the second set consisting of another one of the time series data of the amount of photovoltaic power generation is the time series data of the demand power of the first set and the demand power of the second set Corresponding to the maximum absolute value of the difference between corresponding elements of the time-series data, the time-series data of the first set of photovoltaic power generation amounts, and the time-series data of the second set of photovoltaic power generation amounts The sum of the absolute values of the differences between the elements, ie, l ^∞ norm.

<First embodiment>
The first embodiment will be described with reference to FIGS. In this example, the optimum problem scale is reduced so that the peak cut error is equal to or less than a given value, the optimum number of facilities is calculated with the reduced problem, and is output together with the error.

<< Peak Cut Equipment Scale Estimator >>
FIG. 5 is a diagram illustrating an example of functional blocks of the peak cut facility size estimation apparatus 10 according to the first embodiment.

  The peak cut facility size estimation apparatus 10 includes an input unit 100, a calculation unit 110, and an output unit 120. The calculation unit further includes a time series clustering unit 1102, a representative time series extraction unit 1104, an optimization problem creation unit 1106, and an optimization problem solution unit 1108. The representative time series extraction unit 1104 and the optimization problem creation unit 1106 may be combined to form the optimization problem creation unit 1110. The optimization problem creating unit 1106 and the optimization problem solving unit 1108 may be simply referred to as a problem creating unit 1106 and a problem solving unit 1108, respectively.

The input unit 100 receives various data from the outside for formulating the optimization problem. The data input to the input unit 100 includes a peak cut error e. The data input to the input unit 100 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N). Specifically, T = 365 and N = 16 may be used. Furthermore, the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N) and the target peak cut rate pcr at the date t and time n may be input. Furthermore, the storage battery capacity A, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity charge Mdt per unit of date t (1 kWh) may also be input.

によって定義されたものである。時系列クラスタリング部１１０２は単に、クラスタリング部と呼ぶことがある。 The time-series clustering unit 1102 of the calculation unit 110 includes a power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,..., N) and a date t, where the date is t and the time is n. In order to extract the representative pattern from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N) at time n, as shown in FIG. Data clustering is performed. In this embodiment, clustering is performed so that the diameter is equal to or smaller than the peak cut error e. The distance that determines the size of the diameter e at this time is

Is defined by The time series clustering unit 1102 may be simply referred to as a clustering unit.

As a clustering algorithm used in the time series clustering unit 1102 of the calculation unit 110, the following may be used. That is, given a set of points S = {P1,..., PT} and a distance d between the points, in order to obtain a cluster decomposition of size e and its representative points, first, from the set S Any one Vi is selected. Then, among the elements of the set S, the one having a distance from Pi (e / 2) or less (referred to as Ai) is selected. This process is repeated until the set S becomes empty. Here, {V1,..., VI} is a representative point, and {A1,..., AI} is a cluster of size e. Here, the point Pi is a set of the power demand (estimated) amount D _{t, n} at the time t and the time n, and the PV power generation (estimated) amount P _{t, n} at the time t and the time n, and the distance d is It is defined by the above formula.

In the representative time series extraction unit 1104 of the calculation unit 110, a representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) (i = 0,..., S-1) are extracted.

ただし、ｔ＝０、…、Ｓ−１に対して、

が要請される。上記問題の変数は（Ｎ＋１）×Ｓ＋２、拘束条件の数は(２×Ｎ＋１)×Ｓである。問題規模を縮小することにより処理時間を短縮することができる。 The optimization problem creation unit 1106 of the calculation unit 110 includes S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} , ..., P _{i, N} } (i = 0, ..., S-1), the day when the power consumption and PV power generation pattern are {D _{t, 1} , ..., D _{t, N} ; P _{t, 1} , ..., P _{t, N} } in one year An optimization problem for estimating an optimum facility scale (np, nb) is created with ND (t) (t = 0,..., S-1) as a sum of electricity charges. In this example, the optimization problem is formulated as follows.

However, for t = 0,..., S−1,

Is requested. The variable in question is (N + 1) × S + 2, and the number of constraint conditions is (2 × N + 1) × S. By reducing the problem scale, the processing time can be shortened.

  The optimization problem solving unit 1108 of the calculation unit 110 solves the optimization problem created by the optimization problem creation unit 1106 of the calculation unit 110 using, for example, linear programming. As the linear programming method, for example, a simplex method may be used. Further, the optimization problem created by the optimization problem creation unit 1106 of the calculation unit 110 may be solved using an annealing method, a genetic algorithm, or the like. Then, the optimization problem solving unit 1108 of the calculation unit 110 obtains the optimum facility scale, that is, the number of storage batteries nb and the number np of photovoltaic power generation devices (hereinafter sometimes simply referred to as solar cells). The optimization problem creation unit 1106 may be simply referred to as a problem creation unit, and the optimization problem solution unit 1108 may be simply referred to as a problem solution unit.

  The output unit 120 outputs the optimum facility scale (nb, np) and peak cut error e obtained by the optimization problem solving unit 1108 of the calculation unit 110. The output form may be an image, sound, electronic file, or the like.

  As described above, the facility size estimation device 10 of the power supply facility includes the time series data of the predicted value of the demand power amount per unit period and the time series data of the predicted value of the photovoltaic power generation amount per unit period. When there are a plurality of N-dimensional coordinates expressed in pairs for different periods, a clustering unit 1102 classifies the plurality of N-dimensional coordinates into one or a plurality of groups according to the proximity of the distance between the N-dimensional coordinates. And a representative time series extraction unit 1104 that identifies an N-dimensional representative coordinate representing the group for each of the one or more groups, and the scale of the power supply facility is calculated based on the identified one or more N-dimensional representative coordinates. And a problem solving unit 1108.

  Furthermore, the facility size estimation device 10 of the power supply facility uses the N-dimensional representative coordinates representing the group, the first cost for installing the power supply facility, and the power charge due to the installation of the power supply device A problem creating unit 1106 for defining a problem that reduces the total cost including the second cost due to peak cut may be included. In this case, the problem solving unit 1108 solves the problem and calculates the scale of the power supply facility.

  FIG. 6 is a diagram illustrating an example of the configuration of the peak cut facility scale estimation apparatus 10.

  The computer 400 includes a Central Processing Unit (CPU) 402, a Read Only Memory (ROM) 404, and a Random Access Memory (RAM) 406. The computer 400 further includes a hard disk device 408, an input device 410, a display device 412, an interface device 414, and a recording medium driving device 416. Note that these components are connected via a bus line 418, and various data can be exchanged under the control of the CPU 402.

  A central processing unit (CPU) 402 is an arithmetic processing unit that controls the operation of the entire computer 400, and functions as a control processing unit of the computer 400.

