(19)
(11)EP 2 507 750 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
29.04.2020 Bulletin 2020/18

(21)Application number: 10793093.5

(22)Date of filing:  03.12.2010
(51)Int. Cl.: 
G06N 3/12  (2006.01)
H02J 3/00  (2006.01)
(86)International application number:
PCT/US2010/058881
(87)International publication number:
WO 2011/069061 (09.06.2011 Gazette  2011/23)

(54)

RESTORATION SWITCHING ANALYSIS WITH GENETIC ALGORITHM

WIEDERHERSTELLUNGSSCHALTANALYSE MIT GENETISCHEM ALGORITHMUS

ANALYSE PAR COMMUTATION POUR RÉTABLISSEMENT D'ÉNERGIE PAR ALGORITHME GÉNÉTIQUE


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 04.12.2009 US 266691 P

(43)Date of publication of application:
10.10.2012 Bulletin 2012/41

(73)Proprietor: ABB Schweiz AG
5400 Baden (CH)

(72)Inventors:
  • WANG, Zhenyuan
    Cary NC 27518 (US)
  • LI, Wenping
    Cary NC 27518 (US)

(74)Representative: Gregorj S.r.l. 
Via L. Muratori, 13/b
20135 Milano
20135 Milano (IT)


(56)References cited: : 
EP-A1- 2 507 751
WO-A1-2009/076410
  
  • D. ZHANG, Z. FU, L. ZHANG, Z. SONG: "Network reconfiguration in distribution systems using a modified TS algorithm", PROCEEDINGS OF THE 7TH WSEAS INTERNATIONAL CONFERENCE ON MATHEMATICAL METHODS AND COMPUTATIONAL TECHNIQUES IN ELECTRICAL ENGINEERING (MMACTEE'05), 27 October 2005 (2005-10-27), pages 310-314, XP007916998,
  • N. XIONG, H. CHENG, L. YAO, M. BAZARGAN: "Switch group based Tabu search algorithm for distribution network reconfiguration", PROCEEDINGS OF THE 3RD INTERNATIONAL CONFERENCE ON ELECTRIC UTILITY DEREGULATION AND RESTRUCTURING AND POWER TECHNOLOGIES (DRPT'08), 6 April 2008 (2008-04-06), pages 820-822, XP031253967,
  • F. GLOVER, M. LAGUNA: "Tabu search", HANDBOOK OF COMBINATORIAL OPTIMIZATION, vol. 3, 1998, pages 621-757, XP008132770,
  • X. YANG, Y. ZHANG: "Intelligent real-time fault restoration of the large shipboard power system based on genetic algorithm", INTERNATIONAL JOURNAL OF INFORMATION TECHNOLOGY, vol. 11, no. 12, 2005, pages 129-135, XP007916997,
  • W. P. LUAN, M. R. IRVING, J. S. DANIEL: "Genetic algorithm for supply restoration and optimal load shedding in power system distribution networks", IEE PROCEEDINGS: GENERATION, TRANSMISSION AND DISTRIBUTION, vol. 149, no. 2, 15 March 2002 (2002-03-15), pages 145-151, XP006018050,
  • R. M. V. VITORINO, H. M. M. JORGE, L. M. P. NEVES: "Network reconfiguration using a genetic approach for loss and reliability optimization in distribution systems", PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON POWER ENGINEERING, ENERGY AND ELECTRICAL DRIVES (POWERENG'09), 18 March 2009 (2009-03-18), pages 84-89, XP031453935,
  • J. STOUPIS, Z. WANG, F. YANG, V. DONDE, F. MEKIC, W. PETERSON: "Restoring confidence", ABB REVIEW: THE CORPORATE TECHNICAL JOURNAL OF THE ABB GROUP, vol. 3/2009, 10 September 2009 (2009-09-10), pages 17-22, XP007916987,
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

Background



[0001] An electric power distribution system generally consists of a set of distribution substations, feeders, switches (circuit breakers, reclosers, etc.) electrical loads, and monitoring and control devices. A distribution system delivers electricity from distribution substations via feeders and switches to electrical loads (customers) that connect to feeders. Feeders in a distribution system are usually configured in a radial type to ensure that each load is supplied by only one distribution substation (source) via one feeder at any instant. To maintain the radial configuration, each feeder is linked by normally open (NO) tie-switches to neighboring feeders. The feeder section that is relatively near/far to its source is referred to as upstream/downstream section, while comparing to the feeder section that is relatively far/near to its source. One or more switches in the distribution system may have an associated intelligent electronic device (IED) that has the following monitoring and control functions: (1) measuring and recording electrical and other types of switch related quantities, such as voltage, current, reclosing times (2) monitoring the switch status, (3) operating the switch between open or close, and (4) communicating information to one or more master devices.

[0002] Distribution system reliability can be greatly improved by automating feeder operations such as fault detection, isolation, and load restoration. In such systems, IEDs associated with switches monitor the distribution system and communicate the corresponding information to the feeder automation master controllers located in substations. If a fault occurs in the distribution system, the master controller identifies the fault location, generates fault isolation and service restoration solutions in terms of a sequence of switch operations, and sends switching commands to IEDs to control corresponding switches.

[0003] An example distribution network is shown in Fig. 1, in a normal operation mode wherein loads are omitted for simplicity, sources (S1 to S7) are oval shaped, NO switches (5, 10, 13, 16, 19, 24, 29) are square shaped with a diagonal hatch pattern, and normally closed (NC) switches (other numbers) are square shaped with a vertical hatch pattern. If a fault occurs between switch 1 and 2, the protection function of switch 1 causes switch 1 to open, thereby causing the dashed line circled area to lose power. The boundary switches of the faulted portion of the distribution network include switch 1 and 2. Switch 2 is immediately downstream of the faulted section and thus, is the isolation switch. With reference to Fig. 2, when switches 1 and 2 are open, the faulted portion is isolated and a remaining unserved area is bound by isolation switch 2 and NO switches 5 and 16.

[0004] Switches 5 and 16 are referred to as first layer (Layer 1) restoration switches. If the sources, in this case S4 and S7, respectively, can provide the power to the area that is left unserved due to fault isolatin, a first layer restoration solution is possible. If the first layer sources are not capable of providing power to the unserved area, second or even third layer solutions must be performed to provide power to the unserved area. For example, the power sources for layer 2 are S2 and S5. The power sources for a third layer solution are S3 and S6. As should be evident, the second and third layer restoration switches are topologically more "distant" than the first layer restoration switches.

[0005] The process to obtain a restoration solution beyond layer 1 is called multi-layer (or multiple-layer) restoration service analysis (RSA). This is sometimes also referred to as a multi-tier service restoration problem, which may be classified as one type of network reconfiguration problems. Due to the potentially large number of switches involved in a multi-layer restoration solution, the process to obtain such solution is generally more challenging than a single-layer solution. With any network reconfiguration problem it is desirable to achieve: computational efficiency, avoiding network violations, minimized switching operations and radiality of the restored network topology.

