CN111105025A - Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization - Google Patents

Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization Download PDF

Info

Publication number
CN111105025A
CN111105025A CN201911245120.7A CN201911245120A CN111105025A CN 111105025 A CN111105025 A CN 111105025A CN 201911245120 A CN201911245120 A CN 201911245120A CN 111105025 A CN111105025 A CN 111105025A
Authority
CN
China
Prior art keywords
distribution network
voltage distribution
node
representing
voltage
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.)
Pending
Application number
CN201911245120.7A
Other languages
Chinese (zh)
Inventor
刘畅
陈刚
腾予非
史华勃
范成围
丁理杰
朱童
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Electric Power Research Institute of State Grid Sichuan Electric Power Co Ltd
Original Assignee
Electric Power Research Institute of State Grid Sichuan Electric Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Electric Power Research Institute of State Grid Sichuan Electric Power Co Ltd filed Critical Electric Power Research Institute of State Grid Sichuan Electric Power Co Ltd
Priority to CN201911245120.7A priority Critical patent/CN111105025A/en
Publication of CN111105025A publication Critical patent/CN111105025A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
    • G06N3/065—Analogue means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06313—Resource planning in a project environment
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Biophysics (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Strategic Management (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Neurology (AREA)
  • Development Economics (AREA)
  • Power Engineering (AREA)
  • Educational Administration (AREA)
  • Genetics & Genomics (AREA)
  • Game Theory and Decision Science (AREA)
  • Physiology (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a data-driven heuristic optimization-based urban high-voltage distribution network blocking management method, which comprises the steps of firstly utilizing a good point set theory to generate a large amount of operation section data, and utilizing a plurality of deep neural networks to fit the nonlinear relations among the topological state of a high-voltage distribution network, node loads, node voltages and branch power; secondly, embedding the trained neural network into a heuristic optimization algorithm, taking nonlinear mapping of transmission capacity of a power transmission line, capacity of a transformer substation, upper and lower limits of output of a generator, upper and lower limits of voltage amplitude, upper limit of load shedding and fitting of the neural network as constraint conditions, taking the minimum total amount of load shedding as a target function, constructing a blocking management and control model of the urban power grid, and finally solving by adopting a genetic algorithm. The model replaces the complex 0-1 topological constraint, the alternating current power flow constraint and the N-1 constraint by the neural network, avoids the complex time-consuming N-1 calibration process, has good fitting capability of the deep network to the complex nonlinear relation, and effectively improves the solving efficiency.

