DESCRIPTION
APPARATUS FOR FORECASTING ELECTRIC POWER LOAD OF
INDUSTRIAL PARK, SYSTEM AND METHOD
TECHNICAL FIELD
[000 1 ] The present invention relates to an apparatus, a system, and a method for
forecasting or predicting an electric power load of an industrial park. Particularly, the present
10 invention relates to an apparatus, a system, and a method for predicting an electric power load of
an industrial park required for system stability and realization of an economical operation of
private power generation equipment in which the industrial park is set as a target.
BACKGROUND ART
15 [0002] In recent years, for the purpose of low-carbon emission and realization of
economical electric power operation, research and development of a micro grid and a smart grid
into which an information and communication technology is incorporated is active. On the
other hand, depending on a country or an area, there increases a case example where since
commercial system electric power is weak and unstable, a problem is posed not only for a daily
20 life but also for a production activity of factories. There arises a large problem that when
production bases are moved overseas from Japan, how high-quality electric power is secured.
In this regard, each business initiative is in progress as shown in Non-Patent Literatures 1 and 2.
[0003] In Non-Patent Literature 1, for example, for securing high-quality electric power
required for factory production, a diesel generator is installed in each factory and used for power
25 supply. When a production load is high, cooperation with a commercial system is performed;
however, cooperation with a diesel generator is basically promoted. When production in a
factory is stopped on holidays, a diesel generator is also stopped. However, there is shown a
case example where electric power is interchanged into other factories by using a micro grid
technology in recent years.
30 [0004] In this case, a diesel generator needs to be economically operated. For example,
the above can be performed by solving an optimization problem of diesel generator operation in
which consumption minimization of diesel fitel is taken as an objective fbnction.
[0005] In Non-Patent Literature 1, upon realizing an optimum operation plan of private
power generation equipment, cooperative control in a diesel generator of private power
0 2
generation equipment is shown based on an electric power load prediction. Further, in Non-
Patent Literature 2, an operation plan of the next day is calculated by using electric power record
data up to the previous day.
[0006] In the already proposed schemes for predicting an electric power load, data
5 estimated to be similar is extracted and a rough change in a load is predicted by using past
operation records. Thereafter, when a measured value of the day is acquired, correction is
performed in consideration of prediction errors.
CITATION LIST
10 NON PATENT LITERATURE
[0007] NON PATENT LITERATURE 1 : Daisuke Murai, Ashoku, Ashita, Toshiyuki
Takada, Masahiro Sekoguchi, Shindai Sato, Shuhei Yamano: Development of power source
solutions through introduction of smart grid to India Japanese industrial park: Hitachi Review
VO~9.2 , NO. 3, pp. 50-53 (2010)
15 NON PATENT LITERATURE 2: Yoshihiko Yoshimura, Takenori Kobayashi,
Ryo Yano: Smart grid monitoring and control system @MS: Toshiba Review Vol. 65, No. 9,
pp. 6-9
SUMMARY OF INVENTION
20 TECHNICAL PROBLEM
[OOOS] As described above, in the conventional scheme for predicting an electric power
load, operation record data of the factory in past times is supposed to be used.
[0009] However, when a micro grid is set as a target as in an industrial park, a case where
operation record data of each factory is managed in an integrated fashion is small. Further, a
25 trend of a change in an electric power load is also different in many cases. That is, there are
different a case where a change in a load occurs intermittently as in an assembly line or carrier
electric power and a case where an electric power load continues for a long time as in a chemical
process. In the literatures described above, solutions to such problems are not explicitly
described.
30 [OOlO] From the above, it is considered that there are two problems to be solved. As a
first problem, a prediction scheme is realized based on a change in an electric power load in each
factory. Even if preparing a number of operation record data, when an operation pattern that is
not present in past times is performed, prediction results are not matched with the change in the
electric power load.
6 3
[OOll] As a second problem, for securing high-quality electric power, a scheme for
estimating an electric power load required for forming an operation plan of economical private
power generation equipment is realized. As a result of considering a lot of economical
efficiency and reducing the number of private power generation equipment to be operated, a
5 system imbalance ought not to be caused.
[00 121 To attain the above problem, it is an object of the present invention to construct a
scheme for predicting an electric power load taking into consideration operation patterns of each
factory and provide a means capable of performing a prediction operation even if there are
operation patterns for which there are no past operation records.
10
SOLUTION TO PROBLEM
[OO 131 In the present invention, for attaining the above object, an apparatus for predicting
an electric power load of an industrial park to predict an amount of electric power required on a
fbture specific day by the industrial park including one or a plurality of plants, the apparatus
15 includes, in production equipment contributing to production and non-production equipment not
contributing to production in the plant, a high-quality electric power prediction unit configured
to predict progress of the amount of electric power at each time of the specific day by using a
production plan for a fbture specific day and a device and a usage time of the device used in this
production plan about the production equipment, and a low-quality electric power prediction unit
20 configured to accumulate progress of the amount of electric power of the non-production
equipment in past times and predict progress of the amount of electric power at each time of the
hture specific day based on past records about the non-production equipment, wherein the
apparatus calculates a sum of each unit for the same time and gives the amount of electric power
required by the industrial park on the fbture specific day about the progress of the amount of
25 electric power at each time calculated by each unit.
[OO 141 Further, the high-quality electric power prediction unit has a production
equipment electric power characteristic database for storing a device to be here used and a rated
amount of electric power of the device for each production plan, and the low-quality electric
power prediction unit has an electric power load record database for accumulating progress of an
30 amount of electric power of the non-production equipment in past times.
[00 151 Further, the industrial park includes a plurality of plants, wherein the high-quality
electric power prediction unit and the low-quality electric power prediction unit are installed in
each plant of the industrial park and predict an amount of electric power.
[00 1 61 Further, the industrial park includes a plurality of plants, wherein the high-quality
0 4
electric power prediction unit is installed in each plant of the industrial park and predicts an
amount of electric power, and the low-quality electric power prediction unit is mutually installed
in a plurality of plants of the industrial park and predicts an amount of electric power.
[00 171 In the present invention, for attaining the above object, a system for predicting an
5 electric power load of an industrial park, includes a private generating power source configured
to supply electric power to a plant or an industrial park including one or a plurality of plants
connected via a power line, and an apparatus for predicting an electric power load configured to
predict an amount of electric power required by the industrial park on a future specific day,
wherein the apparatus has, in production equipment contributing to production and non-
10 production equipment not contributing to production in the plant, a high-quality electric power
prediction unit to predict progress of the amount of electric power at each time of the specific
day by using a production plan for a fbture specific day and a device and a usage time of the
device used in this production plan about the production equipment, and a low-quality electric
power prediction unit to accumulate progress of the amount of electric power of the non-
15 production equipment in past times and predcit progress of the amount of electric power at each
time of the future specific day based on past records about the non-production equipment, and
the apparatus calculates a sum of each unit for the same time and gives the amount of electric
power required by the industrial park on the future specific day about the progress of the amount
of electric power at each time calculated by each unit.
20 [0018] Further, the system controls a power generation amount of the private generating
power source on the future specific day based on the progress of the amount of electric power
predicted by the apparatus for predicting an electric power load.
[00 191 In the present invention, for attaining the above object, a method for predicting an
electric power load of an industrial park to predict an amount of electric power required on a
25 hture specific day by the industrial park including one or a plurality of plants, the method
includes the steps of classifying electric power equipment of the plant into production equipment
contributing to production and non-production equipment not contributing to production,
predicting progress of the amount of electric power at each time of the specific day by using a
production plan for the future specific day and a device and a usage time of the device used in
30 this production plan about the production equipment, accumulating the progress of the amount of
electric power of the non-production equipment in past times about the non-production
equipment, and predicting the progress of the amount of electric power at each time of the fbture
specific day based on past records.
[0020] Further, the apparatus calculates a sum of each unit for the same time and gives
0 5
the amount of electric power required by the industrial park on the fbture specific day about the
progress of the amount of electric power at each time calculated by each unit.
ADVANTAGEOUS EFFECTS OF INVENTION
5 [0021] According to an apparatus, a system, and a method for predicting an electric
power load of an industrial park of the present invention, even if there are no past operation
patterns, a change in an electric power load can be estimated and private power generation such
as a diesel generator can be stably operated. When electric power capable of applying past
record data of the previous day such as lighting or office air-conditioning to estimation
10 calculation is added, an electric power load peak can be predicted with high accuracy. As a
result, a stable operation plan of a diesel generator can be performed.
The above and other objects, features and advantages of the invention will
become apparent from the following detailed description taken with the accompanying drawings.