  A Read Only Memory (ROM) 404 is a read-only semiconductor memory in which a predetermined basic control program is recorded in advance. The CPU 402 can control the operation of each component of the computer 400 by reading out and executing the basic control program when the computer 400 is activated.

  A random access memory (RAM) 406 is a semiconductor memory that can be written and read at any time and used as a working storage area as necessary when the CPU 402 executes various control programs.

  The hard disk device 408 is a storage device that stores various control programs executed by the CPU 402 and various data. The CPU 402 can perform various control processes to be described later by reading and executing a predetermined control program stored in the hard disk device 408.

  The input device 410 is, for example, a mouse device or a keyboard device. When operated by a user of the information processing device, the input device 410 acquires input of various information associated with the operation content, and sends the acquired input information to the CPU 402. To do.

  The display device 212 is a liquid crystal display, for example, and displays various texts and images according to display data sent from the CPU 402.

  The interface device 414 manages the exchange of various information with various devices connected to the computer 400.

  The recording medium driving device 416 is a device that reads various control programs and data recorded on the portable recording medium 420. The CPU 402 can read out and execute a predetermined control program recorded on the portable recording medium 420 via the recording medium driving device 216, thereby performing various control processes described later. As the portable recording medium 420, for example, a flash memory equipped with a USB (Universal Serial Bus) standard connector, a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disc Only Only). and so on.

  In order to configure the peak cut facility size estimation apparatus 10 using such a computer 400, for example, a peak cut facility size estimation program (control program) for causing the CPU 202 to perform the processing in each processing unit described above is created. . The created peak cut facility size estimation program is stored in advance in the hard disk device 408 or the portable recording medium 418. Then, a predetermined instruction is given to the CPU 402 to read and execute the peak cut facility size estimation program. By doing so, the functions provided in the peak cut facility size estimation device 10 are provided by the CPU 202.

  With the configuration as described above, the peak cut facility size estimation apparatus 10 reduces the problem scale and estimates the facility scale while considering the mixed integer programming problem for estimating the optimum facility scale while taking into account the error of the peak cut result. The processing load can be reduced.

<< Peak cut facility size estimation process >>
With reference to FIG. 7, the peak cut facility scale estimation process will be described.

  When the peak cut facility size estimation device 10 is a general-purpose computer 200 as shown in FIG. 6, the following description defines a control program that performs such processing. That is, hereinafter, it is also an explanation of a peak cut facility size estimation program (control program) that causes a general-purpose computer to perform the processing described below.

When the process is started, the input unit 100 of the peak cut facility size estimation apparatus 10 receives an input of the peak cut error e in S100. The data input to the input unit 100 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N), date t, PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1, ..., N), the target peak cut rate pcr may be input. Furthermore, the storage battery capacity A, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity charge Mdt per unit of date t (1 kWh) may be input.

In the next S102, the time-series clustering unit 1102 of the peak cut facility size estimation device 10 determines the power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1) at the date t and time n. In order to extract a representative pattern from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,. As shown in FIG. 2, clustering of these data is performed. In this embodiment, clustering is performed so that the diameter is equal to or smaller than the peak cut error e. At this time, the distance that defines the size of the diameter e is the one defined in [Equation 13] above. When the process of this step ends, the process proceeds to S104.

In step S104, the representative time series extraction unit 1104 of the peak cut facility size estimation apparatus 10 generates a representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _i from each cluster. _{, N} }) (i = 0,..., S-1). When the process of this step ends, the process proceeds to S106.

In S106, the optimization problem creating unit 1106 of the peak cut facility size estimation apparatus 10 performs S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1 ,.} } (I = 0,..., S-1), the power consumption and PV power generation pattern in one year are {D _{t, 1} ,..., D _{t, N} ; P _{t, 1 ,.} The sum of the electricity charges on the day _N } is ND (t) (t = 0,..., S−1), and an optimization problem for estimating the optimum facility scale (np, nb) is created. The resulting optimization problem may be, for example, as shown in [Equation 14] to [Equation 15]. When the process of this step ends, the process proceeds to S108.

  In S108, the optimization problem solving unit 1108 of the peak cut facility size estimation apparatus 10 solves the optimization problem created by the optimization problem creation unit 1106 of the calculation unit 110, and sets the optimum facility size, that is, the number of storage batteries as nb. Np is obtained for the number of photovoltaic power generation devices (hereinafter sometimes simply referred to as solar cells).

  In S108, the output unit 120 of the peak cut facility size estimation apparatus 10 outputs the optimum facility scale (nb, np) and the peak cut error e obtained by the optimization problem solving unit 1108 of the calculation unit 110.

  The mixed integer programming problem for estimating the optimum equipment scale by the above processing can be reduced while considering the error of the peak cut result, and the processing load for estimating the equipment scale can be reduced. .

<Second embodiment>
A second embodiment will be described with reference to FIGS. In this example, the optimum problem is reduced to the minimum scale that satisfies the given peak cut error, the optimum number of facilities is calculated with the reduced problem, and the error is output together with the error.

<< Peak Cut Equipment Scale Estimator >>
FIG. 8 is a diagram illustrating an example of functional blocks of the peak cut facility size estimation apparatus 20 according to the second embodiment. The peak cut facility size estimation apparatus 20 may be configured by a general-purpose computer 400 as shown in FIG.

  The peak cut facility size estimation device 20 includes an input unit 200, a calculation unit 210, and an output unit 220. The calculation unit 210 further includes a time series clustering unit 2102, a cluster diameter determining unit 2104, a representative time series extracting unit 2106, an optimization problem creating unit 2108, and an optimization problem solving unit 2110. The representative time series extraction unit 2106 and the optimization problem creation unit 2108 may be combined to form the optimization problem creation unit 2112. The optimization problem creating unit 2108 and the optimization problem solving unit 2110 may be simply referred to as a problem creating unit 2108 and a problem solving unit 2110, respectively.

  The input unit 200 has the same or similar configuration as the input unit 100 of the peak cut facility size estimation device 10 according to the first embodiment. Therefore, detailed description is omitted.

The time series clustering unit 2102 calculates the power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,..., N) and the date t and time n at the date t and time n. Clustering is performed in order to find the minimum number of time-series clusters such that the maximum cluster diameter is equal to or less than e from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N). I do.

  For example, the time-series clustering unit 2102 performs clustering so that the number of time-series clusters is S. The clustering algorithm described above may be used, or another algorithm may be used. The initial value of S may be 1 or any other number.

  The cluster diameter determination unit 2104 determines whether the maximum diameter of the cluster generated by the time series clustering unit 2102 is equal to or less than e. If the maximum diameter is equal to or less than e, the result is sent to the representative time series extraction unit 2106. send. Also, the cluster diameter determining unit 2104 increases S by one and returns the processing to the time-series clustering unit 2102 if the maximum diameter is larger than e.

