JP2005304118A - Controller and control method of distributed energy system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To create and revise the optimal operation schedule of a distributed energy system efficiently in real time with small computational complexity. <P>SOLUTION: An optimal operation scheduling section 22 repeats processing for evaluating a set or revised operation schedule and storing the operation schedule and an evaluation value in an optimal search information storing section 23 until predetermined termination conditions are satisfied while searching and evaluating the operation schedule, and selects an optimal operation schedule when the termination conditions are satisfied while shifting a schedule start time by a current time. An optimal search rule operating section 24 provides an optimal operation schedule search rule to optimal operation scheduling section 22 from the information in the optimal search information storing section 23 or limits a search range. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a control device and a control method for a distributed energy system having a plurality of energy generation devices and a plurality of energy storage devices and connected to an electric power system and an energy load.

  As a method of performing low-cost operation control by effectively using the energy of a plurality of distributed power supply systems, there is a distributed energy community system described in Patent Document 1 and a control method thereof. This is because the control center receives data on the power generation amount of the fuel cell, the energy storage amount of the storage battery, and the power consumption amount of the load from the control device of the distributed power source system via the communication line, and the generated power is sent to each distributed power source system. This is a system for commanding a value and a received / transmitted power value and complementarily controlling power supply and demand through a power line between a plurality of distributed power systems having different daily load characteristics of power demand.

Further, as a method for creating a power generation plan, there are metaheuristic methods such as tabu search and genetic algorithm as described in Non-Patent Document 1, in addition to the mathematical programming method that has been used conventionally. In recent years, with the spread of distributed power sources, etc., the need to consider nonlinear and discontinuous characteristics and constraints has increased, and the metaheuristic method is a high-speed approximate solution of the global optimal solution at a relatively high speed regardless of the function form. Therefore, it has come to be used for power generation planning.
JP 2002-44870 A IEEJ Technical Report No. 923 "Metaheuristics Application Technology to Power Systems" p33-p34

  The control method of the above-described conventional power supply and demand control system increases the amount of calculation required to calculate a low-cost optimal operation plan as the number of distributed power sources increases, and also includes the constraint conditions of energy storage devices such as storage batteries. Since the search for the optimal operation plan becomes difficult as the complexity of the system becomes complicated, there is a problem that the accuracy as the optimal solution of the obtained operation plan is deteriorated.

  Targeting a community with a large number of controllable distributed power sources, and when there is a large fluctuation in demand power, etc., it is necessary to perform control while creating an optimal operation schedule in real time. Therefore, it is necessary to devise in order to always obtain a global optimum solution stably.

  An object of the present invention is to provide a control device and a control method for efficiently creating and correcting an optimum operation plan of a distributed energy system in real time with a small amount of calculation.

In order to achieve the above object, a control device for a distributed energy system according to the present invention provides:
An optimal search information storage unit;
While evaluating the set or changed operation plan and storing the operation plan and the evaluation value in the optimum search information storage unit while searching and evaluating the operation plan until a predetermined end condition is satisfied Repeatedly, when the end condition is satisfied, an optimum operation plan creation unit that selects an optimum operation plan while changing the start time of the plan according to the current time,
An optimum search rule calculation unit is provided that gives the optimum operation plan creation unit a guideline for optimum operation plan search or limits the search range from the information in the optimum search information storage unit.

  The present invention performs rescheduling while changing the start time of the plan according to the current time when creating the optimum operation plan of the distributed energy system, and saves information of the optimum search process when creating and correcting the plan. In addition, it is characterized in that it is effectively used in the subsequent optimum search.

  In addition, by clearing the optimal search information at a predetermined timing or performing an optimal search that does not use this information at predetermined time intervals, it corresponds to the latest information such as demand and power generation prediction that fluctuates depending on the day of the week and season The solution can be searched, and it can be prevented that the solution remains in the local solution.

