JP3332032B2 - Generator operation planning device, generator operation planning method, and storage medium storing generator operation planning program - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a generator operation plan creating apparatus and a storage medium for determining an economical start / stop state and output of a generator while satisfying the generator operation constraints. The generator here is a general term for supplying electric power, and includes a generator for thermal power, pumping, etc., as well as receiving power from the independent electric power company IPP and interchanging electric power between electric power suppliers.

[0002]

2. Description of the Related Art Taking a generator as a typical example of a generator, as a method for determining a start / stop plan of a thermal power generator, a generator start-up using a genetic algorithm is described as a lecture number 1272 of the 1996 IEEJ National Convention. Suspension plan ”and the 1995 IEEJ Power and Energy Division Conference Lecture No. 24
2. “Method of drafting thermal power and pumping week power generation plan” has been proposed. In these methods, it is a unit of time of the expected demand, and at the same time, a unit of time for calculating the operating cost, for example, the start / stop state of all the generators in units of one hour, is used as a variable to apply a combination optimization method to these variables. Are created at once.

[0003]

In the prior art, a genetic algorithm or a problem space search method is employed as a method, but a combination of starting and stopping at all times during a planning period is created at a time. For this reason, for example, the number of combinations of the number of generators and the number of start / stop plans of the generators at the time of 24 / day is as large as 2 to the 240th power, and an enormous amount of time is required to obtain the best plan.

[0004] The proposed methods also have the following problems.

[0005] In the proposed genetic algorithm, when the start / stop state of the generator every hour is represented by a string as described above, the string length becomes very long, and the generator is crossed and mutated by genetic operators. Since a solid that does not satisfy the operation constraint or the operation constraint of the power system is created, it cannot be searched efficiently.

In the proposed method using the problem space search method, the adjacent state is defined as making the start time of one generator earlier or later by one hour, or making the stop time earlier or later by one hour. Therefore, there are 40 adjacent states (= 2 × 2 × 10), and in order to stop the generator operating continuously for 8 hours, the generator is continuously operated eight times out of a combination of 40 to the eighth power. It is necessary to select an event that shortens the operation time. Since such a continuous event has not occurred, the power plant in the initial plan and the final plan described in the 1995 IEEJ Power and Energy Division Conference Lecture No. There is no change in the start / stop plan of the minimum demand time and the maximum demand time, and only the start time or the stop time has changed. It is possible that the search range is narrow.

[0007]

SUMMARY OF THE INVENTION According to the present invention, the expected demand at each time of the entire planning period is characterized by a characteristic point, for example, the power demand has a curve as shown in FIG. , Sampling the time of minimum demand and creating a generator start / stop plan at this sampled time,
Under the condition that the start / stop plan is fixed, a start / stop plan at other times is created.

For example, for a generator start-up / stop plan at 24 times a day with 10 generators, the number of combinations at two time points sampled is 2 to the 10th power and 2 times the combination at 24 time points.
Compared to the 240th power, the number of combinations can be reduced to 1 / 230th power. Considering general operation ideas, such as continuity between generator operation times and economics, if operating at the time of minimum demand, it will operate continuously until the time of maximum demand. If it stops at the time, it means that it stops even at the time of the maximum demand. The reason is that the generator operating at the time of minimum demand has high efficiency, so it is economical to operate even at the time of maximum demand. It is currently difficult to meet this constraint to stop and restart between the minimum demand and the maximum demand, and the restart cost is expensive to start Is to be increased. Utilizing these ideas, the start / stop at a time other than the sampling time is created at high speed by the priority order method in consideration of the following operation constraints. Alternatively, by fixing the start / stop at the time of sampling and using the above concept, the start / stop state of each generator and each generator is determined instead of the start / stop time as a variable. This reduces the number of combinations. That is, for the generator operating at the minimum time, the start time limited to the range from the first time to the time of the minimum demand (the continuous operation is performed when the time is smaller than the first time) becomes a variable,
For generators that stop at the time of minimum demand and operate at the time of maximum demand, the start-up time limited to the range between the time of minimum demand and maximum demand becomes a variable, and the generator that operates at the time of maximum demand As the machine has a variable stop time limited to the time range after the maximum demand (continuous operation when the time is longer than the planned time), it stops at all times when it stops at the maximum demand, Since the variables of the start time and the stop time are unnecessary, the number of combinations can be reduced.
For example, when the minimum demand time is 5:00, the maximum demand time is 17:00, and both times are operated, even when the combination is the maximum,
The number of combinations is 440 (= 5 × 11 × 8), and the number of combinations in the conventional method is 4194304 except for the state at two points in time.
This can be reduced to about 1/1000 compared to (2 to the 22nd power).

That is, when the type of plan is represented in one dimension, the horizontal axis is the plan and the vertical axis is the plan evaluation function, as shown in FIG. The evaluation function has fine irregularities and large irregularities. In the operation plan of the conventional method, the one-hour start / stop plan of one generator is changed and the plan is changed little by little, so that the plan is continuously changed along the fine irregularities in the horizontal axis direction. That is, in this figure, it is considered that a search is made for the vicinity of fine irregularities in one large irregularity including the initial solution. Therefore, changing the start and stop at the maximum and minimum times of demand requires a very large number of searches. On the other hand, in the present invention, since the operation plan is determined at the sampling time of the minimum and maximum demands of the day, a new solution is searched for along the horizontal axis along the large irregularities. Become. Next, deciding a plan at another time under the condition that the start and stop at the sampling time is fixed corresponds to continuously changing the plan in the vicinity thereof.

[0010] By selecting an operation plan with a good evaluation function from the created operation plans, an operation plan that satisfies the operation constraints and has a small operating cost can be created at high speed.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a generator operation plan will be described below with reference to the drawings. A. Device Configuration (1) Hardware Configuration As shown in FIG. 4, an operation plan creation device 1 (FIG. 1) of the present embodiment includes a central processing unit (CPU) 2 and a main storage device 3.
, An input / output device 4 and an external storage device 5.

The input / output device 4 includes a keyboard and a mouse as an input device 19 and a CRT as an output device.
(Cathode ray tubu).
Instead of or in combination with the input / output device 4, input means such as a pointing device and a touch sensor, a liquid crystal display device, a printer,
An output unit such as a speaker may be provided.

As the external storage device 5, a hard disk device, a floppy disk device, a CD-ROM (compact
disc-read only memory) device, DAT (digitalvide
otape) device, RAM (randam access memory) device, D
A VD (digital videodisc) device, a nonvolatile memory, or the like can be used. In this embodiment, as the external storage device 5, a large-capacity storage device for storing the database 21 and the like, a storage medium for storing the processing program and the like, and a reading device for reading the information stored in the storage medium. The device 22 is used, but the present invention is not limited thereto.
A single external storage device may hold both a database and a processing program. As storage media, a floppy disk, CD-ROM, magnetic tape, optical disk, magneto-optical disk, DAT, RAM, D
VD, a nonvolatile memory, or the like can be used.

The receiving system 7 is a facsimile and a computer installed at each power station. Distribution system 6
Is a computer having communication means. Further, if the power plant is an IPP or power interchange, it serves as a means for receiving information on a sales contract. (2) Functional Configuration FIG. 1 shows a functional configuration of the operation plan creation device 1 of the present invention.
The operation plan creation device 1 of the present embodiment includes the input device 1 described above.
9, display device 20, database 21, and reading device 2
2 and a data processing unit 8. In this embodiment, the data processing unit 8 is stored in the storage medium of the external storage device 5 in advance, and is stored in the main storage device 3 via the reading device 22.
The present invention is realized by the CPU 2 executing the program read by the general-purpose processor, but the present invention is not limited to such a programmed general-purpose processor. For example, the data processing unit 8 may be realized by a combination with a specific hardware device including a hard wired logic that executes each step of the present invention.

