JP4115914B2 - Distribution system load accommodation method and apparatus - Google Patents

Description

  The present invention relates to a distribution system load accommodation method and apparatus, and in particular, a distribution system load suitable for creating an optimum system at the time of system switching due to an accident, overload, construction, etc. in a distribution system connected to a distributed power source. The present invention relates to an accommodation method and apparatus.

  In the power distribution system, it is necessary to switch the system due to accidents, overloads, construction, etc. Here, the power distribution system is a power facility composed of a substation, a power plant or a transmission line and a power demand facility or a power line and a switch device (hereinafter referred to as SW or switch) between demand facilities. Say. In the power distribution system, when a power outage section or switching target section (accommodated section) occurs due to an accident, overload, construction, etc., an optimal switching operation procedure for performing re-transmission (load accommodation) in the interchanged section Seeking. In other words, it is determined in what form the power is supplied from the connected peripheral system to the interchanged section excluding the accident section among the power outage sections that occur due to the accident in the distribution system. Alternatively, when an overload occurs in the distribution system, switching of the distribution system is performed, such as in what form the load is distributed and burdened in order to eliminate the overload state.

  Various solutions to this distribution load accommodation problem have been shown. For example, Patent Literature 1 and Patent Literature 2 disclose a method of sequentially determining which interconnection switch supplies power to a plurality of interchanged sections. This sequential determination method was effective for the conventional distribution system.

  Further, as disclosed in Patent Document 3, a technology for determining load accommodation using a genetic algorithm (GA) has been proposed. Here, GA is applied from the viewpoint of how to reduce the number of power outages.

  However, the situation has changed due to changes in power liberalization and the introduction of distributed power sources that have occurred in recent power / distribution systems. A distributed power source is a general term for a power generation system that is installed in a small-scale distributed near a power demand area, unlike conventional large-scale power sources such as nuclear power plants, thermal power plants, and hydropower plants.

  In the past, only specific electric power companies were allowed to generate and sell electric power according to related laws such as the “Electric Power Business Law”. That is, in Japan, a specific electric power company centrally manages a power network of power generation → transmission → distribution. However, the amendment of the law, including the “Revised Electricity Business Law”, which took effect in March 1995, has liberated entry into the wholesale electricity business. As a result, independent electric power providers (IPPs) aiming at private power generation / selling have entered the power business, and many distributed power sources have been introduced into the power system.

JP-A-9-23583 (Overall)

JP-A-9-37462 (Overall) JP-A-7-203630 (Overall)

  The introduction of this distributed power source has a great influence on the operation of the distribution system. Conventionally, the distribution system has been based on the assumption that electricity flows in one direction from a power source called a substation to a consumer who consumes power. However, due to the introduction of distributed power sources, there is a power source in the middle of the distribution network between the substation and the customer, and there is a power source in the customer premises. From consumer to substation (reverse power flow). In addition, due to the difference in the nature (phase, power factor, etc.) of the power source managed by the power company (substation) and distributed power source (phase, power factor, etc.), there are multiple power sources on the same power route (power loop state) May occur, and switching may not be possible.

  Before the distributed power supply was introduced, when it was necessary to switch the distribution system, it was basically possible to switch the system if there was a surplus power (reserved power) to supply power to the interchanged section. . However, if the distributed power supply is connected to the grid, the “distributed power supply presence / absence” of the distributed power supply connected to the distribution line for which the system is changed, “the magnitude of the power generation output value of the distributed power supply” There are conditions such as. For this reason, when it cannot be used as a support distribution line at the time of power transmission to a to-be-accommodated section, or since it will be in a power supply loop state, the case where it is necessary to cancel | release the grid connection of a distributed power source generate | occur | produces. In the conventional distribution load accommodation method, it is impossible to obtain an optimum system switching procedure under the constraints of the distributed power source.

  An object of the present invention is to provide a distribution system load accommodation method and apparatus capable of solving the distribution load accommodation problem in a distribution system interconnected with distributed power sources.

  Another object of the present invention is to solve the distribution load interchange problem with a relatively simple configuration and at high speed in a distribution system in which distributed power sources are interconnected.

  The main feature of the present invention is that when there is a distributed power source connected to the distribution system, the distributed power source is determined based on the information on the connected condition, contract condition, and / or electrical characteristics. In consideration of the classification result, a plurality of power distribution systems to the interchanged section are evaluated using a meta-heuristic method. And based on this evaluation, one power distribution system is selected from a plurality of power distribution systems. Therefore, the distribution system is switched based on this selection, and power is retransmitted to the interchanged section through the switched distribution system.

  In a preferred embodiment of the present invention, first, a distributed power source connected to the system is determined based on the contract conditions of the grid connection, the presence / absence of reverse power flow to the distribution system, the amount of generated power of the distributed power source, and the like. Classify into multiple types according to the nature of switching. For example, the following three are suitable.

  (B) A distributed power source capable of maintaining interconnection with the distribution system.

  (B) A distributed power source that needs to be disconnected from the distribution system.