  Re-scheduling is repeated at predetermined time intervals while changing the start time of the optimal operation plan according to the current time, and the information on the optimal search process when this optimal operation plan is created and corrected is saved. By effectively using the optimum search, it is possible to learn a law for efficiently obtaining an optimum solution corresponding to the purpose such as cost minimization.

  Therefore, even when scheduling a large number of energy generators and energy storage devices in a short time, an output command schedule that is a high-quality solution can be obtained stably and quickly, and an optimum operation plan can be efficiently calculated with a small amount of calculation. Can be created.

  Next, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows the configuration of a distributed energy system and its controller according to an embodiment of the present invention.

Energy systems 1 ₁ , 1 ₂ ,..., 1 _n (n is 2 or more) are installed in each community (district, building, etc.), and solar cells 11, fuel cells 12, energy storage devices as energy generation devices. The storage battery 13 and the heat storage device 14 are included, and the power load 15 and the heat load 16 are included as energy loads. The solar cell 11, the fuel cell 12, the storage battery 13, and the power load 15 are connected to the power system 2 that is a commercial power source through the power line 3, and the fuel cell 12 is connected to the heat storage device 14 such as a hot water tank and the heat load through the heat pipe 17. 16 is connected. The energy generator and the energy storage device are not limited to those described above.

The control device 4 is connected to each energy system 1 ₁ to 1 _n via the communication line 5 and measures the load power of the power load 15 connected to the fuel cell 12 that is an energy generation device and the storage battery 13 that is an energy storage device. On the other hand, information on weather forecasts, energy prices, events, and the like is obtained from the Internet and the like, and the amount of power generated by the solar cell 11, the demand for the power load 15, and the demand for hot water supply as the heat load 16 are predicted using these pieces of information. Then, using this prediction result, an optimum operation plan for the fuel cell 12 and the storage battery 13, that is, a command schedule for power generation of the fuel cell 12 and charge / discharge of the storage battery 13 is created, and the power generation amount and the charge / discharge amount are controlled. This command schedule is created, for example, in units of one day, the current time, or 24 hours after the predetermined time has elapsed, and is updated at predetermined time intervals, for example, every hour.

The control device 4 includes a communication unit 21, an optimum operation plan creation unit 22, an optimum search information storage unit 23, and an optimum search law calculation unit 24 in order to perform the above functions. The communication unit 21 acquires information on weather forecasts, energy prices, events, and the like from the Internet, passes them to the optimum operation plan creation unit 22, and sends the command schedule created by the optimum operation plan creation unit 22 to each of the communication schedules 5 via the communication line 5. It transmits to the fuel cell 12 and the storage battery 13 of the energy system 1 ₁ to 1 _n .

  The optimum operation plan search unit 22 predicts power and heat load demand, evaluates the set or changed operation plan based on the energy cost of the entire system, and stores the operation plan and the evaluation value in the optimum search information. The processing to be stored in the unit 23 is repeated while changing the operation plan until a predetermined end condition is satisfied. When the end condition is satisfied, an optimum operation plan is selected, and each fuel cell 12 and storage battery 13 is selected. Perform output processing.

  The optimum search law calculation unit 24 gives the optimum operation plan creation unit 22 a guide for optimum operation plan search or limits the search range from the information in the optimum search information storage unit 23.

  FIG. 2 shows the overall processing flow in the control device 4 in FIG.