The input device 19 is a means for receiving selection of an option displayed on the display device 20, input of data, and the like, and sending it to the data processing section 8. The display device 20 includes the input device 1
This is a means for displaying the data sent from the control unit 9 and the data sent via the control unit 9. The data processing unit 8 is a means for creating a generator operation plan based on the data input from the input device 19, the data read from the database 21, and the processing program read from the reading device 22. The data processing unit 8 sends the created operation plan to the distribution system 6, and the distribution system 6 processes the data of the operation plan created according to the other party and sends it to the reception system 7.

In this embodiment, the processing result by the data processing unit 8 is sent to the display device 20 for display and stored in the database 21.

The data processing unit 8 of this embodiment includes a control unit 9, a condition reading unit 10, a condition setting unit 11, a sampling unit 12, a sampling time plan creating unit 13, a non-sampling time plan creating unit 14, and a plan creating unit 14. An evaluation unit 15 and a display data creation unit 17 are provided. The data processing unit 8 is connected to an external receiving system 7 via the distribution system 6, and is connected so as to send the created operation plan to each power plant.

The control section 9 processes and processes data for smoothly transmitting and receiving data and processing programs between the distribution system 6 and the processing sections 10 to 17, and controls the transmission and reception. , Means for operating the entire process normally.

The condition reading unit 10 determines, via the control unit 9, the expected demand, the start / stop of the generator, and the load distribution output already stored in the database 21 and / or the storage medium of the reading device 22. This is a means for reading the data of the power system, generator characteristics, transmission and substation characteristics, operation constraint conditions, conditions such as objective functions, processing programs, and the like used for the above. The read data is stored in the data storage unit 17. In the present embodiment, unless otherwise specified, the constraint conditions read by the processing units 10 to 16, the set conditions, and the calculated results are transmitted to the control unit 9 via the control unit 9.
It is stored in the data storage unit 17 and notifies the processing units 10 to 16 via the control unit 9 as necessary.

The condition setting unit 11 sets conditions for creating an operation plan and an evaluation function via the control unit 9. The conditions for creating an operation plan include the conditions for sampling demand, the method for creating a plan at the time of sampling, the method for creating a plan at the time of non-sampling, the termination conditions, and the operator forcibly operating the generator operating state. And conditions for stopping.

The sampling unit 12 includes a control unit 9
, The expected demand and its time are sampled based on the sampling condition input from the condition reading unit 10, and these data are stored in the data storage unit 17 via the control unit 9.

The sampling time plan creator 13 creates a plan candidate by using an optimization method or the like by combining a start / stop plan of only the sampling time from the data storage 17 via the controller 9. These plan candidates are stored in the data storage unit 17 via the control unit 9.

The non-sampling time planning unit 14
For each start / stop plan at the sampling time created by the sampling time plan creation unit 13, the start / stop plan at this sampling time is fixed, and a start / stop plan at other times is created. The non-sampling time plan is created in consideration of the constraints read from the data storage unit 17. The generator output of each generator is calculated based on the start / stop plan and the constraints. The created non-sampling time plan is stored in the data storage unit 17 via the control unit 9 as one plan of the entire planning period, together with the start / stop plan of the sampling time.

The plan evaluator 15 calculates an evaluation function by calculating a power generation cost and a penalty corresponding to a constraint violation for the start / stop plan of the entire plan period via the controller 9. According to the value of the evaluation function, a start / stop plan serving as a source of the next search is selected from the start / stop plans of the entire planning period, and stored in the data storage unit 17 via the control unit 9. Further, the best start-stop plan, load distribution plan and evaluation function in this search are stored in the data storage unit 17 as the best plan. Further, the best plan obtained so far is compared with the best plan, and if the evaluation function is good, the best plan is stored in the data storage unit 17 as the best plan.

The display data creation unit 16 includes the processing units 10 to 10
The display device 20 displays the plan conditions, transitions of plans and evaluation functions, the degree of satisfaction of constraints, data stored in the data storage unit 17, information for supporting data input from the input device 19, and the like calculated in 16. Let it. In the present embodiment, output to the display device 20 is performed via the display data creation unit 16 unless otherwise specified. (3) Data Structure of Data Storage Unit 17 As shown in FIG. 5, the data storage unit 17 includes a storage area 100 for the planning target period, a storage area 101 for the evaluation function, a storage area 102 for the constraint condition, and a storage for the calculation result. Area 103, calculation result storage area 104, and sampling time storage area 10
5 is provided.

The planning target period storage area is a storage area for holding a date and a start time and an end time as a period for which a plan is created.

The evaluation function storage area 101 is a storage area for holding information on the evaluation function. The evaluation function storage area 101 stores a cost function coefficient that increases or decreases according to the output based on the characteristics of the generator or the like, a startup cost according to the stop time, and the like. A generator facility characteristic storage area 1011 for storing a penalty weight coefficient storage area 101 for storing a weight coefficient for adding a penalty cost according to a constraint and a violation amount when a constraint violation occurs.
2 is provided.

The constraint condition storage area 102 is a storage area for retaining constraint conditions to be considered when determining the start / stop and the generator output, and is a demand storage area 1021 for retaining an assumed demand. Upper / lower limit of machine output, minimum continuous stop time, minimum continuous stop time, combination of generators for which simultaneous activation of the same central control unit is prohibited, generator operation stop, forced / forced operation of generator specified by operator, power generation A generator operation characteristic storage area 1022 for retaining generator operation constraints such as supply power characteristics indicating the load by number and facility characteristics indicating generator type by number, and restrictions on the system system such as operating reserve power A system system operation constraint storage area 1023 for retaining the conditions, and a supply power integration constraint storage area 1024 for retaining the fuel consumption, the departure of the pumping basin, the upper and lower water levels, and the final water level for the entire planning period. In order to use in power flow calculation, transmission lines and transformers are represented by branches, and buses are represented by nodes. A system connection state storage area 1025 for storing a node number to which the supply power of each generator is connected and a magnitude of a load applied to each node, and a system facility for storing upper and lower limits of a power flow constraint for each branch And an operation characteristic storage area 1026. The creation condition storage area 103 stores an adopted technique storage area 1031 for storing a technique adopted for a start / stop plan at a sampling time and a technique adopted for a start / stop plan at a non-sampled time, and a parameter storage for storing parameters according to the adopted technique. An area 1032 and an end condition storage area 1033 for storing an end condition of processing such as the number of searches are provided. Applicable methods include a priority optimization method or a combination optimization method such as a tabu search method, a local search method, a genetic algorithm, and simulated annealing. In the parameter storage area 1032, the items to be stored change depending on the method adopted,
In the case of the taboo search method, the definition of the adjacent state, the length of the taboo list, the characteristics (attributes) stored in the taboo list, and the taboo list items stored during the search are stored. Similarly, the definition of the adjacent state is stored. For a genetic algorithm,
The number of individuals, the crossover rate and the mutation rate. In the case of simulated annealing, the initial temperature and the temperature reduction rate are used. The calculation result storage area 104 is a storage area for holding calculation results, calculation histories, and final results formed during the search. The sampling time storage area 105 is a storage area for storing the time extracted by the sampling unit 12. (4) Data Structure of Database 21 The database 21 has a plurality of storage areas of the same format as the data storage unit 17 for storing calculation conditions and calculation results. These stored data can be calculated again as initial values. In addition, the generator characteristics, predicted demand, generator operation characteristics, system system operation constraints, supply power integration constraints, system connection state, system equipment operation characteristics read by the condition reading unit 10 when a new plan is created. Storage area. B. Process Flow The operation plan creation device 1 of the present embodiment creates an operation plan of the generator according to the processing flow of FIG. In step S1, plan creation conditions are set, and in step S2, plan creation conditions are read. In process S3, the forced operation / stop condition of the operator is set. Next, an optimal operation plan is created in steps S4 to S8 based on the set plan creation conditions and the read conditions. Finally, the optimal operation plan and the plan creation result are displayed in step S9.