  In this way, the time-series clustering unit 2102 and the cluster diameter determining unit 2104 obtain the minimum number of time-series clusters such that the maximum cluster diameter is equal to or less than e. The distance that determines the size of the diameter e at this time is defined by [Equation 13].

In the representative time series extraction unit 2106, a representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) (i = 0,. S-1) is extracted. The representative time series extraction unit 2106 has the same or similar configuration as the representative time series extraction unit 1104 of the peak cut facility size estimation device 10 according to the first embodiment. Therefore, detailed description is omitted.

The optimization problem creating unit 2108 includes S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} } (i = 0,..., S− 1), the sum of electricity charges for the day when the power consumption and PV power generation pattern are {D _{t, 1} , ..., D _{t, N} ; P _{t, 1} , ..., P _{t, N} } Is an ND (t) (t = 0,..., S-1), and an optimization problem for estimating the optimum facility scale (np, nb) is created. The optimization problem creation unit 2108 has the same or similar configuration as the optimization problem creation unit 1106 of the peak cut facility size estimation apparatus 10 according to the first embodiment. Therefore, detailed description is omitted.

  The optimization problem solving unit 2110 solves the optimization problem created by the optimization problem creating unit 2106 using, for example, linear programming. The optimization problem solving unit 2108 has the same or similar configuration as the optimization problem solving unit 1106 of the peak cut facility size estimation apparatus 10 according to the first embodiment. Therefore, detailed description is omitted.

  The output unit 220 outputs the optimum facility scale (nb, np) obtained by the optimization problem solving unit 2120 of the calculation unit 110. The output form may be an image, sound, electronic file, or the like.

  The time series clustering unit 2102 and the cluster diameter determining unit 2104 may be combined and simply called a clustering unit.

<< Peak cut facility size estimation process >>
With reference to FIG. 9, the peak cut facility scale estimation process will be described.

  When the peak cut facility size estimation device 20 is a general-purpose computer 200 as shown in FIG. 6, the following description defines a control program for performing such processing. That is, hereinafter, it is also an explanation of a peak cut facility size estimation program (control program) that causes a general-purpose computer to perform the processing described below.

When the process is started, the input unit 200 of the peak cut facility size estimation apparatus 20 receives an input of the peak cut error e in S200. The data input to the input unit 200 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N), date t, PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1, ..., N), the target peak cut rate pcr may be input. Furthermore, the storage battery capacity A, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity charge Mdt per unit of date t (1 kWh) may be input.

  In the next S202, the time-series clustering unit 2102 of the peak cut facility size estimation apparatus 20 sets the number N of clusters to 1. When the process of this step ends, the process proceeds to S204.

In S204, the cluster diameter determining unit 2104 of the peak cut facility size estimation device 20 determines the power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,. N), date t, and PV generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N) at time n, representative patterns for generating S clusters Is extracted, clustering of these data is performed as shown in FIG. In this embodiment, clustering is performed so that the number of clusters is S. When the process of this step ends, the process proceeds to S206.

  In S206, the cluster diameter determination unit 2104 determines whether the maximum diameter of the N clusters generated in S204 is larger than the peak cut error e. If the result of this determination is “Yes”, that is, if the maximum diameter of the N clusters is greater than the peak cut error e, the process proceeds to S208. If the result of this determination is “No”, that is, if the maximum diameter of N clusters is equal to or smaller than the peak cut error e, the process proceeds to S210.

  In S208, the cluster diameter determining unit 2104 increases the number N of clusters by one. When the process of this step ends, the process returns to S204.

In S210, the representative time series extraction unit 2106 obtains a representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) (i = 0, ..., S-1) is extracted. When the process of this step ends, the process proceeds to S212.

In S212, the optimization problem creating unit 2108 of the peak cut facility size estimation apparatus 20 performs S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1 ,.} } (I = 0,..., S-1), the power consumption and PV power generation pattern in one year are {D _{t, 1} ,..., D _{t, N} ; P _{t, 1 ,.} The sum of the electricity charges on the day _N } is ND (t) (t = 0,..., S−1), and an optimization problem for estimating the optimum facility scale (np, nb) is created. The resulting optimization problem may be, for example, as shown in [Equation 14] to [Equation 15]. When the process of this step ends, the process proceeds to S214.

  In S214, the optimization problem solving unit 2110 of the peak cut facility size estimation device 20 solves the optimization problem created by the optimization problem creation unit 2108 of the calculation unit 210, and sets the optimum facility size, that is, the number of storage batteries as nb. The number np of solar power generation devices (hereinafter sometimes simply referred to as solar cells) is obtained.

  In S214, the output unit 220 of the peak cut facility size estimation apparatus 20 outputs the optimum facility size (nb, np) and the peak cut error e obtained by the optimization problem solving unit 2110 of the calculation unit 210.

  The mixed integer programming problem for estimating the optimum equipment scale by the above processing can be reduced while considering the error of the peak cut result, and the processing load for estimating the equipment scale can be reduced. .

<Third embodiment>
A third embodiment will be described with reference to FIGS. In this example, the problem size (allowable variable number: Nv, allowable constraint formula number: Nr) that is allowable in calculation time and the peak cut error e that is allowable in calculation accuracy are input, and the problem can be solved within the problem size and error. Calculate and output the optimal number of facilities.

<< Peak Cut Equipment Scale Estimator >>
FIG. 10 is a diagram illustrating an example of functional blocks of the peak cut facility size estimation device 30 according to the third embodiment. The peak cut facility scale estimation device 30 may be configured by a general-purpose computer 400 as shown in FIG.

  The peak cut facility size estimation device 20 includes an input unit 200, a calculation unit 210, and an output unit 220. The calculation unit 210 further includes a cluster number calculation unit 3102, a time series clustering unit 3104, a cluster diameter determination unit 3106, a representative time series extraction unit 3108, an optimization problem creation unit 3110, and an optimization problem solution unit 3112. The representative time series extraction unit 3108 and the optimization problem creation unit 3110 may be combined to form the optimization problem creation unit 3114.

The input unit 300 receives various data from the outside for formulating the optimization problem. The data input to the input unit 100 includes the allowable variable number Nv, the allowable constraint equation number Nr, and the peak cut error e. The data input to the input unit 100 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N). Specifically, T = 365 and N = 16 may be used. Furthermore, the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N) and the target peak cut rate pcr at the date t and time n may be input. Furthermore, the storage battery capacity A, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity charge Mdt per unit of date t (1 kWh) may also be input.

The cluster number calculation unit 3102 calculates the cluster number S from the allowable variable number Nv and the allowable constraint expression number Nr according to the following equation.

The time-series clustering unit 3104 calculates the power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,..., N) and the date t and time n at the date t and time n. Clustering is performed so that the number of clusters is S from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N). Here, as the clustering algorithm used by the time-series clustering unit 3104, the above-described algorithm may be used, or another algorithm may be used. The initial value of S may be 1 or any other number.