  (C) A distributed power source that is difficult to disconnect from the distribution system and cannot be switched.

  Next, in order to quantitatively evaluate the impact on the distributed power supply due to the system switching, the distributed power supply belonging to the distribution line that is the target of the system switching is based on the result of the above classification. And the points of the degree of influence on the distribution system are calculated. In addition to the evaluation items of “(1) Minimize inoperable section and (2) Minimize switching procedure”, which are basic criteria for evaluating the load interchange system, “(3) Calculated distributed power impact points” In view of the above, the load interchange system with the highest evaluation overall is obtained.

  In the present invention, meta algorithms such as genetic algorithm (GA), tabu search, annealing (simulated annealing), local search (local seach), and ant system algorithm (ant system), which are one of optimization methods, are used. It is desirable to perform optimization using a heuristic (meta-heuristic) method.

  The optimization by the metaheuristic method is characterized by performing random global search without requiring empirical knowledge in the search process. Therefore, it is possible to derive a solution that is difficult to plan with an algorithm that seeks a solution according to a certain rule, such as AI (Artificial Intelligence).

  When applying the GA to the “load accommodation problem” of the distribution system, in order to evaluate a more optimal load accommodation system in a short time, an ark is classified as a unit for each branch of the distribution line. Then, all loops connecting the arcs are extracted, and if there is one switching switch in each loop, pay attention to the fact that there is no power source collision, and assign one switching switch to each loop. It is desirable to make it.

  Specifically, first, a plurality of closed-loop groups with a group of switches belonging to a closed loop or an inter-power supply loop formed by sequentially turning on and off the switch of the distribution system in the initial system configuration as one group To extract. Next, a specific switch in the switch group belonging to each extracted closed loop group is set as a switch for each closed loop group. According to this setting, the initial gene is created by converting the load accommodation system configuration by the switch arrangement of each closed loop group into data. An operation target gene is selected from the switches belonging to each of the closed loop groups, and a new gene that is a load accommodation system configuration candidate is generated based on the selected operation target gene. Next, a combination with a high evaluation value is created as an optimum load accommodation system configuration among the combinations of switch breakers based on the new gene.

  Thereby, the information of each loop and the information of one switch provided to each loop are converted into data.

  ADVANTAGE OF THE INVENTION According to this invention, the distribution load interchange problem can be solved in the distribution system which a distributed power supply connects.

  Further, in a distribution system in which distributed power sources are connected, the distribution load interchange problem can be solved at a high speed with a relatively simple configuration.

  In particular, based on the index data representing the number of loops × 1 switch, it is possible to determine the distribution system without having to convert all the switching states of all switches into a huge amount of load on a wide range of system switching targets. The flexibility problem can be solved at high speed with a simple configuration.

  When configuring the distribution load accommodation system configuration creation device according to the present invention, it is desirable to add the following elements.

  (1) In generating a new gene, an operation target gene is selected from the switches belonging to each closed loop group, and a part of the selected operation target gene is changed to generate a new gene that becomes a candidate for a load accommodation system configuration. (Mutation processing means). Two sets of operation target genes are selected from the switches belonging to each closed loop group, and a part of each selected set of operation target genes is replaced to generate a new gene that is a candidate for load accommodation system configuration (crossover processing means) ). A new gene generated by the mutation process (means) or the crossover process (means) is added to a population that is a set of load accommodation system configuration candidates (addition means). An evaluation value for the new gene generated by the mutation process (means) or the crossover process (means) is calculated (evaluation value calculation means). Depending on the evaluation value calculated by the evaluation value calculation (means), a part of the new genes is excluded from the population with a high probability from the lowest evaluation value (ie, wrinkle processing means).

  (2) In the mutation process (means) and the crossover process (means), a new gene is selected from the closed loop group to which the operation target gene belongs.

  (3) The mutation process (means) and the crossover process (means) indicate whether the generated new gene belongs to any group of switch groups grouped for each branch point of the section constituting the distribution system. At the same time, when two or more switchgears exist in the group to which the generated new gene belongs, the generated new gene is invalidated.

  (4) The evaluation value calculation (means) calculates an evaluation value for switching from the initial system configuration to the load-accommodating system configuration by the new gene (by a mathematical expression described later).

  (5) It is determined whether or not the load accommodation system configuration to be calculated by the evaluation value calculation means violates the constraint condition that the load accommodation system should comply with. When a violation occurs, the load accommodation system configuration that is the calculation target of the evaluation value calculation means is corrected (system configuration correction means), and the system configuration correction (means) selects a switch that satisfies the constraint condition when the violation occurs To do. The switch is registered as a variable-length gene, and the evaluation value calculation (means) is instructed to calculate the evaluation value related to the modified load accommodation system configuration.

  (6) The evaluation value of the optimal load accommodation system configuration candidate having the highest evaluation value among the set of load accommodation system configuration candidates generated by the new gene generation means may not improve for a certain period or more. At this time, the load accommodation system configuration creation means discards the set of load accommodation system configuration candidates generated by the new gene generation means, and operates the new gene generation means with the optimum load accommodation system configuration candidate as an initial system configuration. Command the selection of the gene of interest.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a distribution system load accommodation device showing an embodiment of the present invention. In FIG. 1, the power distribution system load accommodation device includes a computer 100, an input device 20, and a display device 30, and the computer 100 is connected to the power distribution system 50 via a remote monitoring control device 40. Here, the power distribution system 50 means an actual power transmission / distribution facility including a distribution line, a switch, and the like.

  The remote monitoring control device 40 is configured to monitor the state of the power distribution system 50 such as the switching state of the switch and the power flow using communication means such as a communication line, and control the switching operation of the switch. Has been.

  The input device 20 includes a keyboard and a mouse for inputting data to the computer 100, and the display device 30 includes a CRT that displays an image related to data output from the computer 100.

  The computer 100 calculates an optimum load accommodation system configuration based on data input or taken in from the input device 2 or the remote monitoring control device 40, and displays or outputs the result on the display device 30 or the remote monitoring control device 40. It is configured as follows.

  The computer 100 has functions of an overall control unit 101, a load accommodation calculation control unit 102, a system configuration evaluation unit 103, a mutation processing unit 104, a crossover processing unit 105, a system configuration correction unit 106, a survival unit 107, and a database unit 108. It consists of blocks.

  The overall control unit 101 controls data exchange from the input device 20, the display device 30, and the remote monitoring control device 40. In addition, when receiving an instruction to create a load accommodation system from the input device 20 or the remote monitoring control device 40, a load accommodation system configuration search is loaded on the data fetched from the input device 20 or the remote monitoring control device 40. An instruction is given to the accommodation calculation control unit 102. Then, the result is received and output to the display device 30 and the remote monitoring control device 40. Data collected from the input device 20 or the remote monitoring control device 40 is stored in the database unit 108 and referred to as necessary.

  The load accommodation calculation control unit 102 is configured as an initial gene creation unit that creates an initial gene based on data stored in the database 108. Further, when an instruction relating to the creation of a load accommodation system is received from the overall control unit 101 such as at the time of an accident, an optimum load accommodation system configuration is created in cooperation with the function units 103 to 108 in the computer 100. And it is comprised as a load accommodation system structure preparation means which outputs the result to the display apparatus 30 or the remote monitoring control apparatus 40. FIG.

  The system configuration evaluation unit 103 is configured as an evaluation value calculation unit that calculates an evaluation value of the system configuration to be evaluated based on the defined [objective function] and [constraint conditions].

  The mutation processing unit 104 creates a system configuration in which the positions of one or more “OFF” switches are moved for the system configuration instructed by the load accommodation calculation control unit 102, and outputs the creation result. It is configured as a means.

  The crossover processing unit 105 replaces the information about the “switch” switch with respect to the two system configurations instructed from the load accommodation calculation (optimization) control unit 102 by replacing one or more boundary points with each other. It is configured as a crossover processing means for creating a configuration.

  The system configuration correcting unit 106 is configured as a system configuration correcting unit that performs correction processing for eliminating the violation on the system configuration in which the [constraint condition] violation occurs in the system configuration evaluating process in the system configuration evaluating unit 103. Has been. In this case, the corrected system configuration is evaluated again by the system configuration evaluation unit 103, and an evaluation value related to the corrected system configuration is calculated again.

  The survival unit 107 is configured as an addition unit and a trap processing unit that receives a command from the load accommodation calculation control unit 102 at a timing when a specified number of new grid configurations are created and performs trapping of the grid configuration candidate population. Yes. That is, it is configured such that a system configuration with a low evaluation is eliminated from the population of system configuration candidates by the population of the population.

  The database unit 108 manages data input from the input device 2 and the remote monitoring control device 40 and system configuration data held by the optimization control unit 102.

  FIG. 2 is an overall flowchart of load accommodation system configuration optimization processing by the functional units 102 to 108 of the computer 100 in the embodiment of the present invention. An embodiment of the present invention will be described along this flow.

“Step S1”:
First, in step S1, when there is a need to switch the system due to an accident, overload, construction, etc. in the distribution system, the switch on / off information in the initial system configuration indicating the system configuration before the system switching is generized. This is created as an “early gene”. Specifically, an initial gene is created by the following processing, and all of the genes created in the process of searching for the optimal gene become descendants of this early gene.

  FIG. 3 is an example of a power distribution system illustrating the creation of an early gene according to an embodiment of the present invention. This power distribution system includes FCBs (Feeder Circuit Breakers) 1 to 3 corresponding to power sources, switches 4 to 16 for controlling line sections and power supply routes. As the power distribution system, the number of “off” switches necessary for the collision of the power sources or the absence of the closed loop is determined. In this embodiment, a method is adopted in which a combination of “off” switches that do not have a power supply collision and does not have a closed loop is determined, and one “off” switch is determined as a genetic element for each loop. .

  When extracting the closed loop, as shown in FIG. 3 (a), grouping is performed in units of electric wire branch points, and each group is defined as an arc (Ark). Further, in order to extract a closed loop due to a collision between power sources as a thinking process, a virtual arc is assumed as a virtual section connecting the power sources. For example, a section connecting FCB1 and a virtual branch point is defined as a virtual arc PA1, a section connecting FCB2 and a virtual branch point is defined as a virtual arc PA2, and a section connecting FCB3 and a virtual branch point is defined as a virtual arc PA3. In accordance with this definition, arcs are extracted from the target distribution system, virtual arcs are extracted, and an index table T1 is created as shown in FIG. 3B as information about each arc and the switch belonging to the virtual arc. To do. At this time, all the switches belong to one arc. For example, FCB1 belongs to the virtual arc PA1, FCB2 belongs to the virtual arc PA2, and FCB3 belongs to the virtual arc PA3. Further, the switches 4 and 5 belong to the arc A1, the switches 6 and 7 belong to the arc A2, the switch 8 belongs to the arc A3, and the switch 9 belongs to the arc A4. The switches 4, 7, 8, 10, 11, 14, and 16 are switches to be switched and indicate that they are in an on state, and the switches 5, 9, 12, and 15 are switches to be switched. It indicates that it is in the off state. In addition, the switches 6 and 17 are non-switching target switches and are in the on state, and the switch 13 is a non-switching target switch and in the off state.

  Next, the “OFF” switch of the distribution system in the initial system configuration showing the system configuration before the system switching is sequentially virtually turned “ON”. A plurality of closed loop groups are extracted with the switch groups belonging to the closed loop or the loop between power supplies formed as a group. A specific switch in the group of switches belonging to each extracted closed loop group is set as the “OFF” switch for each closed loop group, and the load accommodation system configuration is configured by the arrangement of “OFF” switches for each closed loop group. Data is created and an early gene is created.

  FIG. 4 is a diagram for explaining a loop extraction method of a data distribution method for a distribution system according to an embodiment of the present invention. As shown in FIG. 4A, in order to extract a closed loop, all “OFF” switches are virtually sequentially turned “ON” to extract a closed loop. For example, when the switch 5 is virtually “ON”, a loop L3 passing through the arc A1, the arc A5, the virtual arc PA2, and the virtual arc PA1 is extracted. When such processing is performed, in the case of the power distribution system shown in FIG. 4A, the loops L1 to L4 are extracted, and the index table T2 is created as shown in FIG. 4B. In this case, the loop L1 is composed of arcs A4, A3, A5, virtual arcs PA2, PA1, and arc A2, and the loop L2 is composed of arc A2, virtual arcs PA1, PA3, and arc A9. The loop L3 is composed of arcs A1 and A5, virtual arcs PA2 and PA1, and the loop L4 is composed of arcs A6 and A8 and virtual arcs PA3 and PA2. And the switch which belongs to each loop is registered with the number of a switch in index table T3, as shown in FIG.4 (b). For example, switches 1, 6, 7, 9, 8, 10, 11, and 2 are registered in the loop L1, and switches 1, 6, 7, 15, 16, and 3 are registered in the loop L2. . Further, switches 1, 4, 5, 10, 11, and 2 are registered in the loop L3, and switches 2, 12, 17, and 3 are registered in the loop L4. In addition, since 1, 2, and 3 of the belonging switches are FCBs, they are actually excluded from switching targets.

  Through the above process, gene information is created. This gene element fixes the locus of the gene for each loop number as a switch number. For example, as shown in FIG. 5, in the initial system configuration before system switching, the “OFF” switch is gener- ated as an initial gene. The gene structure thus determined is defined as “fixed length gene part”.

  For each gene locus (loop number), only a switch belonging to the corresponding closed loop is set to suppress the generation of invalid genes. The “fixed-length gene part” defined in this way is a generative “all transmission system” that does not generate an inflexible section.

“Step S2”:
Next, in step S2, a process for calculating the evaluation value of the initial gene is performed. The initial gene evaluation value is calculated by the system configuration evaluation unit 103. When the system configuration evaluation unit 103 calculates an evaluation value related to the initial system configuration, the evaluation value is calculated according to (Equation 1).

  In the present embodiment, the evaluation function shown in (Expression 1) is used in consideration of a plurality of evaluation items such as “constraint violation rate”, “incompatible load amount”, and “distributed power supply influence degree evaluation value”. By optimizing the gene so that the evaluation value of this function is maximized, load accommodation calculation corresponding to the distributed power source is enabled.

  Specifically, the evaluation of the flexible system configuration is calculated by the following three evaluation factors.

(1) Constraint violation rate:
(2) Incompatible load:
(3) Distributed power impact rating:
“(1) Constraint violation rate” is an item that evaluates whether the distribution system expressed by the gene is appropriate and does not violate the constraint. It consists of two elements: current overload violation and appropriate voltage deviation violation. Is done. If the constraint is violated, the goal is to add a penalty according to the degree of violation. Current overload violation is a restriction condition that is intended to prevent the current exceeding the allowable passing current specified in the distribution equipment such as switches and wires constituting the distribution system from flowing. expressed.

  In (Equation 2), if the passing current exceeds the allowable passing current value, the current overload violation rate is calculated by the following (Equation 3).

  Next, the proper voltage deviation violation is a constraint condition that is intended to prevent the ultimate voltage in each section from deviating from the prescribed voltage drop tolerance, and is expressed by (Equation 4).

  In (Equation 4), when the section arrival voltage deviates from the allowable voltage drop, the violation rate is calculated by the following (Equation 5).

  Of the violation rates calculated in (Equation 3) and (Equation 4), the maximum value is extracted in (Equation 6). The value X1 obtained in (Equation 6) is the “constraint violation rate” which is one of the components of the gene evaluation function.

  Next, the “(2) incompatible load amount” is an item for evaluating that the power outage load amount due to an accident, overload, construction, or the like is minimized, and is expressed by the following (Equation 7).

  Finally, the “(3) distributed power supply influence evaluation value” is an evaluation aimed at minimizing the adverse effects on the distributed power supply and the effects of impairing the system stability in the load interchange switching operation. It is an item and is expressed by (Equation 8).

  Since a distributed power source is a generator (power source) even if it is a small scale, it is not a little distributed if it is connected to a power facility (loop between power sources) that has been managed and operated by a power company. Adversely affect the system. In addition, depending on the content of a contract between the distributed power operator and the power company that is the owner of the distribution network, the system may not be the target of system switching at the time of load accommodation. For this reason, in this embodiment, in order to quantitatively evaluate the influence of the distributed power source at the time of load accommodation operation, (1) Contract condition of grid connection, (2) Presence / absence of reverse power flow to distribution system, (3) Depending on the amount of power generated by the distributed power source, etc., it is classified into one of the following from the nature of the distributed power source when system switching is performed.

  (B) A distributed power source capable of maintaining interconnection with the power distribution system.

  (B) A distributed power source that needs to be disconnected from the distribution system.

  (C) A distributed power source that is difficult to disconnect from the distribution system and cannot be switched.

  Here, in (C), a temporary overload occurs in the power distribution system by disconnecting the distributed power source, or unacceptable system fluctuation occurs. Or, there is a situation where it is not possible to request disconnection due to the individual circumstances of the operator of the distributed power supply.

  The degree of influence of the distributed power source is judged to be small in the order of (a) (b) (c). Points are generated according to the degree of influence, and distributed power influence points are calculated for each feeder to be switched. Switching is performed with priority given to the system with the smallest influence point.

  FIG. 6 is a specific example of a simple power distribution system for explaining the concept of switching with priority given to the influence point according to an embodiment of the present invention. In FIG. 6, A to E are power supplies (feeders) in a substation managed by an electric power company, and (1) to (3) indicated by a circle G are distributed power supplies. As shown in the figure, consider a case where an accident occurs in both feeders A and B, and two power outages occur, namely, power outage (1) and power outage (2). Assume that the distributed power source G (1) is connected to the feeder C, the distributed power source (2) is connected to the feeder D, and the distributed power source (3) is connected to the feeder E. As shown in the figure, the reserve capacity of feeders C, D, and E is all 300A, and the reserve condition is sufficient for power transmission to power outages (1) and (2) in feeders C, D, and E. There shall be. In the conventional load accommodation calculation mainly based on “minimization of incompatible section”, since the reserve power condition is the same, it is determined that no problem may occur in any accident section from any feeder. However, in this embodiment, the influence degree of the distributed power sources (1), (2), and (3) linked to the feeders C, D, and E is considered. When the distributed power sources (1), (2), and (3) are classified into the above (a) to (c) according to the above-described conditions, the distributed power source (1) is (a) and the distributed power source (2 ) Is (b), and the distributed power source (3) is (c). The distributed power supply influence levels are (A): 1, (B): 5, and (C): 10. Under this condition, when the power failure (1) is transmitted from the feeder C and the power failure (2) is transmitted from the feeder D, the distributed power sources (1) and (2) are affected. Therefore, the distributed power supply influence evaluation value = {distributed power supply (1)-(b) influence degree} + {distributed power supply (2)-(b) influence degree} = 1 + 5 = 6. Similarly, when the power failure (1) is transmitted from the feeder D and the power failure (2) is transmitted from the feeder E, the distributed power sources (2) and (3) are affected. Therefore, the distributed power supply influence evaluation value = {distributed power supply (2) − (b) influence degree} + {distributed power supply (3) − (c) influence degree} = 5 + 10 = 15. Therefore, in the load accommodation calculation according to the present embodiment, the former switching with a small distributed power source influence degree is selected.

“Step S3”:
Returning to the overall flow of FIG. 2, in step S3, when an evaluation value for the initial gene is calculated, a gene to be subjected to crossover / mutation is selected from all existing genes (gene population). . When selecting the gene to be manipulated, differentiation is performed according to the gene evaluation value determined by the system configuration evaluation unit 103, and in this embodiment, the roulette selection method is applied as the method for selecting the gene to be manipulated. To do. In the roulette selection method, when the total number of genes in a generation t is N (t) and the total gene evaluation value is G, the probability p () that the gene with gene number i and evaluation value g (i) is selected as the operation target gene. i) is defined by (Equation 9).

  That is, when selecting a gene to be manipulated at random, as the area in the roulette, the higher the evaluation value, the larger the area, and the higher the evaluation value is selected.

“Step S4”:
Next, in step S4, after an operation target gene is selected, processing for generating a new gene is performed by the processing of the mutation processing unit 104 and the crossover processing unit 105.

  FIG. 7 is an explanatory diagram and a flowchart of the mutation process according to the embodiment of the present invention. Mutation processing is one of new gene creation techniques in GA. In the present embodiment, as shown in FIG. 7 (a), a new gene is generated by changing the “off” switch at a certain gene locus (= loop) to one of the switches constituting the same loop. To do. In this case, a candidate switch for mutation is a common set of a list of closed loop configuration switches at the sitting position and a list of closed loop configuration switches configured by closing the “off” switch before the mutation. It is said. Moving a “off” switch to a switch other than the common set will inevitably result in either a “closed loop system” or a “loop system between power supplies” and violate non-loop constraints. Become. For this reason, in the present embodiment, in order to avoid this, generation of unnecessary genes and evaluation are avoided to reduce calculation time.

  Specifically, in step S11 of the flow of FIG. 7B, the locus to be mutated is determined using a random number, and in step S12, a list of “OFF” switches that are mutation candidates for the corresponding locus is acquired. To do. That is, when the switch 9 is selected, a list related to the switches in the same loop to which the switch 9 belongs is acquired. Thereafter, in step S13, a mutation candidate switch is determined using a random number. In step S14, it is determined whether or not the candidate switch is the same as the switch before change. If the candidate switch is the same, the process returns to step S13 and the same processing is repeated. When they are different, the process proceeds to step S15, and a process for acquiring a candidate switch belonging arc is performed.

  Thereafter, in step S16, it is determined whether or not the belonging arc is shared by a plurality of loops. In step S17, it is determined whether or not the corresponding arc has been used in the shared loop. When a negative determination is obtained in step S16, the process proceeds to step S18, and processing for replacing the corresponding sitting position with a candidate switch is performed. When a positive determination result is obtained in step S17, the process returns to step S13. That is, in steps S16 and S17, it is determined whether or not there are two or more “off” switches in one arc as a process for preventing a power outage section from occurring. This mutation process is performed as a neighborhood search process for changing a part of the system configuration.

  FIG. 8 is an explanatory diagram and a flowchart of the crossover processing unit according to the embodiment of the present invention. The crossover processing in the crossover processing unit 105 of FIG. 1 is one of new gene creation techniques in GA. In this embodiment, as shown in FIG. 8A, a new gene is generated by replacing two genes with a boundary point called a crossing point as a reference. For example, when the gene selected as the gene to be crossed is a “switch” switch list such as [Gene 1] and [Gene 2], crossover is performed between the gene locus 2 and 3 as a crossover point. . As a result, new genes such as [new gene 1] and [new gene 2] are newly generated.

  Specifically, in the flow of FIG. 8B, a crossing point is determined using a random number in step S21. For example, an intersection point is determined between the second and third sitting positions. Thereafter, in step S22, the used arc is initialized and the sitting position counter is initialized. Next, in step S23, it is determined whether or not the count value of the sitting position counter is smaller than the numerical value of the crossing point. When the intersection point is determined between the sitting positions 2 and 3, when the counter value is 1, 2, the process of step S24 is performed, and when the counter value is 3, 4, the process of step S25 is performed. . In step S24, gene A is changed to gene A, gene B is changed to gene B, that is, without changing the gene, the process proceeds to step S26. On the other hand, in step S25, a process for changing gene A to gene B and gene B to gene A is performed, and the process proceeds to step S26.

  Thereafter, in step S26, it is determined whether or not the belonging arc of the gene switch has been used. That is, since each arc has only one “OFF” switch, it is determined whether or not an “OFF” switch is set for each arc. When the belonging arc has been used, the process returns to step S21 in order to change the position of the crossing point. On the other hand, when the belonging arc is not used, the process proceeds to step S27, the belonging arc of the gene switch is set to used, and the process proceeds to step S28. In step S28, it is determined whether or not the processing related to the crossing points of all sitting positions has been completed. If not all of them have been completed, a process for updating the value of the sitting position counter is performed in step S29, and the process returns to the process of step S23. By performing the crossover process, a gene for dynamically changing the system configuration is selected.

“Step S5”:
After the new gene is generated in the process of step S4 in FIG. 2, the system configuration evaluation unit 103 in FIG. 1 performs the process in step S5. That is, processing for calculating an evaluation value is performed for the generated new gene. This process is the same as the process performed in step S2, and redundant description is avoided here. As described above, the distributed power source that is the main part of the present invention is classified, and the GA processing for selecting a flexible system is performed by evaluating a plurality of power distribution systems from various viewpoints, and processing in step S6 Move on.

“Step S6”:
In step S6, a process is performed as to whether or not the number of generated genes of one generation is satisfied. That is, it is determined whether or not the number of genes to be generated in one generation has reached a set number, for example, 1000. If this number is not satisfied, the process returns to step S3. The process proceeds to S7.

“Step S7”:
In step S7, a process for selecting a gene is performed. In this process, the survival unit 107 performs a process for selecting genes with a high probability in order from the gene with the lowest evaluation value. In this case, the roulette selection method used for the selection of the gene to be manipulated is adopted. However, in the gene selection process, the gene that is difficult to select as the gene to be manipulated is selected, contrary to the selection of the gene to be manipulated. Is going to be selected as. That is, as a process for discarding more as the evaluation value is lower, the total number of genes in a generation t is N (t) and the total gene evaluation value is G, and the gene number i and evaluation value g (i) The probability p ′ (i) that a gene is selected as a selection target gene is defined by the following (Equation 10).

  Thereafter, a combination having a high evaluation value among the combinations of the “switch” switches by the new gene is created as an optimum load accommodation system configuration. For this reason, the interchangeable system configuration (gene) determined by the system configuration evaluation unit 13 as violating the constraint condition cannot be actually adopted as the optimal system, and is therefore normally discarded.

“Steps S8 and S9”:
Next, in step S8, in the process of creating the optimal load accommodation system configuration, it is determined whether or not there is an optimal gene update for a certain period, and if the gene update is not performed for a certain period, the optimization speed is stagnant. It is determined that In this case, a process for discarding the gene population is performed in step S9. For example, if the evaluation value cannot be improved even after 100 generations of processing are continued, the gene population is temporarily discarded, and the process returns to step S1 with the optimal gene as the initial gene. That is, as a characteristic of GA processing, the initial convergence ability of the optimization process is very high, but if the gene population is biased only to genes having a certain gene structure, the convergence speed may stagnate. Therefore, in the present embodiment, in order to compensate for the shortcomings of GA, when the optimal gene is not updated for a certain period of time, the gene population is temporarily discarded, and the current optimal gene is used as the initial gene to form a closed loop configuration. Recreate the switch. As a result, it is possible to reduce the possibility of being caught by the local solution and not reaching the optimal solution, and to improve the optimization convergence speed.

“Step S10”:
Next, in step S10, GA is processed a plurality of times, and it is determined whether or not the GA convergence condition is satisfied. The determination condition for ending the GA process is comprehensively determined according to “the system scale of the calculation target” and “the optimal gene update status”. In other words, if the number of genes that exceeds the maximum number of genes determined according to the “system size to be calculated” is generated or the optimal gene is not updated while a certain number of genes are generated Then, it is determined that the optimization process has converged, and the optimization process is terminated.

  As described above, in this embodiment, each branch of the distribution line is classified as an arc, and all loops connecting the arcs are extracted. Focusing on the fact that there is no power supply collision if there is one “off” switch in each loop, one “off” switch is assigned to each loop to create a database. The information of each loop and the information of one “off” switch assigned to each loop are converted into data, and the switching status of all the switches based on the index data representing the number of loops × one “off” switch The power system can be determined without converting all of the data into data. For this reason, it is possible to solve an enormous load interchange problem for a wide range of system switching targets at high speed with a simple configuration, and it is possible to contribute to shortening of calculation time.

  In addition, since only one “OFF” switch is assigned to each loop, it is possible to prevent the occurrence of an invalid pattern in which two or more switches are “OFF” and a power outage period occurs. it can.

  Further, when the power distribution system in FIG. 3 is described by the conventional data conversion method in which the on / off states of the switches are arranged, 17 switches can take on and off states, respectively. Is 2 to the 17th power = 131,072.

  On the other hand, according to the present embodiment, a data conversion method for describing the number of the extracted closed-loop “off” switch is used. Therefore, it is the number when the product of the switches belonging to each loop of L1 to L4 can be considered. That is, switches 6, 7, 8, 9, 10, 11, switches 6, 7, 15, 16, switches 4, 5, 10, 11 except for FCB 1, 2, 3 among the switches in the index table T 3. , Switches 12 and 17. The number of these switches can be considered, and 6 × 4 × 4 × 2 = 192 can be used.

  Thus, in order to determine the optimum on / off state of the 17 switches, it is necessary to search for the optimum one out of 131,072 ways in the conventional method. On the other hand, in the present embodiment, an optimal one may be searched from 196 types.

  In addition, according to the present embodiment, when the optimal gene cannot be improved for a certain period of time, since the set of load accommodation system configuration candidates is once discarded, the convergence speed is prevented from stagnation due to the bias of the set, Optimization calculation can be performed at high speed.

The block block diagram of the distribution load accommodation system structure production apparatus which shows one Embodiment of this invention. The whole flowchart of the load accommodation system | strain structure optimization process by one Embodiment of this invention. The example figure of the power distribution system explaining preparation of an early stage gene by one embodiment of the present invention. The loop extraction technique explanatory drawing of the data-ized technique of the distribution system by one Embodiment of this invention. The figure explaining the correlation of the loop and gene information in the data collection technique of the power distribution system by one Embodiment of this invention. The specific example figure of the simple power distribution system for demonstrating the view of influence point priority switching by one Embodiment of this invention. An explanatory view and a flow chart of mutation processing by one embodiment of the present invention. The explanatory view and flowchart of a crossing processing part by one embodiment of the present invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 ... Computer, 101 ... Overall control part, 102 ... Load accommodation calculation control part, 103 ... System configuration evaluation part, 104 ... Mutation processing part, 105 ... Crossover processing part, 106 ... System configuration correction part, 107 ... Survival part, 108 ... Database unit, 20 ... Input device, 30 ... Display device, 40 ... Remote monitoring and control device, 50 ... Power distribution system.

Claims (8)

When a power outage section or switching target section occurs due to an accident, overload, construction, etc., one power distribution system is selected from among a plurality of distribution systems to the interchanged section, and power is retransmitted to the interchanged section in the distribution system load accommodation method, when the distributed power sources that interconnection to the power distribution system is present, the distributed power supply, based on the interconnection condition, contract terms, and / or electrical characteristic information, the system When switching, (b) a distributed power source capable of maintaining a connection with the distribution system, (b) a distributed power source that needs to be disconnected from the distribution system, and / or (c) a distribution system. A meta-heuristic method for classifying a plurality of distribution systems to the to-be-accommodated section in consideration of a step of classifying one of a plurality of classifications including a distributed power source that is difficult to disconnect from the grid. Step to use and evaluate and based on this evaluation A step of selecting one distribution system from among a plurality of distribution systems, a step of switching the distribution system based on the selection, and a step of retransmitting power to the interchanged section through the switched distribution system, Distribution system load accommodation method. When a power outage section occurs, in the distribution system load accommodation method that selects one distribution system from among a plurality of distribution systems to the accommodation section and retransmits power to the accommodation section, it is connected to the distribution system. When there is a system switchover based on information on the interconnection conditions, contract conditions, and / or electrical characteristics , (b) the connection with the distribution system Multiple including a distributed power source that can maintain the system, (b) a distributed power source that needs to be disconnected from the distribution system, and / or (c) a distributed power source that is difficult to disconnect from the distribution system A step of classifying the power distribution system into one of the classifications, a step of evaluating a plurality of power distribution systems to the interchanged section in consideration of the classification, and one power distribution among the plurality of power distribution systems based on the evaluation Step to select the grid and switch the distribution system based on this selection Steps and, the power distribution system load accommodation method characterized by comprising the step of re-transmitting said to be flexible section through the switched power distribution system that. 3. The distribution system load accommodation method according to claim 1, wherein the classification of the distributed power source is executed in consideration of the presence or absence of reverse power flow to the distribution system and / or the amount of generated power of the distributed power source. . The step of quantitatively evaluating the degree of influence on the distributed power source and the distribution system in terms of points when the system is switched based on the classification of the distributed power source according to any one of claims 1 to 3. A distribution system load accommodation method characterized by comprising: In any one of Claims 1-4 , the step which considers the evaluation item of minimization of an inflexible area and / or switching procedure minimization, and the some load accommodation system determined by the on / off state of a switch quantitatively A distribution system load accommodation method comprising: evaluating and ranking. The distribution system load accommodation method according to any one of claims 1 to 5 , wherein a genetic algorithm is used for the evaluation. When a power outage section or switching target section occurs due to an accident, overload, construction, etc., one power distribution system is selected from a plurality of power distribution systems to the interchanged section, and power is retransmitted to the interchanged section In a distribution system load accommodation device, a distributed power source that is connected to the distribution system, and when this distributed power source is switched, (a) a distributed power source that can maintain the connection with the distribution system, (B) Means for classifying into one of a plurality of categories including a distributed power source that needs to be disconnected from the distribution system and / or (c) a distributed power source that is difficult to disconnect from the distribution system. Considering this classification result, a means for evaluating a plurality of distribution systems to the interchanged section using a meta-heuristic method, and selecting one distribution system from the plurality of distribution systems based on this evaluation Distribution system load melting characterized by the provision of means Apparatus. In a distribution system load accommodation device that selects one distribution system from among a plurality of distribution systems to the interchanged section when a power outage section or switching target section occurs, When there is an interconnected distributed power supply, when the system is switched on the basis of the information on the interconnected condition, contract condition, and / or electrical characteristics , (a) power distribution Distributed power supply that can maintain the connection to the grid, (b) A distributed power supply that needs to be disconnected from the distribution system, and / or (c) A distributed type that is difficult to disconnect from the distribution system A means for classifying into one of a plurality of classifications including a power supply, a means for evaluating a plurality of distribution systems to the accommodation section taking this classification into account, and a plurality of distribution systems based on the evaluation Distribution means characterized by comprising means for selecting one distribution system from Integrated load accommodation equipment.

Images