[0006] Network reconfiguration problems may also seek to reduce the overall system loss and relieve overloading conditions in the network. Therefore, a network reconfiguration problem may either be formulated as a loss reduction optimization problem or a load balancing optimization problem. Under normal system operation conditions, network configuration allows the periodical transfer of load from heavily loaded portions of the distribution network to relatively lightly loaded ones, and thus takes advantage of the large degree of load diversity that exists in many distribution systems. Under abnormal system operation conditions, such as planned or forced system outages, the network reconfiguration problem becomes a service restoration problem, which is a special load balancing problem, where the main objective is to restore as many out-of-service loads as possible, without violating system operating and engineering constraints.

[0007] As with any other type of network reconfiguration problem, service restoration is a highly complex combinatorial, non-differential, and constrained optimization problem, due to the high number of switching elements in a distribution network, and the non-linear characteristics of the constraints used to model the electrical behavior of the system.

[0008] There is therefore a need in the art for a restoration switching analysis method that properly accounts for a greater number of variables and efficiently processes multiple-layer RSA solutions. In September 2009, Stoupis et al. (ABB Review, vol. 3/2009, pp. 17-22; ISSN:1013-3119) disclosed restoration switching analysis for achieving back-feed power restoration. A genetic algorithm generates multi-layer solutions, but no detail is disclosed. In 2002, Luan et al. (IEE Proc.-Gener. Transm. Distrib, vol. 149(2), pp. 145-151; DOI:10.1049/ip-gtd:20020095) disclosed a genetic algorithm for supply restoration in power system distribution networks. An "integer permutation" encoding scheme is used, wherein each integer is an index into a list of switches usable for restoring the supply, and a sequence represents an ordered operation sequence of switches generating a valid network topology. The fitness function is a weighted sum of five terms penalising solutions that would violate associated constraints.

Summary of the Invention



[0009] According to the present invention, a method is provided for determining a network configuration for back-feed power restoration to one or more out-of-service load areas in an electrical power distribution network after a fault has been isolated, wherein the network comprises a plurality of normally closed switches and a plurality of normally open tie switches, at least one of the normally closed switches has opened to isolate said fault, and the at least one normally closed switch that opened to isolate the fault and at least one of said normally open tie switches collectively create the one or more out-of-service load areas. The method includes defining a chromosome architecture for creating a plurality of unique chromosomes, each said chromosome being a string of characters, each said character corresponding to one of said normally open tie switches and representing one of a list of actions for said corresponding normally open tie switch, said list of actions including (1) said corresponding normally open tie switch remaining open, (2) closing said corresponding normally open tie switch, or (3) opening one of said normally closed switches that is identified by said character and closing said corresponding normally open tie switch. A chromosome list is initialized. The chromosome architecture is used to create initial chromosomes for a first chromosome generation, wherein each chromosome corresponds to a candidate network configuration. A predetermined validity criteria is applied to determine whether each created initial chromosome is valid, wherein said predetermined validity criteria indicates that any chromosome creating an energized loop within said network is not valid, and wherein each new chromosome determined to be not valid is discarded. For each initial chromosome created that is valid and not already in the chromosome list, it is added to the chromosome list. Fitness function values are generated for each chromosome in the initial chromosome population, wherein the fitness function comprises

where idv is an index of a candidate network configuration corresponding to a chromosome; and for the candidate network configuration: Ivionr is a normalized number of back-path current violations in the candidate network configuration; SwOpnr is the normalized switching operations for the candidate network configuration; Pnr is the normalized total unserved load for the candidate network configuration; and wlvio, wSw, wp are user definable weighting factors, and wherein



where base represents a base network configuration corresponding to a network configuration after said at least one of said normally closed switches has opened to isolate said fault; NoTieSwitchesbase is the total number of normally open tie switches in said network; NoSwitchOperations is the number of switch operations, excluding isolation switching operations, needed for a transition from said base network configuration to said candidate network configuration; TotalWeightedLoadbase is a weighted total load of said base network configuration; TotalWeightedLoad(idv) is a weighted total load of the candidate network configuration, and wherein:

where Ivio(idv) is the number of current violations for a candidate network configuration and Iviomax is the largest number of current violations for the candidate network configuration corresponding to any chromosome in a generation. The chromosomes are then sorted by fitness function value, where the chromosome having the lowest fitness function value is considered the best candidate chromosome. If the fitness function value of the best candidate chromosome is below a threshold fitness value the network configuration corresponding to the best candidate chromosome is output. If the fitness function value of the best candidate chromosome is not below the threshold fitness value, genetic manipulation is used to create new chromosomes for a new chromosome population. During the creation of the new chromosome population, each new chromosome in the new generation that is not already in the chromosome list is added to the chromosome list. Fitness function values are again generated for each chromosome in the new chromosome generation. The generation of new populations are repeated in the same manner until the fitness function value of the best candidate chromosome is below the threshold fitness value or until a predetermined number of new populations are created, whereupon a network configuration corresponding to the best candidate chromosome is output.

Brief Description of the Drawings



[0010] 

Figure 1 is a schematic drawing of an exemplary distribution network.

Figure 2 is a schematic drawing of the exemplary distribution network of Fig. 1 wherein a simulated fault is isolated and restoration layers are shown.

Figure 3 is a schematic drawing of a second exemplary distribution network.

Figure 4 is a schematic drawing of the exemplary distribution network of Fig. 3 wherein a fault occurs.

Figure 5 is a schematic drawing of the exemplary distribution network of Fig. 3 wherein the fault is isolated leaving two out-of-service areas.

Fig. 6 is a flow-chart showing an exemplary algorithm of the present invention.

Fig. 7 is a flow-chart showing the step of creating a new generation in greater detail.

Fig. 8 is an exemplary user interface displaying the results of the method.

Fig. 9 is an exemplary user interface according to Fig. 8 showing a different set of exemplary results.


Detailed Description of the Invention



[0011] The methods of the present invention are applicable to both single layer and multi-layer RSA problems. With reference to Fig. 3, an example distribution network is shown in a normal operating topology. Fig. 4 shows the network of Fig. 3 wherein a fault occurred at load LDff-3. As can be seen, in such a scenario upstream recloser RBW85 senses the fault, proceeds through a reclosing sequence, and finally locks-out (remains open). This results in an outage area including loads LDff-1, LDff-2, LDff-3, LDff-5, LDff-6 and LDff-7.

