POWER SUPPRESSION OPTIMIZATION SYSTEM AND POWER SUPPRESSION OPTIMIZATION METHOD

CROSS REFERENCE TO RELATED APPLICATIONS

This application is based upon and claims the benefit of priority from the prior Japanese Patent Application No. 2013-240435, filed on November 20, 2013, the entire contents of which are incorporated herein by reference.

FIELD

Embodiments described herein relate generally to power suppression optimization systems and power suppression optimization methods.

BACKGROUND

To implement a "smart grid" that maintains stable power supply by controlling the power consumption values of consumers, installation of a building energy management system that suppresses peak power demand by remotely controlling devices (air-conditioning and lighting) owned by the consumers is proceeding. By this, a negawatt trading where suppression of power consumption values is requested to large consumers such as factories and commercial facilities, and incentives are given to those consumers responding to the request has started to be commercialized. However, since power demand is expected to grow in the future, too, power supply and demand balance may not be able to be maintained only by demand response for large consumers. Hence, in recent years, demand response for small consumers such as households and stores has been considered.

Demand response for small consumers has a problem that it is uncertain whether a reduction of an amount corresponding to a power shortage which is estimated by an electric power provider such as a power retailer or a power aggregator can be achieved. This is due to high uncertainty of power demand by small consumers unlike large consumers. Therefore, it is important that demand response for small consumers take into account the degree of uncertainty of power demand.

Conventionally, as a technique for selecting consumers to maximize the utility of a power producer, there is proposed a technique in which consumers and the amounts of distribution to the consumers are determined using the degrees of uncertainty of power demand of the consumers, such that a predefined utility function is maximized. However, such a conventional technique is premised on consumer power demand and does not take into account the values of power reduction that can be achieved by consumers. Thus, it is difficult to solve the problem of high uncertainty as to whether a reduction of an amount corresponding to a power shortage which is estimated by an electric power provider can be achieved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a functional configuration of a power suppression optimization system according to a first embodiment.

FIG. 2 is a diagram showing an example of power consumption value data.

FIG. 3 is a diagram showing an example of power consumption value data compiled on a time period basis.

FIG. 4 is a diagram showing an example of expected power reduction values.

FIG. 5 is a diagram showing an example of the degrees of uncertainty σ _η ² .

FIG. 6 is a diagram showing an example of the degrees of uncertainty σ _ηΓ η.

FIG. 7 is a diagram showing an example of a combination of consumers.

FIG. 8 is a flowchart showing the operation of the power suppression optimization system according to the first embodiment. FIG. 9 is a diagram showing an example of an output screen of the power suppression optimization system according to the first embodiment.

FIG. 10 is a block diagram showing a functional configuration of a power suppression optimization system according to a second embodiment.

FIG. 11 is a flowchart showing the operation of the power suppression optimization system according to the second embodiment.

FIG. 12 is block diagram showing a functional configuration of a power suppression optimization system according to a third embodiment.

FIG. 13 is a diagram showing an example of preprocessed power consumption value data.

FIG. 14 is a diagram showing an example of the groups of outside air temperatures.

FIG. 15 is a diagram showing an example of preprocessed outside air temperature data compiled according to the groups of outside air temperatures.

FIG. 16 is a diagram showing an example of expected power reduction value data complied according to the groups of outside air temperatures.

FIG. 17 is a flowchart showing the operation of the power suppression optimization system according to the third embodiment.

DETAILED DESCRIPTION

In one embodiment, a power suppression optimization system includes an expected power reduction calculating unit and an optimizing unit. The calculating unit calculates expected power reduction values, based on power consumption value data. The power consumption data represents a history of power consumption values of a plurality of consumers. The expected power reduction values are power reduction values of the consumers expected when power suppression is requested. The optimizing unit selects a combination of consumers based on the expected power reduction values of the consumers calculated by the calculating unit. The optimizing unit selects such that variation in a sum total value of power reduction values of the consumers against a total expected power reduction value is small, and the total expected power reduction value satisfies a condition regarding the power reduction values for the power suppression. The total expected power reduction value is a sum total of the expected power reduction values of the consumers.

(First Embodiment)

A power suppression optimization system according to a first embodiment will be described below with reference to FIGS. 1 to 9. The power suppression optimization system according to the present embodiment obtains power consumption value data of a plurality of consumers, and selects an optimal combination of consumers to which suppression of power consumption values is requested, based on the obtained power consumption value data. Here, FIG. 1 is a block diagram showing a functional configuration of the power suppression optimization system according to the present embodiment. As shown in FIG. 1, the power suppression optimization system according to the present embodiment includes a storage unit 1 that stores various types of information; a power consumption value data obtaining unit 2 that obtains power consumption value data of consumers; an expected power reduction calculating unit 3 that calculates expected power reduction values of the consumers; a degree-of-uncertainty calculating unit 4 that calculates the degrees of uncertainty of the expected power reduction values of the consumers; an optimizing unit 5 that selects a combination of consumers such that the degree of uncertainty of a total expected power reduction value is minimum; and a total expected power reduction calculating unit 6 that calculates a total expected power reduction value.

