CN113890015B - Power distribution network dynamic reconstruction method based on improved fuzzy C-means clustering algorithm - Google Patents
Power distribution network dynamic reconstruction method based on improved fuzzy C-means clustering algorithm Download PDFInfo
- Publication number
- CN113890015B CN113890015B CN202111124728.1A CN202111124728A CN113890015B CN 113890015 B CN113890015 B CN 113890015B CN 202111124728 A CN202111124728 A CN 202111124728A CN 113890015 B CN113890015 B CN 113890015B
- Authority
- CN
- China
- Prior art keywords
- power
- period
- optimal
- distribution network
- reconstruction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 230000011218 segmentation Effects 0.000 claims abstract description 38
- 238000005457 optimization Methods 0.000 claims abstract description 30
- 230000001580 bacterial effect Effects 0.000 claims abstract description 13
- 230000003068 static effect Effects 0.000 claims abstract description 12
- 230000002431 foraging effect Effects 0.000 claims abstract description 9
- 238000007621 cluster analysis Methods 0.000 claims abstract description 3
- 230000008859 change Effects 0.000 claims description 16
- 230000000694 effects Effects 0.000 claims description 14
- 241000894006 Bacteria Species 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000005012 migration Effects 0.000 claims description 8
- 238000013508 migration Methods 0.000 claims description 8
- 230000035605 chemotaxis Effects 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 6
- 230000003399 chemotactic effect Effects 0.000 claims description 5
- 230000009182 swimming Effects 0.000 claims description 3
- 230000006978 adaptation Effects 0.000 claims description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 8
- 230000009467 reduction Effects 0.000 description 6
- 230000009471 action Effects 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 3
- 238000004220 aggregation Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/004—Generation forecast, e.g. methods or systems for forecasting future energy generation
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/22—The renewable source being solar energy
- H02J2300/24—The renewable source being solar energy of photovoltaic origin
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/28—The renewable source being wind energy
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
- Y02E10/56—Power conversion systems, e.g. maximum power point trackers
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Power Engineering (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Game Theory and Decision Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
According to the power distribution network dynamic reconstruction method based on the improved fuzzy C-means clustering algorithm, the EEMD-SVR is used for combining a prediction model and historical power data, and the DG output power and the load power are subjected to daily front power prediction; inputting Guan Chushi parameters such as initial parameters of a power distribution network, load power prediction quantity, DG predicted force value and the like; constructing a segmentation-loss function according to the power prediction data to determine an optimal segmentation number; the intra-day dynamic reconstruction time interval division is realized by improving a fuzzy C optimal cluster analysis algorithm; determining a time interval division scheme and an equivalent load center of each time interval according to a clustering algorithm; through improving a bacterial foraging algorithm, static reconstruction optimization is respectively carried out on each period of the power distribution network; according to the optimal regulation scheme of each time period of daily reconstruction, the daily operation network loss and voltage fluctuation condition of the power distribution network are calculated and determined, and the solved relevant parameters are output. The method is simple and efficient, can be applied to medium-low voltage distribution networks containing new energy access, and has certain popularization and practical values.
Description
Technical Field
The invention relates to the technical field of dynamic reconfiguration of power distribution networks, in particular to a dynamic reconfiguration method of a power distribution network based on an improved fuzzy C-means clustering algorithm.
Background
In the actual operation of the power distribution network, the reconstruction of the power distribution network is used as an important optimization means of the operation, and the aims of reducing network loss, balancing load and the like are achieved by controlling the switching states of the interconnection switches and the sectionalizing switches in the network. Under the large background that new energy is gradually and massively accessed into a distribution network, the power grid load and the distributed power supply (DG) output are always in a time-varying state, the running condition of the distribution network is complex, and static reconstruction is difficult to realize the running state optimization of the distribution network in multiple periods. In order to ensure safe and reliable operation of the power distribution network, time-varying characteristics of load and DG are required to be fully considered, and dynamic reconstruction of the power distribution network under multiple time sections is researched. Considering the negative influence of the switch action loss and the frequent action on the operation of the power distribution network, the realization of reasonable switch action and the reliable solving of the optimal reconstruction scheme are key to ensuring the dynamic reconstruction effect.
The dynamic reconfiguration time interval division of the power distribution network is an important factor affecting the optimal operation of the power distribution network, and the segmentation thinking is mainly divided into: the pre-segmentation period reduction strategy and the similarity index aggregation segmentation strategy. The time period reduction strategy can reduce unnecessary reconstruction operation, but is greatly influenced by a judgment threshold value of time period division; the aggregation segmentation strategy realizes time-of-day period division through specific indexes, so that the segmentation mode can be effectively prevented from being influenced by human factors, and the objectivity of data segmentation is ensured.