Description

Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization
Technical Field
The invention belongs to the technical field of power system automation, and particularly relates to a data-driven heuristic optimization-based urban high-voltage distribution network blocking management method.
Background
The block management of the urban high-voltage distribution network is generally a large-scale non-convex non-linear combination optimization problem, which comprises non-convex alternating current power flow constraint, topological radial constraint and N-1 constraint. The above constraints result in computationally difficult and time consuming blocking management decisions, which are difficult to cope with rapidly changing urban electrical loads. In recent years, with the development of computer technology, the appearance of deep learning technology provides a new idea for solving various problems of power systems. The good nonlinear characterization capability and the calculation speed of the method can effectively improve the calculation performance of the blocking management model. Therefore, a novel urban high-voltage distribution network blocking management model is constructed by combining data driving with heuristic optimization, the problem that a traditional optimization model processes non-convex problems is avoided, and a dispatcher is helped to quickly generate a decision scheme.
Disclosure of Invention
The invention discloses a data-driven heuristic optimization-based urban high-voltage distribution network blocking management method, which is characterized in that large-scale complex nonlinear power system constraints are fitted through a deep neuron network, the trained neuron network is embedded into a heuristic optimization framework, and an optimal decision scheme is obtained through continuous iteration.
In order to achieve the purpose of the invention, the technical scheme adopted by the invention is as follows:
the method for managing the urban high-voltage distribution network blocking based on data-driven heuristic optimization comprises the following steps:
s1: acquiring a node branch connection relation of an urban high-voltage distribution network, and acquiring a plurality of feasible topologies and node load samples of the high-voltage distribution network;
s2: load flow calculation is carried out on each feasible topology and node load sample, and node voltage and branch power corresponding to each feasible topology and node load sample are obtained;
s3: performing N-1 verification on each feasible topology and node load sample, and calculating an N-1 default quantity;
s4: establishing a node voltage prediction neuron network, fitting the node voltage corresponding to each feasible topology and node load sample, establishing a branch power prediction neuron network, fitting the branch power corresponding to each feasible topology and node load sample, establishing an N-1 default quantity prediction neuron network, and fitting the N-1 default quantity of each feasible topology and node load sample;
s5: the fitted node voltage prediction neuron network, the branch power prediction neuron network, the N-1 default quantity prediction neuron network, the transmission capacity of the power transmission line, the capacity of the transformer substation, the upper and lower limits of the output of the generator, the upper and lower limits of the voltage amplitude and the upper limit of the load shedding are used as constraint conditions, the minimum total load shedding is used as a target function, and the urban high-voltage distribution network blocking management and control model is constructed;
s6: and acquiring the topological state and node load shedding amount of the current high-voltage distribution network, solving the urban high-voltage distribution network blocking management and control model, and outputting the optimized topological state and load shedding amount of the high-voltage distribution network.
Preferably, in step S1, a plurality of feasible topology and node load samples of the high voltage distribution network are obtained based on the theory of good point set.
Preferably, in step S1, the feasible topology of the high voltage distribution network is expressed as follows:
Figure BDA0002307306320000021
Figure BDA0002307306320000022
wherein the content of the first and second substances,
Figure BDA0002307306320000023
represents the topological state of the high-voltage distribution network,
Figure BDA0002307306320000024
for its corresponding reachability matrix,
Figure BDA0002307306320000025
a matrix of connection relations between the power supply points is represented,
Figure BDA0002307306320000026
a connection relation matrix of the power supply point and the high-voltage distribution network node is shown,
Figure BDA0002307306320000027
a high-voltage distribution network node connection relation matrix is shown,
Figure BDA0002307306320000028
representation matrix
Figure BDA0002307306320000029
The (j) th column element of (1),
Figure BDA00023073063200000210
representation matrix
Figure BDA00023073063200000211
Row i and column j.
Preferably, in step S1, the calculation process of the node load sample is as follows:
order to
Figure BDA00023073063200000212
Pn＝{rsk, | k ═ 1,2,3,.., n }, then each node load in the high voltage distribution networkCan be calculated as:
Figure BDA00023073063200000213
wherein s represents the number of nodes of the high-voltage distribution network, p represents a prime number satisfying p ≧ 2s +3, n represents the number of samples, rsRepresenting a point in s-dimensional space, PnRepresenting a set of good points of sample size n, PD,max、PD,minWhich represents the maximum and minimum values of the load,
Figure BDA00023073063200000214
representing the node load of the high voltage distribution network with a sample capacity of n.
Preferably, in step S3, the calculation process of the N-1 default quantity is as follows:
Figure BDA00023073063200000215
Figure BDA00023073063200000216
Figure BDA00023073063200000217
Figure BDA0002307306320000031
Figure BDA0002307306320000032
Figure BDA0002307306320000033
Figure BDA0002307306320000034
Figure BDA0002307306320000035
wherein i represents a sample number; j. k represents a node number; m represents an N-1 scene number; l represents a branch number; cN-1Representing a set of N-1 scenes ξmRepresenting a default quantity for the mth N-1 scene;
Figure BDA0002307306320000036
and
Figure BDA0002307306320000037