The above-described configuration of the present embodiment can be implemented by using a computer apparatus including a CPU and a memory, as basic hardware. For example, the functions of the power consumption value data obtaining unit 2, the expected power reduction calculating unit 3, the degree-of-uncertainty calculating unit 4, the optimizing unit 5, and the total expected power reduction calculating unit 6 can be implemented by the CPU executing a control program. In addition, the power suppression optimization system according to the present embodiment may include a communication means for obtaining information from an external source and outputting information to an external source.

The storage unit 1 stores various types of information used or generated in processes performed by the power suppression optimization system according to the present embodiment, a control program for implementing the above-described functional configurations, and the like. As the storage unit 1, a storage apparatus such as a nonvolatile memory or an external storage apparatus can be used.

The power consumption value data obtaining unit 2 obtains power consumption value data for predetermined ranges from the storage unit 1 or an externally provided power consumption value database. The power consumption value data is data representing power consumption values (30-minute values, 1-hour integrated values, etc.) such as the power consumption or amounts of power consumption of consumers for each time period. The power consumption value of a consumer is a power consumption value of power reduction target devices (an air-conditioning apparatus, a swimming pool pump, a water heater, etc.) owned by the consumer. The predetermined ranges of which the power consumption value data obtaining unit 2 obtains power consumption value data are a consumer range and a date and time range. The power consumption value data obtained by the power consumption value data obtaining unit 2 is stored in the storage unit 1.

FIG. 2 is a diagram showing an example of power consumption value data obtained by the power consumption value data obtaining unit 2. In FIG. 2, the consumer range includes 10000 consumers with consumer IDs from ID000001 to ID010000, and the date and time range is three weeks from December 1, 2012 0:00 to December 21, 2012 23:30. The ranges can be set arbitrarily, and may be pre-stored in the storage unit 1 or may be inputted from a control terminal by an operator that uses the power suppression optimization system.

Note that although the following describes the case in which the sum total of the power consumption values of one or a plurality of devices owned by a consumer is obtained as power consumption value data, power consumption value data may be obtained for each device or each set of devices. For example, when a consumer has a plurality of devices and power consumption value data is obtained for each device, a plurality of pieces of power consumption value data are obtained as the power consumption value data of the consumer. When a plurality of pieces of power consumption value data of a consumer are thus obtained, a device ID for each device or each set of devices is set instead of a consumer ID. In this case, the power suppression request target may be one or a plurality of devices identified by a device ID(s), instead of a consumer.

The expected power reduction calculating unit 3 calculates expected power reduction values, based on the power consumption value data obtained by the power consumption value data obtaining unit 2. An expected power reduction value is the value of power reduction or the amount of power reduction of a consumer which is expected when power suppression is requested. An expected power reduction value can be calculated as, for example, an average value of the power consumption values of each consumer. In this case, the expected power reduction calculating unit 3 compiles power consumption value data of each consumer on a time period basis, and calculates an average value of power consumption values for each time period, as an expected power reduction value for that time period.

FIG. 3 is a diagram showing power consumption value data of FIG. 2 which is compiled on a time period basis. In FIG. 3, the power consumption value data of each consumer is compiled in 30 minute intervals such as 0:00 to 0:29, 0:30 to 0:59, and 1 :00 to 1:29. The expected power reduction calculating unit 3 computes an average value of compiled power consumption values. For example, an expected power reduction value of the consumer ID000001 for a time period of 0:00 to 0:29 can be calculated as an average value of the power consumption values of the consumer ID000001 for three weeks for a time period of 0:00 to 0:29. FIG. 4 is a diagram showing expected power reduction values calculated in the above-described manner. As shown in FIG. 4, the expected power reduction values are calculated for each consumer and each time period.

In addition, the expected power reduction calculating unit 3 may calculate an expected power reduction value by adding up a predetermined operating rate a to the above-described average value of the power consumption values of each consumer. The operating rate a is the operating rate of devices owned by the consumer. An operating rate a of 100% indicates a state in which the consumer allows the devices to operate such that the power consumption value thereof reaches the above-described average value. The operating rate a may be pre-stored in the storage unit 1 or may be inputted by the operator. In addition, the operating rate a may have different values for different time periods. The expected power reduction values thus calculated are stored in the storage unit 1.