The fuzzy C-means clustering algorithm (FCM) can realize the clustering division of the data through the membership degree between the characteristics of the data and the similarity index, and each clustering segment meets the minimum criterion of weighted error square sum in the section. However, when the FCM algorithm performs power data clustering, only the size similarity is considered, and the time sequence of the data is not considered; meanwhile, considering the influence of the number of segments on the reconstruction effect, the optimal time division number is also required to be determined.
Disclosure of Invention
Aiming at the problems existing in the field of distribution network reconstruction under the background of new energy access, the invention provides a dynamic reconstruction method of a distribution network based on an improved fuzzy C-means clustering algorithm, which adopts an improved fuzzy C-means clustering algorithm (FCMC) to realize reconstruction period division, and determines the optimal segmentation number through a segmentation-loss function so as to ensure the reliability of algorithm segmentation; and simultaneously, optimizing the data clustering time sequence and the optimal segmentation number. The method is simple and efficient, can be applied to medium-low voltage distribution networks containing new energy access, and has certain popularization and practical values.
The technical scheme adopted by the invention is as follows:
the dynamic reconfiguration method of the distribution network based on the improved fuzzy C-means clustering algorithm comprises the steps of firstly, predicting the power before the day of DG output power and load power based on historical power data; then, carrying out dynamic reconfiguration time interval division on the operation power data according to a reconfiguration time interval division strategy; and finally, carrying out reconstruction optimization adjustment on different time periods according to the load centers of the time periods, and realizing dynamic reconstruction optimization adjustment in the power distribution network day.
A power distribution network dynamic reconstruction method based on an improved fuzzy C-means clustering algorithm comprises the following steps:
step one, performing daily power prediction on DG output power and load power by combining a prediction model and historical power data through EEMD-SVR;
inputting Guan Chushi parameters such as initial parameters of the power distribution network, load power prediction quantity, DG predicted force value and the like;
step three, constructing a segmentation-loss function according to the power prediction data so as to determine the optimal segmentation number; the intra-day dynamic reconstruction time interval division is realized by improving a fuzzy C optimal cluster analysis algorithm;
step four, determining a time interval division scheme and an equivalent load center of each time interval according to a clustering algorithm;
step five, respectively carrying out static reconstruction optimization on each period of the power distribution network by improving a bacterial foraging algorithm;
and step six, calculating and determining the daily operation network loss and voltage fluctuation condition of the power distribution network according to the daily reconstruction optimization adjustment scheme of each period, and outputting the solved relevant parameters.
Through the steps, dynamic reconstruction of the power distribution network is achieved.
In the first step, the prediction model is combined according to EEMD-SVR, as shown in FIG. 5. See document [1] Li Junjie, dan Jiang, hu Qunyong, he Lixin. Wind power prediction based on EEMD-SVR model [ J ]. Electrical and energy efficiency management techniques 2020 (11): 22-28. Description.
The DG and the load power prediction are determined, and superimposed, the equivalent load power curve of the power distribution system can be determined, and then a daily power data matrix is established; to ensure accuracy of the reconstructed segment, the whole day is divided into 48 time periods at time period dividing intervals of 30 min. Let the equivalent load power of node j at t period be: x is X t,j ＝X L (t,j)+X DG (t, j), the intra-day power data matrix is expressed as:
in the formula (1), X is the equivalent load power of each period of 48 periods in the day, namely active power and reactive power, and n is the number of nodes;
let the intra-day power data matrix be divided into c classes, the fuzzy clustering center M may be expressed as:
in the formula (2), M C The equivalent power value of the node j of the c-th class aggregation center is the number of nodes, and n is the number of nodes;
setting a fuzzy c-means clustering model according to the weighted error square sum minimum criterion in the section:
in the formula (3), u t,j Representing the membership degree of the sample j and the t clustering center; m represents a factor of membership. C is the cluster center.
The objective function J can be gradually stabilized by updating the cluster center and the membership degree of each point, and when the membership degree deviation is judged to meet the convergence criterion, the intra-segment deviation is minimum, and the power data is optimized and clustered.
The specific cluster center updating rule, membership updating rule and algorithm convergence criterion are as follows:
wherein, r represents the iterative times of the algorithm, and epsilon is an error threshold; x is x j Expressed as equivalent load power of node j at time t; m is M i Representing the equivalent load power at the i-th cluster center node j; m is M k Representing the equivalent load power at the kth cluster center node j;representing the membership degree of the sample j and the t clustering center when the membership degree factor is m; />The membership degree of the sample j and the t clustering center is expressed when the iteration number is r+1; />And the membership degree of the sample j and the t clustering center is expressed when the iteration number is r+1.