  The cluster diameter determining unit 3106 determines whether or not the maximum diameter of the cluster generated by the time-series clustering unit 3104 is larger than the peak cut error e. The data is sent to the representative time series extraction unit 3108. In addition, if the maximum diameter is larger than the peak cut error e, the cluster diameter determining unit 3106 determines that calculation is not possible within the allowable scale and error, and ends the processing.

  In this way, the time-series clustering unit 3104 and the cluster diameter determining unit 3106 obtain S time-series clusters such that the maximum cluster diameter is equal to or less than e. The distance that determines the size of the diameter e at this time is defined by [Equation 13].

In the representative time series extraction unit 3108, the representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) (i = 0,. S-1) is extracted. The representative time series extraction unit 3108 has the same or similar configuration as the representative time series extraction unit 1104 of the peak cut facility size estimation apparatus 10 according to the first embodiment. Therefore, detailed description is omitted.

The optimization problem creating unit 3110 includes S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} } (i = 0,..., S− 1), the sum of electricity charges for the day when the power consumption and PV power generation pattern are {D _{t, 1} , ..., D _{t, N} ; P _{t, 1} , ..., P _{t, N} } Is an ND (t) (t = 0,..., S-1), and an optimization problem for estimating the optimum facility scale (np, nb) is created. The optimization problem creation unit 3110 has the same or similar configuration as the optimization problem creation unit 1106 of the peak cut facility size estimation apparatus 10 according to the first embodiment. Therefore, detailed description is omitted.

  The optimization problem solving unit 3112 solves the optimization problem created by the optimization problem creating unit 3110 using, for example, linear programming. The optimization problem solving unit 3112 has the same or similar configuration as the optimization problem solving unit 1106 of the peak cut facility size estimation apparatus 10 according to the first embodiment. Therefore, detailed description is omitted. The optimization problem creating unit 3110 and the optimization problem solving unit 3112 may be simply referred to as a problem creating unit 3110 and a problem solving unit 3112, respectively.

  The output unit 320 outputs the optimum facility scale (nb, np) obtained by the optimization problem solving unit 3112. The output form may be an image, sound, electronic file, or the like.

  The time series clustering unit 3104 and the cluster diameter determining unit 3106 may be combined and simply called a clustering unit.

<< Peak cut facility size estimation process >>
With reference to FIG. 11, the peak cut facility size estimation process will be described.

  When the peak cut facility size estimation device 30 is a general-purpose computer 200 as shown in FIG. 6, the following description defines a control program for performing such processing. That is, hereinafter, it is also an explanation of a peak cut facility size estimation program (control program) that causes a general-purpose computer to perform the processing described below.

When the processing is started, the input unit 300 of the peak cut facility size estimation apparatus 30 receives the allowable variable number Nv, the allowable constraint equation number Nr, and the peak cut error e in S300. The data input to the input unit 300 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N), date t, PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1, ..., N), the target peak cut rate pcr may be input. Furthermore, the storage battery capacity A, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity charge Mdt per unit of date t (1 kWh) may be input.

  In next S302, the cluster number calculation unit 3102 of the peak cut facility size estimation apparatus 30 calculates the cluster number S from the allowable variable number Nv and the allowable constraint equation number Nr according to [Equation 16]. When the process in this step ends, the process proceeds to S304.

In S304, the time-series clustering unit 3104 determines the power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,..., N), the date t, and the time n at the date t and time n. Clustering is performed so that the number of clusters becomes S from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N). When the process of this step ends, the process proceeds to S306.

  In S306, the cluster diameter determining unit 3106 determines whether or not the maximum diameter of the cluster generated in S304 is larger than the peak cut error e. If the result of the determination is “Yes”, that is, if the maximum diameter of the cluster is larger than the peak cut error e, the calculation is impossible within the allowable scale and error, and the process is terminated. If the determination result is “No”, that is, if the maximum diameter of the cluster is equal to or smaller than the peak cut error e, the process proceeds to S308.

In S308, the representative time series extraction unit 2106 obtains a representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) (i = 0, ..., S-1) is extracted. When the process of this step ends, the process proceeds to S310.

In S310, the optimization problem creation unit 3110 of the peak cut facility size estimation device 30 performs S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1 ,.} } (I = 0,..., S-1), the power consumption and PV power generation pattern in one year are {D _{t, 1} ,..., D _{t, N} ; P _{t, 1 ,.} The sum of the electricity charges on the day _N } is ND (t) (t = 0,..., S−1), and an optimization problem for estimating the optimum facility scale (np, nb) is created. The resulting optimization problem may be, for example, as shown in [Equation 14] to [Equation 15]. When the process of this step ends, the process proceeds to S312.

  In S312, the optimization problem solving unit 3112 of the peak cut facility size estimation apparatus 30 solves the optimization problem created by the calculation unit optimization problem creation unit 3110, and sets the optimum facility size, that is, the number of storage batteries as nb and sunlight. Np is obtained for the number of power generation devices (hereinafter sometimes simply referred to as solar cells).

  In S312, the output unit 320 of the peak cut facility size estimation apparatus 30 outputs the optimum facility scale (nb, np) obtained by the optimization problem solving unit 3112 of the calculation unit 310.

  The mixed integer programming problem for estimating the optimum equipment scale by the above processing can be reduced while considering the error of the peak cut result, and the processing load for estimating the equipment scale can be reduced. .

<Fourth embodiment>
A fourth embodiment will be described with reference to FIGS. In this example, the peak cut facility size is estimated in consideration of the aging deterioration of the storage battery. For example, the number of storage batteries required for peak cut facilities and the number of photovoltaic power generations are estimated considering the time series for power demand, time series of photovoltaic (PV) power generation, peak cut rate target value, and storage battery aging. A technique to reduce the scale of the problem of estimating

  When taking into account the aging of the storage battery, the storage battery capacity also changes (decreases). Therefore, only by reducing the demand power time series and PV power generation time series, the demand power and PV power generation coefficient are the same, but there are many constraint conditions with different storage battery capacity coefficients due to aging degradation, etc. It is difficult to reduce the scale of the problem. Below, the peak cut facility size estimation device and the peak cut facility size estimate that can reduce the problem size of the mixed integer programming problem that estimates the optimal facility size considering the aging of the storage battery and thereby shorten the processing time. A method and a peak cut facility size estimation program will be described.

  In the following, it is assumed that the aging of the storage battery only brings about a decrease in the amount of charge in the fully charged state of the storage battery. Therefore, the difference in aging deterioration of the storage battery is measured as the difference in the charge amount in the fully charged state of the storage battery.