The operation plan creation device 1 of this embodiment first displays a process selection menu on the display device 20, and the input device 1
The menu selection is accepted via 9. The process selection menu includes setting of plan creation conditions, reading of plan creation conditions, setting of forced operation / stop of start / stop by an operator, creation of an operation plan, and display of a result. Each process performs a corresponding process by selecting a target process from a process selection menu via the input device 19. (1) Setting of plan creation conditions In the process S1, the following processes are performed. In the process selection menu, the setting screen is displayed on the display device 20 by selecting the setting of the plan creation condition via the input device 19, and the necessary items are input via the input device 19 and confirmed, thereby storing the data. It is stored in the unit 17. The items to be set here are: year, month, day, planning period (days), operating reserve, reduction allowance, sampling conditions, sampling time planning method and variables used in that method, non-sampling time planning method And variables used in that method, search mode, final search method and variables used in that method, end conditions (maximum number of local loops, maximum number of global loops), selection of constraints to be considered, constraint violations used in evaluation functions And the like. Regarding the IPP, the interchange power reception and the interchange power transmission, depending on the operation mode, a method of selecting from a daily pattern operation or a method of determining the output at each time is selected. (A) Setting of sampling conditions Sampling conditions are selected from daily maximum demand, daily minimum demand, daily minimum demand and maximum demand, daily minimum demand, morning maximum demand, maximum demand, lighting maximum demand, and sampling time interval. I do. When the sampling time interval is selected, the time interval or the sampling time is further set. (B) Method of creating start / stop plan at sampling time Select a plan creation method to be used from a taboo search method, a local search method, a genetic algorithm, a simulated annealing, a fixed priority method, a variable priority method, and the like. Items to be set differ depending on each method. When the tabu search method is selected, conditions for applying the tabu search method, such as the adjacent state, characteristics stored in the tabu list, and the tabu list length, are set. In the case of the local search, the definition of the adjacent state is stored as in the tabu search method. When the genetic algorithm is selected, the number of individuals in the population, the crossover rate and the mutation rate are set. In the case of simulated annealing, an initial temperature and a temperature reduction rate are set. In the priority order method, only the method needs to be set, and there is no need to set a new variable. The fixed priority method uses the read priority as a generator characteristic. In the variable priority method, the priority is changed every hour in consideration of demand and continuity with plans before and after. One or more search end conditions are selected from the maximum number of searches, a target evaluation function value, and a calculation time, and values of the number of searches, an evaluation function value, and a calculation time are set, respectively. (C) Start / stop plan creation method at non-sampling time The plan creation method to be used is selected from a tabu search method, a local search method, a genetic algorithm, a simulated annealing, a fixed priority method, a variable priority method, and the like. Items to be set differ depending on each method. When the tabu search method is selected, conditions for applying the tabu search method, such as the adjacent state, characteristics stored in the tabu list, and the tabu list length, are set. In the case of the local search, the definition of the adjacent state is stored as in the tabu search method. When the genetic algorithm is selected, the number of individuals in the population, the crossover rate and the mutation rate are set. In the case of simulated annealing, an initial temperature and a temperature reduction rate are set. In the priority order method, only the method needs to be set, and there is no need to set a new variable. The fixed priority method uses the read priority as a generator characteristic. In the variable priority method, the priority is changed every hour in consideration of demand and continuity with plans before and after. One or more search end conditions are selected from the maximum number of searches, a target evaluation function value, and a calculation time, and values of the number of searches, an evaluation function value, and a calculation time are set, respectively. (D) Final adjustment method of startup / shutdown plan at non-sampling time Select a planning method to be used from tabu search method, local search method, genetic algorithm, simulated annealing, fixed priority method, variable priority method, and the like. Items to be set differ depending on each method. When the tabu search method is selected, conditions for applying the tabu search method, such as the adjacent state, characteristics stored in the tabu list, and the tabu list length, are set. In the case of the local search, the definition of the adjacent state is stored as in the tabu search method. When the genetic algorithm is selected, the number of individuals in the population, the crossover rate and the mutation rate are set. In the case of simulated annealing, an initial temperature and a temperature reduction rate are set. In the priority order method, only the method needs to be set, and there is no need to set a new variable. The fixed priority method uses the read priority as a generator characteristic. In the variable priority method, the priority is changed every hour in consideration of demand and continuity with plans before and after. One or more search end conditions are selected from the maximum number of searches, a target evaluation function value, and a calculation time, and values of the number of searches, an evaluation function value, and a calculation time are set, respectively. (E) Search mode setting The search mode is set by the value of the flag flg. In process 9 and process 11, the route of the process flow is changed according to the value of flg. When Flg is 0, the start / stop plan at the sampling time is fixed, and a plan is created by the final adjustment method of the start / stop plan at the non-sampling time in (d). Flg is 1
In the case of, the start / stop plan at the sampling time and the start / stop plan at the non-sampling time are changed to create an operation plan, and then in step S11, flg is changed to 0, and the start / stop plan at the sampling time is changed. The plan is created by the final adjustment method of the start / stop plan at the non-sampling time in (d) as fixed. When Flg is 2, the operation plan is created by changing the start / stop plan at the sampling time and the start / stop plan at the non-sampling time accordingly. (F) Setting of evaluation function Items of the evaluation function include power generation cost and penalty cost. Power generation costs consist of fuel costs, start-up costs, and stop costs. In operation plans, stop costs are generally zero. The penalty cost is a penalty according to the number of violations and the amount of violation of each constraint when a constraint violation occurs. The evaluation function to be used is a sum total obtained by multiplying each item by a weight coefficient. The constraint conditions include supply and demand balance, upper and lower limits of generator output, minimum continuous operation time, minimum continuous stop time, power flow restriction, power generated during the planning period (fuel consumption restriction), pumping tank water level restriction, and the like. The penalty cost increases as the weight coefficient of constraint violation increases, so when minimizing the evaluation function, it is better to determine the output so as to satisfy the constraint, and it is an index indicating the importance of the constraint . Evaluation function C
Is shown in Equation 1.

The first term is the fuel cost. Uit indicates the operating state of the generator i at the time t. It is 1 when the generator is operating and 0 when it is stopped. Pit is the output of unit i at time t. Ai, bi and ci are coefficients of the fuel cost function of unit i. T is the planned period, and N1 is the number of generators.

The second term indicates the starting cost, and SUi is the generator i
Is a function of the startup cost. The cost of this function varies depending on the length of the stop time.

The third term indicates the stoppage cost, and SPi is the generator i
Is a function of the cost of stopping.

The fourth term is a penalty for violating a constraint determined by the generator output at each time. The target constraints are the upper and lower limits of the output of the generator, the tidal current constraints, the upper and lower limits of the water level, and the like. N2 is the number of target constraints, w
2it is the weighting factor when a violation occurs. For constraints not considered, the weighting factor is zero. The function F2it depends on the output Pit and is a penalty function for violation of the constraint on the time t and the constraint i. A function that increases as the violation amount increases. For example, an absolute value of the violation amount or a value obtained by squaring the violation amount is used. It is zero when no constraint violation occurs.

The fifth term is a penalty for a violation of a constraint satisfied in the entire planning period. The target constraints are fuel consumption constraints, final water level constraints, and the like. N3 is the number of constraints, w3it is the weighting factor when a violation occurs. For constraints not considered, the weighting factor is zero. The function F3it depends on the output Pit and is a penalty function for violating the constraint on the constraint i. A function that increases as the violation amount increases. For example, an absolute value of the violation amount or a value obtained by squaring the violation amount is used. It is zero when no constraint violation occurs.

The sixth term is a penalty for violating the restriction on starting and stopping for each unit. Restrictions to be applied include prohibition of simultaneous start of generators controlled by the same central control unit, minimum continuous operation time of each generator, minimum continuous stop time parallel time difference, disconnection time difference, and the like. N4 is the number of target constraints, w4ij is a weighting factor when a violation occurs. For constraints not considered, the weighting factor is zero. The function F4ij depends on the start / stop state uit, and the generator i
Is a penalty function for violation of the constraint on. A function that increases as the violation amount increases. For example, an absolute value of the violation amount or a value obtained by squaring the violation amount is used. It is zero when no constraint violation occurs.

The seventh term is a penalty when a violation occurs for the entire unit in the time section. The target constraints are supply and demand balance, operating reserve, and the like. w5j is a weight coefficient when a violation occurs. For constraints not considered, the weighting factor is zero. The function F5t is in the start / stop state u
It is a penalty function for violating constraints on the entire generator, depending on it and the generator output Pit. A function that increases as the violation amount increases. For example, an absolute value of the violation amount or a value obtained by squaring the violation amount is used. It is zero when no constraint violation occurs.

When the above settings are completed, the process ends, and the display returns to the menu screen. (2) Reading of plan creation conditions In the process S2, the following process is performed. In the process selection menu, the plan creation conditions are displayed on the display device 20 by selecting the reading of the plan creation conditions via the input device 19, and the necessary items are input via the input device 19 and corrected, thereby obtaining the data. It is stored in the storage unit 17. The year, month, day, and planned period (number of days) set in the process S1 are read from the data storage unit 17, and the following data of the target period stored in the database 21 are read and stored in the data storage unit 17. . That is, the expected demand during the target period, the characteristics of the generator, the plan for stopping the operation of the generator and the system, and the operation constraints of the system equipment and the system. The expected demand is the expected demand at each time. The characteristics of the generator include the type of generator for each generator, the upper and lower limits of the output of the generator equipment, the coefficient of the fuel consumption function, the function of the starting cost, the function of the stopping cost, the minimum continuous operation / stop time, and the priority. Etc. are included. Generators are classified into thermal power, pumping, IPP power reception, interchangeable electricity reception, and interchangeable electricity transmission according to their types. The data read in connection with pumping are the upper and lower limits of the water level of the pumping ponds, fuel consumption constraints for LNG and coal thermal power, and multiple operation patterns for IPP, interchange power reception and interchange power transmission (generators at each time of day). output,
Load). The generator operation stop plan includes the operable state of the generator at each time, that is, the forced stop in which the generator cannot be operated, the forced operation in which the generator is always operated, and the operable state in which the operation can be freely determined. There are three states, and the upper and lower limits of the generator output at each time are read. During the period of forced stop, the upper and lower limits of the output are zero. During the period of forced operation, the upper and lower limits of the output are values according to the work.
When the output is fixed, the upper and lower limits of the output are the same. Otherwise, the upper and lower limits of the output of the generator facility are the upper and lower limits of the output. The system equipment is in a state where equipment such as transmission lines and transformers are connected. The system operation stop plan is data indicating that the system equipment is unusable. The system operation constraints include power flow constraint values of transmission lines and transformers as system facilities.

The sampling unit 12 determines the sampling time based on the sampling conditions set in the setting of the plan creation condition and the read expected demand, and the data storage unit 1
7 is stored. To create an hourly plan of the weekly plan, the forecast demand for the target period is read from the database 21 and the demand at 168 shown in FIG.
To be stored. Assuming that the sampling conditions are daily minimum demand and maximum demand, the sampling time shown in FIG. 8 and the expected demand at that time are calculated from the data at 168 and stored in the data storage unit 17.

When the reading of the plan creation conditions is completed, the process returns to the process selection menu screen. (3) Forcible stop / operation setting by the operator In the process S3, the following process is performed. In the processing selection menu, the operator sets the forced stop / operation setting using the input device 1.
9, the operator's forced stop / operation setting screen shown in FIG. 9 is displayed on the display device 20, and necessary items are input via the input device 19 and confirmed, thereby storing data. It is stored in the unit 17. The plan state 31 shown in FIG. 9 which is displayed by default displays the state at the time of automatically sampling based on the work plan. The plan state 31 in FIG. 9 shows an example in which the sampling condition is a weekly plan (7 days) of daily minimum demand (represented by N) and maximum demand (represented by P). In this example, the generators are the thermal power of G1 to G5 and the pumping of H1 (H1-G generates power, H1-P pumps water). As shown in the legend 36, the status of the thermal power generator is as follows: 0 is a stop due to work, 1 is a forced stop of the operator, 2 is a stop of the balance, 3 is operable, 4 is a balance operation, 5 is a forced operation of the operator, and 6 is a forced operation of the operator. This shows a forced operation due to work or the like. In the equation (1), the variable uit in the start / stop state becomes stop when the state shown in the legend is 2 or less, and becomes 1 when the operation is performed when the value is 4 or more. When the generator uses IPP, interchange power reception, and interchange supply, and the use of the operation pattern is set in the processing 1, this numerical value is a daily operation pattern (output, power transmission to another company) number. The relationship between the operation pattern and the output is a value read under the plan creation conditions. When using the already created plan, the plan stored in the database 21 is read by selecting the button 30 for reading the plan, and the operation state of the generator (start / stop state of the thermal power generator, pumping generator, Is output in the planning state 31. When the operating state of each generator is to be changed from the default, read planned state, the operating state is changed by the input device 19, and is stored in the data storage unit 17 by selecting the confirm button 34. Simultaneously with the selection of the confirmation button, the operating reserve (%) at the time P of the maximum demand and the reduction (%) of the time N of the minimum demand in the planned state 31 are calculated, and the result is displayed. The operating reserve is calculated by adding the upper limit of the output of three or more generators shown in Legend 36, and the reduction is calculated by adding the lower limit of the output of three or less generators shown in Legend 36 to the total demand and subtracting the value. I do. Cancel button 3
When 5 is selected, the result according to the default work plan is displayed. When the setting is completed, the enter button 34 is selected, and the data finally used is stored in the data storage unit 17. Thereafter, by selecting the close button 33, this processing is completed.