[0012] In a next step, the faulted load is isolated from the out-of-service area. With reference to Fig. 5, in the example network, the isolation switches include the immediately downstream switches SW-ifanet-1 and SW-ifanet-2. Once switches SW-ifanet-1 and SW-ifanet-2 are opened, power restoration to the out-of-service area can begin.

[0013] In the present example, two separate out of service areas (OOSA) are formed after fault isolation. OOSA1 includes LDff-5 and OOSA2 includes LDff-1, LDff-2, LDff-6 and LDff-7. Each OOSA can be back-fed by one or more (if the OOSA is separable into different sub-areas) of the remaining back-feed sources. The restoration problem is complex because of the large number of back-feed sourcing combinations. For example, in the network of Fig. 5, source PDN16 may be a single-layer back-feed source for OOSA2 and/or a second-layer back-feed source for OOSA1. Depending on the final network configuration, source 54Z66 could be a single-layer or second-layer back-feed source for OOSA2. The restoration problem is further complicated because the OOSA could be partitioned (by opening switches RC145, SW-ifanet-3 or SW-ifnet-4).

[0014] Solving for a network configuration that back-feeds the OOSAs first requires that a chromosome architecture of a genetic algorithm (GA) be developed. A chromosome represents a particular network topology and is represented by a string of fixed length. The string length is determined by the number of normally open (NO) switches (also known as tie switches) in the pre-fault network. Each character in the chromosome (also referred to as a bit or gene) represents the application options with respect to that tie switch. For each tie switch, three options are available: (1) the tie switch can remain unchanged and stay open; (2) the tie switch can be closed; or (3) the tie switch may be swapped with another normally closed (NC) switch. According to the third option, "swapping" means that the tie switch would close and the identified NC switch would open. A swap switch is an NC switch that is either, upstream of the tie switch or is an NC switch in the OOSA adjacent thereto.

[0015] Each tie switch has a specific set of corresponding swap switches. The set of swap switches may be determined by running a network tracing from the tie switch to its respective source(s). This is otherwise known as back-path tracing. For each of the first-layer tie switches, only one source can be traced to. For each of the higher layer switches, two sources can be traced to. In addition to the back-traced switches, for a tie switch that bounds an OOSA, the set of corresponding swap switches also includes the NC switches in that OOSA.

[0016] In order to simplify the restoration problem solution, the feeder breakers may be excluded from the listing of available swap switches. The listing of available swap switches for a given tie switch may be further limited by limiting the number of switches that are swappable in the back-feed path. For example, for a given tie switch, the available back-feed swap switches may be limited to the first N number of upstream NC switches.

[0017] For the example of Fig. 5, the first layer tie switches are: RCR03, BQ70BL, RC137, RC871, and RBX58. The second layer tie switches are: RCF91 and RCP75. If the number of back-feed switches to be considered is limited to N=1, the respective swap switch sets for the first and second layer tie switches are:
RCR03 - RC145, SW-ifanet-3, SW-ifanet-4, 2WK02 (first layer)
BQ70BL - RC145, SW-ifanet-3, SW-ifanet-4, BP52BL (first layer)
RC137 - RC145, SW-ifanet-3, SW-ifanet-4, RCF65 (first layer)
RC871 - RC145, SW-ifanet-3, SW-ifanet-4, RD618 (first layer)
RBX58 - 13B670 (first layer)
RCF91 - BS73BL, REU99 (second layer)
RCP75 - RBP32, 13B670 (second layer)


[0018] The gene representation of the swap switch sets uses a single character to represent the action for each tie switch. An ASCII character may be used to represent each swap switch or other action (i.e. open or close). Thus, a space character (hereinafter represented as "L" for clarity) (ASCII value 32) may represent "remain open" and character"!" (ASCII value 33) may represent the "close" action. Characters having an ASCII value higher than 33 may represent the "swap status with swapped switch N." Because the ASCII characters are used, a large number of representative variables are available (255-33=222 unique switch swaps).

[0019] To create the chromosomes, the tie switches are arranged in any order. According to the example network of Fig. 5, the tie switches are arranged in the following manner: [RCR03, BQ70BL, RC137, RCF91, RC871, RBX58, RCP75]. The swapped switches are associated with a character in the following manner:
" - RC145
# - SW-ifanet-3
$ - SW-ifanet-4
% - 2WK02
& - BP52BL
' - RCF65
( - RD618
) - 13B670
* - BS73BL
+ - REU99
, - RBP32


[0020] Valid chromosomes include: [! L L L L ! L] which correlates to a network where RCR03 and RBX58 are closed and all other tie switches remain open; [L!L*L!L] which correlates to a network where BQ70BL and RBX58 are closed and RCF91 and BS73BL are swapped (i.e. BS73BL closes and RCF91 opens); [!LLLL)!] which correlates to a network where RCR03 and RCP75 are closed and RBX58 is swapped with 1 3B670 (i.e. RBX58 closes and 1 3B670 opens).

[0021] As will be described later in greater detail, for each candidate network, it's suitability for a final solution is judged by calculating a fitness function value. A fitness function is defined below, wherein the objective is to minimize the value of this function:



[0022] Where idv is the index of the individual network topology (hereafter the candidate system) to be evaluated (corresponds to the chromosomes generated in the genetic algorithm below). Ivionr is the normalized number of back-path current violations in the candidate network, SwOpnr is the normalized switching operations (excluding isolation switching) and Pnr is the normalized total unserved load for the candidate network. The corresponding weighting factors are represented by wlvio, wSw, wp. The weighting factor definition in Eq.1 allows the users of this algorithm to place an emphasis on different optimization variables, thereby increasing the application flexibility. It should be appreciated that additional factors may be considered in a fitness function, such as, for example, minimizing system losses or voltage violations.

[0023] Pnr reduces faster than the other two components of the fitness function, thus after a few generations the contributions from SwOpnr and Ivionr would dominate the final fitness values. In order to achieve a balanced contribution solution, the weight wp may be increased by a factor larger than 1 (for example 1 .05) for each new generation of chromosomes.

[0024] The calculation of the normalized values is defined according to the following:





[0025] Where base represents the network in the post-isolation configuration and (idv) represents the candidate network. NoTieSwitchesbase is the total number of NO tie switches in the network. NoSwitchOperations is the number of switch operations (excluding isolation switching operations) needed for the transition from the fault isolated system state (the base configuration) to the candidate network configuration. TotalWeightedLoad is the weighted total load of the candidate network corresponding to a chromosome:



[0026] As the objective is to minimize the fitness function value, the PaneltyFactor of Eq. 4 is advantageously greater than 1, to prevent load shedding in the back-feeding areas. The calculation of the normalized current violations Ivionr is performed according to the following: for each chromosome in a generation, the corresponding network is created and a balanced or unbalanced load flow analysis is conducted. From the results of the load flow analysis, the current violations from the closed switches and all the feeder sections to their sources are counted. These current violations are referred to as the back-path current violations. Of all the chromosomes in a generation, the network having the largest number of current violations Iviomax is determined, and that number is used according to:



[0027] With reference now to Fig. 6, a flow-chart shows the process according to an example. As discussed above, the process begins after a fault is isolated on a network. Thereafter, the process is evoked at 100. At step 100 new chromosomes are randomly created in accordance with the methods discussed above. Thus, chromosomes are created wherein each NO tie switch is either left open, closed, or swapped with a NC switch. Concurrently, the new chromosomes are validated to ensure no energized loops are created. If so, the chromosome is discarded. A Reactive Tabu Search (RTS) is incorporated, by defining a search list LSTRTS that contains the chromosomes or indices of each candidate network. Thus, each newly created chromosome (whether valid or not) is added to the LSTRTS.

