CN111049159B - Power system transient stability prevention control method embedded in deep belief network - Google Patents
Power system transient stability prevention control method embedded in deep belief network Download PDFInfo
- Publication number
- CN111049159B CN111049159B CN201911330999.5A CN201911330999A CN111049159B CN 111049159 B CN111049159 B CN 111049159B CN 201911330999 A CN201911330999 A CN 201911330999A CN 111049159 B CN111049159 B CN 111049159B
- Authority
- CN
- China
- Prior art keywords
- generator
- transient stability
- belief network
- deep belief
- output
- 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
- 230000001052 transient effect Effects 0.000 title claims abstract description 107
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000002265 prevention Effects 0.000 title claims abstract description 31
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 26
- 238000004364 calculation method Methods 0.000 claims abstract description 20
- 238000004088 simulation Methods 0.000 claims abstract description 20
- 238000011217 control strategy Methods 0.000 claims abstract description 15
- 238000005457 optimization Methods 0.000 claims abstract description 14
- 230000003449 preventive effect Effects 0.000 claims description 19
- 238000005070 sampling Methods 0.000 claims description 11
- 230000002068 genetic effect Effects 0.000 claims description 7
- 230000003828 downregulation Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 12
- 230000001276 controlling effect Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000002222 downregulating effect Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000013097 stability assessment 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
- H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
-
- 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
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
- H02J3/48—Controlling the sharing of the in-phase component
-
- 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
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
- H02J3/50—Controlling the sharing of the out-of-phase component
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to the technical field of power system automation, and aims to provide a power system transient stability prevention control method embedded in a deep belief network, which comprises the following steps of S1: determining the active output fluctuation range and the load fluctuation range of the generator, generating N active output samples of the generator, acquiring a large amount of initial state data, performing time domain simulation calculation on the initial state data, and generating sample data; s2: establishing a deep belief network, training the deep belief network by using sample data, fitting the active output of the generator and the transient stability of the system, and generating a transient stability evaluator of the power system; s3: based on transient stability constraint conditions, adding cost constraint, tide constraint and stable operation constraint for an NSGA-II algorithm, and building an NSGA-II evolution algorithm model; s4: and acquiring a generator output fluctuation range and a load fluctuation range of transient instability under a fault, and carrying out iterative optimization on an NSGA-II evolutionary algorithm model to obtain a prevention control strategy.
Description
Technical Field
The invention belongs to the technical field of power system automation, and particularly relates to a power system transient stability prevention control method embedded in a deep belief network.
Background
With the increasing scale of the power system, the network structure is more complex, the system operating point is closer to the stability limit, and the requirements on the stability prevention and control of the power system are higher. Transient instability is often the main cause of large-scale accidents of the power system, and effective power system transient stability assessment and accident prevention measures are key to solving the problem. The traditional transient stability calculation generally adopts a method of time domain simulation and proper criteria, has the advantages of accurate calculation and high reliability, but the model contains a nonlinear differential algebraic equation, is complex in calculation and long in calculation time, and is difficult to meet the requirement of on-line calculation. The deep learning model has the advantages of automatic feature extraction, strong abstract capability and good convergence, and the network structure is deeper, so that the method is more beneficial to finding the internal rules of data, and is used for transient stability evaluation of the power system. The prevention control means that the potential fault risk of the system is found in advance by identifying the current system state before the system fails, and the system is regulated to a state which can still stably run after the fault by regulating the output force of the generator and changing the load. The prevention control and the transient stability are combined, the transient stability prevention control is provided, and the system is ensured to run in a state meeting the transient stability requirement.