In the example at the sampling time of the example displayed in the plan state 31 of FIG. 9, the generator G1 is in a forced operation for work from 1 to 7 days. G2 is in a forced operation from the P time on the fourth day to the P time on the sixth day, and is operable at any other time in any state of balance stop or balance start. G3 and G4 can be operated at all times, and there is no work stoppage or forced operation due to work. The values of H1-P and H1-G of the pumped water indicate and input the MW of the generator output and the pump value. It does not indicate the start / stop state like thermal power. The reserve shows the calculated value of the operating reserve (%) with respect to demand. The allowance for reduction is also calculated and calculated based on the demand (%) for demand. The operator can change the plan state 31 by using the input device 19 in the operation state where the operation state is 3 (operable), 1 (operator forced stop) or 5 (operator forced operation). Can be changed to 1, 3 or 5. Further, the values of H1-P and H1-G of the pumped water can also be set. (4) Initialization of Search Number K In the process S4, a variable K of the global loop search number is set to zero and initialized. (5) Creation of an operation plan After executing the processes S1, S2, S3 and S4 and setting all the conditions for creating the plan, the creation of the operation plan is input via the input device 19 on the process selection menu screen. When selected, the data processing unit 8 creates an operation plan according to the processes S4 to S9 while cooperating with the control unit 9, the processing units 10 to 17 and the database 21. The data storage unit 17 stores, as a result of the optimal operation plan, a start / stop plan in the process of creating the optimal operation plan, a generator output, an evaluation function, the presence / absence of a constraint violation, the reason for the violation, and the like. Here, a description will be given assuming that the search termination condition is set to the maximum number of searches. In the following processing, only the operable generators other than the generator whose operation state is fixed by the operator setting (referred to as "generator with undetermined state") are determined to be the balance operation and the operation state of the balance stop. Is what you do. When the operation mode is the pattern operation, the numerical value of the start / stop of the sampling time indicates the daily pattern number, and the generator output at the non-sampling time is automatically determined from this number. .
When the operation mode is a method of determining the output at each time, a start / stop plan and a load distribution plan are created as in a general thermal power generator.

In the search mode, the processing flow of FIG. 6 is performed according to the flag flg. Here, the search mode flag flg
Is zero, the start and stop at the sampling time created in step S5 are fixed, and the optimum start and stop times in this plan are determined. The general processing procedure is as follows. Processing S1 to processing 11 are performed in numerical order. Next, processing S12 and processing S6 to processing S11 are repeated. Finally, the result is displayed in step S13, and the process ends. When the search mode flag flg is 2, the start / stop of the sampling time created in step S5 is optimized, and the optimum start time and stop time are determined for each start / stop of the sampling time. The general processing procedure is as follows. Processing S1
Processing is performed in the order of numbers from processing to processing 11. Next, processing S12 and processing S5 to processing S11 are repeated. Finally, the result is displayed in step S13, and the process ends. When the search mode flag flg is 1, the start / stop of the sampling time created in step S5 is optimized, and the optimum start time and stop time are determined for each start / stop of the sampling time. Furthermore,
The optimum start time and stop time are determined for the created start / stop plan at the optimum sampling time. The general processing procedure is as follows. Processing S1 to processing 11 are performed in numerical order. Next, processing S12 and processing S5 to processing S11 are repeated. The result is displayed in step S13. The following processing is performed on the created start / stop plan at the optimal sampling time. Processing S14, processing S15, and processing S12.
Thereafter, the processes S6 to S11 are repeated. Finally, the result is displayed in step S13, and the process ends.

The processing loop of steps S7, S8 and S9 is called a local loop, and the processing loop of steps S5 or S6 to S11 and S12 is called a global loop. (A) Initial Plan Creation of Sampling Time In the creation of the sampling time plan in the process S5, the initial plan is created according to the sampling time plan creation method using the sampling time plan creation unit 13, the data storage unit 17, and the control unit 9. create. In process S5, when the number of searches K is zero, an initial plan is created. Number of searches K
Is larger than 0, a start / stop plan is created at the sampling time of (b). FIG. 10 shows an example of the initial plan. The state of the generator is from 0 to 2 indicating stop or from 4 to 6 indicating operation, and there is no undetermined operation state where the state is 3. (A-1) Tabu Search Method If the sampling time planning method is the tabu search method, generators to be operated in descending order of the priority of the generator facilities are determined until the operation reserve capacity and the supply and demand balance are satisfied. Since the expected demand and the priority are fixed, one start / stop plan can be uniquely created. (A-2) Local Search Method The local search method is the same as the tabu search method in preparing the initial plan. (A-3) Simulated Annealing Simulated annealing has the same initial creation method as the tabu search method. (A-4) Genetic Algorithm In the genetic algorithm, a balance operation or a balance stop of a generator whose state is undetermined is determined for each sampling time by a random number, and an initial plan for the number of individuals is created. The initial planned number to be created is the number of individuals set in the process S1. For example,
When the random number is less than 0.5, the balance is stopped. Otherwise, the balance is activated. However, the state can be determined from the preceding and following times in consideration of the continuity of the state at the N time and the P time. Since the start / stop plan created by the priority method is considered to be good as an initial solution, prepare this solution for the number of individuals and then change the start / stop of one or more places by random numbers for each solution. May be applied to create an initial individual. (A-5) Fixed Priority Method Initial Plan The fixed priority method also has the same method of creating an initial plan as the tabu search method. (A-6) Variable Priority Method The variable priority method has the same method of creating an initial plan as the tabu search method. (B) Creation of start / stop plan at sampling time In the processing S5, when the number of searches K is larger than 0, (b)
Create a start / stop plan at the sampling time. In step S5, the sampling time plan creation unit 14, the data storage unit and the control unit 9 create a start / stop plan at the sampling time based on the current solution, and store the created plan in the data storage unit 17. According to the selected method, the current solution is changed and a new start / stop plan is created. (B-1) Tabu Search Method The case where the tabu search method is selected will be described. The current solution is a start / stop plan shown in FIG. 11, and the adjacent state is a start / stop plan that can be created by changing the start / stop at the sampling time. For example, by changing the N time on the fifth day of the generator G5 from stop to operation with respect to the current solution in FIG. 11, the adjacent state shown in FIG. 12 can be created. However, here, the generator that operates at the N time is operated at the P time before and after the N time, so the P time on the fourth day is also changed from the stop to the operation. Further, by changing the P time on the fifth day of the generator G5 from the operation to the stop with respect to the current solution in FIG. 11, the adjacent state shown in FIG. 13 can be created. However, when stopping at the P time, the N times before and after that time are also changed to the stop.
Here, the items to be stored in the taboo list are the name of the generator whose operation state has been changed, the change time, the change time, and the operation state before and after the change. Basically, the adjacent state is created by changing the operation state of the generator at the sampling time, but is excluded when the following conditions are satisfied. When the changed state is on the taboo list, when the start / stop change target is a generator that is not in the operable state, when the time is changed, and when the start / stop state before and after the start / stop change is changed, a contradiction occurs (eg, forced The driving generator cannot be operated). (B-2) Local Search Method The local search method uses the same adjacent state as the tabu search method based on the current solution. In the case of local search, there are no adjacent states excluded by taboos. (B-3) Simulated Annealing In the case of simulated annealing, a generator and a time at which start / stop is changed using random numbers based on the current solution are determined, and a start / stop plan is changed. (B-4) Genetic Algorithm When the genetic algorithm is selected, two individuals are selected from the individual population of the current solution by the roulette strategy. 14 and 1 show examples of selected individuals (start-stop plan).
It is shown in FIG. A new individual is created by crossing the two selected individuals. The crossover point is determined by a random number. FIG.
In the example of FIG. 4 and FIG. 15, new points shown in FIG. 16 and FIG. 17 are created by crossing the generators G3 and G4.
In the examples of FIGS. 14, 15, 16 and 17, the crossover point is determined by the generator number, but the crossover point may be determined by time. When the first random number is smaller than the mutation probability for the newly created individual, the generator and the time at which the start / stop is changed are further determined using the random numbers, and the start / stop plan is changed. If the first random number is greater than the mutation probability, no mutation occurs. (B-5) Fixed Priority Method Since the priority order is used to create the start / stop plan at the sampling time, the start / stop plan is created in the same manner as (a-5). The same start / stop plan as in (a-5) is created. (B-6) Variable priority method Since the priority order is used for creating the start / stop plan at the sampling time, the start / stop plan is created in the same manner as (a-6). The same start / stop plan as in (a-6) is created.