In the third step, according to the power prediction data, which refers to the 24 hours power data before the day predicted by the EEMD-SVR combined prediction model, a segmentation loss function is constructed, the optimal segmentation number is determined, and the weighted error square sum in the segments is defined as a segmentation loss function S (c), namely:
d(t,j)＝||x j -M t || 2 ＝(x j -M t ) T (x j -M t ) (7)
in the formula (7), j is the j-th node, and t is the t-th period; t is the transposition of the device,
in the formula (8), C is the C-th cluster center, and n is the number of nodes.
Considering that the time interval division is not too much, and simultaneously reducing the calculated amount, setting the segmentation range to be 1-24, and respectively obtaining the loss function value under different segmentation numbers and carrying out data fitting on the loss function by setting different FCMC initial iteration segmentation numbers. The piecewise loss function curve is shown in fig. 1. As can be seen from fig. 1, the number of segments is inversely related to the loss function, and the rate of change of the loss function is smaller as the number of segments increases. In order to ensure the efficiency of time interval division, taking the number of segments corresponding to the extreme point of the change rate of the loss function as the optimal number of segments, converting the solving problem of the optimal time interval number into the solving of the function slope extreme value, and then solving the optimal number of segments F by a formula:
wherein S is d (c) Slope representing the number of adjacent segments; s is S Z (c) A slope change rate representing the number of adjacent segments; argmax represents the value of the number of segments at the maximum value of the slope change rate of the number of adjacent segments. And finally, determining the optimal time interval division number through a segmentation loss function and a slope change rate selection rule.
Wherein S is d (c) Slope representing the number of adjacent segments; s is S Z (c) A slope change rate representing the number of adjacent segments; argmax represents the value of the number of segments at the maximum value of the slope change rate of the number of adjacent segments. And finally, determining the optimal time interval division number through a segmentation loss function and a slope change rate selection rule.
In the formula (9), S (c+1) and S (c) are segment loss functions of the c+1th clustering center respectively; a segment loss function of the c-th cluster center;
in (10)，S d (c)、S d (c-1) an optimal number of segments for a c-th cluster center; the optimal segmentation number of the c-1 th clustering center;
in the formula (11), c E [1,24] is expressed as a value range of 1-24 of a c-th clustering center.
In the fourth step, according to a clustering algorithm, determining a time interval division scheme and an equivalent load center of each time interval, including:
equivalent power timing division and period correction:
and determining an optimal fuzzy C-means clustering result of the equivalent load power according to the optimal segmentation number, classifying and time-sequence arranging the equivalent load power of the clustering result, and dividing the same class quantity of adjacent time periods into one time period, so that t time-sequence clustering time periods can be finally obtained.
Considering that an isolated period possibly exists in the time sequence dividing scheme, carrying out period correction by adopting a loss function minimum principle, fusing the isolated period with an adjacent period with better similarity, carrying out time sequence optimal period correction on the isolated period, and ensuring the time sequence dividing continuity, namely obtaining the corrected F-period equivalent power time sequence dividing scheme by taking the slope of the loss function. In order to ensure the efficiency of time interval division, the number of segments corresponding to the extreme point of the change rate of the loss function is taken as the optimal number of segments, and the problem of solving the optimal time interval number is converted into the problem of solving the slope extreme value of the function.
Because the same type of clustered data has higher similarity, and the clustering center is used as a data center, the clustering center is used as an equivalent load power characterization quantity of different time periods. The power distribution system with the close equivalent load power is considered, and the power flow similarity of the system is higher, so that the daily optimal reconstruction optimization can be converted into the static reconstruction optimization of F time periods, and the overall daily reconstruction adjustment effect is ensured through the static optimal adjustment of different time periods. The time-division flow chart is shown in fig. 2.