and respectively representing the margins of the upper and lower limits of the voltage amplitude and the upper and lower limits of the branch power. Epsilon (x) represents a step function, and the state with the margin being a negative value is screened out through step function mapping and is used as the default quantity in the scene. I isiRepresenting the sum of all N-1 scene violations under the current sample.
Preferably, in step S5, the expression of the constructed urban high voltage distribution network blocking management and control model is as follows:
Figure BDA0002307306320000038
Figure BDA0002307306320000039
Figure BDA00023073063200000310
Figure BDA00023073063200000311
PL,min≤PL≤PL,max
Vmin≤V≤Vmax
0≤Δp≤Δpmax
where Δ p denotes a shear load vector, Δ piRepresenting the elements in the vector Δ p, n representing the total number of nodes, yN-1Indicating that the N-1 check passes, the output being 1 indicates that the check passes, the outputOut of 0 indicates no check, PLRepresenting branch power vector, V representing node voltage vector, PL,min、PL,maxRepresenting the upper and lower limits of the branch power, Vmin、VmaxRepresenting the upper and lower limits of the voltage amplitude, Δ pmaxThe upper limit of the load cut is indicated.
Preferably, in the step S6, a genetic algorithm is used to solve the urban high voltage distribution network blocking management and control model.
Preferably, in step S6, the model solving process includes:
s61, initializing genetic algorithm parameters including a population scale, a cross factor, a variation factor, iteration times and convergence precision, and executing S62;
s62, initializing the topological state of the high-voltage distribution network, initializing the node load shedding amount, and executing S63;
s63, calculating a fitness function according to the current topological state of the high-voltage distribution network and the node load shedding amount, and executing S64;
s64, judging whether an exit condition is met, namely whether the maximum iteration number is reached or the convergence precision is met, if so, executing S66, and if not, executing S65;
s65, performing crossing, selecting and mutation operation, updating the topological state of the high-voltage distribution network and the node load shedding amount, and performing S63;
and S66, completing model solution, and outputting the optimized topological state of the high-voltage distribution network and the node load shedding amount.
Preferably, in step S63, the fitness function is calculated as follows:
Figure BDA0002307306320000041
the first term represents the total load shedding amount of the high-voltage distribution network, the second term represents the punishment on the N-1 default amount, and the third term represents the node voltage out-of-limit punishment amount and the branch power out-of-limit punishment amount; y isN-1Indicating an N-1 parity-check-passed condition, an output of 1 indicating a pass check, an output of 0 indicating a fail check,
Figure BDA0002307306320000042
and
Figure BDA0002307306320000043
respectively representing the margins of the upper and lower limits of the voltage amplitude and the upper and lower limits of the branch power; Δ p denotes the tangential load vector, Δ piRepresenting elements in the vector delta p, epsilon (x) representing a step function, and screening out a state with a negative margin as a default quantity in the scene through step function mapping, IiRepresenting the sum of all N-1 scene violations under the current sample, α being a constant and β being a constant.
Has the advantages that:
the invention provides a data-driven heuristic optimization-based urban high-voltage distribution network blocking control method, which is characterized in that complex non-convex nonlinear constraints of an urban high-voltage distribution network are fitted by adopting a deep neuron network through a large number of samples, and trained neuron network groups are embedded into a heuristic optimization framework for optimization. According to the simulation result, the heuristic optimization based on data driving has the same precision as the traditional heuristic optimization algorithm, but the calculation speed is greatly improved. Meanwhile, DNN network fitting is adopted for N-1 constraint, whether the network meets N-1 verification in different states can be judged quickly, and safety of optimization strategies is improved.
Drawings
Fig. 1 is a schematic step diagram of an urban high-voltage distribution network blocking management method based on data-driven heuristic optimization according to this embodiment;
fig. 2 is a schematic step diagram of S6 of the urban high-voltage distribution network blocking management method based on data-driven heuristic optimization according to the embodiment;
fig. 3 is a schematic diagram for showing a topology structure of an urban high-voltage distribution network in the embodiment;
FIG. 4 is a schematic diagram of the node voltage predicting neuron network, the branch power predicting neuron network, and the N-1 default predicting neuron network training according to the present embodiment;
FIG. 5 is a diagram illustrating the fitting effect of the node voltage prediction neuron network according to the present embodiment;
FIG. 6 is a diagram illustrating the fitting effect of the branch power prediction neural network according to the present embodiment;
FIG. 7 is a diagram illustrating the fitting effect of the N-1 penalty prediction neural network according to the present embodiment;
fig. 8 is a schematic diagram for showing an optimized topology structure of a high-voltage distribution network according to the present embodiment;
FIG. 9 is a diagram illustrating a comparison between the present embodiment and a conventional algorithm without embedded neural network groups;
FIG. 10 is a comparison diagram illustrating the optimized front-to-back branch load ratio of 90% for the present embodiment;
FIG. 11 is a comparison diagram illustrating the optimized front-to-back branch load ratio of 80% for this embodiment;
FIG. 12 is a comparative diagram illustrating the optimized front-to-back branch load ratio of 70% for this embodiment;
FIG. 13 is a comparative diagram illustrating the optimized load ratios of the front and rear branches according to the present embodiment;
fig. 14 is a schematic diagram for showing the optimized network N-1 verification situation in this embodiment.
Detailed Description