[0028] To limit the genetic algorithm processing time, a population size is defined for each generation. The population size = 4N+1; where N=NoTieSwitchesbase/2 (N is an integer) and where N<=10. Thus, for the example network of Fig. 5, 7 normally open tie switches are included in the base network, giving a population size of 13 (where N is rounded from 3.5 to 3).

[0029] At step 102, it is determined whether the number of valid chromosomes has reached the population limit. New chromosomes are generated until the population number is reached. At step 104, a fitness value (see eq. 1) is calculated for each chromosome in the generation. As discussed above, for the population of chromosomes, for each chromosome, the corresponding network is determined and a balanced or unbalanced load flow analysis is conducted to determine current violations and total weighted load which is used in the fitness function value calculation.

[0030] At step 106 the chromosomes are sorted in order of fitness value, wherein the lowest fitness value is "best" or 1st. At step 108 it is determined whether either the generation limit is reached or if the best fitness value is below a threshold value. If either case is true, the network configuration corresponding to the best chromosome is output as the solution. The solution may be sent to a DMS or other distribution control manager for implementation to restore power to some or all of the OOSA. The generation limit is a user defined (or default) maximum generation number that effectively limits the iterations the algorithm proceeds through. The fitness function threshold value is a user defined (or default) value for the fitness function, wherein network configurations having a fitness function value below the threshold are "good" solutions. This is because a low fitness function value indicates; (1) minimized current violations, (2) minimized switching operations and (3) minimized unserved load.

[0031] If the best chromosome does not have a fitness value beneath the threshold value and the max generation number has not yet been reached, the algorithm proceeds to step 110. At step 110 a new generation of chromosomes are created using genetic manipulations. The best N+1 chromosomes (chromosomes with the lowest fitness function values) from the previous generation are directly copied into the new generation. Next, crossover is used to create N new chromosomes. Next, mutation is used to create N new chromosomes. As is known in the art, a crossover is accomplished by merging any two generated chromosomes randomly and mutations are accomplished by randomly redoing one or more of the swappings or open/close actions in a chromosome. The additional chromosomes needed to complete the generation are created by random chromosome creation. Each newly created chromosome (not the direct copies) is checked for validity and against LSTrts. If the chromosome is not valid (energized loops), it is added to the LSTrts and is not added to the new generation. If the chromosome is already in LSTrts the chromosome is not added to the new generation. If the chromosome is valid and not in the LSTrts, it is added to the new generation and to the LSTrts.

[0032] The cross-over is performed using the topmost (lowest fitness function value) chromosome with the next N chromosomes to generate N new chromosomes. Next, the direct copied N chromosomes each mutate once to generate the next N chromosomes.

[0033] Once the new generation of chromosomes is created, the process returns to step 104, and in this manner, the process repeats until a good solution (below the fitness function threshold) is reached or until the maximum generation number is reached.

[0034] With reference now to Fig. 7, a flow-chart shows the step 110 of creating a new generation in greater detail. At 200 the mutation function begins by initializing the NoMutation variable to zero. Next, at 202 a chromosome mutation is performed to create a NewChromosome and the NoMutation variable is increased by one. At 204 it is determined whether the NewChromosome is in the RTSmap (i.e. LSTrts). If so, the algorithm proceeds to step 212 where it is determined whether NoMutation variable is greater than NmutationLimit. If so, the algorithm proceeds to step 214, which will be described in greater detail below. If not, the algorithm proceeds back to step 202. If, at 204 it is determined that the NewChromosome is not in the RTSmap, the algorithm proceeds to step 206 where the NewChromosome is saved in the RTSmap. Next, at 208 a network validation check is performed on the NewChromosome to identify energized loops (i.e. invalid network configurations). If the NewChromosome is valid it is added to the new generation at 246. If the NewChromosome is not valid, the algorithm proceeds to 212, where it is determined whether NoMutation is greater than NmutationLimit.

[0035] The crossover function begins at 216 by initializing the NoBreed variable to zero. Next, at 218 a chromosome crossover is performed to create a NewChromosome and the NoBreed variable is increased by one. At 220 it is determined whether the NewChromosome is in the RTSmap (i.e. LSTrts). If so, the algorithm proceeds to step 228 where it is determined whether NoBreed variable is greater than NbreedLimit. If so, the algorithm proceeds to step 214, which will be described in greater detail below. If not, the algorithm proceeds back to step 218. If, at 220 it is determined that the NewChromosome is not in the RTSmap, the algorithm proceeds to step 222 where the NewChromosome is saved in the RTSmap. Next, at 224 a network validation check is performed on the NewChromosome to identify energized loops (i.e. invalid network configurations). If the NewChromosome is valid it is added to the new generation at 246. If the NewChromosome is not valid, the algorithm proceeds to 228, where it is determined whether NoBreed is greater than NbreedLimit.

[0036] As can be seen, if either the crossover or the mutation functions reach their limit, the algorithm proceeds to step 214, where the NoRandom variable is initialized to zero. The algorithm then proceeds to step 230 where the random chromosome generation function begins. The random chromosome generation function may also be called independently, in which case the process would begin at 230. Next, at 232 a chromosome is created by randomly closing or swapping one or more NO tie switches. The NoRandom variable is increased by one. At 234 it is determined whether the NewChromosome is in the RTSmap (i.e. LSTrts). If so, the algorithm proceeds to step 242 where it is determined whether NoRandom variable is greater than NrandomLimit. If so, the algorithm proceeds to step 244, which returns a null (meaning no additional chromosome can be generated). If not, the algorithm proceeds back to step 232. If, at 234 it is determined that the NewChromosome is not in the RTSmap, the algorithm proceeds to step 236 where the NewChromosome is saved in the RTSmap. Next, at 238 a network validation check is performed on the NewChromosome to identify energized loops (i.e. invalid network configurations). If the NewChromosome is valid it is added to the new generation at 246. If the NewChromosome is not valid, the algorithm proceeds to 242, where it is determined whether NoRandom is greater than NrandomLimit.