Step five, respectively carrying out static reconstruction optimization on each period of the power distribution network by improving a bacterial foraging algorithm,
step1: setting operating parameters and initializing bacterial populations, comprising: population size N, number of swimming times N s Number of chemotaxis N c Number of reproduction N re Number of migration N ed Fix the migration probability P ed A running step length c (i), a chemotactic iteration parameter j, a propagation iteration parameter k and a migration iteration parameter l;
step2: calculating fitness value J (i, J, k, l) of each bacterium initially by taking the bacterium as a solution of the solution space;
step3: judging whether the migration times meet the circulation conditions, if so, continuing to run, and executing the dispelling operation circulation; if the dispelling times are reached, optimizing and stopping;
step4: judging whether the reproduction times meet the circulation conditions, if so, continuing to run, sorting and selecting dominant half bacteria according to the activity of the bacteria, and completing the reproduction operation; returning to Step3 if the reproduction times are reached;
step5: judging whether the chemotaxis times meet the circulation conditions, if so, executing bacterial chemotaxis operation, judging whether the adaptation value is improved after the chemotaxis operation, if so, continuing to move towards the direction, otherwise, executing overturning operation and moving one step by the bacteria; if the number of reproduction has been reached, return to Step4.
And step six, calculating and determining the daily operation network loss and voltage fluctuation condition of the power distribution network according to the daily reconstruction optimization adjustment scheme of each period, and outputting the solved relevant parameters.
The invention discloses a dynamic reconfiguration method of a power distribution network based on an improved fuzzy C-means clustering algorithm, which has the following technical effects:
1) The time interval division strategy in the invention is suitable for medium-low voltage distribution networks accessed by high-proportion new power supplies, and solves the problem that the output of an uncertain power supply in the distribution network affects the division of dynamic reconfiguration time intervals.
2) The improved fuzzy mean value C clustering algorithm adopted by the invention considers the time sequence of data and the influence of the segmentation number on reconstruction, and realizes the optimal time interval division of dynamic reconstruction.
3) The time interval division strategy and the reconstruction method are simple and efficient, can be applied to a medium-low voltage distribution network containing new energy access, and have certain popularization and practical values.
4) According to the invention, a loss function is introduced in the dynamic reconfiguration period dividing process, and the optimal segmentation number is determined through a loss function slope change rate selection rule. The segmentation result is more reasonable and is represented by the minimum loss function.
5) The improved bacterial foraging algorithm has the advantages of wide search field, high convergence speed and the like, and the improved artificial intelligence algorithm can be widely applied to the practical use of power grid engineering.
6) The time interval division strategy and the overall distribution network dynamic reconstruction method complete verification 24H before the day, and can be applied to the time interval division of the low-voltage distribution network in half a month or whole month. The method has wide application prospect in the background of high-proportion access of new energy into the distribution network.
Drawings
Fig. 1 is a graph of a piecewise loss function.
FIG. 2 is a flowchart of optimal time interval partitioning for fuzzy C-optimal mean clustering.
Fig. 3 is a flow chart of a dynamic reconfiguration optimization strategy for a power distribution network.
FIG. 4 is a flow chart of an improved bacterial foraging algorithm.
FIG. 5 is a graph of EEMD-SVR combined prediction model.
FIG. 6 (a) is a flowchart illustrating the operation of an improved bacterial foraging algorithm;
FIG. 6 (b) is a flowchart showing the step5 algorithm for determining the number of chemotaxis.
Fig. 7 is a graph of load, distributed power supply power predictions.
Fig. 8 is an equivalent power segment diagram of a DG-containing power distribution system.
Fig. 9 (a) is a graph showing the comparison of network losses at different moments before and after reconstruction;
fig. 9 (b) is a graph showing the comparison of the intra-day network loss curves before and after reconstruction.
FIG. 10 (a) is a voltage probability density distribution diagram;
fig. 10 (b) is a normal distribution graph.
Detailed Description
The daily dynamic reconstruction optimization method for the power distribution network mainly comprises the following steps: and (3) power prediction of the power distribution system, dynamic time interval division in the day and static reconstruction optimization solution of each time interval. Firstly, predicting the daily power of the multisource DG output power and the load power based on historical power data; then, carrying out dynamic reconfiguration time division on the operation power data according to a reconfiguration time division strategy; and finally, carrying out reconstruction optimization adjustment on different time periods according to the load centers of the time periods, and realizing dynamic reconstruction optimization adjustment in the power distribution network day.
(1) And introducing a loss function in the dynamic reconfiguration period division process, and determining the optimal segmentation number through a loss function slope change rate selection rule.
(2) The invention relates to an optimal time interval division method based on fuzzy C optimal mean clustering, namely, the optimal time interval division is determined by a loss function and an improved fuzzy mean clustering algorithm.
(3) According to the dynamic reconfiguration overall optimization strategy of the power distribution network, the output of a distributed power supply is determined through prediction, the 24H time interval division is completed by adopting the provided dynamic time interval division strategy, after the optimal segmentation number is determined, each segment is subjected to static reconfiguration by adopting an improved bacterial foraging algorithm, and a reconfiguration result is obtained; and finishing dynamic reconstruction of the intra-day distribution network.