Referring to FIG. 1 again, in FIG. 1, points D1, D2,..., D6 indicate power demands at times T1, T2,. Further, M1, M2,..., M6 are solar (PV) power generation amounts at times T1, T2,..., T6, and X1, X2, ..., X6 are storage battery discharge amounts at times T1, T2,. Show. The discharge amounts X1, X2,..., Xn at times T1, T2,..., Tn are determined so that the demand peak Z is minimized when the PV power generation amount is Mi with the capacity (accumulated charge amount) A and the initial charge amount B. . If the value of Xi is positive, it is discharging, and if it is negative, it is charging. That is, estimating the optimum equipment scale results in the problem of determining minZ under the following conditions.

  The condition that the demand peak Z is positive is expressed by a total of n inequalities, and the condition that the capacity A is positive results in a problem of obtaining minZ under 365 × 2n conditions. Is as already described.

  When designing a peak cut facility, it is necessary to determine the scale of the facility (the number of storage batteries and solar power generation) in consideration of the cost required for the facility and the system power saving effect due to the peak cut.

  When formulating the facility size estimation problem for determining the size of the facility, the following assumptions may be made.

  First, the electricity charge system may be a peak charge per day (Critical Peak Pricing). Further, the peak cut target time may be 15 hours from 8:00 to 23:00. Further, the storage battery charge / discharge amount and the PV power generation amount may change every hour. Moreover, it can be assumed that the power demand and the PV power generation amount change in a cycle of one year, and the life of the facility (storage battery) is about 10 years. Moreover, the initial charge amount of each day may be a full charge state. For example, every day at 8:00, the charge amount B of the storage battery may be equal to the capacity A, that is, A = B.

Here, a symbol is defined.
_{Let DT, n be the} amount of power demand (estimated) at date t and time n. t = 1,..., T, n = 1,. Let Dt be the maximum value of power demand (estimated) for the day. The amount of PV power generation (estimated) at date t and time n is _{Pt, n} . The PV power generation (estimated) amount can be assumed to be a positive number _{Pt and n} . Here, t = 1,..., T, n = 1,. Let the target peak cut rate be pcr. The storage battery capacity at date t is At, and At changes from day to day due to aging. This is different from the previous embodiment.

It is assumed that the number of storage batteries is nb, the price per storage battery is Mb, the number of solar power generation devices (hereinafter simply referred to as solar cells) is np, and the price per solar power generation device is Mp. Let Mdt be the electricity charge per unit of date t (1 kWh). It is assumed that the storage battery charge / discharge amount at time t and time n is X _{t, n} , and the maximum power value after peak cut of date t is Zt. Here, t = 1,..., T, n = 1,. The variables are the number of storage batteries nb, the number of solar power generation devices np, the date t, the storage battery charge / discharge amount at time n is Xt _{, n} , and the maximum power value after peak cut is Zt. Here, the values to be obtained are the number nb of storage batteries, the number np of photovoltaic power generation devices, the storage battery charge / discharge amount Xt _{, n,} and the maximum power value Zt after peak cut is an unknown value.

ただし、

である。「Ｍｂ×ｎｂ＋Ｍｐ×ｎｐ」は、蓄電池と太陽光発電装置を設置するためのコスト（第１のコスト）を表す。また、「Ｍｄ×（Ｄ_ｉ−Ｚ_ｉ）」（ｉ＝１、…、Ｔ）は、ピークカットによるコスト（第２のコスト）を表す。上の式と［数３］、［数４］とでは、蓄電池容量が日にちｔごとに変化することが異なっている。上式では、変数が（Ｎ＋１）×Ｔ＋２個、制約条件が、（２Ｎ＋１）×Ｔ個の混合整数計画問題が定式化される。ここで、蓄電池容量Ａｔは、日にちｔに関する関数であり、経年劣化により蓄電池容量は日にちｔに関して単調に減少すると仮定しても良い。つまり、蓄電池容量Ａｔは、日にちｔに関する単調減少関数であると仮定しても良い。 The facility scale that optimizes the total cost (equipment cost-peak cut gain), for example, the number nb of storage batteries and the number np of photovoltaic power generation devices, minimizes (reduces) the following amount of mathematical programming problems Can be formulated as:

However,

It is. “Mb × nb + Mp × np” represents a cost (first cost) for installing the storage battery and the solar power generation device. “Md × (D _i −Z _i )” (i = 1,..., T) represents a cost (second cost) due to peak cut. The above equation differs from [Equation 3] and [Equation 4] in that the storage battery capacity changes every day t. In the above formula, a mixed integer programming problem with (N + 1) × T + 2 variables and (2N + 1) × T constraints is formulated. Here, the storage battery capacity At is a function related to the date t, and it may be assumed that the storage battery capacity decreases monotonously with respect to the date t due to aging. That is, the storage battery capacity At may be assumed to be a monotonically decreasing function related to the date t.

となる。ただし、

ここで、需要電力、ＰＶ発電量時系列（｛Ｄ_{ｓ（ｔ）、ｉ}｝、｛Ｐ_{ｓ（ｔ）、ｉ}｝）（ｉ＝１、…、Ｎ、ｔ＝１、…、Ｔ）代表パタンのどれかで実質Ｓ種類だけ存在する。つまり、需要電力時系列とＰＶ発電量時系列は同じだが、日にちｔでの蓄電池容量Ａｔが異なるものが多数存在する。そこで本例では、そのような変数・拘束条件の削減を図る。 Similar to the previous embodiment, if time series extraction is performed, the equipment optimization problem is that the peak cut error when et (t = 1,..., T) is the demand power / PV power generation time series as a representative pattern ( Estimate) amount, s (t) as the demand power of date t, and PV power generation time series representative pattern number,

It becomes. However,

Here, demand power, PV power generation time series ({D _{s (t), i} }, {P _{s (t), i} }) (i = 1,..., N, t = 1,..., T) There are only S types of patterns. That is, the demand power time series and the PV power generation amount time series are the same, but there are many different storage battery capacities At at the date t. Therefore, in this example, such variables and constraint conditions are reduced.

  FIG. 12 is a diagram for explaining the outline of the peak cut in the fourth embodiment. As shown in FIG. 12, it is known as a general property of the peak cut problem that the power demand after the peak cut is uniform throughout the day, and that constant value tends to become the peak power. Yes. A curve L1 in FIG. 12 is an example of a curve indicating the amount of power demand before peak cut, and a curve L2 is an example of a curve indicating the amount of power demand after peak cut. It can be seen that the curve L2 has a substantially constant value even when the time changes.

であるとする。 For example, it is assumed that there are two peak cut problems in which the storage battery capacities (initial charge amounts) A and A ′ are different from each other with the same demand power and PV power generation amount pattern. Here, generality is not lost even if A> A ′. When using the symbols in the above equation, problem 1 is the problem of minimizing Zt as shown below, min (Zt).