First, a weather forecast, electric power purchase price information, etc. are received by the communication part 21 (step 101). Next, based on these pieces of information, the power, the demand for heat load, and the power generation amount of the solar cell 11 from the current next time zone to 24 hours ahead are predicted (step 102). Next, a command schedule which is an initial solution is set (step 103). This initial solution is created using the command schedule obtained as the optimal solution in the previous search. Next, the set of solutions for the command schedule is evaluated (step 104). In the present embodiment, the energy cost is an objective function, and the schedule that minimizes the value (evaluation value) of the objective function is the optimal solution. The energy cost includes an electric power cost and a fuel cost. The power cost is calculated from the difference between the supply power when the energy generating device and the energy storage device are operated according to the schedule and the demand predicted power of the energy load, and the price of the grid power. The fuel cost is calculated from the fuel flow rate relative to the generated power in the generated power target pattern by modeling the efficiency characteristics, start-up characteristics, response characteristics, etc. of the fuel cell 12. The input / output characteristics are modeled in consideration of charging / discharging loss of the storage battery 13, heat dissipation loss of the heat storage device 14, etc., and the remaining capacity is calculated to balance the input / output amount in the schedule period of the storage battery 13 and hot water tank. For this purpose, the difference between the start and end of the schedule of the amount of stored electricity and the amount of hot water is added to the objective function as a penalty function of the charge cost and hot water storage cost. Further, overcharge, overdischarge, etc. of the storage battery 13 may be treated as a constraint condition or added as a penalty function in the same manner. Based on the evaluation result, the optimum search information is stored in the optimum search information storage unit 23 or replaced with the existing information in the optimum search information storage unit 23 (step 105). The optimal search information is information on the evaluation value (objective function value) of the solution (output command schedule) evaluated in the optimal search process, and is stored as a table database (to be described later) in addition to the solution and the evaluation value. You can also keep it. Subsequently, a schedule to be evaluated next is determined using an optimization method (step 107). As an optimization method, a metaheuristic method such as tabu search or a genetic algorithm can be applied. The above steps 104, 105, and 107 are repeated, and according to a schedule that is a solution of the best evaluation value obtained until the optimum search end condition is satisfied according to the calculation time and the number of repetitions (steps 106 and 108), each energy system 1 The outputs of the ₁ to 1 _n fuel cells 12 and the storage battery 13 are commanded (step 109). Once the optimum search is completed, the optimum search information in the optimum search information storage unit 23 is partially erased or cleared (step 110). Next, if the prediction information is corrected, the correction is made and the process returns to step 103 (steps 111 and 112).

  FIG. 3 shows an example of storing optimal search information (step 105) in FIG. If there is a vacancy in the optimum search information storage unit 23, the evaluated neighborhood solution (output command schedule) and its evaluation value are stored (steps 201 and 202). The optimum search information storage unit 23 can also use information in previous searches for a while by determining a region for each search. If there is no space in the optimum search information storage unit 23, and there is a solution worse than the evaluation value of the evaluated solution in the allocation area of the current search in the optimum search information storage unit 23, the most evaluated value The bad solution and its evaluation value are replaced with the solution and the evaluation value evaluated here (steps 203 and 204). Since demand and power generation prediction may be changed in this way, there is a meaning that schedules of some top patterns having high evaluation values for each search are stored in the allocation area of the optimum search information storage unit 23. is there.

  FIG. 4 shows an example of calculating the search law from the optimum search information (step 113). As described above, for all combinations of the optimum scheduling time range and all the fuel cells and storage batteries to be controlled (steps 303 to 306), the solutions accumulated in the optimum search information storage unit 23 are in each time zone. The ratio of the output command values within the respective ranges is calculated (step 301), and the result is registered / updated in the table database provided in the optimum search information storage unit 23 (step 302). Here, “each range” is an output range in which the output command value is divided into several stages. For example, if there are 1000 solutions accumulated in the optimum search information storage unit 23 and there is 100 schedules (solutions) in which the output command value of the fuel cell i in the time zone t is 95 to 100% of the rated output, “time The ratio of the output command value of the fuel cell i in the band t within the range of 95 to 100% is 100/1000 = 0.1.

  FIG. 5 shows a method for narrowing the search range when the tabu search is applied to the optimization method as an example of effectively utilizing the optimum search information in FIG. In tabu search, after evaluating a solution, the schedule with the highest evaluation value among the previously evaluated solutions is set as the start point of the next neighborhood, and the attribute of adjustment to this solution is registered in the tabu list from the previous time (step 401). Alternatively, in the case of another optimization method, it may be determined probabilistically so that a solution with a better evaluation value is easily selected. Moreover, the said vicinity is the schedule which increased / decreased command values, such as the electric power generation amount in some time slot | zones, and charge / discharge amount, only 1 step, for example. The tabu list is a function provided to escape from the local solution, and is a memory for avoiding returning to the already passed solution. Here, the increase / decrease direction of the fuel cell 12 and the storage battery 13 in the time zone adjusted as the vicinity may be stored.

  Here, the neighborhood solutions to be evaluated are narrowed down among a large number of neighborhood solutions (steps 403 to 411). First of course, the fuel cell or storage battery is limited to those that can output due to the maximum or minimum output of the fuel cell or the storage battery, restrictions on output change, and the like (step 104). Next, in the case of an example of a taboo list, if it is a vicinity in the prohibited adjustment direction registered in the taboo list, it is excluded (step 405). Finally, the search is narrowed down by the optimum search rule using the database of the optimum search information storage unit 23 which is a feature of the present invention (step 406). Of the solutions stored in the optimum search information storage unit 23, if the ratio of the fuel cell and storage battery output command values within each range within the respective ranges is less than or equal to a predetermined value, the next solution candidate By excluding from a certain neighborhood, the neighborhood solution to be evaluated can be narrowed down, so that the amount of calculation can be greatly reduced. Alternatively, random numbers or the like may be used so as to be probabilistically excluded from the vicinity according to the value of the ratio, that is, it may be more easily excluded as the ratio is smaller.

  Another application method of the optimal search law is an initial solution generation method. For example, the initial solution at the start of the search can be obtained by combining the optimal solution candidates that have the maximum ratio of the output command value of each energy generator and energy storage device in each range within each range. Generate. In the case of a genetic algorithm starting from a plurality of initial solutions, an initial schedule may be created by randomly selecting the output ranges of each energy generating device and energy storage device in each time zone. Since the output command value in the optimum solution candidate stored as the optimum search information is more frequently used in the initial solution as it appears more frequently, the optimum search information is effectively utilized. Further, the above solution may be adopted as a solution that becomes a starting point of the neighborhood when the search parameter is changed and the search is restarted when the optimum solution up to that time is not updated.

  FIG. 6 shows an example of the optimum search information erasing method in FIG. Thus, if the condition that the optimal solution candidate is deleted when a predetermined time has elapsed after the search is added, the optimal schedule candidate corresponding to the latest demand prediction or the like is stored. Further, as in this example, all the recording areas may be cleared by changing the search condition at a predetermined timing or evaluation condition. The reason for clearing in a certain time is to prevent a search in which the output command value of a fuel cell or storage battery in a certain time zone is fixed or stay in a local solution. For example, scheduling may be performed once a day by a time-optimal search without using the optimum search information in a time zone in which a demand fluctuation or the like hardly occurs for a while at midnight. Evaluation conditions may include cases where the power and fuel price patterns have changed significantly, demand forecasts or power generation forecasts have changed significantly, specific fuel cells or storage batteries have failed, or have been maintained. It is done.

  When re-scheduling is repeated at relatively short time intervals while changing the start time of the optimal operation plan according to the current time, the schedule that is the optimal solution is the schedule calculated in the optimal search process up to the previous time. Since it is considered that a similar schedule is often an approximate solution of the optimal solution, a high-quality solution can be obtained at high speed by effectively using the search process.

1 is a configuration diagram of a distributed energy system and a control device thereof according to an embodiment of the present invention. It is a flowchart which shows the flow of a process of the control apparatus in the system of FIG. 3 is a flowchart illustrating an example of a method for storing optimum search information in FIG. 2. 3 is a flowchart illustrating an example of a method for calculating a search rule from optimum search information in FIG. 2. It is a flowchart which shows an example of the method of narrowing down a search range with the optimal search information in FIG. 3 is a flowchart illustrating an example of a method for deleting optimum search information in FIG. 2.