Note that although in FIG. 4 the expected power reduction values of each consumer are calculated in 30 minute intervals, the intervals can be set arbitrarily. For example, when the expected power reduction calculating unit 3 calculates expected power reduction values in one hour intervals, the expected power reduction calculating unit 3 calculates an average value of power consumption values for 0:00 to 0:59 in FIG. 3, as the expected power reduction value for 0:00 to 0:59.

The degree-of-uncertainty calculating unit 4 calculates the degrees of uncertainty of the consumers. The degree of uncertainty is variation in power reduction value for each time period against an expected power reduction value of each consumer for each time period. The degree of uncertainty can be computed as follows as, for example, variation (variance) in power consumption value for each time period against an expected power reduction value. [Equation 1]

where n is the consumer (n=l,..., 10000), σ _η ² is the degree of uncertainty of the consumer n, p _nc j is the power consumption value of the consumer n for date d (1st to 21st), p (bar) _n is the expected power reduction value of the consumer n, and D is the number of days (21 days). In this case, the degree of uncertainty σ _η ² of the consumer n for each time period is calculated as the mean square of the difference between the power consumption value p _nc j and the expected power reduction value p (bar) _π · When the expected power reduction value p (bar) _n of the consumer n is the arithmetic mean of the power consumption values p _nc j of the consumer n for each time period, the degree of uncertainty σ _η ² is the variance of the power consumption values p _nc i of the consumer n for that time period. The degrees of uncertainty calculated by the degree-of-uncertainty calculating unit 4 are stored in the storage unit 1. FIG. 5 is a diagram showing the degrees of uncertainty σ _η ² calculated based on the power consumption value data of FIG. 2 and the expected power reduction values of FIG. 4. Although in FIG. 5 the degrees of uncertainty are calculated in 30 minute intervals, the intervals can be set arbitrarily.

In addition, the degrees of uncertainty between two consumers can also be calculated as follows.

[Equation 2] where n, m is the consumer (n, m=l,..., 10000), a _nm is the degree of uncertainty between the consumer n and the consumer m, Pnd, Pmd is the power consumption value of the consumer n, m for date d (1st to 21st), p (bar) _n , P (bar) _m is the expected power reduction value of the consumer n, m, and D is the number of days (21 days). In this case, the degree of uncertainty a _nm is calculated as the average of the product of the difference between the expected power reduction value p (bar) _n and the power consumption value p _nc j of the consumer n and the difference between the expected power reduction value p (bar) _m and the power consumption value p _m( j of the consumer m. The degree of uncertainty o _nm thus calculated indicates the degree of association between the difference between the expected power reduction value p (bar) _n and the power consumption value p _nc j of the consumer n and the difference between the expected power reduction value p (bar) _m and the power consumption value p _mc i of the consumer m. When the expected power reduction value of the consumer n, m is the arithmetic mean of the power consumption values of the consumer n, m for each time period, the degree of uncertainty σ _ηΓ η is the covariance between the power consumption values of the consumers n and m for that time period. In addition, when n=m, the degree of uncertainty a _nm coincides with the above-described degree of uncertainty σ _η ² .

FIG. 6 is a diagram showing the degrees of uncertainty a _nm calculated based on the power consumption value data of FIG. 2 and the expected power reduction values of FIG. 4. Although in FIG. 6 the degrees of uncertainty are calculated in 30 minute intervals, the intervals can be set arbitrarily. Note that the degree of uncertainty is not limited to the above-described degrees of uncertainty a _nm and σ _π ² . For example, as the degree of uncertainty, the value which is a constant multiple of the degree of uncertainty cfnm _/ σ _η ² , or the square root of the degree of uncertainty a _nm , σ _η ² can be used.

The optimizing unit 5 selects a combination of consumers to which power suppressed is requested, such that a total expected power reduction value is greater than or equal to a planned power reduction value or is in a predetermined range, and the number of consumers is less than or equal to the upper limit of the number of consumers or is in a predetermined range. Here, the total expected power reduction value is the sum total of the expected power reduction values of consumers selected as consumers to which power suppression is requested. The planned power reduction value is a power reduction value planned by an electric power provider, etc., and is, for example, a power shortage predicted at peak times. The upper limit of the number of consumers is the upper limit of the number of consumers to which power suppression is requested. The planned power reduction value and the upper limit of the number of consumers may be pre-stored in the storage unit 1 or may be inputted by the operator.

The optimizing unit 5 selects a combination of consumers to which power suppression is requested, such that the degree of uncertainty indicating variation in total expected power reduction value is the minimum, with the above-described conditions satisfied. The degree of uncertainty of the total expected power reduction value can be calculated, for example, based on the degrees of uncertainty of consumers calculated by the degree-of-uncertainty calculating unit 4. In this case, the optimizing unit 5 obtains expected power reduction values for those time periods included in a time period during which implementation of power suppression is planned by the electric power provider, etc. (hereinafter, referred to as a "power suppression time period"), and creates a vector such as that shown below.

[Equation 3]

The above-described vector is an N x 1 row vector having, as elements, the expected power reduction values p (bar) _n of consumer n for each time period. A number of such vectors that is equal to the number of time periods included in the power suppression time period are created. When the optimizing unit 5 obtains the expected power reduction values of FIG. 4, N = 10000. In addition, the optimizing unit 5 obtains the degrees of uncertainty for the time periods included in the power suppression time period from the storage unit 1, and creates an N x N matrix having, as elements, the degrees of uncertainty of the consumers for each time period. When the optimizing unit 5 obtains the degrees of uncertainty σ _η ² such as those shown in FIG. 5, the optimizing unit 5 creates a diagonal matrix such as that shown below.

On the other hand, when the optimizing unit 5 obtains the degrees of uncertainty a _nm such as those shown in FIG. 6, the optimizing unit 5 creates a matrix such as that shown below.

[Equation 5]

Furthermore, the optimizing unit 5 creates a vector such as that shown below.

[Equation 6]

w = w _lr w ₂ , ...,w _N ) ^T

The above-described vector is an N ^χ 1 row vector having, as elements, variables w _n which take either value 0 or 1. When the consumer n is selected as a consumer to which power suppression is requested, w _n is 1, and when not selected, w _n is 0. Therefore, the above-described vector indicates a combination of consumers to which power suppression is requested. At this time, the total expected power reduction value can be represented by the following equation:

[Equation 7]

N

→T -» \^

In the present embodiment, the degree of uncertainty of the total expected power reduction value can be represented by the following equation:

[Equation 8]

N N

The optimizing unit 5 searches for a combination of w _n such that the degree of uncertainty of the total expected power reduction value is minimum, under the constraint conditions where the total expected power reduction value is greater than or equal to a planned power reduction value, and the number of consumers selected (the number of w _n which is 1) is less than or equal to the upper limit of the number of consumers. This can be formulated as shown below when the planned power reduction value is p, and the upper limit of the number of consumers is M (< N).

[Equation 9]

w _e e {0,1}

The optimizing unit 5 selects a combination of consumers by solving a 0-1 convex quadratic programming problem such as that described above. The optimizing unit 5 can solve the above-described problem by, for example, solving a quadratic programming problem where the 0-1 integer constrains of w _n are relaxed, by metaheuristic solutions such as an interior-point method and GA, and rounding up or down the obtained solutions to integer solutions. The solutions obtained thereby such as those shown below are stored in the storage unit 1.

[Equation 10]

..., w _N *y

FIG. 7 is a diagram showing an example of the above-described solutions (a combination of consumers) stored in the storage unit 1. As shown in FIG. 7, the results of a combination of consumers are indicated by the value of w _n (0 or 1). Note that the upper limit M of the number of consumers does not necessarily need to be set. The total expected power reduction calculating unit 6 calculates a total expected power reduction value, based on the combination of consumers selected by the optimizing unit 5. The total expected power reduction value can be calculated by the following equation. The calculated total expected power reduction value is stored in the storage unit 1.

[Equation 11]

Next, the operation of the power suppression optimization system according to the present embodiment will be described with reference to FIG. 8. FIG. 8 is a flowchart showing the operation of the power suppression optimization system. Note that the following describes the operation performed when the operator selects a combination of consumers using the power suppression optimization system according to the present embodiment.

First, the operator allows the power suppression optimization system to start selection of a combination of consumers. The power suppression optimization system determines whether the expected power reduction values and degrees of uncertainty of consumers included in a consumer range, i.e., a population of a combination of consumers, which is preset or inputted by the operator, have been calculated (step SI). If the expected power reduction values and degrees of uncertainty of all of the consumers included in the population have been calculated (YES at step SI), the process proceeds to step S5.

On the other hand, if there is a consumer whose expected power reduction values and degrees of uncertainty have not been calculated (NO at step SI), processing proceeds to step S2. In this case, the power consumption value data obtaining unit 2 obtains power consumption value data of the consumer whose expected power reduction values and degrees of uncertainty have not been calculated (step S2). The power consumption value data obtaining unit 2 obtains power consumption value data for a date and time range which is preset or inputted by the operator. Then, the expected power reduction calculating unit 3 calculates expected power reduction values of the consumer for each time period, based on the power consumption value data obtained by the power consumption value data obtaining unit 2 (step S3). The expected power reduction calculating unit 3 can calculate an expected power reduction value, based on an average value of the power consumption values of the consumer, etc.

Then, the degree-of-uncertainty calculating unit 4 calculates the degrees of uncertainty for each time period, based on the power consumption value data obtained by the power consumption value data obtaining unit 2 and the expected power reduction values calculated by the expected power reduction calculating unit 3 (step S4). The degree-of-uncertainty calculating unit 4 can calculate the degree of uncertainty σ _η ² , based on the mean square of the difference between a power consumption value and an expected power reduction value.

When the calculation of the degrees of uncertainty of the consumer is completed, processing returns to step SI. Then, the power suppression optimization system repeats the above-described steps S2 to S4 until the expected power reduction values and degrees of uncertainty of all of the consumers are calculated. If the expected power reduction values and degrees of uncertainty of all of the consumers are calculated (YES at step SI), the process proceeds to step S5.

In the above description, at steps S2 to S4, power consumption value data is obtained on a per consumer basis, and expected power reduction values and the degrees of uncertainty are calculated. However, the power suppression optimization system may obtain, at step S2, power consumption value data of all those consumers whose expected power reduction values and degrees of uncertainty are determined at step SI to have not been calculated, and calculate, at step S3, expected power reduction values of all of the consumers, and then calculate, at step S4, the degrees of uncertainty of all of the consumers. In this case, after processing proceeds to step S4, the process proceeds to step S5 instead of returning to step SI. In addition, when the degree of uncertainty a _nm is used as the degree of uncertainty, the power suppression optimization system first calculates expected power reduction values of all of the consumers or obtains the expected power reduction values from the storage unit 1. Thereafter, the degree-of-uncertainty calculating unit 4 calculates the degrees of uncertainty σ _ηη ν

Then, the optimizing unit 5 optimizes a combination of consumers to which power suppression is requested (step S5). Specifically, the optimizing unit 5 selects a combination of w _n such that the degree of uncertainty (e.g., σ _η ² or a _mn ) of a total expected power reduction value is minimum, under the constraint conditions where the total expected power reduction value is greater than or equal to a planned power reduction value, and the number of consumers selected is less than or equal to the upper limit of the number of consumers.

The power suppression optimization system may further calculate a total expected power reduction value by the total expected power reduction calculating unit 6, or may output the results of the optimization in response to a request from the operator. For example, the power suppression optimization system can display visualized optimization results on a monitor of the control terminal, etc.

FIG. 9 is a diagram showing an example of an output screen of optimization results. As shown in FIG. 9, the output results may include a predicted power consumption value, a power consumption value predicted when power suppression is requested to an optimized combination of consumers, the optimized combination of consumers, etc.

As described above, according to the power suppression optimization system according to the present embodiment, of combinations of consumers where the total expected power reduction value is greater than or equal to the planned power reduction value, a combination of consumers with minimum variation (the degree of uncertainty σ _η ² , σ _ηΓ η _/ or the like) in total expected power reduction value can be selected. Therefore, by performing demand response, etc., according to a combination of consumers selected by the power suppression optimization system, the electric power provider can reduce a predicted power reduction value with high probability. In addition, the electric power provider can also perform, based on the degrees of uncertainty of consumers, efficient demand response where the consumers are narrowed down to those with low degrees of uncertainty. Furthermore, the electric power provider can also specify, based on optimization results, the power reduction values of consumers by a mechanism such as DLC (Direct Load Control) that directly controls devices owned by the consumers. Moreover, the electric power provider can use the optimization results to specify behavior change targets. Specifically, the electric power provider considers consumers selected by optimization, as consumers with high efficiency of demand response and can specify the consumers as behavior change targets to promote behavior change by issuing coupons, etc.

(Second Embodiment)

Next, a power suppression optimization system according to a second embodiment will be described with reference to FIGS. 10 and 11. Here, FIG. 10 is a block diagram showing a functional configuration of the power suppression optimization system according to the present embodiment. As shown in FIG. 10, the power suppression optimization system according to the present embodiment includes a storage unit 1, a power consumption value data obtaining unit 2, an expected power reduction calculating unit 3, an optimizing unit 5, and a total expected power reduction calculating unit 6. The above-described configuration is the same as that of the first embodiment. Note, however, that unlike the first embodiment, the power suppression optimization system according to the present embodiment does not need to include a degree-of-uncertainty calculating unit 4. In the present embodiment, as the degree of uncertainty indicating the degree at which a power reduction value of a total expected power reduction value extremely falls below, β-CVaR which is a representative risk measure is used. Note that a configuration using, as the degree of uncertainty, VaR (Value-at-Risk) or absolute deviation instead of β-CVaR is also possible.

The optimizing unit 5 calculates β-CVaR (Conditional Value at Risk). β-CVaR is an expected value of a total power reduction value for when the total power reduction value falls below β-VaR where a is such minimum power that the total power reduction value does not fall below the predetermined power a with a predetermined probability β. To calculate β-CVaR, the optimizing unit 5 first obtains power consumption value data of consumers for a power suppression time period which is preset or specified by an operator, and creates a matrix such as that shown below.

[Equation 12]

In addition, as in the first embodiment, the optimizing unit 5 creates a vector such as that shown below.

[Equation 13]

w = (w _lf w ₂ , ... , w _N ) ^r

w„ e {0,1}

Then, the optimizing unit 5 calculates β-CVaR which is the degree of uncertainty of a total expected power reduction value by the following equation:

[Equation 14]

wnPnd + 0- R + u _d ≥ 0,η _ά ≥ 0; d

= 1, ... ,D

The optimizing unit 5 selects a combination of w _n such that the degree of uncertainty (β-CVaR) of the total expected power reduction value is minimum, under the constraint conditions where the total expected power reduction value is greater than or equal to a planned power reduction value, and the number of consumers selected (the number of w _n which is 1) is less than or equal to the upper limit of the number of consumers. This can be formulated as shown below when the planned power reduction value is p, and the upper limit of the number of consumers is M (< N).

[Equation 15]

V = ( i>Pz> -· > Ρχ) ^Τ minimize t

subject to > Q; d = 1, ..., 0,

u _d ≥ 0; d = 1, ..,, D, w ^T p = ^ w _n p _n ≥ p,

n—1

^ w _n < M, w _B e {0,1}

The optimizing unit 5 determines a combination of consumers by solving a 0-1 linear programming problem such as that described above. The optimizing unit 5 can solve the above-described problem by, for example, solving a linear programming problem where the 0-1 integer constrains of w _n are relaxed, by a simplex method, an interior-point method, or the like, and rounding up or down the obtained solutions to integer solutions. The solutions obtained thereby such as those shown below are stored in the storage unit 1.

[Equation 16]

The total expected power reduction calculating unit 6 calculates a total expected power reduction value, based on the combination of consumers optimized by the optimizing unit 5. The total expected power reduction value can be calculated by the following equation. The calculated total expected power reduction value is stored in the storage unit 1. [Equation 17]

Note that the degree of uncertainty β-CVaR of the total expected power reduction value at this time is α _β which is obtained as a solution to the above-described 0-1 linear programming problem. The power suppression optimization system can output the thus obtained degree of uncertainty of the total expected power reduction value, as the optimization result.

Next, the operation of the power suppression optimization system according to the present embodiment will be described. Here, FIG. 11 is a flowchart showing the operation of the power suppression optimization system according to the present embodiment. As shown in FIG. 11, the power suppression optimization system determines whether the expected power reduction values of consumers included in a consumer range which is preset or inputted by the operator have been calculated (step SI). Steps S2 and S3 which are the same as those of the first embodiment are repeated until the expected power reduction values of all of the consumers are calculated. If the expected power reduction values of all of the consumers are calculated, processing proceeds to step S5.

Note that power consumption value data of all those consumers whose expected power reduction values are determined at step SI to have not been calculated may be obtained at step S2, and expected power reduction values of all of the consumers may be calculated at step S3. In this case, after the process proceeds to step S3, the process proceeds to step S5 instead of returning to step SI.

Then, the optimizing unit 5 optimizes a combination of consumers to which power suppression is requested (step S5). Specifically, the optimizing unit 5 searches for a combination of w _n where the degree of uncertainty (β-CVaR) of a total expected power reduction value is minimum, under the constraint conditions where the total expected power reduction value is greater than or equal to a planned power reduction value, and the number of consumers selected is less than or equal to the upper limit of the number of consumers.

As described above, according to the power suppression optimization system according to the present embodiment, a combination of consumers where the total expected power reduction value is greater than or equal to the planned power reduction value, and variation (the degree of uncertainty β-CVaR) in total expected power reduction value is minimum can be selected.

(Third Embodiment)

A power suppression optimization system according to a third embodiment will be described below with reference to FIGS. 12 to 17. In the present embodiment, power consumption value data of consumers is subjected to preprocessing. In addition, expected power reduction values of the consumers are calculated using outside air temperature data. Here, FIG. 12 is block diagram showing a functional configuration of the power suppression optimization system according to the present embodiment. As shown in FIG. 12, the power suppression optimization system according to the present embodiment includes a storage unit 1, a power consumption value data obtaining unit 2, an expected power reduction calculating unit 3, a degree-of-uncertainty calculating unit 4, an optimizing unit 5, and a total expected power reduction calculating unit 6. The above-described configuration is the same as that of the first embodiment. In the present embodiment, the power suppression optimization system further includes a preprocessing unit 7.

The preprocessing unit 7 performs preprocessing, such as a smoothing process, an interpolation process, and an abnormal value removing process, on power consumption value data obtained by the power consumption value data obtaining unit 2. The smoothing process is the process of smoothing power consumption value data and outside air temperature data. The smoothing process can be implemented by, for example, calculating the moving average value and moving median of power consumption value data stored in the storage unit 1, or applying the Nadaraya-Watson estimate, a spline function, or the like. The preprocessing unit 7 may determine whether to perform a smoothing process, by calculating a variance of power consumption value data stored in the storage unit 1 and comparing the calculated variance with a predetermined threshold value.

The interpolation process is the process of interpolating missing power consumption value data. The interpolation process can be implemented by, for example, interpolating missing data using data adjacent to the missing data or data estimated from adjacent data. The preprocessing unit 7 may determine whether to perform an interpolation process, by determining whether power consumption value data stored in the storage unit 1 has a missing portion.

The abnormal value removing process is the process of removing data including an abnormal value from power consumption value data. The abnormal value removing process can be implemented by comparing power consumption values with a predetermined threshold value and removing those power consumption value data exceeding the threshold value. The preprocessing unit 7 may determine whether to perform an abnormal value removing process, by comparing the maximum and minimum values of power consumption values with predetermined threshold values. Note that it is preferred that, when an abnormal value removing process is performed, an interpolation process be performed to interpolate removed data.

The functions of the preprocessing unit 7 such as those described above can be implemented by a CPU executing a control program. The preprocessing unit 7 may perform preprocessing only once or may perform preprocessing a plurality of times. Alternatively, preprocessing does not need to be performed where unnecessary. Whether or not to perform preprocessing and the number of times preprocessing is performed may be inputted from a control terminal by an operator or may be automatically determined by the power suppression optimization system. The power consumption value data having been subjected to preprocessing is stored in the storage unit 1, as preprocessed power consumption value data.

The preprocessing unit 7 further combines the preprocessed power consumption value data with outside air temperature data. The outside air temperature data is data representing outside air temperatures for each time period which are measured in areas where the consumers are located, and may be pre-stored in the storage unit 1 or may be obtained from an external outside air temperature database, etc., by the power suppression optimization system. The preprocessing unit 7 combines the preprocessed power consumption value data with the outside air temperature data, according to the dates and times of those pieces of data. The preprocessing unit 7 may combine together power consumption value data and outside air temperature data with the same date and time, or may combine power consumption value data with outside air temperature data whose measurement time is shifted by a predetermined period of time. For example, power consumption value data may be combined with outside air temperature data whose measurement time is earlier by one to two hours than the power consumption value data. By this, a time lag before the outside air temperature influences on the power consumption value can be taken into account.

FIG. 13 is a diagram showing an example of preprocessed power consumption value data combined with outside air temperature data. The preprocessed power consumption value data of FIG. 13 is combined with outside air temperature data with the same dates and times as those of the preprocessed power consumption value data. Note that the preprocessing unit 7 may also perform the same preprocessing as that for the power consumption value data, on the outside air temperature data.

The expected power reduction calculating unit 3 calculates expected power reduction values of each consumer for each time period, based on the preprocessed power consumption value data and the outside air temperature data. First, the expected power reduction calculating unit 3 compiles outside air temperatures for each time period from preprocessed power consumption value data such as that shown in FIG. 13, according to predetermined groups. The groups of outside air temperatures can be set on an arbitrary outside air temperature interval basis. FIG. 14 is a diagram showing an example of the groups of outside air temperatures. In FIG. 14, the groups of outside air temperatures are set for every 10°C, and are labeled with A for -0°C, B for 0°C-10°C, C for 10°C-25°C, and D for 25°C-. Such groups of outside air temperatures may be preset or may be inputted by the operator. FIG. 15 is a diagram showing outside air temperature data of FIG. 13 which is complied according to the groups of outside air temperatures of FIG. 14. In FIG. 15, the groups of outside air temperatures are indicated by the labels of FIG. 14.

Then, the expected power reduction calculating unit 3 calculates expected power reduction values for each group of outside air temperatures, based on the outside air temperature data compiled as shown in FIG. 14 and the preprocessed power consumption value data of FIG. 13. Specifically, the expected power reduction calculating unit 3 complies the preprocessed power consumption value data on a group of outside air temperatures basis, and calculates an average value of power consumption values for each of the compiled group of outside air temperatures, or a value where a predetermined operating rate a is added up to the average value.

FIG. 16 is a diagram showing expected power reduction values which are calculated for each group of outside air temperatures, based on the preprocessed power consumption value data of FIG. 13 and the outside air temperature data of FIG. 15. As shown in FIG. 16, the expected power reduction values are calculated for each consumer, each time period, and each group of outside air temperatures.

In the present embodiment, the degree-of-uncertainty calculating unit 4 computes the degrees of uncertainty of each consumer for each time period and each group of outside air temperatures, based on expected power reduction value data calculated for each group of outside air temperatures, such as that shown in FIG. 16. The calculated degrees of uncertainty are stored in the storage unit 1.

Next, the operation of the power suppression optimization system according to the present embodiment will be described. FIG. 17 is a flowchart showing the operation of the power suppression optimization system according to the present embodiment. As shown in FIG. 17, the power suppression optimization system determines whether the expected power reduction values and degrees of uncertainty of consumers included in a consumer range which is preset or inputted by the operator have been calculated (step SI).

If there is a consumer whose expected power reduction values and degrees of uncertainty have not been calculated (NO at step SI), the power consumption value data obtaining unit 2 obtains power consumption value data of the consumer (step S2), and the preprocessing unit 7 performs at least one of preprocessing including a smoothing process, an interpolation process, and an abnormal value removing process, on the obtained power consumption value data (step S7). When the preprocessing unit 7 has performed preprocessing on the power consumption value data, the preprocessing unit 7 combines the preprocessed power consumption value data with outside air temperature data.

Then, the expected power reduction calculating unit 3 calculates expected power reduction values of each consumer (step S3). The expected power reduction calculating unit 3 calculates expected power reduction values of each consumer for each time period and each group of outside air temperatures. Then, the degree-of-uncertainty calculating unit 4 calculates the degrees of uncertainty of each consumer (step S4). The degree-of-uncertainty calculating unit 4 calculates the degrees of uncertainty of each consumer for each time period and each group of outside air temperatures.

The above-described processes at steps S2 to S4 are repeated until the expected power reduction values and degrees of uncertainty of all of the consumers are calculated. If the expected power reduction values and degrees of uncertainty of all of the consumers are calculated (YES at step SI), the process proceeds to step S5.

Note that power consumption value data of all those consumers whose expected power reduction values are determined at step SI to have not been calculated may be obtained at step S2, preprocessing may be performed on the power consumption value data of all of the consumers at step S7, expected power reduction values of all of the consumers may be calculated at step S3, and the degrees of uncertainty of all of the consumers may be calculated at step S4. In this case, after the process proceeds to step S4, the process proceeds to step S5 instead of returning to step SI.

Then, the optimizing unit 5 optimizes a combination of consumers to which power suppression is requested (step S5). Specifically, the optimizing unit 5 searches for a combination of w _n where the degree of uncertainty (e.g., σ _η ² or a _nm ) of a total expected power reduction value is minimum, under the constraint conditions where the total expected power reduction value is greater than or equal to a planned power reduction value, and the number of consumers selected is less than or equal to the upper limit of the number of consumers.

At this time, the optimizing unit 5 obtains predicted outside air temperature data for a power suppression time period. The predicted outside air temperature data is data representing outside air temperatures predicted for the power suppression time period in areas where the consumers are located. The predicted outside air temperature data may be inputted by the operator or may be obtained from an external predicted outside air temperature database by the power suppression optimization system. The optimizing unit 5 calculates the degrees of uncertainty of total expected power reduction values for each time period, using expected power reduction values and the degrees of uncertainty for outside air temperatures corresponding to the predicted outside air temperatures for each time period included in the power suppression time period. The optimizing unit 5 selects a combination of consumers such that the calculated degrees of uncertainty of total expected power reduction values are the minimum.

Note that in the present embodiment β-CVaR in the second embodiment may be used as the degree of uncertainty. In this case, the degree-of-uncertainty calculating unit 4 and step S4 are not necessary. In addition, the optimizing unit 5 calculates β-CVaR, based on expected power reduction values for outside air temperatures corresponding to the predicted outside air temperatures.

As described above, according to the power suppression optimization system according to the present embodiment, preprocessed power consumption value data is used instead of power consumption value data. Therefore, a combination of consumers can be selected based on power consumption value data where missing , data and abnormal values are removed. By this, expected power reduction values and the degrees of uncertainty can be more accurately calculated. Therefore, a combination of consumers can be selected such that variation in total expected power reduction value becomes smaller.

In addition, the expected power reduction calculating unit 3 calculates expected power reduction values for each outside air temperature. By this, the expected power reduction values and the degrees of uncertainty can be changed on a predicted outside air temperature basis for a power suppression time period. Therefore, the expected power reduction values and the degrees of uncertainty can be more accurately predicted.

While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel methods and systems described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.