Suppose that

であるとする。 Further, it is assumed that Problem 2 is a problem of minimizing Z′t as described below, min (Z′t).

Suppose that

  FIG. 13 shows an outline of the change in power demand after peak cut for two peak cut problems that differ only in the storage battery capacity (initial charge amount A × nb> A ′ × nb).

として良い。上式は、蓄電池の満充電状態の充電量の違いに起因するコストを表す。ここで、Ｎは１日のうち、ピークカットをする時刻の数である。たとえば、ピークカット対象時間は８：００〜２３：００の１５時間であれば、Ｎ＝１６である。 As shown in FIG. 13, the peak power of the power demand (estimated) amount obtained by the peak cut problem 2 is larger than that of the power demand (estimated) amount obtained by the peak cut problem 1 from the difference in the initial charge amount. . In FIG. 13, a curve L1 is an example of a curve indicating the amount of power demand before peak cut, as in the case of FIG. Curve L3 is an example of a curve showing the amount of power demand after peak cut when the initial charge amount is A ′ × nb, and curve L4 is the amount of power demand after peak cut when the initial charge amount is A × nb. It is an example of the curve shown. And the difference in the (peak) demand power amount between the curves L3 and L4 is considered to be caused only by the difference between the initial charge amounts A × nb and A ′ × nb, and the difference is approximately

As good. The above formula represents the cost due to the difference in the charged amount of the storage battery in the fully charged state. Here, N is the number of times of peak cut in one day. For example, if the peak cut target time is 15 hours from 8:00 to 23:00, N = 16.

が近似的に成立すると考えることが出来る。すなわち、蓄電池の満充電状態の充電量の違いは、蓄電池の満充電状態の充電量の違いに起因するコスト（第３のコスト）によって考慮され得る。これにより多数の同じ需要電力、ＰＶ発電量パタンを持つ拘束条件を１組にまとめることで、（近似的に）問題規模の縮小が可能となる。 Therefore, the peak power problem (A> A ′), min (Zt), and min (Z′t) with the same power demand / PV power generation amount pattern but different storage battery capacities

Can be considered to be established approximately. That is, the difference in the charged amount of the storage battery in the fully charged state can be considered by the cost (third cost) caused by the difference in the charged amount of the fully charged state of the storage battery. As a result, the problem scale can be reduced (approximately) by combining the constraint conditions having a large number of the same demand power and PV power generation amount patterns into one set.

ただし、

である。［数２６］において、「Ｍｂ×ｎｂ＋Ｍｐ×ｎｐ」は、蓄電池と太陽光発電装置を設置するためのコスト（第１のコスト）を表す。また、「Ｍｄ_ｉ×Ｋ（ｉ）×（Ｄ_ｉ−Ｚ_ｉ）」（ｉ＝１、…、Ｔ）は、ピークカットによるコスト（第２のコスト）を表す。さらに、［数２６］では、第１、第２のコストに加え、［数２４］の形をした蓄電池の満充電状態の充電量の違いに起因するコスト（第３のコスト）を含んでいる。上記問題の変数は（Ｎ＋１）×Ｓ＋２、拘束条件数は（２Ｎ＋１）×Ｓである。 Eventually, optimum equipment scale (np, nb) the optimization problem to estimate the cost function D _t, the section on e _t is omitted because there is no impact on the optimization, the power demand of the day t, PV power generation amount time series The representative pattern number is s (t), the demand power, the storage battery charge amount representing the PV power generation amount pattern i is T (i), the power consumption in one year, the PV power generation amount pattern is {D _{i, 1} ,. D _{i, N} ; P _{i, 1} ,..., P _{i, N} }, that is, the number of days that is the i-th demand power / PV power generation amount pattern is K (i).

However,

It is. In [Equation 26], “Mb × nb + Mp × np” represents the cost (first cost) for installing the storage battery and the solar power generation device. Further, “Md _i × K (i) × (D _i −Z _i )” (i = 1,..., T) represents a cost (second cost) due to peak cut. Further, in [Equation 26], in addition to the first and second costs, a cost (third cost) caused by a difference in the charged amount in the fully charged state of the storage battery in the form of [Equation 24] is included. . The variable in question is (N + 1) × S + 2, and the number of constraint conditions is (2N + 1) × S.

Once formulated in this way, applying the techniques of the first to third embodiments can reduce the problem scale of the mixed integer programming problem that estimates the optimum facility scale considering the aging of the storage battery. The processing load for estimating the equipment scale can be reduced.
The problem formulated as the above equation is also called “optimization problem considering aging degradation”.

<< Peak Cut Equipment Scale Estimator >>
FIG. 14 is a diagram illustrating an example of functional blocks of the peak cut facility size estimation device 40 according to the fourth embodiment. FIG. 14 shows an example of a peak cut facility size estimation device that takes into account the aging of the storage battery while using the clustering technique described in the third embodiment. However, the clustering techniques described in the first and second embodiments may be used.

  The peak cut facility size estimation device 40 includes an input unit 400, a calculation unit 410, and an output unit 420. The calculation unit further includes a time series clustering unit 4102, a representative time series extraction unit 4104, an optimization problem creation unit 4106, and an optimization problem solution unit 4108. The representative time series extraction unit 4104 and the optimization problem creation unit 4106 may be combined to form the optimization problem creation unit 4110. The optimization problem creating unit 4106 and the optimization problem solving unit 4108 may be simply referred to as a problem creating unit 4106 and a problem solving unit 4108, respectively.

The input unit 400 receives various data from the outside for formulating the optimization problem. The data input to the input unit 400 includes the allowable variable number Nv, the allowable constraint equation number Nr, and the peak cut error e. The data input to the input unit 400 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N). Specifically, T = 365 and N = 16 may be used. Furthermore, the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N) and the target peak cut rate pcr at the date t and time n may be input. Furthermore, the storage battery capacity At at the date t, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity rate Mdt per unit of date t (1 kWh) may also be input. Considering the deterioration of the storage battery, the storage battery capacity At is a function of the date t.

で与えられる。 The time-series clustering unit 4102 of the calculation unit 410 includes a power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,..., N) and a date t, where the date is t and the time is n. Clustering is performed so that the number of clusters is N from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N) at time n. Here, assuming that the allowable variable number Nv and the allowable constraint equation number Nr, the cluster number N is

Given in.

  Here, as the clustering algorithm used by the time-series clustering unit 4104, the above-described algorithm may be used, or another algorithm may be used. The initial value of N may be 1 or any other number. The time series clustering unit 4102 may be simply referred to as a clustering unit.