The following description of the embodiments of the present invention will be provided in conjunction with the accompanying drawings 1 to 14 to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined by the appended claims, and all of the inventions utilizing the inventive concept are intended to be protected.
Referring to fig. 1, the data-driven heuristic optimization-based urban high-voltage distribution network blocking management method comprises the following steps:
s1: acquiring a node branch connection relation of an urban high-voltage distribution network, and acquiring a plurality of feasible topologies and node load samples of the high-voltage distribution network based on a good point set theory;
s2: load flow calculation is carried out on each feasible topology and node load sample, and node voltage and branch power corresponding to each feasible topology and node load sample are obtained;
s3: performing N-1 verification on each feasible topology and node load sample, and calculating an N-1 default quantity;
s4: establishing a node voltage prediction neuron network, fitting the node voltage corresponding to each feasible topology and node load sample, establishing a branch power prediction neuron network, fitting the branch power corresponding to each feasible topology and node load sample, establishing an N-1 default quantity prediction neuron network, and fitting the N-1 default quantity of each feasible topology and node load sample;
s5: the fitted node voltage prediction neuron network, the branch power prediction neuron network, the N-1 default quantity prediction neuron network, the transmission capacity of the power transmission line, the capacity of the transformer substation, the upper and lower limits of the output of the generator, the upper and lower limits of the voltage amplitude and the upper limit of the load shedding are used as constraint conditions, the minimum total load shedding is used as a target function, and the urban high-voltage distribution network blocking management and control model is constructed;
s6: and acquiring the topological state and node load shedding amount of the current high-voltage distribution network, solving the urban high-voltage distribution network blocking management and control model, and outputting the optimized topological state and load shedding amount of the high-voltage distribution network.
Specifically, in step S1, the feasible topology of the high voltage distribution network is expressed as follows:
Figure BDA0002307306320000061
Figure BDA0002307306320000062
wherein the content of the first and second substances,
Figure BDA0002307306320000063
represents the topological state of the high-voltage distribution network,
Figure BDA0002307306320000064
for its corresponding reachability matrix,
Figure BDA0002307306320000065
a matrix of connection relations between the power supply points is represented,
Figure BDA0002307306320000066
a connection relation matrix of the power supply point and the high-voltage distribution network node is shown,
Figure BDA0002307306320000067
a high-voltage distribution network node connection relation matrix is shown,
Figure BDA0002307306320000068
representation matrix
Figure BDA0002307306320000069
The (j) th column element of (1),
Figure BDA00023073063200000610
representation matrix
Figure BDA00023073063200000611
Row i and column j.
Specifically, in step S1, the calculation process of the node load sample is as follows:
order to
Figure BDA00023073063200000612
Pn＝{rsk, | k ═ 1,2,3,.., n }, then the load at each node in the high voltage distribution network can be calculated as:
Figure BDA00023073063200000613
wherein s represents the number of nodes of the high-voltage distribution network, p represents a prime number satisfying p ≧ 2s +3, n represents the number of samples, rsRepresenting a point in s-dimensional space, PnRepresenting a set of good points of sample size n, PD,max、PD,minWhich represents the maximum and minimum values of the load,
Figure BDA00023073063200000614
representing the node load of the high voltage distribution network with a sample capacity of n.
Specifically, in step S3, the calculation procedure of the N-1 default is as follows:
Figure BDA00023073063200000615
Figure BDA00023073063200000616
Figure BDA00023073063200000617
Figure BDA00023073063200000618
Figure BDA00023073063200000619
Figure BDA00023073063200000620
Figure BDA0002307306320000071
Figure BDA0002307306320000072
wherein i represents a sample number; j. k represents a node number; m represents an N-1 scene number; l represents a branch number; cN-1Representing a set of N-1 scenes ξmRepresenting a default quantity for the mth N-1 scene;
Figure BDA0002307306320000073
and
Figure BDA0002307306320000074
and respectively representing the margins of the upper and lower limits of the voltage amplitude and the upper and lower limits of the branch power. Epsilon (x) represents a step function, and the state with the margin being a negative value is screened out through step function mapping and is used as the default quantity in the scene. I isiRepresenting the sum of all N-1 scene violations under the current sample.
It should be noted that, in this embodiment, the network structures of the node voltage prediction neuron network, the branch power prediction neuron network, and the N-1 default quantity prediction neuron network are all x × 10 × 10 × 10 × y, and include an input layer x, 10 hidden layers, and an output layer. Wherein x is input, and in the three neuron networks used in the invention, the dimension of x is equal to the sum of the number of load nodes in the network and the total number of branches; y represents an output, for the N-1 default quantity prediction neuron network, the dimension of y is 1, and the output is the N-1 default quantity; for a node voltage prediction neuron network, the dimension of y is equal to the total number of nodes in the network; for a branch power prediction neuron network, the dimension of y is equal to the total number of branches in the network.
Specifically, in step S5, the expression of the constructed urban high voltage distribution network blocking management and control model is as follows:
Figure BDA0002307306320000075
Figure BDA0002307306320000076
Figure BDA0002307306320000077
Figure BDA0002307306320000078
PL,min≤PL≤PL,max
Vmin≤V≤Vmax
0≤Δp≤Δpmax