In the above manner, the adjacent state or the newly created start / stop plan is stored in the data storage unit 17. (C) Initialization of the Number of Searches G In process S6, a variable G of the number of local searches is set to zero and initialized. (D) Non-sampling time plan creation method In step S7, the non-sampling time start / stop plan corresponding to each sampling time start / stop plan is performed by the non-sampling time plan creation unit 14, the data storage unit 17, and the control unit 9. And the load distribution plan of (e) is created. The non-sampling time plan determines the start and stop times with the sampling time plan fixed.

It is necessary to determine the start time when the vehicle is stopped at the N time and the vehicle is operating at the P time on the same day. It is necessary to determine the stop time when the vehicle is stopped at the N time and the vehicle is operating at the P time on the previous day. For other combinations of states, the running and stopping states are constant between N time and P time on the same day, and between N time and P time on the previous day, so it is not necessary to determine the starting time and stopping time as variables. Thereby, the search range can be narrowed.

When the local loop variable G is zero, N
Assuming that the start and stop plan at the time and the P time is fixed, the start time of the generator started at the N time to the P time is determined by the fixed priority method or the variable priority method. Similarly, the stop time of the generator stopped at the time P to the time N is determined by the fixed priority method or the variable priority method. As a result, an initial start / stop plan for all times can be created. Clear the taboo list data when creating a plan for non-sampling times.

When the variable G is larger than 0, a new start / stop plan at a non-sampling time is created from a start / stop plan at a non-sampling time according to a method selected as described below. The operation plan created when the variable G is 0 is set as the optimum operation plan in the search for the local loop, and the start / stop plan, the load distribution plan, and the evaluation function value are stored.

In this processing, a start / stop plan at one non-sampling time is determined by searching for a start / stop time at one sampling time. For example, the start-stop plan at the sampling time is as many as the number of individuals in the genetic algorithm, but the purpose is to determine the optimum start time and stop time in each start-stop plan at the sampling time.
That is, no crossover or mutation occurs between the start / stop plans at different sampling times.
The start / stop of the sampling time does not change. (D-1) Tabu search method When the local loop variable G is zero, the start / stop plan at N time and P time is fixed, and the generator started from N time to P time is fixed priority method or variable. The start time is determined by the priority method. Similarly, the stop time of the generator stopped at the time P to the time N is determined by the fixed priority method or the variable priority method. As a result, an initial start / stop plan for all times can be created. Clear the taboo list data when creating a plan for non-sampling times.

When the local loop variable G is larger than zero, the start / stop plan at the non-sampling time corresponding to the start / stop plan at the sampling time has already been made.
The adjacent state is created by changing the start / stop time based on the start / stop plan at the non-sampling time. The adjacent state is created by setting the start time earlier or later by a unit time or the stop time earlier or later by a unit time. However, the start / stop plan at the N time and the P time is fixed. Also, states in the taboo list are excluded from adjacent states. (D-2) Local search method When the local loop variable G is zero, an initial start / stop plan for all times is created by the same method as the tabu search method.

When the local loop variable G is larger than zero, an adjacent state is created by the same method as the tabu search method. However, in the case of the local search, there is no adjacent state excluded by the taboo. (D-3) Simulated annealing When the local loop variable G is zero, an initial start / stop plan for all times is created by the same method as the tabu search method.

When the local loop variable G is greater than zero, the start / stop plan at the non-sampling time corresponding to the start / stop plan at the sampling time has already been made.
By changing the start / stop time based on the start / stop plan at the non-sampling time, a new start / stop plan is created.
Create one. Generator to change by random number, the day,
Decide the start time or stop time and advance or delay.
However, the start / stop plan at the N time and the P time is fixed.

In the case of the simulated annealing, the generator and the time to change the start / stop are determined using random numbers based on the current solution, and the start / stop plan is changed. (D-4) Genetic Algorithm When the local loop variable G is zero, an initial start / stop plan for all times is created by the same method as the tabu search method. This plan is prepared for the number of individuals, and for each individual, the generator to be changed, the number of days, the start time or the stop time, and the earlier or later are determined by random numbers. However, the start / stop plan at the N time and the P time is fixed. As a result, an initial start / stop plan for all times for the number of individuals can be created.

When the local loop variable G is larger than zero, the start / stop plan at the non-sampling time corresponding to the start / stop plan at the sampling time has already been made.
Based on the start / stop plan at the non-sampling time, new start / stop plans are created for the number of individuals by crossover and mutation. For the crossover, a method of dividing the two plans in the time axis direction or a method of dividing the two plans in the direction of the generator number is applied. Mutation is caused by the random number determining which generator to change, the day of the week, the start or stop time, and the earlier or later. However, the start / stop plan at the N time and the P time is fixed. (D-5) Fixed priority method When the fixed priority method is selected, the generator of the balance operation is operated so as to satisfy the operation reserve according to the priority stored in the data storage unit 17 as the generator characteristic. decide. Generators that are operable and did not perform balance operation are stopped in balance. FIG. 10 shows an example of a start / stop plan at a sampling time of seven days (weeks). Here, it is assumed that the priorities are higher in the order of G1 to G5. At this time, G2 and G5 are activated on the seventh day of the last day, G2 is activated first, and then G5 is activated. The start time is uniquely determined by the expected demand and reserve capacity. As shown in FIG. 8, the minimum demand time N and the maximum demand time P on the seventh day are 149 hours and 160 hours, respectively. FIG. 18 shows an example of the start / stop plan at the time 168, and FIG. 19 shows an example of the operation state between the minimum demand time 149 and the maximum demand time 160 on the seventh day. 149 o'clock is 5 o'clock on the seventh day,
16:00 is 16:00 on the seventh day. In this example, G2 having a higher priority starts the balance at 153, and then G5
Is activated at 154. For the created start / stop plan at all times, the constraint conditions that can be checked according to the start / stop state are checked, and if the violation is violated, the start / stop time is changed to eliminate the violation as much as possible. Here, the constraint conditions to be checked are: simultaneous start prohibition at the same center, parallel time difference, paralleling time difference, minimum continuous stop time, minimum continuous operation time, and the like. (D-6) Variable Priority Method In the variable priority method, the priority of the planned time is calculated by using the start / stop plan at the time when the start / stop plan next to the planned start / stop plan is created. The fuel cost for the generator output when the load is distributed, that is, the order in which the fuel cost unit price is small is determined. The variable priority is calculated for each time, and the priority is used to create a start / stop plan at the non-sampling time in the same manner as the fixed priority. For the created start / stop plan at all times, the constraint conditions that can be checked according to the start / stop state are checked, and if the violation is violated, the start / stop time is changed to eliminate the violation as much as possible. Here, the constraint conditions to be checked are: simultaneous start prohibition at the same center, parallel time difference, paralleling time difference, minimum continuous stop time, minimum continuous operation time, and the like.