[0037] Thus, for each generation, N crossover calls, N mutation calls and at least N random calls will be made. The ultimate number of random calls may be higher depending on the results of the crossover and mutation calls. (i.e. if NoMutation or NoBreed exceeds NmutationLimit or NbreedLimit respectively.)

[0038] With reference now to Figs. 8 and 9, an exemplary graphical user interface is shown. The interface shown displays the results of the algorithms discussed above. As can be seen, the "Restoration Switching Plans" area displays the final generation, wherein the chromosomes are sorted in order of fitness value (lowest being first). As can also be seen, when a chromosome is highlighted, (chromosome 1 in Fig. 8 and chromosome 3 in Fig. 9) the corresponding switching sequence is displayed in the "Switching Sequences for the Selected Restoration Switching Plan" area. As can be seen, for completeness and clarity, the switching sequences provided in the "Switching Sequences for the Selected Restoration Switching Plan" includes the switching operations for both isolation (the determination of which is not the subject of the present invention) and restoration, which is determining in the manner discussed above.

[0039] The method requires network load flow calculations (in order to evaluate the fitness function) only for the valid chromosomes in question, as opposed to many more load flow calculations if approaches such as classical genetic algorithm, network tracing or deterministic optimization methods are used. This increases the speed of solution finding, thus making it appropriate for real-time restoration switching applications. It is especially practical for multi-layer RSA, when the network topology is complex (for example, many tie switches between adjacent feeders) and many different alternatives for back-feed restoration exist. The functionality of multi-layer RSA would reside at either at DMS or in a sub-station. Thus, the network configuration (solution) is generated by the DMS or sub-station control system. The DMS or sub-station control system then implements the switching operations to convert the network topology to match the network configuration solution. In this manner, power is restored to the OOSAs.

[0040] As will be appreciated by one of ordinary skill in the art, the present invention may be embodied as or take the form of the method and system previously described, as well as of a computer readable medium having computer-readable instructions stored thereon which, when executed by a processor, carry out the operations of the present inventions as previously described. The computer-readable medium may be any medium that can contain, store, communicate, propagate, or transport the user-interface program instruction for use by or in connection with the instruction execution system, apparatus, or device and may by way of example but without limitation, be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium or other suitable medium upon which the program is printed. More specific examples (a non-exhaustive list) of the computer-readable medium would include: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a transmission media such as those supporting the Internet or an intranet, or a magnetic storage device. Computer program code or instructions for carrying out operations of the present invention may be written in any suitable programming language provided it allows achieving the previously described technical results.


Claims

1. A method for determining a network configuration for back-feed power restoration to one or more out-of-service load areas in an electrical power distribution network after a fault has been isolated in the network, wherein said network comprises a plurality of normally closed switches and a plurality of normally open tie switches, at least one of said normally closed switches has opened to isolate said fault, and the at least one normally closed switch that opened to isolate said fault and at least one of said normally open tie switches collectively create said one or more out-of-service load areas, the method comprising:

i. defining a chromosome architecture for creating a plurality of unique chromosomes, each said chromosome being a string of characters, each said character corresponding to one of said normally open tie switches and representing one of a list of actions for said corresponding normally open tie switch, said list of actions including (1) said corresponding normally open tie switch remaining open, (2) closing said corresponding normally open tie switch, or (3) opening one of said normally closed switches that is identified by said character and closing said corresponding normally open tie switch;

ii. initializing a chromosome list;

iii. creating (100) initial chromosomes for a first chromosome generation using said chromosome architecture, wherein each chromosome corresponds to a candidate network configuration;

iv. applying a predetermined validity criteria to determine whether each created initial chromosome is valid, wherein said predetermined validity criteria indicates that any chromosome creating an energized loop within said network is not valid, and wherein each new chromosome determined to be not valid is discarded;

v. adding (102) to said chromosome list each created initial chromosome that is valid and not already in said chromosome list;

vi. generating (104) a fitness value for the corresponding candidate network configuration for each chromosome in said first chromosome generation using a fitness function, wherein the fitness function comprises:

where idv is an index of a candidate network configuration corresponding to a chromosome; and for the candidate network configuration: Ivionr is a normalized number of back-path current violations in the candidate network configuration; SwOpnr is a normalized number of switching operations needed for transition to the candidate network configuration, excluding isolation switching; Pnr is a normalized total unserved load for the candidate network configuration; and wlvio, wSw, wp are user definable weighting factors;

vii. sorting (106) said chromosomes by their corresponding fitness values and identifying the chromosome having the lowest fitness value as the best candidate chromosome;

viii. determining (108) if the fitness value of said best candidate chromosome is below a threshold fitness value, and if so, outputting the candidate network configuration corresponding to said best candidate chromosome;

ix. using genetic manipulation to create (110) new chromosomes for a new chromosome generation if said fitness value of said best candidate chromosome is not below said threshold fitness value, and adding (246) to said chromosome list each new chromosome in said new chromosome generation that is not already in said chromosome list;

x. generating (104) a fitness value for the corresponding candidate network configuration for each new chromosome in said new chromosome generation using said fitness function;

xi. repeating steps vii-x until the fitness value for said best candidate chromosome is below said threshold fitness value or until a predetermined number of new chromosome generations have been created (108), whereupon the candidate network configuration corresponding to said best candidate chromosome is output.


 
2. The method according to claim 1 wherein weighting value wp is increased by a factor larger than 1 for each new chromosome generation.
 
3. The method of claim 1 wherein:



where base represents a base network configuration corresponding to a network configuration after said at least one of said normally closed switches has opened to isolate said fault and (idv) represents the candidate network configuration; NoTieSwitchesbase is the total number of normally open tie switches in said network; NoSwitchOperations is the number of switch operations, excluding isolation switching operations, needed for a transition from said base network configuration to said candidate network configuration; TotalWeightedLoadbase is a weighted total load of said base network configuration; TotalWeightedLoad(idv) is a weighted total load of the candidate network configuration.
 
4. The method of claim 3 wherein:

where LoadinOOSAreas is the load in the out-of-service load areas of said network and LoadinBackFeedingAreas is the remaining load in the non-isolated or faulted portions of said network; the PaneltyFactor being greater than 1.
 
5. The method of claim 1 wherein:

where Ivio(idv) is the number of current violations for a candidate network configuration and Iviomax is the largest number of current violations for the candidate network configuration corresponding to any chromosome in a generation.
 
6. The method of claim 1 wherein said step of creating initial chromosomes includes, for each said normally open tie switch, randomly selecting one of said list of actions.
 
7. The method of claim 1 wherein, for each said normally open tie switch, said normally closed switch identified by said character is upstream of said normally open tie switch and, if said normally open tie switch bounds one of said out-of-service load areas, said normally closed switch identified by said character is also in that out-of-service load area.
 
8. The method of claim 1 wherein said characters are ASCII characters.
 
9. The method of claim 3 wherein each of said first chromosome generation and said new chromosome generations have a population size equaling 4N+1, where N=NoTieSwitchesbase/2 and where N is an integer.
 
10. The method of claim 1 wherein said step of using genetic manipulation to create new chromosomes for a new chromosome generation further comprises:

directly copying (110) a plurality of chromosomes from a previous chromosome generation to the new chromosome generation; and

performing crossovers and mutations (110) on said plurality of chromosomes from said previous chromosome generation to the new chromosome generation.


 
11. The method of claim 1 wherein said step of using genetic manipulation to create new chromosomes for a new chromosome generation further comprises:

adding (110) the top N chromosomes having the N lowest fitness values from the previous generation to the new generation unchanged;

crossing-over (110) the best candidate chromosome with the next N chromosomes to generate N new chromosomes;

mutating (110) the next N chromosomes after the top N to create N new chromosomes in the new generation; and

if additional new chromosomes are required to fill said new generation, randomly selecting (110) one of said list of actions for each said normally open tie switch.


 
12. The method of claim 8 wherein said ASCII characters representing said action of opening one of said normally closed switches that is identified by said character and closing said corresponding normally open tie switch have an ASCII value higher than 33.
 


Ansprüche

1. Verfahren zum Bestimmen einer Netzwerkkonfiguration für die Rückspeisungs-Spannungswiederherstellung einer oder mehrerer außer Betrieb befindlicher Lastbereiche in einem Stromverteilungsnetz, nachdem ein Fehler im Netzwerk isoliert wurde, wobei das Netzwerk mehrere, normalerweise geschlossene Schalter und mehrere, normalerweise offene Kuppelschalter umfasst, wobei mindestens einer der normalerweise geschlossenen Schalter geöffnet wurde, um den Fehler zu isolieren, und wobei der mindestens eine normalerweise geschlossene Schalter, der geöffnet wurde, um den Fehler zu isolieren, und mindestens einer der normalerweise offenen Kuppelschalter gemeinsam den einen oder die mehreren, außer Betrieb befindlichen Lastbereiche schaffen, wobei das Verfahren umfasst:

i. Definieren einer Chromosomenarchitektur zum Erzeugen mehrerer eindeutiger Chromosomen, wobei jedes Chromosom eine Zeichenfolge ist, wobei jedes Zeichen einem der normalerweise offenen Kuppelschalter entspricht und eine aus einer Liste von Aktionen für den entsprechenden, normalerweise offenen Kuppelschalter darstellt, wobei die Liste von Aktionen (1) das Offenbleiben des entsprechenden, normalerweise offenen Kuppelschalters, (2) das Schließen des entsprechenden, normalerweise offenen Kuppelschalters oder (3) das Öffnen eines der normalerweise geschlossenen Schalter, der durch das Zeichen identifiziert ist, und das Schließen des entsprechenden, normalerweise offenen Kuppelschalters enthält;

ii. Initialisieren einer Chromosomenliste;

iii. Erzeugen (100) anfänglicher Chromosomen für eine erste Chromosomengeneration unter Verwendung der Chromosomenarchitektur, wobei jedes Chromosom einer in Frage kommenden Netzwerkkonfiguration entspricht;

iv. Anwenden eines vorgegebenen Gültigkeitskriteriums, um zu bestimmen, ob jedes erzeugte anfängliche Chromosom gültig ist, wobei das vorgegebene Gültigkeitskriterium anzeigt, dass jedes Chromosom, das eine angeregte Schleife innerhalb des Netzwerks erzeugt, nicht gültig ist, und wobei jedes neue Chromosom, das als ungültig bestimmt wurde, verworfen wird;

v. Hinzufügen (102) jedes erzeugten anfänglichen Chromosoms, das gültig ist und nicht bereits in der Chromosomenliste enthalten ist, zu der Chromosomenliste;

vi. Erzeugen (104) eines Fitnesswerts für die entsprechende, in Frage kommende Netzwerkkonfiguration für jedes Chromosom in der ersten Chromosomengeneration unter Verwendung einer Fitnessfunktion, wobei die Fitnessfunktion umfasst:

wobei idv ein Index einer in Frage kommenden Netzwerkkonfiguration in Entsprechung zu einem Chromosom ist; und
für die in Frage kommende Netzwerkkonfiguration:

Ivionr eine normalisierte Anzahl von Rückwegstromverstößen in der in Frage kommenden Netzwerkkonfiguration ist;

SwOpnr eine normalisierte Anzahl von Schaltvorgängen ist, die für den Übergang zur in Frage kommenden Netzwerkkonfiguration erforderlich sind, ausgenommen Isolationsschaltung;

Pnr eine normalisierte, nicht bereitgestellte Gesamtlast für die in Frage kommende Netzwerkkonfiguration ist; und

wlvio, wSw, wp benutzerdefinierbare Gewichtungsfaktoren sind;

vii. Sortieren (106) der Chromosomen nach ihren entsprechenden Fitnesswerten und Identifizieren des Chromosoms mit dem niedrigsten Fitnesswert als besten Chromosomkandidaten;

viii. Bestimmen (108), ob der Fitnesswert des besten Chromosomkandidaten unter einem Fitness-Schwellenwert liegt, und wenn ja, Ausgeben der in Frage kommenden Netzwerkkonfiguration, die dem besten Chromosomkandidaten entspricht;

ix. Verwenden genetischer Manipulation, um neue Chromosomen für eine neue Chromosomengeneration zu erzeugen (110), wenn der Fitnesswert des besten Chromosomkandidaten nicht unter dem Fitness-Schwellenwert liegt, und Hinzufügen (246) jedes neuen Chromosoms in der neuen Chromosomengeneration, das noch nicht in der Chromosomenliste enthalten ist, zu der Chromosomenliste;

x. Erzeugen (104) eines Fitnesswerts für die entsprechende in Frage kommende Netzwerkkonfiguration für jedes neue Chromosom in der neuen Chromosomengeneration unter Verwendung der Fitnessfunktion;

xi. Wiederholen der Schritte vii-x, bis der Fitnesswert für das beste Kandidatenchromosom unter dem Fitness-Schwellenwert liegt oder bis eine vorgegebene Anzahl neuer Chromosomengenerationen erzeugt wurde (108), woraufhin die in Frage kommende Netzwerkkonfiguration ausgegeben wird, die dem besten Chromosomkandidaten entspricht.


 
2. Verfahren nach Anspruch 1, wobei der Gewichtungswert wp um einen Faktor größer als 1 für jede neue Chromosomengeneration erhöht wird.
 
3. Verfahren nach Anspruch 1, wobei:



wobei base eine Basis-Netzwerkkonfiguration darstellt, die einer Netzwerkkonfiguration entspricht, nachdem der mindestens eine der normalerweise geschlossenen Schalter geöffnet wurde, um den Fehler zu isolieren, und (idv) die in Frage kommende Netzwerkkonfiguration darstellt;

NoTieSwitchesbase die Gesamtzahl der normalerweise offenen Kuppelschalter in diesem Netzwerk ist;

NoSwitchOperations die Anzahl von Schaltvorgängen ist, ausgenommen Isolations-Schaltvorgänge, die für einen Übergang von der Basisnetzwerkkonfiguration zur in Frage kommenden Netzwerkkonfiguration erforderlich sind;

TotalWeightedLoadbase eine gewichtete Gesamtlast der Basisnetzwerkkonfiguration ist;

TotalWeightedLoad (idv) eine gewichtete Gesamtlast der in Frage kommenden Netzwerkkonfiguration ist.


 
4. Verfahren nach Anspruch 3, wobei:

wobei LoadinOOSAreas die Last in den außer Betrieb befindlichen Lastbereichen des Netzwerks ist und LoadinBackFeedingAreas die restliche Last in den nicht isolierten oder fehlerhaften Teilen des Netzwerks ist;

wobei der paneltyFactor größer als 1 ist.


 
5. Verfahren nach Anspruch 1, wobei:

wobei Ivio(idv) die Anzahl der aktuellen Verstöße für eine in Frage kommende Netzwerkkonfiguration und Iviomax die größte Anzahl der aktuellen Verstöße für die in Frage kommende Netzwerkkonfiguration mit Bezug auf ein Chromosom in einer Generation ist.
 
6. Verfahren nach Anspruch 1, wobei der Schritt der Schaffung anfänglicher Chromosomen für jeden normalerweise offenen Kuppelschalter das zufällige Auswählen einer aus der Liste von Aktionen enthält.
 
7. Verfahren nach Anspruch 1, wobei für jeden normalerweise offenen Kupplungsschalter der durch das Zeichen identifizierte, normalerweise geschlossene Schalter stromaufwärts des normalerweise offenen Kuppelschalters ist, und, wenn der normalerweise offene Kuppelschalter an eine der außer Betrieb befindlichen Lastbereiche grenzt, der von dem Zeichen identifizierte, normalerweise geschlossene Schalter sich ebenfalls in diesem außer Betrieb befindlichen Lastbereich befindet.
 
8. Verfahren nach Anspruch 1, wobei die Zeichen ASCII-Zeichen sind.
 
9. Verfahren nach Anspruch 3, wobei jede aus der ersten Chromosomengeneration und den neuen Chromosomengenerationen eine Populationsgröße gleich 4N+1 aufweisen, wobei N=NoTieSwitchesbase/2 und wobei N eine ganze Zahl ist.
 
10. Verfahren nach Anspruch 1, wobei der Schritt der Verwendung genetischer Manipulation zum Erzeugen neuer Chromosomen für eine neue Chromosomengeneration ferner umfasst:

direktes Kopieren (110) mehrerer Chromosomen von einer früheren Chromosomengeneration in die neue Chromosomengeneration; und

Durchführen von Überkreuzungen und Mutationen (110) an den mehreren Chromosomen von der früheren Chromosomengeneration zur neuen Chromosomengeneration.


 
11. Verfahren nach Anspruch 1, wobei der Schritt der Verwendung genetischer Manipulation zum Erzeugen neuer Chromosomen für eine neue Chromosomengeneration ferner umfasst:

Hinzufügen (110) der obersten N Chromosomen mit den N niedrigsten Fitnesswerten von der früheren Generation unverändert zur neuen Generation;

Überkreuzen (110) des besten Chromosomenkandidaten mit den nächsten N Chromosomen, um N neue Chromosomen zu erzeugen;

Mutieren (110) der nächsten N Chromosomen nach den obersten N, um N neue Chromosomen in der neuen Generation zu erzeugen; und,

wenn zusätzliche neue Chromosomen erforderlich sind, um die neue Generation zu füllen, zufälliges Auswählen (110) einer aus der Liste von Aktionen für jeden normalerweise offenen Kuppelschalter.


 
12. Verfahren nach Anspruch 8, wobei die ASCII-Zeichen, die die Aktion des Öffnens eines der normalerweise geschlossenen Schalter, der durch das Zeichen identifiziert wird, und des Schließens des entsprechenden, normalerweise offenen Kuppelschalters darstellen, einen ASCII-Wert größer als 33 aufweisen.
 


Revendications

1. Procédé permettant de déterminer une configuration de réseau pour le rétablissement de l'énergie de rétro-alimentation d'une ou de plusieurs zones de charge hors service dans un réseau de distribution d'énergie électrique après isolement d'un dysfonctionnement dans le réseau, dans lequel ledit réseau comprend une pluralité de commutateurs normalement fermés et une pluralité de commutateurs de connexion normalement ouverts, au moins l'un desdits commutateurs normalement fermés s'est ouvert pour isoler ledit dysfonctionnement, et l'au moins un commutateur normalement fermé qui s'est ouvert pour isoler ledit dysfonctionnement et au moins l'un desdits commutateurs de sectionnement normalement ouverts crée collectivement lesdites une ou plusieurs zones de charge hors service, le procédé comprenant :

i. l'établissement d'une architecture chromosomique dans le but de créer une pluralité de chromosomes uniques, chacun desdits chromosomes étant une chaîne de caractères, chacun desdits caractères correspondant à l'un desdits commutateurs de sectionnement normalement ouverts et représentant l'une d'une liste d'actions pour ledit commutateur de sectionnement normalement ouvert correspondant, ladite liste d'actions comprenant (1) ledit commutateur de sectionnement normalement ouvert correspondant restant ouvert, (2) la fermeture dudit commutateur de sectionnement normalement ouvert correspondant, ou (3) l'ouverture de l'un desdits commutateurs normalement fermés qui est identifié par ledit caractère et la fermeture dudit commutateur de sectionnement normalement ouvert correspondant ;

ii. l'initialisation d'une liste de chromosomes ;

iii. la création (100) des chromosomes initiaux pour une première génération de chromosomes en utilisant ladite architecture chromosomique, dans laquelle chaque chromosome correspond à une configuration de réseau candidate ;

iv. l'application d'un critère de validité prédéfini pour déterminer si chaque chromosome initial créé est valide, dans lequel lesdits critères de validité prédéfinis indiquent que tout chromosome créant une boucle sous tension dans ledit réseau n'est pas valide, et dans lequel chaque nouveau chromosome déterminé comme non valide est rejeté ;

v. l'adjonction (102) à ladite liste de chromosomes de chaque chromosome initial créé qui est valide et qui ne se trouve pas déjà dans ladite liste de chromosomes ;

vi. la génération (104) d'une valeur d'aptitude pour la configuration de réseau candidate correspondante pour chaque chromosome dans ladite première génération de chromosomes en utilisant une fonction d'aptitude, dans lequel la fonction d'aptitude comprend :

où idv représente un index d'une configuration de réseau candidate correspondant à un chromosome ; et
pour la configuration de réseau candidate :

Ivionr représente un nombre normalisé de violations de courant de réalimentation dans la configuration de réseau candidate ;

SwOpnr représente un nombre normalisé d'opérations de commutation nécessaires pour la transition vers la configuration de réseau candidate, à l'exclusion de la commutation d'isolement ;

Pnr représente une charge totale non desservie normalisée pour la configuration de réseau candidate ; et

wlvio, wSw, wp représentent des facteurs de pondération pouvant être définis par l'utilisateur ;

vii. le triage (106) desdits chromosomes par leurs valeurs d'aptitude correspondantes et l'identification du chromosome ayant la valeur d'aptitude la plus basse comme meilleur chromosome candidat ;

viii. la détermination (108) du fait que la valeur d'aptitude dudit meilleur chromosome candidat est inférieure à une valeur d'aptitude seuil, et si c'est le cas, l'émission de la configuration de réseau candidate correspondant audit meilleur chromosome candidat ;

ix. l'utilisation de la manipulation génétique pour créer (110) de nouveaux chromosomes pour une nouvelle génération de chromosomes si ladite valeur d'aptitude dudit meilleur chromosome candidat n'est pas inférieure à ladite valeur d'aptitude seuil, et l'adjonction (246) à ladite liste de chromosomes de chaque nouveau chromosome dans ladite nouvelle génération de chromosomes qui ne se trouve pas déjà dans ladite liste de chromosomes ;

x. la génération (104) d'une valeur d'aptitude pour la configuration de réseau candidate correspondante pour chaque nouveau chromosome dans ladite nouvelle génération de chromosomes en utilisant ladite fonction d'aptitude ;

xi. la répétition des étapes vii à x jusqu'à ce que la valeur d'aptitude pour ledit meilleur chromosome candidat soit inférieure à ladite valeur d'aptitude seuil ou jusqu'à ce qu'un nombre prédéfini de nouvelles générations de chromosomes ait été créé (108), après quoi la configuration de réseau candidate correspondant audit meilleur chromosome candidat est émise.


 
2. Procédé selon la revendication 1, dans lequel la valeur de pondération wp est augmentée d'un facteur supérieur à 1 pour chaque nouvelle génération de chromosomes.
 
3. Procédé selon la revendication 1, dans lequel :



où base représente une configuration de réseau de base correspondant à une configuration de réseau après que ledit au moins l'un desdits commutateurs normalement fermés s'est ouvert pour isoler ledit dysfonctionnement et (idv) représente la configuration de réseau candidate ;

NoTieSwitchesbase représente le nombre total de commutateurs de sectionnement normalement ouverts dans ledit réseau ;

NoSwitchOperations représente le nombre d'opérations de commutation, à l'exclusion des opérations de commutation d'isolement, nécessaires pour une transition de ladite configuration de réseau de base à ladite configuration de réseau candidate ;

TotalWeightedLoadbase représente une charge totale pondérée de ladite configuration de réseau de base ;

TotalWeightedLoad (idv) représente une charge totale pondérée de la configuration de réseau candidate.


 
4. Procédé selon la revendication 3, dans lequel :

où LoadinOOSAreas représente la charge dans les zones de charge hors service dudit réseau et LoadinBackFeedingAreas représente la charge restante dans les parties non isolées ou défaillantes dudit réseau ;

le paneltyFactor étant supérieur à 1.


 
5. Procédé selon la revendication 1, dans lequel :

où Ivio (idv) représente le nombre de violations actuelles pour une configuration de réseau candidate et Iviomax représente le plus grand nombre de violations actuelles pour la configuration de réseau candidate correspondant à n'importe quel chromosome dans une génération.
 
6. Procédé selon la revendication 1, dans lequel ladite étape de création de chromosomes initiaux comprend, pour chaque dit commutateur de sectionnement normalement ouvert, la sélection aléatoire d'une de ladite liste d'actions.
 
7. Procédé selon la revendication 1, dans lequel, pour chaque dit commutateur de sectionnement normalement ouvert, ledit commutateur normalement fermé identifié par ledit caractère est en amont dudit commutateur de sectionnement normalement ouvert et, si ledit commutateur de sectionnement normalement ouvert délimite l'une desdites des zones de charge hors service, ledit commutateur normalement fermé identifié par ledit caractère se trouve également dans cette zone de charge hors service.
 
8. Procédé selon la revendication 1 dans lequel lesdits caractères sont des caractères ASCII.
 
9. Procédé selon la revendication 3, dans lequel chacune de ladite première génération de chromosomes et desdites nouvelles générations de chromosomes a une taille de population égale à 4N+1, où N=NoTieSwitchesbase/2 et où N est un nombre entier.
 
10. Procédé selon la revendication 1, dans lequel ladite étape d'utilisation de la manipulation génétique pour créer de nouveaux chromosomes pour une nouvelle génération de chromosomes comprend en outre :

la copie directe (110) d'une pluralité de chromosomes d'une génération de chromosomes précédente à la nouvelle génération de chromosomes ; et

la réalisation de croisements et de mutations (110) sur ladite pluralité de chromosomes de ladite génération de chromosomes précédente à la nouvelle génération de chromosomes.


 
11. Procédé selon la revendication 1, dans lequel ladite étape d'utilisation de la manipulation génétique pour créer de nouveaux chromosomes pour une nouvelle génération de chromosomes comprend en outre :

l'adjonction (110) des N chromosomes supérieurs ayant les N valeurs d'aptitudes les plus basses de la génération précédente à la nouvelle génération inchangés ;

le croisement (110) du meilleur chromosome candidat avec les N chromosomes suivants pour générer N nouveaux chromosomes ;

la mutation (110) des N chromosomes suivants après les N supérieurs pour créer N nouveaux chromosomes dans la nouvelle génération ; et

si de nouveaux chromosomes supplémentaires sont nécessaires pour remplir ladite nouvelle génération, la sélection (110) au hasard d'une de ladite liste d'actions pour chaque dit commutateur de sectionnement normalement ouvert.


 
12. Procédé selon la revendication 8, dans lequel lesdits caractères ASCII représentant ladite action d'ouverture de l'un desdits commutateurs normalement fermés qui est identifié par ledit caractère et de fermeture dudit commutateur de sectionnement normalement ouvert correspondant ont une valeur ASCII supérieure à 33.
 




Drawing






























REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Non-patent literature cited in the description