where Δ p denotes a shear load vector, Δ piRepresenting the elements in the vector Δ p, n representing the total number of nodes, yN-1Indicating that N-1 parity passed, an output of 1 indicating pass, an output of 0 indicating fail, and PLRepresenting branch power vector, V representing node voltage vector, PL,min、PL,maxRepresenting the upper and lower limits of the branch power, Vmin、VmaxRepresenting the upper and lower limits of the voltage amplitude, Δ pmaxThe upper limit of the load cut is indicated.
Referring to fig. 2, in step S6, the model solving process is as follows:
s61, initializing genetic algorithm parameters including a population scale, a cross factor, a variation factor, iteration times and convergence precision, and executing S62;
s62, initializing the topological state of the high-voltage distribution network, initializing the node load shedding amount, and executing S63;
s63, calculating a fitness function according to the current topological state of the high-voltage distribution network and the node load shedding amount, and executing S64;
s64, judging whether an exit condition is met, namely whether the maximum iteration number is reached or the convergence precision is met, if so, executing S66, and if not, executing S65;
s65, performing crossing, selecting and mutation operation, updating the topological state of the high-voltage distribution network and the node load shedding amount, and performing S63;
and S66, completing model solution, and outputting the optimized topological state of the high-voltage distribution network and the node load shedding amount.
Specifically, in step S63, the fitness function is calculated as follows:
Figure BDA0002307306320000081
the first term represents the total load shedding amount of the high-voltage distribution network, the second term represents the punishment on the N-1 default amount, and the third term represents the node voltage out-of-limit punishment amount and the branch power out-of-limit punishment amount; y isN-1Indicating an N-1 parity-check-passed condition, an output of 1 indicating a pass check, an output of 0 indicating a fail check,
Figure BDA0002307306320000082
and
Figure BDA0002307306320000083
respectively representing the margins of the upper and lower limits of the voltage amplitude and the upper and lower limits of the branch power; Δ p denotes the tangential load vector, Δ piRepresenting elements in the vector delta p, epsilon (x) representing a step function, and screening out a state with a negative margin as a default quantity in the scene through step function mapping, IiRepresenting the sum of all N-1 scene violations under the current sample, α being a constant and β being a constant.
The method for managing the congestion of the urban high-voltage distribution network based on data-driven heuristic optimization, which is provided by the invention, is specifically described by taking the original urban high-voltage distribution network shown in fig. 3 as an example.
In fig. 3, the black dashed lines indicate open high voltage distribution network branches and the solid lines indicate closed high voltage distribution network branches. According to the collected node information of the urban high-voltage distribution network, a great number of node loads and branch topological state samples are generated by adopting a good point set theory, load flow calculation is carried out on each sample to obtain the whole network state, namely node voltage and branch current, and then N-1 verification is carried out on the node voltage and branch current.
Referring to fig. 4 and 5, a node voltage prediction neuron network is established, and node voltages corresponding to each feasible topology and node load sample are fitted; with reference to fig. 4 and 6, a branch power prediction neuron network is established, and branch power corresponding to each feasible topology and node load sample is fitted; and (4) establishing an N-1 default quantity prediction neuron network by combining the graphs in the figures 4 and 7, and fitting the N-1 default quantity of each feasible topology and node load sample.
Referring to fig. 8, the trained node voltage prediction neuron network, branch power prediction neuron network and N-1 default amount prediction neuron network are embedded into the urban high-voltage distribution network blocking management model, the node voltage prediction neuron network, the branch power prediction neuron network and the N-1 default amount prediction neuron network respectively calculate node voltage, branch power and N-1 default amount according to the topological structure and node load of the current high-voltage distribution network, and the urban high-voltage distribution network blocking management model is solved by adopting a genetic algorithm to obtain the optimized topological state of the high-voltage distribution network. In fig. 8, the dashed line indicates that the originally closed branch is opened compared to the initial topological state; the solid line indicates closing an otherwise open leg.
Specifically, it can be seen from the figure that the optimized loads are transferred from the substations S1 and S2 to the substations S3 and S4, so that the load rate of the heavy-load S1 is controlled to 80%. The percentage of the total load shedding of the four substations is shown in table 1.
Transformer substation Percentage of load shedding
S1
0％
S2
0％
S3 0.98％
S4
0％
TABLE 1
Referring to fig. 9, in contrast to the conventional algorithm without embedded neural network population, it can be seen that the conventional heuristic optimization and the heuristic optimization based on data driving can converge to the same objective function value. The method is continuously operated for 10 times, and compared with the traditional heuristic optimization algorithm for calculating time to obtain the result shown in the table 2.
Figure BDA0002307306320000091
TABLE 2
From table 2, it can be seen that the computation time required by the data-driven heuristic optimization algorithm proposed by the present invention is much shorter than that of the conventional heuristic optimization method.
Referring to fig. 10 and 11, in conjunction with fig. 12, the reconstruction of the high voltage distribution network by the method of the present disclosure can effectively alleviate the substation overload problem compared to the initial state.
Referring to fig. 13, it can be seen that the branch load rate can be controlled below 80% by the method herein.
Referring to fig. 14, it can be known that after network optimization, the branch load rate is controlled within the allowable capacity range in each N-1 scenario.