In the representative time series extraction unit 4104 of the calculation unit 410, time series data that differs only in the initial charge amount At is obtained by using the same representative pattern, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,. _{i, N} (i = 0} , ..., S-1) Summary, power consumption in one year, PV power generation pattern is _{{D S (t), 1} , ..., D s (t), N; P s _{(T), 1} ,..., P _{s (t), N} }, the sum of the electricity charges on the day is ND (t) (t = 0,..., S−1), and the optimal equipment scale (np, nb Create an optimization problem to estimate In this example, since aging of the storage battery is taken into account, each time series is accompanied by the initial charge amount At of the storage battery. In other words, when viewed as time series data, there are many cases where the same time series pattern is obtained when the demand power and PV power generation amount time series representative patterns are corrected in consideration of the initial charge amount At of the storage battery. They are demand power, PV power generation time series ({D _{s (t), i} }, {P _{s (t), i} }) (i = 1,..., N, t = 1,..., T) There are only S types of patterns.

The optimization problem creating unit 4110 includes S representative patterns, {D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} } (i = 0,..., S− 1), the sum of electricity charges for the day when the power consumption and PV power generation pattern are {D _{t, 1} , ..., D _{t, N} ; P _{t, 1} , ..., P _{t, N} } Is an ND (t) (t = 0,..., S-1), and an optimization problem for estimating the optimum facility scale (np, nb) is created.

  The optimization problem creation unit 4106 of the calculation unit 410 has a configuration similar to that of the optimization problem creation unit 3106 of the calculation unit 310 of the third embodiment, but differs from the optimization problem creation unit 3106 over time. A consideration optimization problem creation unit 4112 is included. The aged deterioration-considering optimization problem creating unit 4112 may also be referred to as an aged deterioration processing unit. The optimization problem creating unit 4112 considering aging degradation may be prepared separately from the optimization problem creating unit 4106.

In the aging deterioration-consideration optimization problem creating unit 4112, first, the representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1 ,. , N} }), the series of only the difference in the initial charge amount of the storage battery is collected, the demand power of the date t, the PV power generation time series representative pattern number is specified as s (t), the demand power, the PV power generation pattern i calculation of representative storage battery state of charge T (i), power consumption in one year, the PV power generation pattern _{{D i, 1, ···,} D i, N; P i, 1, ···, P i, _{N (} }) is calculated for the number of days that is _N }.

ただし、

［数２９］において、「Ｍｂ×ｎｂ＋Ｍｐ×ｎｐ」は、蓄電池と太陽光発電装置を設置するためのコスト（第１のコスト）を表す。また、「１０×Ｍｄ_ｉ×Ｋ（ｉ）×（Ｄ_ｉ−Ｚ_ｉ）」（ｉ＝１、…、Ｔ）は、ピークカットによるコスト（第２のコスト）を表す。係数「１０」は１０年分を仮定することに起因する係数である。さらに、［数２９］では、第１、第２のコストに加え、［数２４］の形をした蓄電池の満充電状態の充電量の違いに起因するコスト（第３のコスト）を含んでいる。 Then, the obtained representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) _{D i, 1, ···, D} i, N; P i, 1, ···, using K (i) the number of days is _{P i, N},} the aging considering optimization problem creation unit 4112 Get the following optimization problem.

However,

In [Equation 29], “Mb × nb + Mp × np” represents the cost (first cost) for installing the storage battery and the solar power generation device. Further, “10 × Md _i × K (i) × (D _i −Z _i )” (i = 1,..., T) represents a cost (second cost) due to peak cut. The coefficient “10” is a coefficient resulting from assuming 10 years. Furthermore, in [Equation 29], in addition to the first and second costs, the cost (third cost) resulting from the difference in the charged amount in the fully charged state of the storage battery in the form of [Equation 24] is included. .

  The optimization problem solving unit 4108 of the calculation unit 410 solves the optimization problem created by the optimization problem creation unit 4106 of the calculation unit 410 using, for example, linear programming. As the linear programming method, for example, a simplex method may be used. The optimization problem created by the optimization problem creation unit 4106 of the calculation unit 410 may be solved using an annealing method, a genetic algorithm, or the like. Then, the optimization problem solving unit 1108 of the calculation unit 410 obtains the optimum facility scale, that is, the number of storage batteries nb and the number np of photovoltaic power generation devices (hereinafter simply referred to as solar cells).

  The output unit 420 outputs the optimum facility scale (nb, np) and peak cut error e obtained by the optimization problem solving unit 4108 of the calculation unit 410. The output form may be an image, sound, electronic file, or the like.

  The peak cut facility size estimation device having such a configuration is necessary for the peak cut facility in consideration of the demand power time series, the photovoltaic (PV) power generation time series, the peak cut rate target value, and the aging of the storage battery. The scale of the problem of estimating the number of storage batteries and the number of photovoltaic power generations can be reduced.

  The time series clustering unit 4104 and the cluster diameter determining unit 4106 may be combined and simply called a clustering unit.

<< Peak cut facility size estimation process >>
With reference to FIG. 15, the peak cut facility size estimation process will be described.

  When the peak cut facility size estimation device 40 is a general-purpose computer 200 as shown in FIG. 6, the following description defines a control program for performing such processing. That is, hereinafter, it is also an explanation of a peak cut facility size estimation program (control program) that causes a general-purpose computer to perform the processing described below.

When the process is started, the input unit 400 of the peak cut facility size estimation device 40 receives the allowable variable number Nv, the allowable constraint equation number Nr, and the peak cut error e in S400. The data input to the input unit 400 includes the number of days T, the amount of storage battery charge / discharge per day, the number of PV power generation data N, the date t, and the power demand (estimated) amount D _{t at} time n. _{, N} (t = 1,..., T, n = 1,..., N), date t, PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1, ..., N), the target peak cut rate pcr may be input. Furthermore, the storage battery capacity A, the price Mb per storage battery, the price Mp per solar power generation device, and the electricity charge Mdt per unit of date t (1 kWh) may be input.

  In next step S402, the cluster number calculation unit 4102 of the peak cut facility size estimation apparatus 40 calculates the cluster number N from the allowable variable number Nv and the allowable constraint equation number Nr according to [Equation 28]. When the process of this step ends, the process proceeds to S404.

In S404, the time-series clustering unit 4104 determines the power demand (estimated) amount D _{t, n} (t = 1,..., T, n = 1,..., N), the date t, and the time n when the date is t and the time is n. Clustering is performed so that the number of clusters becomes S from the PV power generation (estimated) amount P _{t, n} (t = 1,..., T, n = 1,..., N). When the process in this step ends, the process proceeds to S406.

In S406, the representative time series extraction unit 4106 obtains a representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) (i = 0, ..., S-1) is extracted. When the process of this step ends, the process proceeds to S408.

In S408, the optimization problem creation unit 4110 of the peak cut facility size estimation device 40 extracts the representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1)} extracted by the representative time series extraction unit 4104. ,..., _{Pi, N} }), the series of only the difference in the initial charge amount of the storage battery is collected, the demand power of the date t, the PV power generation amount time series representative pattern number is specified as s (t), the demand power, PV Calculation of storage battery charge amount T (i) representative of power generation amount pattern i, power consumption in one year, PV power generation amount pattern {D _{i, 1} ,..., D _{i, N} ; P _{i, 1} ,.・ Calculate K (i) for the number of days P _{i, N} }. Then, the obtained representative time series ({D _{i, 1} ,..., D _{i, N} }, {P _{i, 1} ,..., P _{i, N} }) _{D i, 1, ···, D} i, N; P i, 1, ···, using K (i) the number of days is _{P i, N},} the aging considering optimization problem creation unit 4112 The optimization problem formulated as [Equation 29] to [Equation 30] is obtained. In [Equation 29], “Mb × nb + Mp × np” represents the cost (first cost) for installing the storage battery and the solar power generation device. Further, “10 × Md _i × K (i) × (D _i −Z _i )” (i = 1,..., T) represents a cost (second cost) due to peak cut. The coefficient “10” indicates the sum of costs for 10 years. Furthermore, in [Equation 29], in addition to the first and second costs, the cost (third cost) resulting from the difference in the charged amount in the fully charged state of the storage battery in the form of [Equation 24] is included. . When the process of this step ends, the process proceeds to S410.

  In S410, the optimization problem solving unit 4112 of the peak cut facility size estimation device 40 solves the optimization problem created by the calculation unit optimization problem creation unit 4110, and the optimum facility size, that is, the number of storage batteries is nb and sunlight. Np is obtained for the number of power generation devices (hereinafter sometimes simply referred to as solar cells).

  In S410, the output unit 420 of the peak cut facility size estimation apparatus 40 outputs the optimum facility scale (nb, np) obtained by the optimization problem solving unit 4112 of the calculation unit 410.

  The mixed integer programming problem for estimating the optimum equipment scale by the above processing can be reduced while considering the error of the peak cut result, and the processing load for estimating the equipment scale can be reduced. .

10, 20, 30, 40 Peak cut facility size estimation device 100.200, 300, 400 Input unit 110, 210, 310, 410 Calculation unit 120, 220, 320, 420 Output unit 1102, 2102, 3104, 4102 Time series clustering Part 1104, 2106, 3108, 4104 Representative time series extraction part 1106, 2108, 3110, 4106 Optimization problem creation part (problem creation part)
1108, 2110, 3112, 4108 Optimization problem solving section (problem solving section)
2104, 3106 Cluster diameter determination unit 4112 Aged deterioration consideration optimization problem creation unit 3102 Cluster number calculation unit

Claims (10)

There are a plurality of N-dimensional coordinates expressed by a set of time-series data of predicted values of power demand per unit period and time-series data of predicted values of photovoltaic power generation per unit period for different periods. A clustering unit that classifies a plurality of N-dimensional coordinates into one or a plurality of groups according to the proximity of the distance between the N-dimensional coordinates,
A representative time series extraction unit that specifies N-dimensional representative coordinates representing the group for each of the one or more groups;
A problem solving unit for calculating the scale of the power supply facility based on the identified one or more N-dimensional representative coordinates;
An equipment size estimation device for power supply equipment, comprising: further,
Using the N-dimensional representative coordinates representing the group, a total cost including a first cost for installing the power supply facility and a second cost due to a peak cut in power charges due to the installation of the power supply device A problem creation section for defining problems to be reduced;
Including
The facility size estimation apparatus according to claim 1, wherein the problem solving unit solves the problem and calculates a scale of a power supply facility.   A first set of one of the time series data of the predicted value of the demand power amount, one of the time series data of the forecast value of the photovoltaic power generation amount, and the time series of the forecast value of the demand power amount The distance between another one of the data and the second set composed of another one of the time series data of the predicted value of the photovoltaic power generation amount is the prediction of the demand power amount of the first set Maximum value of the absolute value of the difference between the time series data of the value and the corresponding elements of the time series data of the predicted value of the demand power amount of the second set, and the photovoltaic power generation amount of the first set Or the maximum value of the absolute value of the difference between the corresponding elements of the time series data of the predicted value of the photovoltaic power generation amount of the second set and the predicted value time series data of the second set or The equipment scale estimation apparatus according to 2.   The equipment scale estimation device according to any one of claims 1 to 3, wherein the power supply facility includes a solar power generation device and a storage battery. N-dimensional representative coordinates representing the two groups that differ by the amount of charge in the fully charged state of the storage battery are determined to be the same, and a third cost due to the difference in the amount of charge in the fully charged state of the storage battery is calculated. Aged deterioration processing section,
Including
5. The peak cut facility size estimation device according to claim 4, wherein the problem creating unit defines the problem including the third cost in the total cost in addition to the first cost and the second cost. 6.   The clustering unit selects one group from the one or more groups, and selects at least one of the one or more groups in which the distance from the one group is less than or equal to half of the peak cut error. The equipment size estimation device according to any one of claims 1 to 5, wherein the cluster is selected to obtain the cluster in which the maximum value of the distance is equal to or less than a peak cut error.   6. The clustering unit according to claim 1, wherein the clustering unit is configured to gradually increase the number of the one or a plurality of groups to obtain the cluster in which the maximum value of the distance is equal to or less than a peak cut error. Equipment scale estimation equipment. further,
A cluster number calculator that calculates the number of the one or more groups from the number of variables in question and the number of constraints,
In the clustering unit, the maximum value of the distance is equal to or less than a peak cut error when the power supply from the outside is peak cut by supplying a part of the demand power of the power load from the solar power generation device and the storage battery. The equipment scale estimation apparatus according to any one of claims 1 to 5, wherein the one or the plurality of groups are formed in a number corresponding to the number of the one or the plurality of groups. A facility size estimation method executed by a computer,
There are a plurality of N-dimensional coordinates expressed by a set of time-series data of predicted values of power demand per unit period and time-series data of predicted values of photovoltaic power generation per unit period for different periods. A plurality of existing N-dimensional coordinates according to the proximity of the distance between the respective N-dimensional coordinates, and classifying the group into one or more groups,
Identifying an N-dimensional representative coordinate representing the group for each of the one or more groups;
Calculating the scale of the power supply facility based on the identified one or more N-dimensional representative coordinates;
A method for calculating the scale of a power supply facility, comprising: There are a plurality of N-dimensional coordinates expressed by a set of time-series data of predicted values of power demand per unit period and time-series data of predicted values of photovoltaic power generation per unit period for different periods. In some cases, a plurality of N-dimensional coordinates are classified into one or a plurality of groups according to the proximity of the distance between the N-dimensional coordinates,
Identifying an N-dimensional representative coordinate representing the group for each of the one or more groups;
A facility size estimation program that causes a computer to execute a process of calculating the scale of a power supply facility based on one or more specified N-dimensional representative coordinates.

Images