Disclosure of Invention
The invention aims to provide a method for preventing and controlling transient stability of a power system embedded in a deep belief network, which introduces the deep belief network and NSGA-II into transient stability prevention control, and realizes quick and stable solving of a transient stability prevention control optimization strategy aiming at faults.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a method for preventing and controlling transient stability of an electric power system embedded in a deep belief network comprises the following steps:
s1: determining the active output fluctuation range and the load fluctuation range of the generator, generating N active output samples of the generator, acquiring a large amount of initial state data, performing time domain simulation calculation on the initial state data, generating sample data, and executing S2;
s2: establishing a deep belief network, training the deep belief network by using sample data, fitting the active output of the generator and the transient stability of the system, generating a transient stability estimator of the power system, embedding the transient stability estimator of the power system into a non-dominant ordering genetic algorithm NSGA-II as a transient stability constraint condition, and executing S3;
s3: based on transient stability constraint conditions, adding cost constraint, tide constraint and stable operation constraint for an NSGA-II algorithm, building an NSGA-II evolution algorithm model, and executing S4;
s4: and acquiring a generator output fluctuation range and a load fluctuation range of transient instability under a fault, and carrying out iterative optimization on an NSGA-II evolutionary algorithm model to obtain a prevention control strategy.
Preferably, in the step S1, N power generator active output samples are obtained according to the power generator active output fluctuation range and the load fluctuation range by using a Latin hypercube sampling algorithm.
Preferably, the step of S1 generating sample data includes the steps of,
s11: selecting M lines in a system as an expected fault set, selecting one line as a fault line in each time domain simulation calculation, wherein the fault type of the fault line is three-phase short-circuit fault, and executing S12;
s12: combining M fault lines in the expected fault set and N generator active output samples two by two to generate M-N transient stability simulation initial conditions, and executing S13;
s13: performing time domain simulation calculation on M.N transient stability simulation initial conditions, solving a power angle curve of the generator, calculating M.N transient stability coefficients TSI, wherein each generator has M corresponding active power output, and executing S14;
s14: and selecting the minimum TSI in the M TSIs and the corresponding generator active power to be combined into sample data of a training deep belief network, and generating N sample data.
Preferably, the formula of the TSI is:
wherein delta max For the maximum power angle difference between any two generators of the system, when TSI>0, the system transient stability, and the larger the TSI value is, the higher the system transient stability is; when TSI<0, system transient instability.
Preferably, the embedding of the power system transient stability estimator in the non-dominant ranking genetic algorithm NSGA-II in S2 is expressed as:
wherein, the active output of the generator is given for preventive control; phi (P) G ) For the trained DBN model, the input is the active output of all generators of the system, the output is the TSI estimated by the model, and when the estimated TSI>And 0, the DBN model is considered to be stable in system transient state under the condition of active force, otherwise, the system transient state is considered to be unstable, and prevention and control are needed.
Preferably, the cost constraint is expressed as
Wherein C is Ui The cost is adjusted up for the output of the generator; c (C) Di The cost is adjusted downwards for the output of the generator; p (P) Oi To prevent the output of the generator before control; p (P) Pi To prevent post-control generator output; ΔP Ui The output of the generator is adjusted upwards; ΔP Di And (5) down-regulating the output of the generator.
Preferably, the power flow constraint is expressed as
Wherein P is Ni And Q Ni Injecting power for node active and reactive power; p (P) Di And Q Di Active and reactive output power for the node; v (V) i And V j The node voltage amplitude; alpha ij Is the node voltage phase angle difference; g ij And B ij Real and imaginary parts of node admittance; s is S n Is a collection of nodes.
Preferably, the steady operation constraint is expressed as
Wherein, and->The upper limit and the lower limit of the active output of the generator are set; />And->Outputting an upper limit and a lower limit for the reactive power source;and->The upper and lower limits of the node voltage; />And->Constraining upper and lower limits for line thermal stability; s is S l Is a collection of lines.
Preferably, said S4 comprises the steps of,
s41: acquiring the active output of the generator, inputting the active output to a deep belief network, and executing S42;
s42: judging whether the system is unstable or not and needs preventive control, if yes, executing S43, and if not, executing S43;
s43: performing cross and inheritance iterative optimization on the NSGA-II algorithm, and executing S43;
s44: if the iteration number reaches the maximum iteration number, executing S45, and if not, executing S43;
s45: and outputting a preventive control strategy.
Preferably, the deep belief network is based on a Keras framework of TensorFlow, and the number of layers of the deep belief network is four, including two limited Boltzmann machine layers and one full connection layer.
In summary, the beneficial effects of the invention are as follows:
according to the invention, the deep belief network and NSGA-II are introduced into transient stability prevention control, so that the rapid and stable solving of the transient stability prevention control optimization strategy aiming at faults is realized.
Drawings
FIG. 1 is a schematic diagram showing the steps of a method for transient stability prevention control of a power system embedded in a deep belief network according to the present invention;
FIG. 2 is a schematic diagram of the present invention for demonstrating Latin hypercube sampling principles;
FIG. 3 is a schematic diagram of the step S1 of the present invention for showing a method for transient stability prevention control of a power system embedded in a deep belief network;
FIG. 4 is a schematic diagram of the present invention for showing a deep belief network;
FIG. 5 is a schematic diagram of the present invention for showing a transient stability estimator of a power system;
FIG. 6 is a schematic diagram showing the deep belief network embedded NSGA-II according to the present invention;
FIG. 7 is a flow chart of S4 for showing a method for controlling transient stability prevention of a power system embedded in a deep belief network according to the present invention;
FIG. 8 is a schematic diagram of the step S4 of the method for controlling transient stability prevention of a power system embedded in a deep belief network according to the present invention;
FIG. 9 is a schematic diagram showing an IEEE39, 68, 140 node system in accordance with an embodiment of the invention;
FIG. 10 is a schematic diagram showing the generation of 1000 generator active power cases from Latin hypercube sampling in accordance with an embodiment of the invention;
FIG. 11 is a schematic diagram showing the active power output of a generator after preventive control according to an embodiment of the present invention;
FIG. 12 is a schematic diagram showing costs after preventive control according to an embodiment of the present invention;
FIG. 13 is a schematic diagram showing TSI after preventative control according to an embodiment of the present invention;
FIG. 14 is a graph showing the pre-and post-prevention control angle of a 39-node system under fault 7, in accordance with an embodiment of the present invention;
FIG. 15 is a graph showing the pre-and post-prevention control angle of a 39-node system under fault 10, in accordance with an embodiment of the present invention;
FIG. 16 is a graph showing the pre-and post-prevention control angle of a 68-node system under fault 8 in accordance with an embodiment of the present invention;
FIG. 17 is a graph showing the pre-and post-prevention control angle of a 68-node system under fault 9 in accordance with an embodiment of the present invention;
FIG. 18 is a graph showing the pre-and post-prevention control angle of a 140 node system under fault 1 in accordance with an embodiment of the present invention;
fig. 19 is a graph showing the pre-and post-prevention control angle of a 140-node system under fault 2 according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made more clearly and fully with reference to the accompanying drawings 1-19, in which it is evident that the embodiments described are only some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Examples
Referring to fig. 1, a method for preventing and controlling transient stability of an electric power system embedded in a deep belief network includes the steps of:
s1: determining the active output fluctuation range and the load fluctuation range of the generator, generating N active output samples of the generator, acquiring a large amount of initial state data, performing time domain simulation calculation on the initial state data, generating sample data, and executing S2;
s2: establishing a deep belief network, training the deep belief network by using sample data, fitting the active output of the generator and the transient stability of the system, generating a transient stability estimator of the power system, embedding the transient stability estimator of the power system into a non-dominant ordering genetic algorithm NSGA-II as a transient stability constraint condition, and executing S3;
s3: based on transient stability constraint conditions, adding cost constraint, tide constraint and stable operation constraint for an NSGA-II algorithm, building an NSGA-II evolution algorithm model, and executing S4;
s4: and acquiring a generator output fluctuation range and a load fluctuation range of transient instability under a fault, and carrying out iterative optimization on an NSGA-II evolutionary algorithm model to obtain a prevention control strategy.
S1 is described in detail below.
Specifically, in S1, N power generator active output samples are obtained according to the power generator active output fluctuation range and the load fluctuation range by using a Latin hypercube sampling algorithm.
Referring to fig. 2, the latin hypercube sampling principle is: according to the sampling number N, the sample value range is divided into N equal parts, and a point is selected in each equal part, so that the sample is spread over the whole sample space and has certain randomness. The method comprises the steps of setting the active output of a generator to fluctuate within a range of 90% -110%, sampling the active output of the generator based on Latin hypercube sampling, generating N uniformly distributed active output samples of the generator within the active output range of the generator, and enabling the load active and reactive power to fluctuate up and down along with the active of the generator according to the principles of active balance and load power factor constancy.
Specifically, referring to fig. 3, S1 generating sample data includes the steps of,
s11: selecting M lines in a system as an expected fault set, selecting one line as a fault line in each time domain simulation calculation, wherein the fault type of the fault line is three-phase short-circuit fault, and executing S12;
s12: combining N generator active output samples generated by M fault lines in an expected fault set and Latin hypercube sampling in pairs to generate M.N transient stability simulation initial conditions, and executing S13;
s13: performing time domain simulation calculation on M.N transient stability simulation initial conditions, solving a power angle curve of the generator, calculating M.N transient stability coefficients TSI, wherein each generator has M corresponding active power output, and executing S14;
s14: and selecting the minimum TSI in the M TSIs and the corresponding generator active power to be combined into sample data of a training deep belief network, and generating N sample data.
Specifically, the formula of TSI is:
wherein delta max For the maximum power angle difference between any two generators of the system, when TSI>0, the system transient stability, and the larger the TSI value is, the higher the system transient stability is; when TSI<0, system transient instability.
S2 is described in detail below.
Specifically, training a deep belief network by using sample data, fitting a nonlinear relation between the active output of a generator and the transient stability of the system, and generating a transient stability estimator of the power system based on the deep belief network.
Specifically, referring to fig. 4, the deep belief network is built based on a Keras framework of a TensorFlow, and the number of layers of the built deep belief network is 4, and the deep belief network is formed by stacking 2 limited boltzmann machine layers and a full-connection layer. The deep belief network training is divided into two stages, wherein the first stage is pre-training, and each layer of limited boltzmann machine performs greedy layer-by-layer unsupervised learning by using sample data without labels. Through pre-training, the deep belief network is near the optimal solution, so that the problem that the deep neural network cannot be trained due to gradient loss or gradient explosion is solved. The second stage is to train the model integrally by using sample data with labels, and fine-tune the weights and the biases on the basis of pre-training by a random gradient descent algorithm and back propagation to achieve the best fitting effect.
The energy function and the joint probability function of the deep belief network are respectively:
E(v,h 1 ,h 2 ,h 3 )＝-v T W 1 h 1 -
let the number of generators be L, then the neuron number of each layer of deep belief network is in proper order: l-100-50-1. The learning rate is set to be 0.0001, the batch processing number is 50, the pre-training times are 50, the activation function is ReLU, the iteration times of the whole training is 1000, a value range can be set for the selection of the super parameters, then the super parameters are optimized through an optimizing algorithm such as a particle swarm algorithm, the accuracy of the deep belief network is used as a standard for judging the performance of the super parameters, and the optimal super parameters are found, so that the deep belief network achieves the best fitting effect.
Specifically, a learning rate attenuation method and MSE and L2 regularization are adopted in the model training process. The learning rate attenuation method can improve learning speed in the early stage of training and improve evaluation accuracy in the later stage of training; the gradient of MSE loss is reduced along with the reduction of the loss, the gradient is very small when the loss tends to 0, and the MSE is more accurate than the average absolute error MAE calculation result at the end of training; l2 regularization can prevent model overfitting. The mathematical expressions are respectively:
wherein L is r Is the learning rate; d is a learning rate attenuation coefficient; e is training times; phi (x) is a DBN model; x is x i Is a training set; y is i Is equal to x i A corresponding tag value; m is the number of training set samples; n is the number of DBN layers; omega i Is a weight coefficient; lambda is a regularization parameter.
It should be noted that, referring to fig. 5, the transient stability estimator of the power system is embedded in the non-dominant ranking genetic algorithm NSGA-II, as a transient stability constraint condition, specifically:
referring to fig. 6, the transient stability estimator of the power system is embedded in the non-dominant ordered genetic algorithm NSGA-II, and instead of solving the time domain equation, for determining the transient stability of the system, the transient stability estimator can be represented by the following formula:
in the method, in the process of the invention,the active output of the generator is given for preventive control; phi (P) G ) For the trained DBN model, the input is the active output of all generators of the system, the output is the TSI estimated by the model, and when the estimated TSI>And 0, the DBN model is considered to be stable in system transient state under the condition of active force, otherwise, the system transient state is considered to be unstable, and prevention and control are needed.
S3 is specifically described below.
Specifically, NSGA-II is a multi-objective optimization algorithm, and can find an optimization result which simultaneously meets a plurality of constraint conditions. In the invention, NSGA-II has four optimization targets, namely: and the control and adjustment cost, the tide constraint, the stable operation constraint and the transient stability constraint are prevented. The NSGA-II algorithm can consider four optimization target calculation results at the same time, and gives out an optimal prevention control strategy in a crossing, mutation and loop iteration mode.
In particular, the cost constraint is expressed as
Wherein C is Ui The cost is adjusted up for the output of the generator; c (C) Di The cost is adjusted downwards for the output of the generator; p (P) Oi To prevent the output of the generator before control; p (P) Pi To prevent post-control generator output; ΔP Ui The output of the generator is adjusted upwards; ΔP Di And (5) down-regulating the output of the generator.
Specifically, the flow constraint is expressed as
Wherein P is Ni And Q Ni Injecting power for node active and reactive power; p (P) Di And Q Di Active and reactive output power for the node; v (V) i And V j The node voltage amplitude; alpha ij Is the node voltage phase angle difference; g ij And B ij Real and imaginary parts of node admittance; s is S n Is a collection of nodes.
Specifically, the steady operation constraint is expressed as
Wherein, and->The upper limit and the lower limit of the active output of the generator are set; />And->Outputting an upper limit and a lower limit for the reactive power source;and->The upper and lower limits of the node voltage; />And->Constraining upper and lower limits for line thermal stability; s is S l Is a collection of lines.
S4 will be described in detail below.
Referring to fig. 7 and 8, S4 includes the steps of,
s41: acquiring real-time generator active output, inputting the real-time generator active output to a deep belief network, and executing S42;
s42: judging whether the system is unstable or not and needs preventive control, if yes, executing S43, and if not, executing S43;
it is worth to say that, the transient stability estimator of the power system calculates TSI, when the estimated TSI is more than 0, the DBN model is regarded as the transient stability of the system under the condition of the active output, otherwise, the transient instability of the system is regarded as the transient instability of the system, and the prevention control is needed;
s43: performing cross and inheritance iterative optimization on the NSGA-II algorithm, and executing S43;
the method is worthy of explanation, wherein the power flow constraint and the stable operation constraint are judged whether to be met or not according to a power flow calculation result, the transient stability constraint is judged whether to be met or not according to a deep belief network evaluation result, and the prevention control adjustment cost is obtained by calculating the difference value between the active output of the generator and the original active output of the generator according to a prevention control strategy;
s44: if the iteration number reaches the maximum iteration number, executing S45, and if not, executing S43;
s45: and outputting a preventive control strategy.
It is worth to say that, step S45 specifically includes, after obtaining the preventive control strategy, using a power system tool box to calculate a power angle curve of the generator under the preventive control strategy, calculating a transient stability coefficient TSI, and judging the accuracy of the preventive control strategy; the dispatcher adjusts the original active output of the generator of the system according to the active output of the generator given by the preventive control strategy, so that the system meets the transient stability requirement.
Referring to fig. 9, the following describes a control method for preventing transient stability of a power system embedded in a deep belief network according to the present embodiment, taking an IEEE39, 68, 140 node system as an example.
10 lines among all lines of each system were selected as expected failure lines, and three-phase short-circuit failure among the lines was set to constitute an expected failure set, as shown in table 1.
TABLE 1
Referring to fig. 10, the generator active force is set to fluctuate in the range of 90% -110% for the expected failure concentration failure line and failure type. The Latin hypercube sampling is used for generating 1000 active output conditions of the generator, and according to the principles of active power balance and constant load power factor, the active and reactive loads of the system float up and down integrally along with the change of the sum of the active outputs of the generator. The 1000 power generator active power conditions are combined with 10 expected fault lines in the expected fault set to form 10000 expected fault data. And performing time domain simulation calculation on the expected fault data by using a power system toolbox PST, setting the fault removal time to be 0.1 second, setting the total simulation time to be 20 seconds, and solving the corresponding TSI. Each generator output condition corresponds to 10 TSIs, and the minimum TSI in the 10 TSIs is screened to form 1000 sample data for training the transient stability estimator together with the generator output.
Specifically, 1000 sample data are divided into a training set containing 800 data and a test set containing 200 data, the transient stability evaluator is trained using the training set, and model accuracy is verified by the test set. The invention uses Keras framework based on TensorFlow to build the deep belief network, the number of layers of the built deep belief network is 4, and the built deep belief network comprises 2 limited Boltzmann machine layers and a full connection layer. Let the number of generators be L, then the neuron number of each layer of deep belief network is in proper order: l-100-50-1. The learning rate is set to be 0.0001, the batch processing number is 50, the pre-training times are 50, the activation function is ReLU, the iteration times of the whole training is 1000, a value range can be set for the selection of the super parameters, then the super parameters are optimized through an optimizing algorithm such as a particle swarm algorithm, the accuracy of the deep belief network is used as a standard for judging the performance of the super parameters, and the optimal super parameters are found, so that the deep belief network achieves the best fitting effect. The different test system identification accuracy pairs are shown in table 2:
TABLE 2
Referring to fig. 7, the output fluctuation range and the load fluctuation range of the generator with transient instability under the fault are obtained, and the NSGA-II evolutionary algorithm model is subjected to iterative optimization to obtain a prevention control strategy.
Referring to fig. 11 to 13, the preventive control strategy was verified using the power system tool box PST, and the TSI before and after preventive control for each expected failure was calculated by a time domain simulation method. The transient unstable faults before the prevention and control are changed into transient stability through the prevention and control, the transient stable faults before the prevention and control still keep the transient stability, and TSI is slightly improved, so that the strategy is shown to properly increase the transient stability margin of the system facing the full fault set.
Referring to fig. 14-19, comparison of the angles of work before and after the preventive control shows that the system returns to transient stability from transient instability through predictive control, thereby verifying the feasibility of the method.
In the description of the present invention, it should be understood that the terms "counterclockwise," "clockwise," "longitudinal," "transverse," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, are merely for convenience in describing the present invention, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention.
Claims (10)
1. A method for preventing and controlling transient stability of an electric power system embedded in a deep belief network is characterized by comprising the following steps,
s1: determining the active output fluctuation range and the load fluctuation range of the generator, generating N active output samples of the generator, acquiring a large amount of initial state data, performing time domain simulation calculation on the initial state data, generating sample data, and executing S2;
s2: establishing a deep belief network, training the deep belief network by using sample data, fitting the active output of the generator and the transient stability of the system, generating a transient stability estimator of the power system, embedding the transient stability estimator of the power system into a non-dominant ordering genetic algorithm NSGA-II as a transient stability constraint condition, and executing S3;
s3: based on transient stability constraint conditions, adding cost constraint, tide constraint and stable operation constraint for an NSGA-II algorithm, building an NSGA-II evolution algorithm model, and executing S4;
s4: and acquiring a generator output fluctuation range and a load fluctuation range of transient instability under a fault, and carrying out iterative optimization on an NSGA-II evolutionary algorithm model to obtain a prevention control strategy.
2. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 1, wherein N samples of the generator active power output are obtained according to the generator active power output fluctuation range and the load fluctuation range by using a latin hypercube sampling algorithm in S1.
3. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 1 or 2, wherein the step of S1 generating sample data comprises the steps of,
s11: selecting M lines in a system as an expected fault set, selecting one line as a fault line in each time domain simulation calculation, wherein the fault type of the fault line is three-phase short-circuit fault, and executing S12;
s12: combining M fault lines in the expected fault set and N generator active output samples two by two to generate M-N transient stability simulation initial conditions, and executing S13;
s13: performing time domain simulation calculation on M.N transient stability simulation initial conditions, solving a power angle curve of the generator, calculating M.N transient stability coefficients TSI, wherein each generator has M corresponding active power output, and executing S14;
s14: and selecting the minimum TSI in the M TSIs and the corresponding generator active power to be combined into sample data of a training deep belief network, and generating N sample data.
4. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 3, wherein the formula of calculation of the TSI is:
wherein, for the maximum power angle difference between any two generators of the system, when TSI>0, the system transient stability, and the larger the TSI value is, the higher the system transient stability is; when TSI<0, system transient instability.
5. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 1, wherein the embedding of the transient stability estimator of the power system in the non-dominant ranking genetic algorithm NSGA-II is expressed as:
wherein, the active output of the generator is given for preventive control;for the trained DBN model, the input is the active output of all generators of the system, the output is the TSI estimated by the model, and when the estimated TSI>And 0, the DBN model is considered to be stable in system transient state under the condition of active force, otherwise, the system transient state is considered to be unstable, preventive control is needed,nindicating the number of generators.
6. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 1, wherein said cost constraint is expressed as
Wherein, C iU the cost is adjusted up for the output of the generator;C iD the cost is adjusted downwards for the output of the generator;P iO to prevent the output of the generator before control;P iP to prevent post-control generator output;ΔP iU the output of the generator is adjusted upwards;ΔP iD for the generator output down-regulation value,Sgrepresenting a set of generators.
7. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 1, wherein the power flow constraint is expressed as
Wherein, P iN andQ iN injecting power for node active and reactive power;P iD andQ iD active and reactive output power for the node;V i andV j the node voltage amplitude;is the node voltage phase angle difference; />And->Real and imaginary parts of node admittance;S n for a set of nodes,jthe value range of (2) is [1, n ]]。
8. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to claim 1, wherein the steady operation constraint is expressed as
Wherein, andthe upper limit and the lower limit of the active output of the generator are set;andoutputting an upper limit and a lower limit for the reactive power source;andthe upper and lower limits of the node voltage;andconstraining upper and lower limits for line thermal stability;S l for the collection of lines,a set of nodes is represented and,representing a set of generators.
9. The method for preventing and controlling transient stability of a power system embedded in a deep belief network according to any one of claims 4 to 8, wherein S4 comprises the steps of,
s41: acquiring the active output of the generator, inputting the active output to a deep belief network, and executing S42;
s42: judging whether the system is unstable or not and needs preventive control, if yes, executing S43, and if not, executing S41;
s43: performing cross and inheritance iterative optimization on the NSGA-II algorithm, and executing S44;
s44: if the iteration number reaches the maximum iteration number, executing S45, and if not, executing S43;
s45: and outputting a preventive control strategy.
10. The method for preventing and controlling transient stability of an electric power system embedded in a deep belief network according to any one of claims 4 to 8, wherein the deep belief network is based on a kenasframe of a TensorFlow, and the number of layers of the deep belief network is four, including two limited boltzmann machine layers and a fully connected layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911330999.5A CN111049159B (en) | 2019-12-20 | 2019-12-20 | Power system transient stability prevention control method embedded in deep belief network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911330999.5A CN111049159B (en) | 2019-12-20 | 2019-12-20 | Power system transient stability prevention control method embedded in deep belief network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111049159A CN111049159A (en) | 2020-04-21 |
| CN111049159B true CN111049159B (en) | 2023-09-29 |
Family
ID=70238189
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911330999.5A Active CN111049159B (en) | 2019-12-20 | 2019-12-20 | Power system transient stability prevention control method embedded in deep belief network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111049159B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111950765B (en) * | 2020-07-06 | 2024-04-19 | 四川大川云能科技有限公司 | Probabilistic transient stability prediction method based on stacked noise reduction self-encoder |
| CN113591379B (en) * | 2021-07-27 | 2023-11-14 | 四川大学 | Auxiliary decision-making method for transient stability prevention and emergency coordination control of power system |
| CN113904384B (en) * | 2021-11-09 | 2023-06-09 | 国网四川省电力公司电力科学研究院 | Power grid transient stability coordination control method and system based on gradient elevator |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS63181618A (en) * | 1987-01-22 | 1988-07-26 | 中部電力株式会社 | Preventive controller of power system |
| CN107590564A (en) * | 2017-09-08 | 2018-01-16 | 四川大学 | Power system active power output method of adjustment based on Transient Stability Constraints |
| CN109086913A (en) * | 2018-07-11 | 2018-12-25 | 山东大学 | A kind of transient stability evaluation in power system method and system based on deep learning |
| CN110163540A (en) * | 2019-06-28 | 2019-08-23 | 清华大学 | Electric power system transient stability prevention and control method and system |
-
2019
- 2019-12-20 CN CN201911330999.5A patent/CN111049159B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS63181618A (en) * | 1987-01-22 | 1988-07-26 | 中部電力株式会社 | Preventive controller of power system |
| CN107590564A (en) * | 2017-09-08 | 2018-01-16 | 四川大学 | Power system active power output method of adjustment based on Transient Stability Constraints |
| CN109086913A (en) * | 2018-07-11 | 2018-12-25 | 山东大学 | A kind of transient stability evaluation in power system method and system based on deep learning |
| CN110163540A (en) * | 2019-06-28 | 2019-08-23 | 清华大学 | Electric power system transient stability prevention and control method and system |
Non-Patent Citations (3)
| Title |
|---|
| A systematic approach for dynamic security assessment and the corresponding preventive control scheme based on decision trees;Liu C,etc;《IEEE Transactions on Power Systems》;第717-730页 * |
| 基于灵敏度分析和时域仿真的暂态稳定预防控制优化方法;田芳等;《电力自动化设备》(第07期);第160-166页 * |
| 基于神经网络预测校核的暂态稳定预防控制;杨跃，等;《电网技术》;第264-272页;第264-272页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111049159A (en) | 2020-04-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Saxena et al. | Reactive power control in decentralized hybrid power system with STATCOM using GA, ANN and ANFIS methods | |
| CN111049159B (en) | Power system transient stability prevention control method embedded in deep belief network | |
| Bevrani et al. | An intelligent droop control for simultaneous voltage and frequency regulation in islanded microgrids | |
| CN111523785A (en) | Power system dynamic security assessment method based on generation countermeasure network | |
| CN112310980B (en) | Safety and stability evaluation method and system for direct-current blocking frequency of alternating-current and direct-current series-parallel power grid | |
| CN105429134B (en) | A kind of Network Voltage Stability Forecasting Methodology based on electric power big data | |
| CN116245033B (en) | Artificial intelligent driven power system analysis method and intelligent software platform | |
| CN112947672B (en) | Maximum power point tracking method and device for photovoltaic cell | |
| Zhang et al. | Cyber-physical cooperative response strategy for consensus-based hierarchical control in micro-grid facing with communication interruption | |
| Su et al. | An optimized algorithm for optimal power flow based on deep learning | |
| Wang et al. | Dynamic equivalent method of PMSG‐based wind farm for power system stability analysis | |
| CN112001066A (en) | Deep learning-based method for calculating limit transmission capacity | |
| CN114429248A (en) | Transformer apparent power prediction method | |
| Darabian et al. | A UPFC‐based robust damping controller for optimal use of renewable energy sources in modern renewable integrated power systems | |
| Hossain et al. | Machine learning accelerated real-time model predictive control for power systems | |
| Razmi et al. | Neural network based on a genetic algorithm for power system loading margin estimation | |
| Wei et al. | A combination forecasting method of grey neural network based on genetic algorithm | |
| Zhang et al. | CNN‐LSTM based power grid voltage stability emergency control coordination strategy | |
| Pepiciello et al. | Artificial neural Network-based small signal stability analysis of power systems | |
| CN111950765B (en) | Probabilistic transient stability prediction method based on stacked noise reduction self-encoder | |
| Andreoiu et al. | Lyapunov's method based genetic algorithm for multi-machine PSS tuning | |
| Aththanayake et al. | Performance Analysis of Regression and Artificial Neural Network Schemes for Dynamic Model Reduction of Power Systems | |
| CN113241793A (en) | Prevention control method for power system with IPFC (intelligent power flow controller) considering wind power scene | |
| CN106682760A (en) | Wind power climbing prediction method | |
| Zarkovic et al. | ANN for solving the harmonic load flow in electric power systems with DG |
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 |