Examples:
in order to verify the dynamic reconstruction effect of DG, a photovoltaic and wind power distributed power supply is connected into an IEEE33 node system, and the connection nodes are 5 nodes and 31 nodes. Meanwhile, the prediction original data of wind power and photovoltaic DG are obtained by taking the literature Lu Yang. Active distribution network reconstruction research [ D ] containing a distributed power source, beijing university of transportation, 2016. Related data as references, and carrying out power prediction solution through a combined prediction model, wherein the related power prediction value is shown in fig. 7 and 8.
According to wind power, photovoltaic DG output and load power prediction data, an equivalent power curve graph of the power distribution network can be made, intra-day reconstruction time interval division is achieved through a time interval division strategy, and an equivalent power segmentation graph is shown in fig. 8. As can be seen from the sectional view of fig. 8, the reconfiguration period of the power distribution network is divided into 6 periods, which are respectively: 0-4, 4-6, 6-8.5, 8.5-16, 16-22, 22-24. By respectively carrying out dynamic optimization adjustment on each time period, the following reconstruction optimization result can be obtained.
TABLE 1 time-phased reconstruction results with DG
The table shows that the dynamic reconstruction strategy still has a good effect on the optimization of the active network loss of the DG-containing power distribution system reconstruction, the active network loss is reduced to 600.8kW from 989.7kW, the network loss reduction rate is 64.73%, and the loss reduction effect is obviously improved compared with the dynamic reconstruction without DG.
According to the network loss comparison diagrams and the intra-day network loss curve comparison diagrams of the power distribution network before and after the reconstruction of the power distribution network in the fig. 9 (a) and the fig. 9 (b), the loss reduction optimization effect of most of the reconstruction period is improved, and the active loss is obviously reduced; the power loss optimization effect of a part of sections is not ideal due to the influence of uncertainty output of the distributed power supply, the loss reduction rate is poor, and the overall active loss of the power distribution system still presents optimization characteristics in a day. The dynamic reconstruction active loss can be found to be obviously optimized compared with the dynamic reconstruction active loss before reconstruction through the network loss curve comparison diagrams before and after the dynamic reconstruction.
In order to evaluate the voltage stability optimizing effect of dynamic reconstruction on the DG-containing power distribution system, a node voltage probability density distribution diagram and a voltage normal distribution diagram before and after reconstruction are constructed, as shown in fig. 10 (a) and 10 (b).
As can be seen from the probability distribution diagram of fig. 10 (a), the concentration and the voltage deviation of the overall voltage distribution after reconstruction are still effectively improved compared with those before reconstruction, the voltage after reconstruction is between 0.94 and 1.0, and the node voltages are distributed on two sides of 0.96, so that the concentration degree is obviously improved, namely: the voltage stability optimizing effect of the dynamic reconfiguration strategy on the DG-containing power distribution system is proved.
According to the analysis, the dynamic reconstruction optimization method provided by the invention can realize the optimization of the active network loss and the voltage offset of the system of the DG-containing power distribution system, ensure the economic and reliable operation of the power distribution network and prove the effectiveness and the feasibility of the reconstruction strategy.