Explanation of symbols

1 ₁ to 1 _n energy system 2 power system 3 power line 4 control device 5 communication line 11 solar cell 12 fuel cell 13 storage battery 14 heat storage device 15 power load 16 heat load 17 heat pipe 21 communication unit 22 optimum operation plan creation unit 23 optimum search Information storage unit 24 Optimal search law calculation unit 101-113, 201-204, 301-306, 401-411, 501-504 steps

Claims (12)

An optimal output command schedule of the energy generator and energy storage device of a distributed energy system having a plurality of energy generation devices and a plurality of energy storage devices and connected to a plurality of energy loads and a power system; In a distributed energy system controller that obtains and outputs an operation plan,
An optimal search information storage unit;
While evaluating the set or changed operation plan and storing the operation plan and the evaluation value in the optimum search information storage unit while searching and evaluating the operation plan until a predetermined end condition is satisfied Repeatedly, when the end condition is satisfied, an optimum operation plan creation unit that selects an optimum operation plan while changing the start time of the plan according to the current time,
An optimal search law calculation unit that gives an optimal operation plan search guideline to the optimal operation plan creation unit from the information in the optimal search information storage unit or limits the search range. Control device.   The said optimal operation plan preparation part accumulate | stores and correct | amends the output command schedule which is the said operation plan in the said optimal search information storage part only as the optimal solution candidate from the thing with the said favorable evaluation value. Control device for distributed energy system.   The optimal search law calculation unit calculates a ratio of output command values of each energy generation device and energy storage device in each time zone among the optimal solution candidates, and the optimal search information storage unit The control apparatus for a distributed energy system according to claim 2, which is stored in   The optimal operation plan creation unit omits or evaluates the evaluation of the output command schedule in which the output command value of each energy generating device and energy storage device in each time zone is within a range of a predetermined value or less. The control device of the energy system according to claim 3, wherein the control device is limited.   The optimum operation plan creation unit generates an initial solution at the start of a search by combining the output command values of each energy generation device and energy storage device in each time zone that have a high ratio within each range. The control apparatus for a distributed energy system according to claim 3.   The optimal operation plan creation unit clears the information in the optimal search information storage unit at a predetermined timing, or performs an optimal search without using the information in the optimal search information storage unit at predetermined time intervals. The control apparatus of the distributed energy system in any one of Claims 5-5. An optimal output command schedule of the energy generator and energy storage device of a distributed energy system having a plurality of energy generation devices and a plurality of energy storage devices and connected to a plurality of energy loads and a power system; In a control method for a distributed energy system that obtains and outputs an operation plan,
While evaluating the set or changed operation plan and storing the operation plan and the evaluation value in the optimum search information storage unit while searching and evaluating the operation plan until a predetermined end condition is satisfied Repetitively, when the end condition is satisfied, an optimum operation plan creation step is performed in which an optimum operation plan is selected while the start time of the plan is changed according to the current time;
An optimum search law calculation step for giving an optimum operation plan search guideline or limiting a search range to the optimum operation plan creation step from the information of the optimum search information storage unit. Control method.   The optimal operation plan creation step includes accumulating and correcting an output command schedule, which is the operation plan, in the optimal search information storage unit as a predetermined number of optimal solution candidates from those having a good evaluation value. 8. A control method of a distributed energy system according to 7.   The optimal search law calculation step calculates a ratio of output command values of each energy generation device and energy storage device in each time zone among the optimal solution candidates, and the optimal search information storage unit The method for controlling a distributed energy system according to claim 8, comprising storing in a storage system.   In the optimum operation plan creation step, evaluation is omitted or probability is not evaluated for an output command schedule in which the output command value of each energy generating device and energy storage device in each time zone is within a certain range. The method for controlling an energy system according to claim 9, further comprising:   The optimal operation plan creation step generates an initial solution at the start of a search by combining the output command values of the energy generators and energy storage devices in each time zone that have a high ratio of being in each range. The method for controlling a distributed energy system according to claim 9, comprising:   The optimal operation plan creation unit clears the information in the optimal search information storage unit at a predetermined timing, or performs an optimal search without using the information in the optimal search information storage unit at predetermined time intervals. The control method of the distributed energy system in any one of Claims 11-11.

Images