As described above, a start / stop plan can be created. (E) Creation of load distribution plan When the generator is IPP, interchange power reception, interchange power transmission, the operation state and output at each time are automatically determined uniquely from the operation state at N time and P time. When the generator is pumping water, the generator output is set to an output necessary only when the supply power is insufficient or the operating reserve is insufficient. For the pump, the load is added by the shortage of the lowering allowance. However, the output is below the upper limit of the output and the pump is below the upper limit of the pump. This is because the upper and lower limits of the output of the generator are required constraints.

With respect to the start / stop plan created at all times, a load distribution plan is created by an equal lambda method, a linear programming method or a quadratic programming method. FIG. 20 shows an example of the output of each generator when the load is distributed. Check the constraints determined from the generator output when the load is distributed. If there is a constraint violation, it is adjusted by increasing or decreasing the generator output to eliminate the violation or reduce the amount of violation as much as possible. The constraint conditions checked here are fuel consumption constraints, upper and lower limits of the pumping tank water level, target water levels, and the like. In addition, when the output is adjusted for pumping, the amount of pumping power and pumping pump at each time is determined in consideration of economy. In other words, when it is necessary to increase the amount of pumped water generation in order to reach the target water level of the pumping basin at the last time, the pumped water generation amount at the time when the incremental fuel cost (fuel cost required to change the unit output) is large The target water level will be achieved by increasing it. When the amount of pumped-storage power generation is increased, the incremental fuel cost at that time decreases. Further, when it is necessary to increase the number of pumps, the pump is increased at a time when the incremental fuel cost is small so that the target water level is obtained. In order to reduce the increasing cost, pumping power generation and pumping pumps at the time when the incremental fuel cost is the largest, reduce the thermal fuel cost,
The pump is operated at the time when the incremental fuel cost is the lowest to minimize the increase in thermal fuel cost. From the above,
Generator output is determined, including planning for non-sampling times. (F) Calculation of Evaluation Function of Plan of Local Loop In step S8, the evaluation unit 15 of the plan, the data storage unit 17, and the control unit 9 calculate the evaluation function of each start-stop plan, and calculate the value of the evaluation function. Then, one or more start / stop plans corresponding to the number of individuals are selected according to the plan creation method. The plan evaluation unit 15 calculates the evaluation function by starting and stopping the generator, substituting the generator output and the constraint (constraint value) into the equation (1). (F-1) Tabu Search Method In the tabu search method, one having the best evaluation function is selected from the start / stop plans of the adjacent states. This is stored in the data storage unit 17. Further, the taboo list stores the generator name and the time at which the start / stop state has changed at the non-sampling time. If the taboo list contains more data than the taboo list length, delete the oldest data. (F-2) Local Search Method In the local search method, one with the best evaluation function is selected from the start / stop plans of adjacent states. This is stored in the data storage unit 17. (F-3) Simulated Annealing In the simulated annealing, a newly created start / stop plan is selected whenever the evaluation function value of the newly created start / stop plan is better than the original start / stop plan.
When the value of the random number is equal to or less than the value determined by the temperature and the number of searches, even if the evaluation function value of the newly created start / stop plan is worse than the original start / stop plan, the newly created start / stop plan is selected. . (F-4) Genetic Algorithm In the genetic algorithm, the next generation of individuals (by roulette strategy and elite strategy) is selected so that the created start / stop plan and the original start / stop plan with good evaluation function values are easily selected. Start and stop plan) for each individual. (F-5) Fixed priority method Since there is only one start / stop plan, the created start / stop plan is selected as the next start / stop plan. (F-6) Variable priority method Since there is only one start / stop plan, the created start / stop plan is selected as the next start / stop plan.

The start / stop plan, evaluation function values, and the like selected as described above are stored in the data storage unit 17. Also, for each start / stop plan at the sampling time, the plan with the best evaluation function in the G-th operation plan of the number of local loop searches is stored. Furthermore, comparing this operation plan with the current optimum operation plan in the local loop search, if the evaluation function is good, the optimum solution in the local loop search, its start / stop plan, and the load distribution plan are stored in the data storage unit. 17 is stored. (G) Increasing the Number of Searches of the Local Loop In the process S9, after increasing the count of the number G of searches of the local loop by one, the process returns to the process S7 if it is smaller than the maximum number, and moves to the process S10 if it is smaller. (H) Selection of Global Loop Operation Plan In the process S10, if the optimal operation plan created by the local loop search is better than the optimal operation plan in the global loop search, if the evaluation function is better, the global loop Store as the optimal operation plan for search. Stored are a start / stop plan, a load distribution plan, an evaluation function value, and the like. Further, the optimum operation plan created by searching for the local loop is stored in the data storage unit 17 as the current solution. FIG. 21 shows an example of an evaluation function for each search in the global loop, and these data are stored in the data storage unit 17. Similarly, for the optimal solution, see FIG.
As shown in (1), the number of times of search to obtain the optimal solution, and the evaluation function and the like at that time are stored in the database 21. (I) Increasing the Number of Searches for the Global Loop In the process S11, after increasing the count of the number of times K for searching the global loop by one, if the number is smaller than the maximum number, the process returns to the process S11; (J) Whether or not to fix the start / stop plan at the sampling time In the process S14, if the value of the flag flg in the search mode is 1, the process proceeds to the process S15; otherwise, all processes are ended. (K) Condition Setting for Neighborhood Search with Fixed Start / Stop Plan at Sampling Time In step S15, the value of the search mode flag flg is changed to 0, and the number of global loops K is initialized to zero.
Further, the maximum number of searches Lmax when the flag flg of the search mode is 0 is set. (L) Judgment as to whether the start / stop plan at the sampling time is a fixed condition or not Step S12 is a step S12 if the search mode flag flg is 0.
The process proceeds to S6 if not. (6) Display of Result In the process S13, the display data creation unit 16, the data storage unit 17, and the control unit 9 display the calculation result of the display device 20. Further, the calculation results stored in the data storage unit 17 and the data being calculated are stored in the database 21.
With the file name. The contents displayed on the display device are an evaluation function, a power generation cost, a fuel cost, a start-up cost, and a penalty cost in a global loop, which are displayed in a table or a graph. In addition, a start / stop plan at all times of the obtained optimal solution, a start / stop plan at sampling time, and a generator output are displayed. 24 shows an example of a graph of the evaluation function displayed in FIG. Also, examples of the generator state, the generator output, the history of the evaluation function, and the evaluation function of the optimal solution displayed on the display device are shown in FIGS. 18, 20, 21, and 22. FIG. 9 also shows an example of the display screen of the start / stop plan at the sampling time. Here, the operating state is indicated by a numeral, but the operating state may be indicated only by the numeral color, the background color, or the display color. As for the classification of the operating state, the plan result can be classified into two types of operation and stop, and three types including an undetermined operation state in the initial state.

In the above description, an example of creating an operation plan at 168 times a week has been described. However, the present invention can be applied to a plan for 24 times a day, a month, and a year. Although the sampling time is set to the minimum demand N time and the maximum demand P time, only one of them may be used, or a time arbitrarily designated may be used.