Claims (4)
1. The dynamic reconfiguration method of the power distribution network based on the improved fuzzy C-means clustering algorithm is characterized by comprising the following steps of:
step one, performing daily power prediction on DG output power and load power by combining a prediction model and historical power data through EEMD-SVR;
step two, inputting initial parameters of a power distribution network, load power prediction and DG prediction out force value related initial parameters;
step three, constructing a segmentation loss function according to the power prediction data so as to determine the optimal segmentation number; the intra-day dynamic reconstruction time interval division is realized by improving a fuzzy C optimal cluster analysis algorithm;
in the third step, according to the power prediction data, which refers to the 24 hours power data before the day predicted by the EEMD-SVR combined prediction model, a segmentation loss function is constructed, the optimal segmentation number is determined, and the weighted error square sum in the segments is defined as a segmentation loss function S (c), namely:
d(w,j)＝||x j -M w || 2 ＝(x j -M w ) T (x j -M w ) (7)
in the formula (7), x j Representing the equivalent load power of the j-th node; m is M w Cluster center load parameters representing w-time periods; j is the j-th node, w represents the w-th period; t is the transpose;
in the formula (8), c represents a c-th cluster center; n is the number of nodes; d (w, j) represents the weighted sum of squares of errors of the jth node over the w period;
setting the segmentation range to be 1-24, and respectively obtaining the loss function values under different segmentation numbers and performing data fitting on the loss function by setting different initial iteration segmentation numbers of the improved fuzzy C-means clustering algorithm;
the number of the segments is inversely related to the loss function, and the smaller the rate of change of the loss function is with the increase of the number of the segments; in order to ensure the efficiency of time interval division, taking the number of segments corresponding to the extreme point of the change rate of the loss function as the optimal number of segments, converting the solving problem of the optimal time interval number into the solving of the function slope extreme value, and then solving the optimal number of segments F by a formula:
wherein S is d (c) Slope representing the number of adjacent segments; s is S Z (c) A slope change rate representing the number of adjacent segments; argmax represents the value of the segment number when the slope change rate of the adjacent segment number is maximum; the optimal time interval dividing number can be finally determined through a segmentation loss function and a slope change rate selection rule;
in the formula (9), S (c+1) and S (c) are the segment loss function of the c+1th clustering center and the segment loss function of the c-th clustering center respectively;
in the formula (10), S d (c)、S d (c-1) the optimal number of segments of the c-th cluster center and the optimal number of segments of the c-1-th cluster center, respectively;
in the formula (11), c epsilon [2,24] is represented as the value range of the c clustering center is 2-24;
step four, determining a time interval division scheme and an equivalent load center of each time interval according to a clustering algorithm;
step five, respectively carrying out static reconstruction optimization on each period of the power distribution network by improving a bacterial foraging algorithm;
step six, calculating and determining the daily operation network loss and voltage fluctuation condition of the power distribution network according to the daily reconstruction optimization adjustment scheme of each period, and outputting the solved relevant parameters;
through the steps, dynamic reconstruction of the power distribution network is achieved.
2. The power distribution network dynamic reconstruction method based on the improved fuzzy C-means clustering algorithm of claim 1, wherein the method comprises the following steps: determining DG and load power prediction according to an EEMD-SVR combined prediction model, superposing the DG and the load power prediction, and determining an equivalent load power curve of a power distribution system so as to establish a daily power data matrix; to ensure the accuracy of the reconstruction segmentation, dividing the whole day into 48 time periods by taking 30min as a time period dividing interval; let the equivalent load power of node j in w period be: x is X w,j ＝X L (w,j)+X DG (w, j), wherein: x is X L (w, j) represents the load power of node j at w period; x is X DG (w, j) represents the distributed energy active power of node j at w time period;
the intra-day power data matrix is expressed as:
in the formula (1), X represents an intra-day power data matrix divided into 48 time periods for 24 hours a day, and n represents the number of nodes;
setting the number of the clustering centers as c, and converting the intra-day power data matrix into a fuzzy clustering center matrix M, wherein the fuzzy clustering center matrix M can be expressed as:
in the formula (2), M c,n An equivalent power value of the c-th cluster center node n; n is the number of nodes;
setting a fuzzy c-means clustering model according to the weighted error square sum minimum criterion in the section:
in the formula (3), u t,j Representing the membership degree of the sample j and the t clustering center; c is the number of cluster centers;representing the membership degree of the sample j and the t clustering center when the membership degree factor is m;
the objective function J can gradually tend to be stable by updating the cluster center and the membership degree of each point, when the membership degree deviation is judged to meet the convergence criterion, the intra-segment deviation is minimum, and the power data completes the optimized cluster;
the specific cluster center updating rule, membership updating rule and algorithm convergence criterion are as follows:
wherein, r represents the iterative times of the algorithm, and epsilon is an error threshold; x is x j Equivalent load power, denoted node j; m is M i Representing the equivalent load power at the i-th cluster center node j; m is M k Representing the equivalent load power at the kth cluster center node j;representing the membership degree of the sample j and the t clustering center when the membership degree factor is m; />The membership degree of the sample j and the t clustering center is expressed when the iteration number is r+1; />And the membership degree of the sample j and the t clustering center is expressed when the iteration number is r.