Claims (9)

1. The method for managing the urban high-voltage distribution network blocking based on data-driven heuristic optimization is characterized by comprising the following steps of:
s1: acquiring a node branch connection relation of an urban high-voltage distribution network, and acquiring a plurality of feasible topologies and node load samples of the high-voltage distribution network;
s2: load flow calculation is carried out on each feasible topology and node load sample, and node voltage and branch power corresponding to each feasible topology and node load sample are obtained;
s3: performing N-1 verification on each feasible topology and node load sample, and calculating an N-1 default quantity;
s4: establishing a node voltage prediction neuron network, fitting the node voltage corresponding to each feasible topology and node load sample, establishing a branch power prediction neuron network, fitting the branch power corresponding to each feasible topology and node load sample, establishing an N-1 default quantity prediction neuron network, and fitting the N-1 default quantity of each feasible topology and node load sample;
s5: the fitted node voltage prediction neuron network, the branch power prediction neuron network, the N-1 default quantity prediction neuron network, the transmission capacity of the power transmission line, the capacity of the transformer substation, the upper and lower limits of the output of the generator, the upper and lower limits of the voltage amplitude and the upper limit of the load shedding are used as constraint conditions, the minimum total load shedding is used as a target function, and the urban high-voltage distribution network blocking management and control model is constructed;
s6: and acquiring the topological state and node load shedding amount of the current high-voltage distribution network, solving the urban high-voltage distribution network blocking management and control model, and outputting the optimized topological state and load shedding amount of the high-voltage distribution network.
2. The urban high-voltage distribution network congestion management method based on data-driven heuristic optimization according to claim 1, wherein in step S1, a plurality of high-voltage distribution network feasible topology and node load samples are obtained based on a good point set theory.
3. The urban high-voltage distribution network congestion management method based on data-driven heuristic optimization according to claim 1 or 2, wherein in step S1, the expression of the feasible topology of the high-voltage distribution network is as follows:
Figure FDA0002307306310000011
Figure FDA0002307306310000012
wherein the content of the first and second substances,
Figure FDA0002307306310000013
represents the topological state of the high-voltage distribution network,
Figure FDA0002307306310000014
for its corresponding reachability matrix,
Figure FDA0002307306310000015
a matrix of connection relations between the power supply points is represented,
Figure FDA0002307306310000016
a connection relation matrix of the power supply point and the high-voltage distribution network node is shown,
Figure FDA0002307306310000017
a high-voltage distribution network node connection relation matrix is shown,
Figure FDA0002307306310000018
representation matrix
Figure FDA0002307306310000019
The (j) th column element of (1),
Figure FDA00023073063100000110
representation matrix
Figure FDA00023073063100000111
Row i and column j.
4. The urban high-voltage distribution network congestion management method based on data-driven heuristic optimization according to claim 1 or 2, wherein in step S1, the calculation process of the node load samples is as follows:
order to
Figure FDA0002307306310000021
Pn＝{rsk, | k ═ 1,2,3,.., n }, then the load at each node in the high voltage distribution network can be calculated as:
Figure FDA0002307306310000022
wherein s represents the number of nodes of the high-voltage distribution network, p represents a prime number satisfying p ≧ 2s +3, n represents the number of samples, rsRepresenting a point in s-dimensional space, PnRepresenting a set of good points of sample size n, PD,max、PD,minWhich represents the maximum and minimum values of the load,
Figure FDA0002307306310000023
representing the node load of the high voltage distribution network with a sample capacity of n.
5. The urban high-voltage distribution network blocking management method based on data-driven heuristic optimization according to claim 1, wherein in step S3, the calculation process of the N-1 default quantity is as follows:
Figure FDA0002307306310000024
Figure FDA0002307306310000025
Figure FDA0002307306310000026
Figure FDA0002307306310000027
Figure FDA0002307306310000028
Figure FDA0002307306310000029
Figure FDA00023073063100000210
Figure FDA00023073063100000211
wherein i represents a sample number; j. k represents a node number; m represents an N-1 scene number; l represents a branch number; cN-1Representing a set of N-1 scenes ξmRepresenting a default quantity for the mth N-1 scene;
Figure FDA00023073063100000212
and
Figure FDA00023073063100000213
respectively representing the margins of the upper and lower limits of the voltage amplitude and the upper and lower limits of the branch power; epsilon (x) represents a step function, and the condition with the margin being a negative value is screened out through step function mapping and is used as a default quantity in the scene, IiRepresenting the sum of all N-1 scene violations under the current sample.
6. The method for managing blocking of an urban high-voltage distribution network based on data-driven heuristic optimization according to claim 1, wherein in step S5, the urban high-voltage distribution network blocking management model is expressed as follows:
Figure FDA0002307306310000031
Figure FDA0002307306310000032
Figure FDA0002307306310000033
Figure FDA0002307306310000034
PL,min≤PL≤PL,max
Vmin≤V≤Vmax
0≤Δp≤Δpmax
where Δ p denotes a shear load vector, Δ piRepresenting the elements in the vector Δ p, n representing the total number of nodes, yN-1Indicating that N-1 parity passed, an output of 1 indicating pass, an output of 0 indicating fail, and PLRepresenting branch power vector, V representing node voltage vector, PL,min、PL,maxRepresenting the upper and lower limits of the branch power, Vmin、VmaxRepresenting the upper and lower limits of the voltage amplitude, Δ pmaxThe upper limit of the load cut is indicated.
7. The urban high-voltage distribution network blockage management method based on data-driven heuristic optimization according to claim 1, wherein the urban high-voltage distribution network blockage management model is solved using a genetic algorithm in step S6.
8. The urban high-voltage distribution network blocking management method based on data-driven heuristic optimization according to claim 1 or 7, wherein in step S6, the model solving process is as follows:
s61, initializing genetic algorithm parameters including a population scale, a cross factor, a variation factor, iteration times and convergence precision, and executing S62;
s62, initializing the topological state of the high-voltage distribution network, initializing the node load shedding amount, and executing S63;
s63, calculating a fitness function according to the current topological state of the high-voltage distribution network and the node load shedding amount, and executing S64;
s64, judging whether an exit condition is met, namely whether the maximum iteration number is reached or the convergence precision is met, if so, executing S66, and if not, executing S65;
s65, performing crossing, selecting and mutation operation, updating the topological state of the high-voltage distribution network and the node load shedding amount, and performing S63;
and S66, completing model solution, and outputting the optimized topological state of the high-voltage distribution network and the node load shedding amount.
9. The method for managing blocking of an urban high-voltage distribution network based on data-driven heuristic optimization according to claim 8, wherein in step S63, the fitness function is calculated as follows:
Figure FDA0002307306310000035
the first term represents the total load shedding amount of the high-voltage distribution network, the second term represents the punishment on the N-1 default amount, and the third term represents the node voltage out-of-limit punishment amount and the branch power out-of-limit punishment amount; y isN-1Indicating an N-1 parity-check-passed condition, an output of 1 indicating a pass check, an output of 0 indicating a fail check,
Figure FDA0002307306310000041
and
Figure FDA0002307306310000042
respectively representing the margins of the upper and lower limits of the voltage amplitude and the upper and lower limits of the branch power; Δ p denotes the tangential load vector, Δ piRepresenting elements in the vector delta p, epsilon (x) representing a step function, and screening out a state with a negative margin as a default quantity in the scene through step function mapping, IiRepresenting the sum of all N-1 scene violations under the current sample, α being a constant and β being a constant.
CN201911245120.7A 2019-12-06 2019-12-06 Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization Pending CN111105025A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911245120.7A CN111105025A (en) 2019-12-06 2019-12-06 Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911245120.7A CN111105025A (en) 2019-12-06 2019-12-06 Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization

Publications (1)

Publication Number Publication Date
CN111105025A true CN111105025A (en) 2020-05-05

Family

ID=70422065

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911245120.7A Pending CN111105025A (en) 2019-12-06 2019-12-06 Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization

Country Status (1)

Country Link
CN (1) CN111105025A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115545477A (en) * 2022-10-08 2022-12-30 广东电力交易中心有限责任公司 Power transmission line blocking risk probability assessment method and product based on incremental interpolation

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101734148B1 (en) * 2016-11-16 2017-05-11 이엔테크놀로지 주식회사 Congestion Management System For Transmission Line Overload Mitigation
CN107887908A (en) * 2017-11-20 2018-04-06 四川大学 Power transmission network based on VSC and high voltage distribution network topology reconstruction blocks management-control method
CN108964037A (en) * 2018-07-19 2018-12-07 四川大学 Based on the reconstitution model of high voltage distribution network
CN109449934A (en) * 2018-12-10 2019-03-08 四川大学 A kind of power transmission network obstruction management-control method based on the reconstruct of high voltage distribution network subregion

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101734148B1 (en) * 2016-11-16 2017-05-11 이엔테크놀로지 주식회사 Congestion Management System For Transmission Line Overload Mitigation
CN107887908A (en) * 2017-11-20 2018-04-06 四川大学 Power transmission network based on VSC and high voltage distribution network topology reconstruction blocks management-control method
CN108964037A (en) * 2018-07-19 2018-12-07 四川大学 Based on the reconstitution model of high voltage distribution network
CN109449934A (en) * 2018-12-10 2019-03-08 四川大学 A kind of power transmission network obstruction management-control method based on the reconstruct of high voltage distribution network subregion

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
YOUBO LIU等: "Data-driven Heuristic Optimization to Manage Congestion of Urban Power Grid", 《2019 IEEE INNOVATIVE SMART GRID TECHNOLOGIES – ASIA(ISGT ASIA)》 *
唐伦等: "计及N-1安全约束的输电网阻塞管控数据驱动模型", 《电测与仪表》 *
高心等: "《电力市场营销》", 31 May 2011 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115545477A (en) * 2022-10-08 2022-12-30 广东电力交易中心有限责任公司 Power transmission line blocking risk probability assessment method and product based on incremental interpolation
CN115545477B (en) * 2022-10-08 2023-07-14 广东电力交易中心有限责任公司 Power transmission line blocking risk probability assessment method and product based on incremental interpolation

Similar Documents

Publication Publication Date Title
Narimani et al. Enhanced gravitational search algorithm for multi‐objective distribution feeder reconfiguration considering reliability, loss and operational cost
Moeini et al. Fuzzy rule-based model for hydropower reservoirs operation
Saraswat et al. Multi-objective optimal reactive power dispatch considering voltage stability in power systems using HFMOEA
CN107256441B (en) Power distribution network planning construction scheme design method based on non-dominated sorting genetic algorithm
Shayeghi et al. Robust modified GA based multi-stage fuzzy LFC
Bhati et al. Robust fuzzy logic power system stabilizer based on evolution and learning
Shayeghi et al. Studying effect of bundle lines on TNEP considering network losses using decimal codification genetic algorithm
CN112310980B (en) Safety and stability evaluation method and system for direct-current blocking frequency of alternating-current and direct-current series-parallel power grid
Gao et al. Multi-objective dynamic reconfiguration for urban distribution network considering multi-level switching modes
CN111917134A (en) Power distribution network dynamic autonomous reconstruction method and system based on data driving
CN115940294A (en) Method, system, equipment and storage medium for adjusting real-time scheduling strategy of multi-stage power grid
Rao et al. Multi-objective adaptive clonal selection algorithm for solving optimal power flow problem with load uncertainty
CN114429248A (en) Transformer apparent power prediction method
CN111105025A (en) Urban high-voltage distribution network blocking management method based on data-driven heuristic optimization
CN108334950A (en) A kind of Distribution Network Reconfiguration using partheno genetic algorithm
Wu et al. A genetic approach for simultaneous design of membership functions and fuzzy control rules
Razmi et al. Neural network based on a genetic algorithm for power system loading margin estimation
Modi et al. Voltage stability evaluation of power system with FACTS devices using fuzzy neural network
Madeiro et al. Gradient-based algorithms for the automatic construction of fuzzy cognitive maps
Wai et al. Short-term load forecasting via fuzzy neural network with varied learning rates
Shterev et al. Time series prediction with neural networks: a review
CN111832836B (en) Power distribution network reconstruction method and system considering load power utilization characteristics
Sadegheih Optimization of network planning by the novel hybrid algorithms of intelligent optimization techniques
CN113240072A (en) Deep learning-based prediction method for direct-current micro-grid cluster
Li et al. A novel approach for dynamic reconfiguration of the distribution network via multi-agent system

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20200505

RJ01 Rejection of invention patent application after publication