15 BRIEF DESCRIPTION OF DRAWINGS
[0022] FIG 1 shows a system for predicting an electric power load of the present
invention;
FIG 2 shows a configuration example of an industrial park to which the present
invention is applied;
20 FIG 3 shows an example where an operation is performed during the day as one
example of a production plan;
FIG 4 shows one example of a production equipment electric power characteristic
database in the example where an operation is performed during the day;
FIG 5 shows a relationship between electric power progress and an operating
25 time of a manufacturing device in the example where an operation is performed during the day;
FIG 6 shows a concept of a low-quality electric power prediction;
FIG 7 shows an electric power load prediction result of high quality and low
quality in the example where an operation is performed during the day;
FIG. 8 shows an example of performing a continuous operation as one example of
30 a production plan;
FIG. 9 shows one example of the production equipment electric power
characteristic database in the example of performing a continuous operation;
FIG 10 shows a relationship between electric power progress and an operating
time of a manufacturing device in the example of performing a continuous operation;
FIG 11 shows an electric power load prediction result of high quality and low
quality in the example of performing a continuous operation;
FIG 12 shows an example where an operation is mainly performed at night as one
example of the production plan;
FIG 13 shows one example of the production equipment electric power
characteristic database in the example where an operation is mainly performed at night;
FIG 14 shows a relationship between electric power progress and an operating
time of a manufacturing device in the example where an operation is mainly performed at night;
and
10 FIG 15 shows an electric power load prediction result of high quality and low
quality in the example where an operation is performed during the day;
DESCRIPTION OF EMBODIMENTS
[0023] Hereinafter, one embodiment of an apparatus and a system for predicting an
15 electric power load of the present invention will be described with reference to the
accompanying drawings.
The above description has been made about the embodiment, but the present
invention is not limited thereto, it is possible to various changes and modifications within the
scope of the appended claims and the spirit of the present invention will be apparent to those
20 skilled in the art.
EMBODIMENT
[0024] First, a configuration of an industrial park 100 to which the present invention is
applied will be described with reference to FIG. 2. In this industrial park 100, a plurality of
25 plants 10 (3 plants of 104 10B, and 10C) and a power source 20 are installed. The plurality of
plants 10 are connected to the power source 20 via a power line 40 and receive electric power.
The power source 20 may also be installed off the premises of the industrial park 100.
However, the power source 20 is used as private power generation equipment to be exclusively
supplied to the plants in the industrial park in many cases, and may be configured by a plurality
30 of diesel generators having good response. Further, as countermeasures in the case of being
short of supply from the power source 20, the plurality of plants 10 can be connected to an
external commercial power source 30 and receive electric power as needed.
[0025] In each plant 10, various types of equipment are used as an electric power load,
and are roughly divided into production equipment L1 and non-production equipment L2. The
production equipment L1 is devices or equipment to be directly operated or shut down in
response to a production plan or an operation plan of a factory. In the non-production
equipment L2, devices in the plant, lighting devices in the plant or factory, and utility for the
production equipment are included. Note, however, that the utility for the production
equipment may be divided into the production equipment L1.
[0026] Here, electric power to be supplied to the production equipment is referred to as
electric power for production equipment PI, and electric power to be supplied to the nonproduction
equipment is referred to as electric power for non-production equipment P2.
Further, since the electric power for production equipment P1 may be generally called highquality
electric power, the electric power for non-production equipment P2 is called low-quality
electric power in response thereto in the present invention.
[0027] An apparatus for predicting an electric power load 50 of the present invention is
installed in any of the inside and the outside of the industrial park 100 of FIG. 2, and performs an
electric power demand prediction of the industrial park for a hture specific day (hereinafter,
referred to as an X day).
[0028] For this purpose, a past record value of the electric power for non-production
equipment P2 measured in each plant 10 is taken and accumulated in the apparatus for predicting
an electric power load 50. In addition, a production plan S of each plant for the fbture specific
day (X day) is given. As a result, an electric power load prediction value Y (X) for the hture
specific day is acquired from the apparatus for predicting an electric power load 50. The
electric power load prediction value Y (X) may be given as reference information to
administrators 60 that manage the private generating power source 20, or as a direct control
signal to a diesel generator DEG that configures the power source 20.
[0029] The apparatus for predicting an electric power load 50 of the present invention is
configured as shown in FIG. 1. In FIG. 1, a reference number 50A denotes an apparatus for
predicting an electric power load (hereinafter, referred to as a partial electric power load
prediction apparatus) for a plant 104 and this partial electric power load prediction apparatus is
installed in each plant of the industrial park. Reference numbers 50B and 50C denote partial
electric power load prediction apparatus for plants 10B and 10C, respectively, and they have the
same configuration as that of 5OA. In these partial electric power load prediction apparatus,
data to be held is different but basic processing contents are the same. Therefore, descriptions
will be made hereinafter with reference to an example of the partial electric power load
prediction apparatus 50A for the plant 10A. Note that the electric power load prediction value
Y (X) of this industrial park for the fbture specific day is finally acquired from the apparatus for
43 8
predicting an electric power load 50. The electric power load prediction value Y (X) is a timeseries
total electric power load amount of the industrial park.
[003 01 The partial electric power load prediction apparatus 50A includes a prediction
unit 1 of an electric power load for the production equipment (hereinafter, referred to as a high-
5 quality electric power load prediction unit) and a prediction unit 2 of an electric power load for
the non-production equipment (hereinafter, referred to as a low-quality electric power load
prediction unit). Respective prediction values YA1 (X) and YA2 (X) are added and a partial
electric power load prediction value Y (X) of the plant 10A is acquired. In the partial electric
power load prediction apparatus 50A of the present invention, the electric power load prediction
10 value YA1 (X) for the production equipment and the electric power load prediction value YA2
(X) for the non-production equipment are calculated by using a best method suitable for each
value.
[003 11 For this purpose, in the high-quality electric power load prediction unit 1, a day
and time of the hture specific day X is set in a prediction day and time setting unit 13. A
15 production plan SA for the day and time of the specific day X acquired from the plant 10A is
held in a production plan setting unit 16. Further, an amount of electric power required by each
production equipment is stored in a production equipment electric power characteristic database
11.
[0032] One example of the production plan SA in which an operation is performed
20 during the day will be described with reference to FIG 3. This plant 10A is, for example, a
factory that manufactures a resin compound, and is supposed to be basically in operation until
18:OO from 8:00 in the daytime on weekdays and shutdown at night and during weekends. In
addition, in manufacturing of the resin compound, each process of a kneading process Prl, a
granulation process Pr2, a cooling process Pr3, and a cutter weighing process Pr4 of materials is
25 executed. Each process Pr and an execution time of the process Pr are given in advance as the
production plan. In FIG 3, an example where all the processes Pr are executed at the same time
until 18:OO fkom 8:00 is shown.
[0033] In the production equipment electric power characteristic database 1 I, the
production equipment for executing each process and an electric power load capacity thereof
30 (rated power consumption) are stored and one specific example of the database 11 is shown. In
FIG. 4, for the purpose of executing the kneading process Prl of materials, for example, electric
power devices having device numbers from MI1 to M14 need to be operated as the production
equipment. In the electric power load capacity of these electric power devices, the device
numbers MI 1 and M12 correspond to 200kW and 1000kW, respectively. In this example,
0 9
power consumption (1000kW) of an extruder (device number M12) used in the kneading process
Prl of materials is large and it is considered to be a main factor of electric power variation. In
the production equipment electric power characteristic database 1 1, a relationship between a
device to be used and the amount of electric power of the device is similarly stored with regard
5 to the other processes Pr2 to Pr4.
[0034] Here, production equipment electric power characteristic data has a rated electric
power value based on an equipment capacity peculiar to manufacturing equipment and a time
variation of a load during an operation. Specifically, the production equipment electric power
characteristic data has an operating time until end from start during an equipment operation and a
10 corresponding variation in an electric power load. In the present invention, since a peak value
of an electric power load is confirmed as an object, electric power characteristics in an operating
time range may be a constant value equal to the equipment capacity.
[003 51 FIG. 5 shows a relationship between electric power progress and a specific
manufacturing device operating time in the kneading process Prl of materials. The
15 manufacturing device operating time represents an operation plan of the manufacturing device,
namely, an operation stop schedule of the manufacturing device related to the production plan.
The manufacturing device operating time may be stored to be associated with the production
plan 16 in relation to the process. Alternatively, since a device to be used and an operating time
of the device are uniquely determined upon deciding the process, the manufacturing device
20 operating time may be stored in the production equipment electric power characteristic 1 1.
[0036] In an example of FIG 5, as shown in a middle stage of FIG 5, the device MI1
operates for the entire period of the process Prl, and the device M12 being a largest electric
power load operates until 11:OO from 9:OO. The device M13 is supposed to operate until 16:OO
from 11 :00, and the operating device M14 is supposed to operate until 18:OO from 16:OO.
25 [0037] An operation plan equipment search unit 12 of FIG. 1 acquires a name of a
process executed at each time of the X day fi-om the production plan setting unit 16. Further,
the production equipment used (operated) by the acquired process name and an electric power
load capacity thereof are acquired from the production equipment electric power characteristic
database 1 1.
30 [0038] A time and power combining unit 14 continuously acquires an electric power load
capacity at each time of the X day, and calculates progress of the electric power load capacity for
24 hours of the X day. That is, the time and power combining unit 14 extracts a device
operating at each time and adds the electric power load capacity of the device, thereby acquiring
electric power progress for 24 hours of the X day.
0 10
[0039] This case example is shown as an electric power load in a lower stage of FIG 5.
In the case example of the kneading process Prl of this material, the device MI1 is used until
9:00 from 8:00 and the electric power load at this time is 200kW. The devices Mll and M12
are used until 11:OO from 9:00 and the electric power load at this time is 1200kW. HereinaRer,
5 detailed descriptions are omitted; further, the electric power load progresses to 400kW and
300kW, and an operation of this day is stopped.
[0040] FIG 5 shows an example where the electric power progress for 24 hours of the X
day is acquired in the case example of the kneading process Prl of materials. In the same
manner, the electric power progress for 24 hours of the X day is acquired about the other
10 processes. Moreover, the electric power progress for the 24 hours of the X day of the total of
each process is acquired. This progress of the total amount of electric power is not shown;
however, for ease of explanation, a progress pattern of the electric power load amount shown in
the lower stage of FIG. 5 will be described below as the prediction value YAl (X) of the highquality
electric power load prediction unit 1. The progress of the electric power load capacity
15 for 24 hours of the X day acquired in this manner is produced as the prediction value YA1 (X)
from a high-quality electric power prediction value output unit 15.
[004 11 Next, the low-quality electric power load prediction unit 2 of FIG. 1 sets the
fbture specific day X in a prediction month and day setting unit 22 and includes an operation
calendar 24. Further, the low-quality electric power load prediction unit 2 takes in a past record
20 value of the electric power for non-production equipment P2 measured in the plant 10A and
accumulates it in an electric power load record database 2 1.
[0042] Here, in the low-quality electric power (electric power load having no connection
relation to a manufacturing device), equipment not directly involved in manufacture such as
office lighting and air-conditioning electric power is targeted as described above. Voltage or
25 frequency characteristics directly have no influence on product quality, and therefore are referred
to as low-quality electric power. In many cases, in addition to calendar information such as the
day and time, the low-quality electric power may be considered to have characteristics affected
by weather factors such as temperature and humidity.
[0043] A concept of the low-quality electric power prediction will be described with
30 reference to FIG 6. An electric power load record in units of day accumulated in the electric
power load record database 21, for example, a variation in the electric power load for 24 hours is
shown as in FIG 6. This electric power load record indicates a past record value of the electric
power for non-production equipment P2, and therefore is mainly generated by the electric power
load such as lighting, and heating and cooling. As a whole, the electric power load record
electric power load record shows a different trend in operation days and non-operation days, and
also fluctuates depending on the day and the season. In addition, the electric power load record
is affected also by weather of the day. In the electric power load record database 21, a past
5 record value of the electric power for non-production equipment P2 measured under various
types of these conditions is stored.
[0044] FIG 6 of upper stage shows a standard pattern of the electric power load of office
lighting and air conditioning in the factory. As a feature of the factory, the low-quality electric
power load is supposed to be an electric power load smaller than or equal to that of the
10 manufacturing process. In a standard pattern of the upper stage of FIG 6, the low-quality
electric power load is supposed to be an electric power load of 500kW at a maximum at an
operating time zone and an electric power load of 200kW in general at the other time zone. In
addition to this standard pattern, the low-quality electric power prediction is performed with
reference to the record value of the previous day or the electric power load record of the same
15 day in the previous week.
[0045] Similarly to the standard pattern in the upper stage of FIG 6, the record value of
the previous day in the middle stage shows the electric power load of 500kW at a maximum, and
shows data in which a load rise is changed rapidly in the morning. With reference to the
standard pattern and the record value of the previous day, a result of predicting the low-quality
20 electric power load is shown in the lower stage of FIG 6. Various methods for acquiring a
prediction value based on the standard pattern and the record value of the previous day are used.
Generally, a weighted average is calculated at each time and is set as a prediction value. As a
result, a result of a lower solid line can be acquired.
[0046] A similar load search unit 25 of the low-quality electric power load prediction unit
25 2 acquires a past record value of the electric power for non-production equipment P2 matched
with conditions of the fbture specific day X from the electric power load record database 2 1. In
addition, the past record value to be here acquired may be acquired not only by extracting one
data but also by calculating an average of several similar data matched with the conditions.
[0047] It suffices that the similar load search of the low-quality electric power load
30 prediction unit 2 is performed based on the above-described several concepts, and one typical
example thereof will be described with reference to FIG. 6. In this case, the standard pattern of
the upper stage of FIG. 6 is acquired from past records as the average of several similar data
matched with the conditions. In this pattern, the amount of electric power increases or
decreases relatively slowly during the day when persons are present, and reaches a peak around
0 12
noon. In contrast, in data (in the middle stage of FIG. 6) of the previous day, the amount of
electric power noticeably tends to increase rapidly before start of a plant and decrease swiftly
with a clock-out time. By taking the two patterns into consideration, as a final electric power
prediction, there is acquired the electric power prediction shown in a thick solid line of the lower
5 stage of FIG. 6, in which a trend of the record of the previous day is reflected on a clock-in time
and a clock-out time and the amount of electric power reaches a peak around noon as a feature of
the standard pattern. The progress of the electric power load capacity for 24 hours of the X day
acquired in this manner is produced as the prediction value YA2 (X) from a low-quality electric
power prediction value output unit 23.
10 [0048] Here, in the low-quality electric power load prediction unit 2 of the apparatus for
predicting an electric power load 50A of FIG. 1, the low-quality electric power load of the plant
10A is predicted is described. The low-quality electric power load is not predicted in each plant
but predicted collectively as the entire industrial park. The process permits a system to be
simply configured as compared to a case of separately performing the low-quality electric power
15 load prediction in the plant. That is, preferably, the embodiment has a configuration in which
the low-quality electric power load prediction unit 2 is installed only in the partial electric power
load prediction apparatus 50A and the low-quality electric power load is predicted collectively as
the entire industrial park, and in which the low-quality electric power load prediction unit 2 is
not installed in the other prediction units 50B and 50C.
20 [0049] FIG 7 shows a result in which the high-quality electric power progress YA1 (X)
and the low-quality electric power YA2 (X) is predicted along the time direction at the time when
the plant 10A of FIG 2 performs an operation on the hture specific day X based on the
production plan (daytime operation of the processes Prl to Pr4). Note that, as described above
for convenience of explanation, as shown in FIG 7, a pattern as a prediction result (in the lower
25 stage of FIG 5) of the process Prl of FIG 5 is the same pattern as the prediction result of the
total high-quality electric power including the other processes. In addition, it suffices that for
actually acquiring a total electric power load prediction value YA (X) of the final factory, the
high-quality electric power is acquired about each of the other processes P2, P3, and P4 similarly
to FIG 5 for cumulative addition.
30 [0050] Referring back to FIG 1, the total electric power load prediction value YA (X) of
the factory is similarly acquired with respect to the other plants 10B and 10C in this industrial
park. Since a method for acquiring it can be easily understood basically from the above
description, one example will be simply described here.
[005 11 The plant 10B is supposed to be an assembly factory such as air conditioning and
0 13
car in the case of a 24-hour full production. FIG 8 shows the production plan in the case of a
24-hour full production. Here, the processes Pr5 to Pr8 are continuously operated.
[0052] FIG 9 shows one example of the production equipment electric power
characteristic database 11 in a continuous operation example, and particularly shows a
5 relationship between rated power consumption and manufacturing devices in the process Pr5.
In this table, devices M5 1, M52, and M53 in the process Pr5 and rated power consumption W5 1,
W52, and W53 thereof are written. Also, the other processes Pr6 to Pr8 are also set in the same
manner.
[0053] FIG 10 shows in the middle stage an operation stop schedule of the
10 manufacturing device in the process Pr5 of the continuous operation example. In the process
Pr5, the devices M5 1 and M52 are put in a full operating condition day and night, and the device
M53 is operated until 10:OO from 2:OO. In the lower stage of FIG 10, the corresponding electric
power load prediction value is shown and the electric power load of 950kW or 1000kW is
acquired as a prediction value through a 24-hour period. In 21 :00 to 23:00, the electric power
15 load is reduced to 930kW, which is due to the fact that a production load is reduced at night.
When partial load operation data of the manufacturing device is not written, there is no problem
in that the electric power load is predicted to be 950kW.
[0054] FIG 11 shows an example in which the high-quality electric power prediction
value YE3 1 (X) and the low-quality electric power prediction value YB2 (X) as a prediction result
20 are written along the time direction. Here, in the description, a prediction result waveform of
the process Pr5 of FIG 10 is written as the final high-quality electric power prediction value YB1
(X) in consideration of operation results in the other processes.
[0055] Further, similarly to the embodiment in which an operation is performed during
the day, since the low-quality prediction value changes without relation to an operation schedule
25 of the manufacturing device, a prediction value collectively acquired in the industrial park of
FIG 6 is here used.
[0056] In the plant 10C of FIG 2, an example where devices are put in a full operating
condition day and night is shown. As shown in FIG 12, in addition to the process Pr9 of a 24-
hour operation, the plant 10C takes a night subject-type operating mode containing the processes
30 Prl 0, Prl 1, and Pr 12 of night operations. In FIG. 12, since the next data of 23 :00 is data of the
next day, the process PrlO is indicated in 0:00 to 6:OO. As an actual operation, a process in
which an operation is started at 20:OO and stopped at 6:00 of the next day is adopted.
[0057] FIG. 13 is a data table of the rated power consumption corresponding to the
manufacturing device number in the process 10. In the device M102, an equipment start-up
0 14
time is here written in parentheses. As in an induction heating furnace, for example, in the
device M102, the rated electric power capacity is large and much time is necessary to start it.
Further, an increase in power consumption pattern at each time needs to be written.
[0058] FIG 14 shows in the middle stage a manufacturing device operation stop schedule
5 in the process PrlO. There is adopted a schedule in which the devices MlOl and M102 are
started at 20:OO at night, the device M102 is stopped at 4:00 the next day, the device M103 is
started in place of the device M102, and the devices MlOl and M103 are stopped at 6:OO. The
electric power load prediction value corresponding to the above schedule is shown in the lower
stage of FIG. 14. Since the start-up time of 1.5 hour is required by the device MI02 of the
10 equipment that is operated at 20:OO at night, a prediction result in which the device M102
reaches the rated power consumption at 21:30 is acquired. After that, a result in which the
power consumption prediction value is reduced to 250kW from 1100kW at 4:00 is produced in
the next day prediction value calculation.
[00 5 91 FIG 15 is an example where in the process Pr 10, the high-quality electric power
15 prediction value and the low-quality electric power prediction value are shown in the middle
stage and the lower stage along the time direction, respectively. The low-quality electric power
prediction value is supposed to be acquired similarly to the case of FIG 6. The electric power
load prediction value of the entire factory is acquired by cumulatively adding the process Pr9
during the operation for 24 hours and the other processes Pr 1 1, Pr 12, and Pr 1 3.
20 [0060] As described above, the electric power load prediction in the factory is indicated
in the case where an operation is performed during the day, a 24-hour full production is
performed, and an operation is performed at night. The electric power load prediction can
correspond to a factory holiday by using the low-quality electric power prediction value shown
in FIG 6. Note that the standard pattern in the case of a factory holiday needs to be prepared.
25 [0061] As described in detail above, the high-quality electric power load prediction and
the low-quality electric power load prediction are performed in the electric power load prediction
of the present invention, respectively. Further, prediction values based on the above predictions
are added and a final electric power load prediction value is acquired. The former is an electric
power load prediction performed based on the hture production plan, and the latter is an electric
30 power load prediction performed based on the past records.
[0062] In FIG 1, the high-quality electric power prediction value YA1 (X) and the lowquality
electric power prediction value YA2 (X) for the same time of the plant 10A are added by
using an addition means 3, and as a result an electric power load prediction value 4 (YA (X)) is
acquired.
0 15
100631 In the present invention, electric power load prediction values YB (X) and YC (X)
are hrther received as notification values from the other factories SOB and 50C. Further, they
are added by using an addition means 5 and the entire electric power load prediction total value
Y (X) is acquired.
5 [0064] As described previously, prediction results are automatically reflected on direct
control in a plurality of diesel generators DEG of the private generating power source 20, or
reflected on manual setting based on determinations of the administrator.
[0065] High-quality electric power acquired in this manner is directly connected to a
diesel generator group DEG being private power generation equipment. Even if an external
10 system power source is unstable, stable high-quality electric power can be acquired. On the
other hand, low-quality electric power is supposed to allow a voltage variation and is directly
connected to an external system power source. Note that a diesel generator and an external
system can also be connected and the low-quality electric power allows a diesel generator to
supply electric power.
15 [0066] As described above, based on a production equipment operation plan, a scheme
for predicting an electric power load of the present invention can be constructed for performing a
load prediction with high accuracy to electric power load by which high quality such as voltage
stability is required. In addition, an electric power load prediction is separated from a
prediction of low-quality electric power as in miscellaneous electric power and a load prediction
20 value is acquired. Therefore, a load change with periodicity can also be considered.
REFERENCE SINGS LIST
[0067] 1 Prediction unit of electric power load for production equipment
2 Prediction unit of electric power load for non-production equipment
5 Apparatus for predicting electric power load
104 10B, 10C A plurality of plants
11 Production equipment electric power characteristic database
13 Prediction day and time setting unit
14 Time and power combining unit
16 Production plan setting unit
20 Power source
2 1 Electric power load record database
24 Operation calendar
25 Similar load search unit
3 0 Commercial power source
40 Power line
50A Apparatus for predicting an electric power load for plant 10A
50B, 50C Partial electric power load prediction apparatus for plants 10B and
5 10C
100 Industrial park
L1 Production equipment
L2 . Non-production equipment
P 1 Electric power for production equipment
P2 Electric power for non-production equipment
DEG Diesel generator
M Device

CLAIMS
[Claim 1]
An apparatus for predicting an electric power load of an industrial park to predict
an amount of electric power required on a hture specific day by the industrial park including one
or a plurality of plants, the apparatus comprising:
in production equipment contributing to production and non-production
equipment not contributing to production in the plant,
a high-quality electric power prediction unit configured to predict progress of the
amount of electric power at each time of the specific day by using a production plan for a hture
specific day and a device and a usage time of the device used in this production plan about the
production equipment; and
a low-quality electric power prediction unit configured to accumulate progress of
the amount of electric power of the non-production equipment in past times and predict progress
of the amount of electric power at each time of the hture specific day based on past records
about the non-production equipment, wherein
the apparatus calculates a sum of each unit for the same time and gives the
amount of electric power required by the industrial park on the hture specific day about the
progress of the amount of electric power at each time calculated by each unit.
[Claim 2]
The apparatus for predicting an electric power load of an industrial park
according to claim 1, wherein
the high-quality electric power prediction unit has a production equipment electric
power characteristic database for storing a device to be here used and a rated amount of electric
power of the device for each production plan, and
the low-quality electric power prediction unit has an electric power load record
database for accumulating progress of an amount of electric power of the non-production
equipment in past times.
[Claim 3]
The apparatus for predicting an electric power load of an industrial park
according to claim 1, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit and the low-quality electric power
prediction unit are installed in each plant of the industrial park and predicts an amount of electric
power.
[Claim 4]
The apparatus for predicting an electric power load of an industrial park
according to claim 1, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit is installed in each plant of the
industrial park and predicts an amount of electric power, and
the low-quality electric power prediction unit is mutually installed in a plurality of
plants of the industrial park and predicts an amount of electric power.
[Claim 5]
A system for predicting an electric power load of an industrial park, comprising:
a private generating power source configured to supply electric power to a plant
or an industrial park including one or a plurality of plants connected via a power line; and
an apparatus for predicting an electric power load configured to predict an amount
of electric power required by the industrial park on a fbture specific day, wherein
the apparatus includes:
in production equipment contributing to production and non-production
equipment not contributing to production in the plant,
a high-quality electric power prediction unit to predict progress of the amount of
electric power at each time of the specific day by using a production plan for a hture specific
day and a device and a usage time of the device used in this production plan about the production
equipment and
a low-quality electric power prediction unit to accumulate progress of the amount
of electric power of the non-production equipment in past times and predict progress of the
amount of electric power at each time of the hture specific day based on past records about the
non-production equipment, and
the apparatus calculates a sum of each unit for the same time and gives the
amount of electric power required by the industrial park on the hture specific day about the
progress of the amount of electric power at each time calculated by each unit.
[Claim 6]
The system for predicting an electric power load of an industrial park according to
claim 5, wherein
the high-quality electric power prediction unit has a production equipment electric
power characteristic database for storing a device to be here used and a rated amount of electric
power of the device for each production plan, and
the low-quality electric power prediction unit has an electric power load record
0 19
database for accumulating progress of an amount of electric power of the non-production
equipment in past times.
[Claim 7]
The system for predicting an electric power load of an industrial park according to
claim 5, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit and the low-quality electric power
prediction unit are installed in each plant of the industrial park and predicts an amount of electric
power.
[Claim 8]
The system for predicting an electric power load of an industrial park according to
claim 5, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit is installed in each plant of the
industrial park and predicts an amount of electric power, and
the low-quality electric power prediction unit is mutually installed in a plurality of
plants of the industrial park and predicts an amount of electric power.
[Claim 9]
The system for predicting an electric power load of an industrial park according to
claim 5, wherein
the system controls a power generation amount of the private generating power
source on the hture specific day based on the progress of the amount of electric power predicted
by the apparatus for predicting an electric power load.
[Claim 10]
A method for predicting an electric power load of an industrial park to predict an
amount of electric power required on a kture specific day by the industrial park including one or
a plurality of plants, the method comprising the steps of
classifying electric power equipment of the plant into production equipment
contributing to production and non-production equipment not contributing to production,
predicting progress of the amount of electric power at each time of the specific day by using a
production plan for the hture specific day and a device and a usage time of the device used in
this production plan about the production equipment, accumulating the progress of the amount of
electric power of the non-production equipment in past times about the non-production
equipment, and predicting the progress of the amount of electric power at each time of the hture
specific day based on past records.
The method for predicting an electric power load of an industrial park according
to claim 10, wherein
the apparatus calculates a sum of each unit for the same time and gives the
amount of electric power required by the industrial park on the fbture specific day about the
progress of the amount of electric power at each time calculated by each unit.
[Claim 12]
The apparatus for predicting an electric power load of an industrial park
according to claim 2, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit and the low-quality electric power
prediction unit are installed in each plant of the industrial park and predicts an amount of electric
power.
[Claim 13]
The apparatus for predicting an electric power load of an industrial park
according to claim 2, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit is installed in each plant of the
industrial park and predicts an amount of electric power, and
the low-quality electric power prediction unit is mutually installed in a plurality of
plants of the industrial park and predicts an amount of electric power.
[Claim 14]
The system for predicting an electric power load of an industrial park according to
claim 6, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit and the low-quality electric power
prediction unit are installed in each plant of the industrial park and predicts an amount of electric
power.
[Claim 15]
The system for predicting an electric power load of an industrial park according to
claim 6, the industrial park including a plurality of plants, wherein
the high-quality electric power prediction unit is installed in each plant of the
industrial park and predicts an amount of electric power, and
the low-quality electric power prediction unit is mutually installed in a plurality of
plants of the industrial park and predicts an amount of electric power.