The program for the processing flow for creating the operation plan shown in FIG. 6 is stored in a floppy disk or CD-RO.
It is stored in a medium such as M. A specific device for executing the above processing program is a general-purpose computer system including an input / output device, a display device, a reading device, and the like in addition to a dedicated device configured by the device shown in FIG. It can also be realized by a processing program. When such a general-purpose computer system is implemented by adding a processing program, the processing program is recorded on a storage medium, distributed, interpolated, mounted, read by a reading device provided in the computer main body, and taken into the computer main body. It is. In the case where the processing program delivered via the communication network is fetched and realized by the input unit, the fetched processing program is stored and stored in a storage medium such as a magnetic disk so that it can be used repeatedly.

As described above, by changing the priorities and creating the operation plan, the start and stop priorities can be optimized, and the start / stop plan and the generator with the minimum objective function can be achieved. Output can be calculated at high speed.

[0059]

As described above, according to the present invention, planning is performed by a combination optimization method such as a tabu search method, a local search, a simulated annealing, a genetic algorithm, or a fixed priority method or a variable priority method. A start / stop plan for a part of the time is created, the start / stop plan is fixed, the start / stop time is determined, and a load distribution plan is created for the created start / stop plan. Among these created operation plans, the operation plan with the smallest evaluation function, that is, the operation plan that satisfies the constraints and has the lowest power generation cost can be automatically calculated.

[Brief description of the drawings]

FIG. 1 is a diagram showing an apparatus configuration of the present invention.

FIG. 2 is a diagram showing a curve of daily power demand.

FIG. 3 is an example of a graph of an evaluation function for an operation plan of an operation plan problem.

FIG. 4 is a diagram showing a hardware configuration of the present invention.

FIG. 5 is a diagram showing an internal configuration of a data storage unit.

FIG. 6 is a diagram showing a processing flow of the present invention.

FIG. 7 is a diagram showing an example of expected demand at 168;

FIG. 8 is a diagram showing an example of expected demand at a sampling time.

FIG. 9 is a diagram showing an operator's forced stop / operation setting screen.

FIG. 10 is a diagram showing an example of a start / stop plan at two times a day.

FIG. 11 is a diagram showing an example of a current solution.

FIG. 12 is a diagram illustrating an example of an adjacent state that can be created by changing a stop to an operation.

FIG. 13 is a diagram illustrating an example of an adjacent state that can be created by changing operation to stop.

FIG. 14 is a diagram illustrating an example of a current solution of an intersection object.

FIG. 15 is a diagram illustrating an example of a current solution of an intersection object.

FIG. 16 is a diagram showing an example of a plan created by crossover.

FIG. 17 is a diagram showing an example of a plan created by crossover.

FIG. 18 is a diagram showing a start / stop state of the generator.

FIG. 19 is a diagram showing a start / stop state of the generator.

FIG. 20 is a diagram showing a generator output.

FIG. 21 is a diagram showing values of an evaluation function and the like for each search.

FIG. 22 is a diagram showing values of an evaluation function and the like of an optimal solution.

FIG. 23 is a diagram showing a graph of a change in an evaluation function.

[Explanation of symbols]

1 ... Operation plan creation device, 2 ... CPU, 3 ... Main storage device,
4 input / output device, 5 external storage device, 6 distribution system, 7 reception system, 9 control unit, 10 condition reading unit, 11 condition setting unit, 12 sampling unit, 13 sampling time planning Creation unit, 14: Non-sampling time plan creation unit, 15: Plan evaluation unit, 1
6 ... display data creation unit, 17 ... data storage unit, 19 ... input device, 20 ... display device, 21 ... database, 22 ...
Reader.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-10-56734 (JP, A) JP-A-5-15071 (JP, A) JP-A 2-136034 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) H02J 3/00-5/00

Claims (12)

(57) [Claims] 1. A generator operation plan creation device for creating an operation plan of a generator based on expected demand and operation constraints of an operation plan, wherein a generator characteristic reading / or setting means and / or a generator operation plan Means for sampling a part of the time or / and demand from the expected demand at each time, and means for starting and stopping the generator with respect to the time or / and the demand sampled by the means; Evaluating the means for generating and stopping the generator at an undetermined time or / and demand, and fixing the generated generator operation plan with the fixed condition of fixing the start / stop of the generator at the specified time or / and demand An operation plan creation device for a generator, comprising: 2. The generator operation plan creating apparatus according to claim 1, wherein the sampling means includes at least a local maximum value or a local minimum value of the demand, or a generation time of the local maximum value or the local minimum value, or a time specified by the operator. A generator operation plan creating apparatus characterized in that sampling is performed based on time or a sampling time interval specified by operation. 3. The generator operation planning device according to claim 1, wherein the sampling means samples at least a maximum demand or a maximum demand time for a certain period such as one day, week, season, or month. A generator operation plan creation device characterized by the above-mentioned. 4. A generator operation planning device according to claim 1, wherein the means for generating the start and stop of the generator at the time or / and the demand sampled includes a local search, a tab search method, and a genetic algorithm. An operation plan creation device for a generator, wherein metaheuristics such as simulated annealing or a priority method is used. 5. The generator operation planning device according to claim 1, wherein the power generation is performed at an undetermined time or / and demand under the condition that the start / stop is generated at the time or / and the demand is fixed. A generator operation plan creation device characterized by using a meta-heuristic or a priority method such as a local search, a tab search method, a genetic algorithm, and simulated annealing as a means for creating start and stop of a machine. 6. The generator operation planning device according to claim 4, wherein the adjacent state when using local search, tab search method or simulated annealing is at least one time or at least one generator. Changing the start / stop state of the generator in response to demand, that is, changing the generator to stop when the generator is operating at the time or the demand; A generator operation plan creation device characterized in that, when stopped, the operation is changed to start. 7. The generator operation plan creating apparatus according to claim 4, wherein when using a genetic algorithm, the generator structure has a genetic structure representing an operation state at a sampling time. 8. The generator operation plan creating device according to claim 1, wherein the evaluation function used in the means for evaluating the generated generator operation plan is based on the term relating to the generator operation cost and the violation of the operation constraint. An operation plan creation device for a generator, which is a sum of penalty terms. 9. A generator operation plan creating apparatus according to claim 1, wherein a screen for displaying a generator start / stop state at a sampling time or a generator start / stop state at a sampling time is forcibly operated or stopped. A generator operation plan creation device, characterized in that it has a screen for fixing the generator. 10. The generator operation planning device according to claim 1, wherein the balance is stopped, the balance is started, the operation is forcibly stopped, the operation is forcibly stopped by a test, the start / stop is undetermined, the operator is forcibly stopped, and the operator is forcibly stopped. An operation plan creation device for a generator, wherein at least two types of starting and stopping or three types of starting, stopping and undetermined starting and stopping are displayed in different colors with respect to the state of the generator in forced operation. 11. A recording medium for storing a program for creating a generator operation plan by a computer, wherein the generator operation plan creation device according to claim 1 stores a program for creating a generator operation plan. A function of creating a part of time representing the characteristic of demand from the time series data of demand forecast, a function of creating a start / stop of the generator at the time determined by the function, and a function of creating a start / stop of the time already created A memory storing a generator operation plan program for executing a function of generating and starting a generator at an undetermined time and a function of evaluating the generated generator operation plan with the generator start and stop fixed. Medium. 12. A generator operation plan creation method for creating an operation plan of a generator based on demand data and operation constraints of the generator, comprising: Sampling at least one of them; creating a generator operation plan at that time based on the sampled demand and time; and fixing the start or stop of the generator at the planned sampling time. Creating an operation plan of the generator at a time other than the time included in the constraint condition.