3. The power distribution network dynamic reconstruction method based on the improved fuzzy C-means clustering algorithm of claim 1, wherein the method comprises the following steps: in the fourth step, according to a clustering algorithm, determining a time interval division scheme and an equivalent load center of each time interval, including: equivalent power timing division and period correction:
determining an optimal fuzzy C-means clustering result of the equivalent load power according to the optimal segmentation number, classifying and time-sequence arranging the equivalent load power of the clustering result, dividing the same class quantity of adjacent time periods into one time period, and finally obtaining z time-sequence clustering time periods;
considering that an isolated period possibly exists in the time sequence dividing scheme, carrying out period correction by adopting a loss function minimum principle, fusing the isolated period with an adjacent period with better similarity, carrying out time sequence optimal period correction on the isolated period, and ensuring time sequence dividing continuity, namely obtaining a corrected F-period equivalent power time sequence dividing scheme by taking a loss function slope;
because the similarity of the clustered data of the same type is higher, and the clustering center is used as a data center, the clustering center is used as an equivalent load power representation quantity of different time periods; the power distribution system with the close equivalent load power is considered, and the power flow similarity of the system is higher, so that the daily optimal reconstruction optimization can be converted into the static reconstruction optimization of F time periods, and the overall daily reconstruction adjustment effect is ensured through the static optimal adjustment of different time periods.
4. The power distribution network dynamic reconstruction method based on the improved fuzzy C-means clustering algorithm of claim 1, wherein the method comprises the following steps: step five, respectively carrying out static reconstruction optimization on each period of the power distribution network by improving a bacterial foraging algorithm:
step1: setting operating parameters and initializing bacterial populations, comprising: population sizep, number of swimming times N s Number of chemotaxis N c Number of reproduction N re Number of migration N ed Fix the migration probability P ed A running step length c (v), a chemotactic iteration parameter q, a propagation iteration parameter r and a migration iteration parameter l;
step2: calculating fitness value J (v, q, r, l) of each bacterium initially by taking the bacterium as a solution of the solution space;
step3: judging whether the migration times meet the circulation conditions, if so, continuing to run, and executing the dispelling operation circulation; if the dispelling times are reached, optimizing and stopping;
step4: judging whether the reproduction times meet the circulation conditions, if so, continuing to run, sorting and selecting dominant half bacteria according to the activity of the bacteria, and completing the reproduction operation; returning to Step3 if the reproduction times are reached;
step5: judging whether the chemotactic times meet the circulation conditions, if so, executing bacterial chemotactic operation, judging whether the adaptation value after the chemotactic operation is improved, if so, continuing to swim towards the original direction, otherwise, executing overturning operation and swimming by one step; if the number of reproduction has been reached, return to Step4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111124728.1A CN113890015B (en) | 2021-09-25 | 2021-09-25 | Power distribution network dynamic reconstruction method based on improved fuzzy C-means clustering algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111124728.1A CN113890015B (en) | 2021-09-25 | 2021-09-25 | Power distribution network dynamic reconstruction method based on improved fuzzy C-means clustering algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113890015A CN113890015A (en) | 2022-01-04 |
| CN113890015B true CN113890015B (en) | 2023-08-25 |
Family
ID=79006470
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111124728.1A Active CN113890015B (en) | 2021-09-25 | 2021-09-25 | Power distribution network dynamic reconstruction method based on improved fuzzy C-means clustering algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113890015B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115065057A (en) * | 2022-03-25 | 2022-09-16 | 国网宁夏电力有限公司银川供电公司 | Power grid dynamic power supply margin analysis method based on data driving |
| CN114817865A (en) * | 2022-06-28 | 2022-07-29 | 深圳市傲立电子有限公司 | Big data-based operation fault prediction system for radio frequency power amplifier |
| CN115833116B (en) * | 2023-02-06 | 2023-05-23 | 广东电网有限责任公司东莞供电局 | Power distribution network reconstruction optimization method based on multi-objective optimization |
| CN116432443B (en) * | 2023-04-03 | 2024-08-23 | 海南电网有限责任公司 | Power grid simulation method and device, electronic equipment and storage medium |
| CN116975731B (en) * | 2023-08-08 | 2024-02-20 | 山东大学 | Cross-domain cutter damage monitoring method and system based on transfer learning |
| CN116826847B (en) * | 2023-08-24 | 2023-11-28 | 国网山西省电力公司运城供电公司 | Dynamic network reconstruction and reactive voltage adjustment collaborative optimization method, device and equipment |
