CN115034493A - Wind power plant black start path optimization method considering unit operation state - Google Patents
Wind power plant black start path optimization method considering unit operation state Download PDFInfo
- Publication number
- CN115034493A CN115034493A CN202210721339.5A CN202210721339A CN115034493A CN 115034493 A CN115034493 A CN 115034493A CN 202210721339 A CN202210721339 A CN 202210721339A CN 115034493 A CN115034493 A CN 115034493A
- Authority
- CN
- China
- Prior art keywords
- black start
- wind power
- path
- power plant
- unit
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 241001672018 Cercomela melanura Species 0.000 title claims abstract description 46
- 238000005457 optimization Methods 0.000 title claims abstract description 18
- 238000010977 unit operation Methods 0.000 title claims abstract description 9
- 230000008569 process Effects 0.000 claims abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000011156 evaluation Methods 0.000 claims abstract description 15
- 238000012360 testing method Methods 0.000 claims abstract description 12
- 230000001413 cellular effect Effects 0.000 claims abstract description 9
- 238000012216 screening Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000010219 correlation analysis Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 7
- 238000003491 array Methods 0.000 claims description 5
- 238000010248 power generation Methods 0.000 claims description 5
- 238000011217 control strategy Methods 0.000 claims description 3
- 238000007562 laser obscuration time method Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 12
- 238000011084 recovery Methods 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/001—Methods to deal with contingencies, e.g. abnormalities, faults or failures
-
- 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
- H02J3/388—Islanding, i.e. disconnection of local power supply from the network
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/06—Wind turbines or wind farms
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/28—The renewable source being wind energy
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/76—Power conversion electric or electronic aspects
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- Power Engineering (AREA)
- Marketing (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Geometry (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Control Of Eletrric Generators (AREA)
Abstract
The invention discloses a wind power plant black start path optimization method considering unit operation state, which comprises the steps of obtaining SCADA operation data of an offshore wind power plant, screening key parameters, generating an input/output cellular array to obtain a training set and a test set, and constructingPredicting a model and carrying out model training; sending the test set into a prediction model to obtain the deviation degree between the actual value and the prediction value; obtaining the unit state evaluation coefficientProviding guidance for path optimization in the black start process; obtaining a generating capacity target function and a black start electrical distance target function of a wind power plant; and integrating the generating capacity objective function and the electrical distance objective function, solving mixed integer linear programming to obtain a pareto front, and selecting an optimal black start path of the wind power plant. According to the method, a reasonable wind power plant black start path optimization method is adopted, the running state of a wind turbine generator is considered, the electrical distance and the total generated energy are balanced to obtain the black start path, and the stability of the isolated grid black start initial stage with poor robustness is improved.
Description
Technical Field
The invention belongs to the technical field of black start of a power grid, and particularly relates to a wind power plant black start path optimization method considering a unit running state.
Background
The offshore wind power grid-connected capacity is expected to continue to increase, but the influence of large-scale wind power access on the power grid is more remarkable. A large number of power electronic devices make the structure of a power grid more and more complex, the problems of thermal stability and transient stability of a power transmission channel are prominent, the possibility of power failure is obviously increased, and the uncertainty of wind power will aggravate the frequency fluctuation of the power grid.
The evaluation of the black start path scheme is to form a sequence on the basis of comprehensively evaluating various indexes in the recovery process and the running process so as to make and execute a recovery plan by a dispatching center. The research targets of the power system recovery path are as follows: the voltage conversion times of the recovery system are reduced; reducing the switching operation times of system equipment; the system starting time is reduced; increasing the system capacity; shortening the length of the system cable; the quality of each parameter of the system is improved. The number of return wires in the wind power plant is small and fixed, the starting candidate routes are different from those in the traditional electric field, the stability of the wind power plant at the initial stage of isolated grid black starting is poor, and if the wind turbine generator fails at the initial stage of black starting, the system state is further deteriorated, and even the black starting fails.
For the started wind power plant, because the type and the capacity of the fan, the generator, the grid-side controller and the step-up transformer of each fan are the same, the switching operation procedure in the black start process is the same, and the start time of each unit is similar. Therefore, the black start method of the offshore wind power system mainly considers the balance of the electrical distance and the total power generation amount in the unit operation state and considers the influence of the state of the wind turbine on the start of the wind power plant.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a method for optimizing a black start path of a wind power plant by considering the running state of a wind turbine generator, to solve the technical problem of the prior art, wherein the method for optimizing the black start path of the wind power plant by considering the running state of the wind turbine generator is provided for balancing the electrical distance and the total generated energy, and the purpose of optimizing the black start path by considering the cooperation of the wind turbine generator and an auxiliary power supply in an isolated network is achieved.
In order to achieve the technical purpose, the technical scheme adopted by the invention is as follows:
a wind power plant black start path optimization method considering unit operation states comprises the following steps:
step S1, obtaining SCADA operation data generated in the normal operation process of the offshore wind farm, carrying out correlation analysis on the data to obtain a Pearson correlation coefficient matrix, screening key parameters according to the correlation coefficients and generating an input/output cellular array in the normal operation time period and a latest input/output cellular array before failure;
step S2, using the input and output cell array in the normal operation time period as a training set, using the latest input and output cell array before the fault as a test set, then constructing a prediction model, wherein the prediction model is an LSTM neural network comprising an input layer, an LSTM layer, a dropout layer, a full-connection layer and a regression layer, and the training set is sent to the LSTM neural network for prediction model training;
step S3, sending the test set into a prediction model, and carrying out moving index weighted average calculation on residual errors of prediction data and actual data to obtain the deviation degree of the actual data and the prediction data;
step S4, calculating and controlling the upper and lower limits by using a control chart theory, and obtaining a unit state evaluation coefficient independent of the numerical value of the index according to the deviation degree of the actual and predicted valuesBefore starting, the operation burden and the starting weight of the unit in a bad state are reduced according to the coefficient, and guidance is provided for path optimization in the black starting process;
step S5, considering the running state of the wind turbine generator and combining the wind power predicted value to issue a power instruction to the generator to obtain a generated energy target function;
step S6, the Dijkstra algorithm is adopted to obtain the shortest path L between the balance node and each node i Forming an electrical distance matrix to obtain a black start electrical distance target function of the wind power plant;
and step S7, integrating the generating capacity objective function and the electrical distance objective function by adopting a linear weighting method, solving mixed integer linear programming to obtain a pareto front, and selecting an optimal black start path of the wind power plant.
In order to optimize the technical scheme, the specific measures adopted further comprise:
the step S1 is specifically:
1) obtaining SCADA operation data generated in the normal operation process of the offshore wind farm, and performing correlation analysis on the data to obtain a Pearson correlation coefficient matrix;
2) counting the frequency counting result of the classifiable faults of the wind turbine generator according to the operation data, taking the parameter with high fault probability as a key parameter, and screening out a data group of which the correlation coefficient with the key parameter is greater than a set value;
3) and generating an input/output cell array in the normal operation time period and the latest input/output cell array before the fault.
The above 3) generation manner of the input/output cell array is:
and taking the first n variables with higher correlation in the screened data sets as input quantities, namely, the input feature dimension is n, taking the key parameter variable as output quantity, setting the width of a sliding window to be m, generating a first input cell array by respectively continuously sampling points at m moments of the n features with high correlation, wherein the predicted value is an output value at the m moment, then moving the window to continuously generate new input and output cell arrays by sampling points at 2 to m +1 moments, and the like.
The step S4 includes:
1) the calculation method of the actual deviation degree from the prediction at each moment of each unit, namely the out-of-limit degree of the prediction residual error is as follows:
wherein A is a mass coefficient, U CL And L CL To control the upper and lower limits, e i (t) is a moving exponential weighted average;
2) the operation state evaluation coefficients corresponding to the units are obtained as follows:
classifying the unit by using the evaluation coefficient of the running state, 1-MAX (epsilon) i (t)) and 0 are respectively full, derated and overhaul and outage.
The operation burden and the starting weight of the unit with the bad state are reduced according to the coefficient before starting, and the guidance for path optimization in the black starting process is as follows:
coefficient of performanceNot 1, during black start controlAnd setting the wind turbine generator as a PQ control strategy, carrying out load shedding operation according to the coefficient, and adjusting the pitch angle according to the power instruction and the output power difference value.
The step S5 is:
the running state of the unit and the wind power predicted value are considered to be combined and then a power instruction is issued to the unitThe calculation formula of the power generation amount objective function is as follows:
N G the total number of the units; p Gi The predicted value of the ultra-short-term wind power is obtained; t is a unit of ist For each unit start-up time, i.e. a it Time of transition from 0 to 1;and evaluating the coefficient for the unit state.
In the above step S6, d is set ij Is the resistance value R of the submarine cable ij +jX ij Die of (2), without cables, set d ij Is equal to ∞ and furthermore d is equal to j when i is equal to j ij Is 0, thereby obtaining a contact matrix D of the wind power plant w ;
Connection matrix D w For an undirected graph with weights, if a path K has the shortest distance among all paths from a node i to a node j, then K is the shortest path from i to j, and the sum L of the weights of all arcs passed by the path is called the shortest distance from the node i to the node j.
The step S6 includes:
solving the shortest path L between the balance node and each node by adopting Dijkstra algorithm i Forming an electrical distance matrix;
the method comprises the following steps of obtaining a black start electrical distance target function of the wind power plant as follows:
in the formula, N G The total number of the units; a is a it Is the optimization variable of the startup matrix; l is i Balancing the shortest paths between the nodes; t is N The starting time of each unit is.
The invention has the following beneficial effects:
the method utilizes the advantage of generating a large amount of SCADA data in the operation process of the offshore wind turbine generator system, performs correlation analysis on the data, preprocesses a data set with high correlation, and obtains a training data set in a normal operation mode and a test data set for testing; in the training stage, establishing a normal behavior model of the wind turbine generator by using an LSTM neural network, taking the sampling data of the wind turbine generator in a stable running state before a system fault causes total darkness as a training set in a normal running state, putting the training set in the model for learning, and obtaining the normal behavior model of the wind turbine generator after training; in the testing stage, inputting a testing data set into a trained normal behavior model for wind power prediction, calculating the offset of a residual error between a wind power actual value and a model predicted value by using an exponential weighted moving average, and calculating the state evaluation coefficient of the wind power generator set according to the offset amplitude exceeding the upper and lower control limits; combining wind power prediction information and a state evaluation coefficient as power instructions of each set after starting, and increasing the total power generation amount of the wind power plant as much as possible in the black starting process to serve as a first starting target; meanwhile, the started units are close to the balance node as much as possible, and the electrical distance between each wind turbine and the balance node is used as a second evaluation target; and integrating the two targets by adopting a linear weighting method, solving mixed integer linear programming to obtain a pareto front edge, and selecting an optimal black start path of the wind power plant.
According to the method, a reasonable wind power plant black start path optimization method is adopted, the running state of a wind turbine generator is considered, the electrical distance and the total generated energy are balanced to obtain the black start path, and the stability of the isolated grid black start initial stage with poor robustness is improved.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
fig. 2 is a wiring structure diagram of an offshore wind power system in a black start scene.
Detailed Description
Embodiments of the present invention are described in further detail below with reference to the accompanying drawings.
A wind power plant black start path optimization method considering unit operation states is characterized by comprising the following steps:
step S1, acquiring a large amount of SCADA operation data generated in the normal operation process of the offshore wind farm, performing correlation analysis on the data to obtain a Pearson correlation coefficient matrix, screening key parameters according to the correlation coefficients, and generating an input and output cell array in the normal operation time period and a latest input and output cell array before failure;
1) acquiring a large amount of SCADA operation data generated in the normal operation process of the offshore wind farm, and performing correlation analysis on the data to obtain a Pearson correlation coefficient matrix;
2) counting the frequency counting result of the classifiable faults of the wind turbine generator according to the operation data, taking the parameter with high fault probability as a key parameter, and screening out a data group of which the correlation coefficient with the key parameter is greater than a set value;
if the temperature of the coil of the generator in the engine room is taken as a key parameter of the running state of the unit, screening out a data group of which the correlation coefficient with the temperature of the coil of the generator is greater than 0.6;
3) and generating an input/output cell array in the normal operation time period and the latest input/output cell array before the fault.
The generation mode of the input and output cellular array is as follows:
the first n variables with higher correlation in the screened data sets are used as input quantity, namely the input characteristic dimension is n, the unit coil temperature variable is used as output quantity, the width of a sliding window is set to be m, the n characteristics with high correlation are respectively continuous to sampling points at m moments to generate a first input cellular array, the predicted value is an output value at m moments, namely the cabin generator coil temperature is used as a first output cellular array, then the window is moved to enable the sampling points at 2 nd to m +1 th moments to continuously generate new input and output cellular arrays, and the like.
And the part with higher correlation in the SCADA data is utilized to carry out many-to-one prediction, so that the prediction precision of the system is improved.
Step S2, using the input and output cell arrays in the normal operation time period as a training set, using the latest unit operation data before the fault, such as the last 10% input and output cell arrays as a test set, then constructing a prediction model, wherein the prediction model is an LSTM neural network comprising an input layer, an LSTM layer, a dropout layer, a full connection layer and a regression layer, and the training set is sent to the LSTM neural network for prediction model training;
step S3, the test set is sent to a prediction model, and the residual error of the prediction data and the actual data is subjected to moving index weighted average calculation to obtain the deviation degree of the actual data and the prediction data, so that the offset detection capability of the non-stationary sequence is improved;
step S4, calculating control upper and lower limits by using a control chart theory, and obtaining a unit state evaluation coefficient independent of the numerical value of the index according to the actual and predicted deviation degree
Coefficient of performanceIf not 1, setting the wind turbine generator set to be a PQ control strategy in the black start control process, carrying out load shedding operation according to a coefficient, and adjusting the size of a pitch angle according to a power instruction and an output power difference value, namely reducing the operation burden and the start weight of a unit in a bad state according to the coefficient before starting, and providing guidance for path optimization in the black start process;
the method comprises the following steps:
1) the calculation method of the actual deviation degree from the prediction at each moment of each unit, namely the out-of-limit degree of the prediction residual error is as follows:
wherein A is a mass coefficient, U CL And L CL To control the upper and lower limits, e i (t) is a moving exponential weighted average;
2) the operation state evaluation coefficients corresponding to the units are obtained as follows:
classifying the unit by using the evaluation coefficient of the running state, 1-MAX (epsilon) i (t)) and 0 are respectively full, derated and overhaul and outage.
Step S5, considering the running state of the wind turbine generator and combining the wind power predicted value to issue a power instruction to the generator to obtain a generated energy target function;
the running state of the unit and the predicted value of the wind power are combined and a power instruction is issued to the unitThe calculation formula of the power generation amount objective function is as follows:
N G the total number of the units; p Gi The predicted value of the ultra-short-term wind power is obtained; t is ist For each unit start-up time, i.e. a it Time of transition from 0 to 1;and evaluating the coefficient for the unit state.
Step S6, the Dijkstra algorithm is adopted to obtain the shortest path L between the balance node and each node i Forming an electrical distance matrix to obtain a black start electrical distance target function of the wind power plant;
set d ij Is the resistance value R of the submarine cable ij +jX ij Die of (2), without cables, set d ij Is equal to ∞ and furthermore d is equal to j when i is equal to j ij Is 0, thereby obtaining a contact matrix D of the wind power plant w ;
Contact matrix D w For an undirected graph with weights, if a path K has the shortest distance among all paths from a node i to a node j, then K is the shortest path from i to j, and the sum L of the weights of all arcs passed by the path is called the shortest distance from the node i to the node j.
Solving the shortest path L between the balance node and each node by adopting Dijkstra algorithm i Forming an electrical distance matrix;
the method comprises the following steps of obtaining a black start electrical distance target function of the wind power plant as follows:
in the formula, N G The total number of the units; a is it Is the optimization variable of the startup matrix; l is i Balancing the shortest paths between the nodes; t is N The starting time of each unit is.
And step S7, integrating the generated energy objective function and the electrical distance objective function by adopting a linear weighting method, solving mixed integer linear programming to obtain the pareto front, and selecting an optimal black start path of the wind power plant.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above-mentioned embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and adaptations to those skilled in the art without departing from the principles of the present invention may be apparent to those skilled in the relevant art and are intended to be within the scope of the present invention.
Claims (8)
1. A wind power plant black start path optimization method considering unit operation states is characterized by comprising the following steps:
step S1, obtaining SCADA operation data generated in the normal operation process of the offshore wind farm, carrying out correlation analysis on the data to obtain a Pearson correlation coefficient matrix, screening key parameters according to the correlation coefficients and generating an input/output cellular array in the normal operation time period and a latest input/output cellular array before failure;
step S2, using the input and output cell array in the normal operation time period as a training set, using the latest input and output cell array before the fault as a test set, then constructing a prediction model, wherein the prediction model is an LSTM neural network comprising an input layer, an LSTM layer, a dropout layer, a full-connection layer and a regression layer, and the training set is sent to the LSTM neural network for prediction model training;
step S3, sending the test set into a prediction model, and carrying out moving index weighted average calculation on residual errors of prediction data and actual data to obtain the deviation degree of the actual data and the prediction data;
step S4, calculating control upper and lower limits by using a control chart theory, and obtaining a unit state evaluation coefficient independent of the numerical value of the index according to the actual and predicted deviation degreeBefore starting, the operation burden and the starting weight of the unit in a bad state are reduced according to the coefficient, and guidance is provided for path optimization in the black starting process;
step S5, considering the running state of the wind turbine generator and combining the wind power predicted value to issue a power instruction to the generator to obtain a generated energy target function;
step S6, the Dijkstra algorithm is adopted to obtain the shortest path L between the balance node and each node i Forming an electrical distance matrix to obtain a black start electrical distance target function of the wind power plant;
and step S7, integrating the generating capacity objective function and the electrical distance objective function by adopting a linear weighting method, solving mixed integer linear programming to obtain a pareto front, and selecting an optimal black start path of the wind power plant.
2. The method for optimizing the black start path of the wind farm according to claim 1, wherein the step S1 specifically comprises:
1) obtaining SCADA operation data generated in the normal operation process of the offshore wind power plant, and performing correlation analysis on the data to obtain a Pearson correlation coefficient matrix;
2) counting the frequency counting result of the classifiable faults of the wind turbine generator according to the operation data, taking the parameter with high fault probability as a key parameter, and screening out a data group of which the correlation coefficient with the key parameter is greater than a set value;
3) and generating an input/output cell array in the normal operation time period and the latest input/output cell array before the fault.
3. The method for optimizing the black start path of the wind farm in consideration of the running state of the unit as claimed in claim 2, wherein 3) the input and output cell array is generated in a manner that:
and taking the first n variables with higher correlation in the screened data sets as input quantities, namely, the input feature dimension is n, taking the key parameter variable as output quantity, setting the width of a sliding window to be m, generating a first input cell array by respectively continuously sampling points at m moments of the n features with high correlation, wherein the predicted value is an output value at the m moment, then moving the window to continuously generate new input and output cell arrays by sampling points at 2 to m +1 moments, and the like.
4. The method for optimizing the black start path of the wind farm according to claim 1, wherein the step S4 comprises:
1) the calculation method of the actual deviation degree from the prediction at each moment of each unit, namely the out-of-limit degree of the prediction residual error is as follows:
wherein A is a mass coefficient, U CL And L CL To control the upper and lower limits, e i (t) is a moving exponential weighted average;
2) the operation state evaluation coefficients corresponding to the units are obtained as follows:
classifying the unit by using the evaluation coefficient of the running state, 1-MAX (epsilon) i (t)) and 0 are respectively full, derated and overhaul stopped.
5. The method for optimizing the black start path of the wind power plant in consideration of the unit operation state according to claim 1, wherein the step of reducing the operation burden and the start weight of the unit in a bad state according to a coefficient before starting is performed, and the step of providing guidance for path optimization in the black start process comprises the following steps:
coefficient of performanceAnd if not, setting the wind turbine generator to be a PQ control strategy in the black start control process, carrying out load shedding operation according to the coefficient, and adjusting the pitch angle according to the power instruction and the output power difference value.
6. The method for optimizing the black start path of the wind farm according to claim 1, wherein the step S5 is:
the running state of the unit and the wind power predicted value are considered to be combined and then a power instruction is issued to the unitThe calculation formula of the power generation amount objective function is as follows:
7. The method for optimizing the black start path of the wind farm according to the claim 1, wherein in the step S6, d is set ij Is the resistance value R of the submarine cable ij +jX ij Die of (2), without cable then set d ij Is equal to ∞ and furthermore d is equal to j when i is equal to j ij Is 0, thereby obtaining a contact matrix D of the wind power plant w ;
Contact matrix D w If the distance of the path K in all paths between the node i and the node j is the shortest, the path K is the shortest path from the node i to the node j, and the sum L of the weights of all arcs passed by the path is called the shortest distance from the node i to the node j.
8. The method for optimizing the black start path of the wind farm according to claim 7, wherein the step S6 comprises:
solving the shortest path L between the balance node and each node by adopting Dijkstra algorithm i Forming an electrical distance matrix;
the method comprises the following steps of obtaining a black start electrical distance target function of the wind power plant as follows:
in the formula, N G The total number of the units is; a is it Is the optimization variable of the startup matrix; l is i Balancing the shortest paths between the nodes; t is N And starting time for each unit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210721339.5A CN115034493B (en) | 2022-06-24 | 2022-06-24 | Wind farm black start path optimizing method considering unit running state |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210721339.5A CN115034493B (en) | 2022-06-24 | 2022-06-24 | Wind farm black start path optimizing method considering unit running state |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115034493A true CN115034493A (en) | 2022-09-09 |
CN115034493B CN115034493B (en) | 2024-05-14 |
Family
ID=83126109
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210721339.5A Active CN115034493B (en) | 2022-06-24 | 2022-06-24 | Wind farm black start path optimizing method considering unit running state |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115034493B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117039895A (en) * | 2023-10-09 | 2023-11-10 | 西安热工研究院有限公司 | Wind power prediction method and system for energy storage auxiliary black start |
CN117057258A (en) * | 2023-10-12 | 2023-11-14 | 西安热工研究院有限公司 | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105281375A (en) * | 2015-11-26 | 2016-01-27 | 江苏方天电力技术有限公司 | Power matching method in power grid black start involving wind power station |
CN109995036A (en) * | 2017-12-29 | 2019-07-09 | 中国电力科学研究院有限公司 | A kind of the black starting-up path automatic optimizing method and system of wind-light storage isolated network |
WO2021164112A1 (en) * | 2020-02-18 | 2021-08-26 | 山东大学 | Frequency control method and system during using wind farm as black-start power source by means of optimal configuration of energy storage |
AU2021106593A4 (en) * | 2021-08-23 | 2021-11-11 | Northwest Branch Of State Grid Corporation Of China | Dynamic Path Restoration Method for Black Start of Power System Based on New Energy and Energy Storage |
CN113644653A (en) * | 2021-08-20 | 2021-11-12 | 西安交通大学 | New energy and energy storage cooperative power system black start path recovery method |
WO2022022101A1 (en) * | 2020-07-30 | 2022-02-03 | 国网甘肃省电力公司电力科学研究院 | Method for controlling coordinated peak regulation of emergency source network in sending-end grid fault state |
CN114530844A (en) * | 2022-01-12 | 2022-05-24 | 西安理工大学 | Black start method after blackout accident of low-inertia power system |
-
2022
- 2022-06-24 CN CN202210721339.5A patent/CN115034493B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105281375A (en) * | 2015-11-26 | 2016-01-27 | 江苏方天电力技术有限公司 | Power matching method in power grid black start involving wind power station |
CN109995036A (en) * | 2017-12-29 | 2019-07-09 | 中国电力科学研究院有限公司 | A kind of the black starting-up path automatic optimizing method and system of wind-light storage isolated network |
WO2021164112A1 (en) * | 2020-02-18 | 2021-08-26 | 山东大学 | Frequency control method and system during using wind farm as black-start power source by means of optimal configuration of energy storage |
WO2022022101A1 (en) * | 2020-07-30 | 2022-02-03 | 国网甘肃省电力公司电力科学研究院 | Method for controlling coordinated peak regulation of emergency source network in sending-end grid fault state |
CN113644653A (en) * | 2021-08-20 | 2021-11-12 | 西安交通大学 | New energy and energy storage cooperative power system black start path recovery method |
AU2021106593A4 (en) * | 2021-08-23 | 2021-11-11 | Northwest Branch Of State Grid Corporation Of China | Dynamic Path Restoration Method for Black Start of Power System Based on New Energy and Energy Storage |
CN114530844A (en) * | 2022-01-12 | 2022-05-24 | 西安理工大学 | Black start method after blackout accident of low-inertia power system |
Non-Patent Citations (1)
Title |
---|
汤奕;戴剑丰;冯鑫;王琦;徐钢;杜先波;: "基于虚拟惯量的风电场黑启动频率协同控制策略", 电力系统自动化, no. 03, 10 February 2017 (2017-02-10) * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117039895A (en) * | 2023-10-09 | 2023-11-10 | 西安热工研究院有限公司 | Wind power prediction method and system for energy storage auxiliary black start |
CN117039895B (en) * | 2023-10-09 | 2024-01-26 | 西安热工研究院有限公司 | Wind power prediction method and system for energy storage auxiliary black start |
CN117057258A (en) * | 2023-10-12 | 2023-11-14 | 西安热工研究院有限公司 | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient |
CN117057258B (en) * | 2023-10-12 | 2023-12-12 | 西安热工研究院有限公司 | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient |
Also Published As
Publication number | Publication date |
---|---|
CN115034493B (en) | 2024-05-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN115034493A (en) | Wind power plant black start path optimization method considering unit operation state | |
CN111950900A (en) | Power system source network load storage intrinsic safety risk assessment method | |
Han et al. | A multi-view and multi-scale transfer learning based wind farm equivalent method | |
CN114336608B (en) | Unit blocking optimization method and system considering dynamic capacity increase and reconstruction | |
CN114006410B (en) | Large-scale offshore wind power access point optimization method based on opportunity constraint planning | |
CN113344283B (en) | Energy internet new energy consumption capability assessment method based on edge intelligence | |
Liang et al. | Adaptive critic design based dynamic optimal power flow controller for a smart grid | |
GAVRILAS et al. | Flexible Alternating Current Transmission System Optimisation In The Context Of Large Disturbance Voltage Stability | |
CN113991652B (en) | Data-driven multi-output calculation method for short-circuit current of distribution network containing IIDG | |
CN117882261A (en) | Predicting grid frequency | |
CN114465276A (en) | Power distribution network wind storage combined optimization configuration method and related device thereof | |
Iwata et al. | Multi-population differential evolutionary particle swarm optimization for distribution state estimation using correntropy in electric power systems | |
Dozein et al. | Optimal distribution voltage control through a sub-framework in the reactive power management on the Smart Grid | |
CN113241793A (en) | Prevention control method for power system with IPFC (intelligent power flow controller) considering wind power scene | |
Kisengeu et al. | Under Voltage Load Shedding using Hybrid Metaheuristic Algorithms for Voltage Stability Enhancement: A Review | |
Dipp et al. | Training of Artificial Neural Networks Based on Feed-in Time Series of Photovoltaics and Wind Power for Active and Reactive Power Monitoring in Medium-Voltage Grids | |
Lv et al. | A very short-term online forecasting model for photovoltaic power based on two-stage resource allocation network | |
Abdel-Salam et al. | Optimal sizing and locations of DG sources in distribution systems: A review of different techniques | |
Bagchi et al. | Composite system adequacy assessment incorporating virtual power plants | |
Li et al. | Wind farms reactive power optimization using genetic/tabu hybrid algorithm | |
Lv et al. | Contribution Analysis of Influential Factors for Wind Power Curtailment Caused by Lack of Load-Following Capability Based on BP-MIV | |
Ma et al. | Cyber-physical Modeling Technique based Dynamic Aggregation of Wind Farm Considering LVRT Characteristic | |
Kuo | Increasing grid-connected capacity of offshore wind farms for isolated power systems-a case study of the Taiwan power system | |
CN114825418B (en) | Multifunctional flexible complementary digital power supply management method | |
Sarwar et al. | Deep reinforcement learning framework for short-term voltage stability improvement |
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 |