| CN117408394B (en) * | 2023-12-14 | 2024-05-31 | 国网天津市电力公司电力科学研究院 | Carbon emission factor prediction method and device for electric power system and electronic equipment |
| CN118378899A (en) * | 2024-06-24 | 2024-07-23 | 浙江省金融综合服务平台管理有限公司 | LSTM model-based financial comprehensive risk prediction method, system and medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101232180A (en) * | 2008-01-24 | 2008-07-30 | 东北大学 | Power distribution system load obscurity model building device and method |
| JP2015202022A (en) * | 2014-03-31 | 2015-11-12 | 株式会社デンソー | energy management system |
| CN106329516A (en) * | 2015-07-09 | 2017-01-11 | 中国电力科学研究院 | Typical scene recognition based dynamic reconstruction method of power distribution network |
| CN111342458A (en) * | 2020-03-25 | 2020-06-26 | 四川大学 | Method and device for two-stage reconstruction of power distribution network based on ordered optimization algorithm |
| CN112330042A (en) * | 2020-11-17 | 2021-02-05 | 合肥工业大学 | Power distribution network reconstruction method based on self-adaptive fuzzy C-means clustering scene division |
-
2021
- 2021-09-25 CN CN202111124728.1A patent/CN113890015B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101232180A (en) * | 2008-01-24 | 2008-07-30 | 东北大学 | Power distribution system load obscurity model building device and method |
| JP2015202022A (en) * | 2014-03-31 | 2015-11-12 | 株式会社デンソー | energy management system |
| CN106329516A (en) * | 2015-07-09 | 2017-01-11 | 中国电力科学研究院 | Typical scene recognition based dynamic reconstruction method of power distribution network |
| CN111342458A (en) * | 2020-03-25 | 2020-06-26 | 四川大学 | Method and device for two-stage reconstruction of power distribution network based on ordered optimization algorithm |
| CN112330042A (en) * | 2020-11-17 | 2021-02-05 | 合肥工业大学 | Power distribution network reconstruction method based on self-adaptive fuzzy C-means clustering scene division |
Non-Patent Citations (1)
| Title |
|---|
| 采用改进细菌觅食算法的含分布式电源配电网动态重构;唐浩;周步祥;彭章刚;周海忠;李世新;王精卫;;电力系统及其自动化学报(第04期);122-128 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113890015A (en) | 2022-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113890015B (en) | Power distribution network dynamic reconstruction method based on improved fuzzy C-means clustering algorithm | |
| CN109768573B (en) | Power distribution network reactive power optimization method based on multi-target differential gray wolf algorithm | |
| Zhang et al. | Multi-objective optimal reactive power dispatch of power systems by combining classification-based multi-objective evolutionary algorithm and integrated decision making | |
| CN108448620A (en) | High permeability distributed generation resource assemblage classification method based on integrated performance index | |
| Niknam et al. | Multi-objective daily operation management of distribution network considering fuel cell power plants | |
| CN110535118B (en) | Active power distribution network multi-period dynamic reconstruction method based on improved recursion ordered clustering | |
| CN111463778A (en) | Active power distribution network optimization reconstruction method based on improved suburb optimization algorithm | |
| CN111490552A (en) | Reactive power optimization method for power distribution network | |
| CN111932012A (en) | Energy storage system-distributed power supply-capacitor comprehensive control reactive power optimization method | |
| CN113673065B (en) | Loss reduction method for automatic reconstruction of power distribution network | |
| CN109390971A (en) | A kind of power distribution network multiple target active reconstructing method based on genetic algorithm well-matched in social and economic status | |
| CN117439203A (en) | Active power distribution network comprehensive optimization method based on multi-objective bald eagle search algorithm | |
| CN112531789A (en) | Dynamic reconfiguration strategy for power distribution network with distributed power supplies | |
| CN116882105A (en) | Method for analyzing boundaries of power supply modes of large-scale photovoltaic access AC/DC power distribution network | |
| CN110135640A (en) | A kind of wind-powered electricity generation distribution Optimization Scheduling improving harmony algorithm based on fuzzy clustering | |
| CN115034293B (en) | Dynamic reconstruction method of power distribution network based on improved double-scale spectral clustering algorithm | |
| CN115983428A (en) | Cluster division method and device based on improved particle swarm optimization algorithm | |
| Zhao et al. | Distribution network reconfiguration digital twin model based on bi-level dynamical time division | |
| CN115018140A (en) | Active power distribution network light storage multi-target planning method and system | |
| CN114662256A (en) | Power distribution network dynamic reconstruction method considering wind power uncertainty | |
| CN111130150B (en) | IPM-intPSO-based two-stage dynamic reactive power optimization method | |
| Pemmada et al. | A novel covid-19 herd immunity-based optimizer for optimal accommodation of solar PV with battery energy storage systems including variation in load and generation | |
| CN114417566A (en) | MOEA/D-based active power distribution network multi-region division optimization method | |
| CN112734590A (en) | Two-stage regional power grid reactive power optimization method based on dynamic planning method of interior point method and variable field search | |
| Wang et al. | Distribution network loss calculation and reactive power optimization loss reduction analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |