CN113659630A - Wind power plant power optimization scheduling method and system based on fatigue damage value estimation - Google Patents
Wind power plant power optimization scheduling method and system based on fatigue damage value estimation Download PDFInfo
- Publication number
- CN113659630A CN113659630A CN202110841483.8A CN202110841483A CN113659630A CN 113659630 A CN113659630 A CN 113659630A CN 202110841483 A CN202110841483 A CN 202110841483A CN 113659630 A CN113659630 A CN 113659630A
- Authority
- CN
- China
- Prior art keywords
- fatigue damage
- value
- preset time
- power
- 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
- 238000005457 optimization Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 36
- 239000002245 particle Substances 0.000 claims abstract description 47
- 238000003062 neural network model Methods 0.000 claims abstract description 40
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 23
- 238000005452 bending Methods 0.000 claims abstract description 16
- 239000013598 vector Substances 0.000 claims description 70
- 239000003550 marker Substances 0.000 claims description 23
- 230000001133 acceleration Effects 0.000 claims description 12
- 239000000463 material Substances 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 238000005096 rolling process Methods 0.000 claims description 6
- 238000010977 unit operation Methods 0.000 claims description 6
- 239000013307 optical fiber Substances 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 125000004122 cyclic group Chemical group 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims 1
- 238000012886 linear function Methods 0.000 claims 1
- 238000010248 power generation Methods 0.000 abstract 1
- 238000013461 design Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Images
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
- H02J3/38—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
- H02J3/466—Scheduling the operation of the generators, e.g. connecting or disconnecting generators to meet a given demand
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
-
- 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
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/381—Dispersed generators
-
- 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
-
- 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
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Power Engineering (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Wind Motors (AREA)
Abstract
The invention discloses a wind power plant power optimization scheduling method and system based on fatigue damage value estimation, wherein accurate bending moment load is obtained through a load sensor arranged on a benchmark set, fatigue damage values of all components of the benchmark set are obtained through a rain flow counting method and a linear accumulated damage theory, the operation data of the benchmark set is used as input, the fatigue damage values of all components of the benchmark set are used as output, a neural network model for optimizing the fatigue damage values is trained, and the power generation power of all the units is determined by adopting a particle swarm optimization algorithm, so that the power optimization scheduling in a wind power plant is realized; the invention can obtain the accurate fatigue damage value of the whole set with lower hardware cost, realize the dynamic balance of the fatigue damage value of the whole set, the accurate tracking of the total power set value of the whole set and the optimized dispatching of power, exert the generating capacity of the set with small fatigue damage, reduce the pressure of the set with large fatigue damage and achieve the aim of fatigue balance.
Description
Technical Field
The invention relates to the technical field of wind power plant power scheduling, in particular to a wind power plant power optimal scheduling method and system based on fatigue damage value estimation.
Background
With the arrival of the wind power fair era, on the premise of ensuring the load safety of a wind power generator, the whole generating capacity of the wind power plant is improved, and the important direction of cost reduction and efficiency improvement is achieved. The load safety of the wind turbine generator is mainly divided into limit load safety and fatigue load safety. For limit load safety, various limit load control strategies may be employed, such as power limit, shutdown, and the like. For fatigue load safety, the annual average wind speed and the turbulence intensity of the worst machine position under the full-field wind condition are adopted as load calculation input parameters in the unit design stage, so that the fatigue load safety of the full-field unit is ensured; in fact, the design method is conservative, for a machine position with low turbulence intensity, the fatigue life of each component of the unit is longer than the design life, if the fatigue life of the components of the unit can be fully utilized, the rated power is set to be a value exceeding the design rated power, and the generated energy of the wind power plant can be improved under the condition of no electricity limitation;
the existing method for performing power optimization scheduling according to the fatigue damage of the unit generally obtains wind resource data and a unit fatigue damage database under a set power value through load calculation software simulation, obtains an estimated fatigue damage value according to database query when the unit actually runs, and performs power scheduling optimization according to the estimated fatigue damage. The method has the limitation that a large amount of simulation data is needed, and the database is not optimized and updated after being determined, so that the fatigue damage value estimated according to the database is inaccurate; the accurate fatigue damage value of the component can be obtained by mounting a load sensor on the unit, measuring the actual component load and combining a rain flow counting method and a linear accumulated damage accumulation theory; but installing load sensors for each unit increases the cost;
in addition, the existing scheduling method in the wind power plant generally adopts a mode of stopping a part of units to track a full-field power set value under the condition of power limitation of a power grid, and has the defects of frequent starting and stopping of the part of units and delayed response of full-field power tracking.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides a wind power plant power optimal scheduling method and system based on fatigue damage value estimation.
In order to achieve the purpose, the technical scheme provided by the invention is as follows: the wind power plant power optimization scheduling method based on the fatigue damage value estimation comprises the following steps:
s1, collecting and processing data
Selecting one unit from each wind power plant as a benchmark unit, respectively installing optical fiber load sensors at the blade root, the tower drum top and the tower drum bottom of the benchmark unit, measuring bending moment loads at the blade root, the tower drum top and the tower drum bottom of the benchmark unit, simultaneously measuring operation data of all units in the wind power plant, storing the operation data into a high-performance server installed in a booster station of the wind power plant, and performing sectional processing on the operation data by the high-performance server according to preset time to obtain a preset time statistic value of the operation data of all the units;
s2, calculating the fatigue damage value of the marker post unit on line
Respectively carrying out sectional processing on the measured bending moment loads of the blade root, the tower drum top and the tower drum bottom of the marker post unit according to preset time and converting the bending moment loads into rain flow counting results by using a rain flow counting method, and calculating fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the marker post unit in the preset time according to S-N curves corresponding to materials of the blade root, the tower drum top and the tower drum bottom of the marker post unit and the rain flow counting results;
s3 training fatigue damage value neural network model
The method comprises the steps that a high-performance server presets a period, a preset time statistic value of operation data of the benchmarking unit in the preset period is used as input, fatigue damage values of a blade root, the top of a tower barrel and the bottom of the tower barrel in the preset time of the benchmarking unit in the preset period are used as output, and a fatigue damage value optimization neural network model is trained;
s4, calculating the fatigue damage value of each unit in the wind power plant
The high-performance server utilizes the fatigue damage value neural network model and the preset time statistic of all the unit operation data to calculate the preset time fatigue damage value vector D of all the units in the wind power plant on linei,jAnd accumulating the fatigue damage value vectors of all the units at preset time intervals to obtain the accumulated fatigue damage value D of all the unitssum,i,j,Dsum,i,j=Dsum,i,j-1+Di,j;
S5, calculating the predicted fatigue damage values of all units in the wind power plant
For each unit in the wind power plant, a preset time statistic value vector A of the current operation data of each uniti,jReplacing the average power value in the process with a future preset time power set value P of each unitset,i,jTo give A'i,j(ii) a A'i,jAs the input of the fatigue damage value neural network model, calculating the predicted fatigue damage value vector D of the future preset time of all the unitspred,i,Dpred,i=fANN(A'i,j) Is provided with DpredSum,iCumulative fatigue damage values after a predetermined time for all units, DpredSum,i=Dsum,i,j+Dpred,i;
S6 optimizing power set value combination of all units in wind power plant
Optimizing the future preset time power set value combinations of all the units in the wind power plant by utilizing a particle swarm optimization algorithm to obtain the optimized future preset time power set value combinations of all the units, and outputting a global optimal solution xgbest,KVector quantity;
s7, issuing set value of optimal power of whole plant set
Solving the global optimum xgbest,KThe value of each element of the vector is taken as the power set value, i.e. P, of all the units in the wind farmset=xgbest,KAnd the power is transmitted to each unit through the high-performance server, so that the power in the wind power plant is optimized and scheduled.
Further, in step S1, the operational data includes generator power, generator speed, pitch angle of the three blades, nacelle fore-aft acceleration and nacelle lateral acceleration; the preset time statistic of all the unit operation data comprises the maximum value, the minimum value, the variation range, the average value and the standard deviation of each operation data.
Further, in step S2, the method includes the steps of:
s201, for the bending moment loads of the blade root, the tower drum top and the tower drum bottom of the benchmark set,the rain flow counting result is the load cycle amplitude S in the preset time load time sequence of each bending moment loadpCorresponding number N of rain flow cyclesrainflow,p,p=1,2,…,128;
S202, regarding materials of the blade root, the tower drum top and the tower drum bottom of the benchmark set, the abscissa of the S-N curve is different load cycle amplitude values SpThe vertical coordinate is the cyclic amplitude S of different loads at the root, the top and the bottom of the tower of the marker post unitpMaximum number of cycles N that can be tolerated before fatigue failure occursp(ii) a The S-N curve is calculated by the following formula:
Sum is the ultimate load of the component, and is related to the material of the component;
s203, calculating fatigue damage values D of the blade root, the tower drum top and the tower drum bottom of the benchmark set within preset time according to S-N curves and rain flow counting results corresponding to materials of the blade root, the tower drum top and the tower drum bottom of the benchmark set, wherein the fatigue damage values D are obtained by calculation according to a linear accumulated damage theory:
according to the linear accumulated damage theory, when the fatigue damage value of the blade root, the top of the tower drum or the bottom of the tower drum of the benchmark set is more than or equal to 1, the fatigue life is reached, and the fatigue damage occurs.
Further, in step S3, the method includes:
the high-performance server takes the preset time statistic value of the operation data of the marker post unit as input and records the preset time statistic value as A1,jThe j-th statistical value vector in the preset time of the operation data of the marker post unit is represented, the vector dimension is mx 1, namely, the j-th statistical value vector contains m statistical values of the operation data in the preset time, and the j-th statistical value vector represents the fatigue loss of the blade root, the tower top and the tower bottom of the marker post unitThe damage value is recorded as D1,jRepresenting a fatigue damage value vector of the benchmark set within the jth preset time, marking the benchmark set as 1, and training a fatigue damage value neural network model, wherein the vector dimension is n multiplied by 1, namely the fatigue damage value vector contains n preset time fatigue damage values; the fatigue damage value neural network model is a nonlinear function of fatigue damage values of all parts in the preset time of all the units relative to the preset time statistic of all the unit operation data, and is recorded as Di,j=fANN(Ai,j) And i is the number of the unit, and j represents the jth preset time since all the units are put into operation in a grid-connected mode.
Further, the preset time statistic of the operation data includes a maximum value of wind speed within preset time, a variation range of wind speed within preset time, an average value of power within preset time, a minimum value of generator rotation speed within preset time, a maximum value of pitch angle within preset time, a variation range of pitch angle within preset time, a standard deviation of front and rear acceleration of the nacelle within preset time, and a standard deviation of lateral acceleration of the nacelle within preset time.
Further, the fatigue damage values of the blade root, the tower top and the tower bottom of the benchmark set include the in-plane direction fatigue damage value of the first blade root of the benchmark set, the out-of-plane direction fatigue damage value of the first blade root, the in-plane direction fatigue damage value of the second blade root, the out-of-plane direction fatigue damage value of the second blade root, the in-plane direction fatigue damage value of the third blade root, the out-of-plane direction fatigue damage value of the third blade root, the pitch direction fatigue damage value of the tower top, the rolling direction fatigue damage value of the tower top, the pitch direction fatigue damage value of the tower bottom and the rolling direction fatigue damage value of the tower bottom.
Further, the hidden layer of the fatigue damage value neural network model is 1 layer, the number of units of the hidden layer is 50, the activation function of the fatigue damage value neural network model is selected as a linear rectification function, and the weight optimization adopts an L-BFGS algorithm.
Further, in step S6, the method includes the steps of:
s601, theP for combining future preset time power set values of all unitssetVector designation, Pset=[Pset,1,Pset,2,…,Pset,N]T,Pset,NPresetting a power set value of the future preset time for the Nth set of the wind power plant;
s602, taking accumulated fatigue damage value vector D of each unit after preset timepredSum,iMaximum value of, i.e. max (D)predSum,i) Recording the equivalent accumulated fatigue damage value after the preset time of the unit; calculating the average value of the equivalent accumulated fatigue damage values of all the units after the preset timeCalculating the standard deviation of the equivalent accumulated fatigue damage values of all the units after the preset timeK1A weight factor that is the standard deviation of fatigue damage; the standard deviation index of fatigue damage is
S603, calculating the sum of the future preset time power set values of all the units, namelyCalculating a total power tracking indexK2A weight factor that is a total power trace; the total power tracking index is
S604, according to the fatigue damage standard deviation index and the total power tracking index, solving a target function adopted by the particle swarm optimization algorithm as follows:
wherein, Pset,farmPresetting a total power set value of the whole time field in the future;
the particle swarm optimization algorithm adopts the following constraint conditions:
Pset,i<Ppred,i
Pset,i<Pboost,i
wherein, the fatigue damage value of a single unit is restricted to be that the accumulated fatigue damage value after the preset time of each unit is smaller than the expected fatigue damage value, namelytjTime elapsed since the wind farm was put into operation for grid connection, TlifeThe expected life of the wind turbine generator; the predicted power set value of a single unit is restricted to be the power set value P of each unitset,iLess than the predicted power P of a single unitpred,iI.e. Pset,i<Ppred,i(ii) a The over-sending power set value of a single unit is restricted to be the power set value P of each unitset,iOver-power P less than maximum single unitboost,iI.e. Pset,i<Pboost,i(ii) a The full field power set value is restricted to be the sum of the power set values of all the units in the full fieldLess than the power set value P dispatched by the power grid dispatchingset,farmI.e. by
S605, iterating the particle swarm optimization algorithm, wherein the dimension of a position vector and a velocity vector of each particle of the particle swarm optimization algorithm is Nx 1, and N is the number of wind power plant sets;
the iterative formula of the particle swarm optimization algorithm is as follows:
vm,k=ωvm,k-1+C1r1(xpbest,m,k-1-xi-1)+C2r2(xgbest,k-1-xm,k-1)
xm,k=xm,k-1+vm,k
k=k+1
wherein K is the maximum iteration number, K is the current iteration round, omega is an inertia weight factor, the value of the inertia weight factor is linearly reduced along with the increase of the iteration round, and omega isiniIs an initial inertial weight factor, ωendTo end the inertial weight factor, vm,kVelocity vector, x, for the m-th particle at the k-th iterationm,kFor the position vector of the m-th particle at the k-th iteration, vm,k-1Velocity vector, x, for the m-th particle at the k-1 th iterationm,k-1Is the position vector, C, of the m-th particle at the (k-1) th iteration1And C2Is a constant of a learning factor, r1And r2Is an N x 1-dimensional random vector, i.e. each vector element is random data between (0,1), xpbest,m,k-1For the local optimal solution, x, of the m-th particle at the k-1 iterationgbest,k-1The global optimal solution of the k-1 iteration is obtained; when the iteration number K is equal to the maximum iteration number K, the particle swarm optimization calculation is finished, and a global optimal solution x is outputgbest,KAnd (5) vector quantity.
The invention provides a wind power plant power optimization scheduling system based on fatigue damage value estimation, which comprises:
the operation data acquisition and statistics module is used for acquiring operation data of each unit in the wind power plant;
the fatigue damage value on-line calculation module is used for calculating the fatigue damage value of the benchmark set according to the load data of the benchmark set in the wind power plant;
the fatigue damage value neural network model module is used for outputting fatigue damage value vectors of all units in the wind power plant within preset time;
the neural network model offline training module is used for performing offline training on the fatigue damage value neural network model according to the input operation data and the fatigue damage value of the benchmark set;
and the full-field power particle swarm optimization module is used for calculating iterative particle positions corresponding to the iterative times, so that the fatigue damage value neural network model is optimized and the power scheduling of all the units in the wind power field is optimized.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the method calculates the actual fatigue damage value according to the actual measurement bending moment load of the benchmark set, trains the neural network model according to the operation data statistic value and the fatigue damage value of the benchmark set, optimizes and trains the model again every 3 months, and ensures the estimation accuracy of the model;
2. the fatigue damage neural network model is adopted to estimate the fatigue damage values of other units, only the optical fiber load sensor is needed to be arranged on the benchmark unit, and the accurate fatigue damage value of the whole unit is obtained with lower hardware cost;
3. the optimal power combination is determined by adopting a particle swarm optimization algorithm, so that the dynamic balance of the fatigue damage value of the whole unit and the accurate tracking of the total power set value of the whole unit are realized, and the condition that an individual unit is damaged in advance or the residual fatigue life is too long is avoided;
4. the method and the device reduce the starting and stopping times of the units under the condition of power grid power limitation by issuing the optimal power setting combination to each unit; when the power grid is not limited, the rated power set value of the unit is properly improved, and the unit is allowed to overrun, so that the generating capacity of the unit with small fatigue damage is exerted to the maximum extent, the pressure of the unit with large fatigue damage is relieved, and the aim of fatigue balance is fulfilled.
Drawings
FIG. 1 is a block diagram of the present invention.
Detailed Description
The present invention will be further described with reference to the following specific examples.
Referring to fig. 1, the wind farm power optimization scheduling method based on fatigue damage value estimation provided by the present embodiment includes the following steps:
s1, collecting and processing data
Selecting one unit from each wind power plant as a benchmark unit, respectively installing optical fiber load sensors at the blade root, the tower drum top and the tower drum bottom of the benchmark unit, measuring bending moment loads at the blade root, the tower drum top and the tower drum bottom of the benchmark unit, simultaneously measuring operation data of all units in the wind power plant, storing the operation data into a high-performance server installed in a booster station of the wind power plant, and performing sectional processing on the operation data by the high-performance server every 10 minutes to obtain a 10-minute statistical value of the operation data of all the units; the operation data comprises generator power, generator rotating speed, pitch angles of three blades, forward and backward acceleration of the engine room and lateral acceleration of the engine room; and the 10-minute statistical values of all the unit operation data comprise the maximum value, the minimum value, the variation range, the average value and the standard deviation of each operation data.
S2, calculating the fatigue damage value of the marker post unit on line
By utilizing a rain flow counting method, respectively carrying out sectional processing on the bending moment loads of the blade root, the tower drum top and the tower drum bottom of the measured marker post unit every 10 minutes and converting the bending moment loads into rain flow counting results, and calculating fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the marker post unit within 10 minutes according to S-N curves and the rain flow counting results corresponding to materials of the blade root, the tower drum top and the tower drum bottom of the marker post unit, wherein the method comprises the following steps:
s201, moment load for blade root, tower top and tower bottom of marker post unitThe rain flow counting result is the load cycle width S in the 10-minute load time sequence of each bending moment loadpCorresponding number N of rain flow cyclesrainflow,p,p=1,2,…,128;
S202, regarding materials of the blade root, the tower drum top and the tower drum bottom of the benchmark set, the abscissa of the S-N curve is different load cycle amplitude values SpThe vertical coordinate is the cyclic amplitude S of different loads at the root, the top and the bottom of the tower of the marker post unitpMaximum number of cycles N that can be tolerated before fatigue failure occursp(ii) a The S-N curve is calculated by the following formula:
Sum is the ultimate load of the component, and is related to the material of the component;
s203, calculating fatigue damage values D of the blade root, the tower drum top and the tower drum bottom of the benchmark set within 10 minutes according to S-N curves and rain flow counting results corresponding to materials of the blade root, the tower drum top and the tower drum bottom of the benchmark set, wherein the fatigue damage values D are obtained by calculation according to a linear accumulated damage theory:
according to the linear accumulated damage theory, when the fatigue damage value of the blade root, the top of the tower drum or the bottom of the tower drum of the benchmark set is more than or equal to 1, the fatigue life is reached, and the fatigue damage occurs.
S3 training fatigue damage value neural network model
The high-performance server takes 3 months as a period, takes 10-minute statistical values of operation data of the marker post unit in the past 3 months as input, and takes fatigue damage values of a blade root, the top of a tower drum and the bottom of the tower drum in 10 minutes of the marker post unit in the past 3 months as output, and trains and optimizes a fatigue damage value neural network model, and the neural network model comprises the following components:
high performanceThe server takes the 10-minute statistic value of the operation data of the marker post unit as input and is marked as A1,jRepresenting the statistic value vector of the operation data of the benchmark set within the jth 10 minutes, wherein the vector dimension is mx 1, namely the statistic value of the operation data of the benchmark set within m 10 minutes is output by taking the fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the benchmark set as D1,jRepresenting a fatigue damage value vector of the flagpole unit within the jth 10 minutes, marking the flagpole unit as 1, wherein the vector dimension is n multiplied by 1, namely the flagpole unit contains n fatigue damage values within 10 minutes, and training a fatigue damage value neural network model; the fatigue damage value neural network model is a nonlinear function of the fatigue damage value of each part in 10 minutes of all the units relative to the 10-minute statistical value of the running data of all the units, and is recorded as Di,j=fANN(Ai,j) Wherein i is the number of the unit, and j represents the jth 10 minutes since all the units are put into operation in a grid-connected mode. The 10-minute statistic value of the operation data comprises a maximum value of wind speed in 10 minutes, a variation range of the wind speed in 10 minutes, an average value of power in 10 minutes, a minimum value of the rotating speed of the generator in 10 minutes, a maximum value of a pitch angle in 10 minutes, a variation range of the pitch angle in 10 minutes, a standard deviation of front and rear acceleration of the cabin in 10 minutes and a standard deviation of lateral acceleration of the cabin in 10 minutes. The fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the benchmark set comprise an in-plane direction fatigue damage value of a first blade root of the benchmark set, an out-plane direction fatigue damage value of the first blade root, an in-plane direction fatigue damage value of a second blade root, an out-plane direction fatigue damage value of the second blade root, an in-plane direction fatigue damage value of a third blade root, an out-plane direction fatigue damage value of the third blade root, a pitching direction fatigue damage value of the tower drum top, a rolling direction fatigue damage value of the tower drum top, a pitching direction fatigue damage value of the tower drum bottom and a rolling direction fatigue damage value of the tower drum bottom. The hidden layer of the fatigue damage value neural network model is selected from 1 layer, the number of units of the hidden layer is 50, the activation function of the fatigue damage value neural network model is selected as a linear rectification function, and the weight optimization adopts an L-BFGS algorithm.
S4, calculating the fatigue damage value of each unit in the wind power plant
The high-performance server utilizes the fatigue damage value neural network model and the 10-minute statistical value of the running data of all the units to calculate the 10-minute fatigue damage value vector D of all the units in the wind power plant on linei,jAnd accumulating the fatigue damage value vectors of all the units every 10 minutes to obtain the accumulated fatigue damage value D of all the unitssum,i,j,Dsum,i,j=Dsum,i,j-1+Di,j;
S5, calculating the predicted fatigue damage values of all units in the wind power plant
For each unit in the wind power plant, a 10-minute statistic value vector A of the current operation data of each uniti,jReplacing the average power value in the process with the future 10-minute power set value P of each unitset,i,jTo give A'i,j(ii) a A'i,jAs the input of the fatigue damage value neural network model, calculating the predicted fatigue damage value vector D of all the units in the future 10 minutespred,i,Dpred,i=fANN(A'i,j) Is provided with DpredSum,iThe cumulative fatigue damage value after 10 minutes for all units, DpredSum,i=Dsum,i,j+Dpred,i;
S6 optimizing power set value combination of all units in wind power plant
Optimizing the future 10-minute power set value combination of all the units in the wind power plant by utilizing a particle swarm optimization algorithm to obtain the optimized future 10-minute power set value combination of all the units, and outputting a global optimal solution xgbest,KVector, comprising the steps of:
s601, P for combining future 10-minute power set values of all the unitssetVector designation, Pset=[Pset,1,Pset,2,…,Pset,N]T,Pset,NThe power set value of the Nth set of the wind power plant in the future 10 minutes is obtained;
s602, taking accumulated fatigue damage value vector D of each unit after 10 minutespredSum,iMaximum value of, i.e. max (D)predSum,i) Recording as an equivalent accumulated fatigue damage value of the unit after 10 minutes; calculating the average value of the equivalent accumulated fatigue damage values of all the units after 10 minutesCalculating the standard deviation of the equivalent accumulated fatigue damage value of all the units after 10 minutesK1A weight factor that is the standard deviation of fatigue damage; the standard deviation index of fatigue damage is
S603, obtaining the sum of the power set values of all the units in the future 10 minutes, namelyCalculating a total power tracking indexK2A weight factor that is a total power trace; the total power tracking index is
S604, according to the fatigue damage standard deviation index and the total power tracking index, solving a target function adopted by the particle swarm optimization algorithm as follows:
wherein, Pset,farmTotal power setpoint for a future 10 minute full field
The particle swarm optimization algorithm adopts the following constraint conditions:
Pset,i<Ppred,i
Pset,i<Pboost,i
wherein, the fatigue damage value of a single unit is restricted to be that the accumulated fatigue damage value of each unit after 10 minutes is smaller than the expected fatigue damage value, namelytjTime elapsed since the wind farm was put into operation for grid connection, TlifeThe expected life of the wind turbine generator; the predicted power set value of a single unit is restricted to be the power set value P of each unitset,iLess than the predicted power P of a single unitpred,iI.e. Pset,i<Ppred,i(ii) a The over-sending power set value of a single unit is restricted to be the power set value P of each unitsetiOver-power P less than maximum single unitboositI.e. Pset,i<Pboost,i(ii) a The full field power set value is restricted to be the sum of the power set values of all the units in the full fieldLess than the power set value P dispatched by the power grid dispatchingset,farmI.e. by
S605, iterating the particle swarm optimization algorithm, wherein the dimension of a position vector and a velocity vector of each particle of the particle swarm optimization algorithm is Nx 1, and N is the number of wind power plant sets;
the iterative formula of the particle swarm optimization algorithm is as follows:
vm,k=ωvm,k-1+C1r1(xpbest,m,k-1-xi-1)+C2r2(xgbest,k-1-xm,k-1)
xm,k=xm,k-1+vm,k
k=k+1
wherein K is the maximum iteration number, K is the current iteration round, omega is an inertia weight factor, the value of the inertia weight factor is linearly reduced along with the increase of the iteration round, and omega isiniIs an initial inertial weight factor, ωendTo end the inertial weight factor, vm,kVelocity vector, x, for the m-th particle at the k-th iterationm,kFor the position vector of the m-th particle at the k-th iteration, vm,k-1Velocity vector, x, for the m-th particle at the k-1 th iterationm,k-1Is the position vector, C, of the m-th particle at the (k-1) th iteration1And C2Is a constant of a learning factor, r1And r2Is an N x 1-dimensional random vector, i.e. each vector element is random data between (0,1), xpbest,m,k-1For the local optimal solution, x, of the m-th particle at the k-1 iterationgbest,k-1The global optimal solution of the k-1 iteration is obtained; when the iteration number K is equal to the maximum iteration number K, the particle swarm optimization calculation is finished, and a global optimal solution x is outputgbest,KAnd (5) vector quantity.
S7, issuing set value of optimal power of whole plant set
Solving the global optimum xgbest,KThe value of each element of the vector is taken as the power set value, i.e. P, of all the units in the wind farmset=xgbest,KAnd the power is transmitted to each unit through the high-performance server, so that the power in the wind power plant is optimized and scheduled.
The following wind farm power optimization scheduling system based on fatigue damage value estimation provided by this embodiment includes:
the operation data acquisition and statistics module is used for acquiring operation data of each unit in the wind power plant;
the fatigue damage value on-line calculation module is used for calculating the fatigue damage value of the benchmark set according to the load data of the benchmark set in the wind power plant;
the fatigue damage value neural network model module is used for outputting fatigue damage value vectors of all units in the wind power plant within preset time;
the neural network model offline training module is used for performing offline training on the fatigue damage value neural network model according to the input operation data and the fatigue damage value of the benchmark set;
and the full-field power particle swarm optimization module is used for calculating iterative particle positions corresponding to the iterative times, so that the fatigue damage value neural network model is optimized and the power scheduling of all the units in the wind power field is optimized.
The above-mentioned embodiments are merely preferred embodiments of the present invention, and the scope of the present invention is not limited thereto, so that the changes in the shape and principle of the present invention should be covered within the protection scope of the present invention.
Claims (9)
1. The wind power plant power optimization scheduling method based on fatigue damage value estimation is characterized by comprising the following steps:
s1, collecting and processing data
Selecting one unit from each wind power plant as a benchmark unit, respectively installing optical fiber load sensors at the blade root, the tower drum top and the tower drum bottom of the benchmark unit, measuring bending moment loads at the blade root, the tower drum top and the tower drum bottom of the benchmark unit, simultaneously measuring operation data of all units in the wind power plant, storing the operation data into a high-performance server installed in a booster station of the wind power plant, and performing sectional processing on the operation data by the high-performance server according to preset time to obtain a preset time statistic value of the operation data of all the units;
s2, calculating the fatigue damage value of the marker post unit on line
Respectively carrying out sectional processing on the measured bending moment loads of the blade root, the tower drum top and the tower drum bottom of the marker post unit according to preset time and converting the bending moment loads into rain flow counting results by using a rain flow counting method, and calculating fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the marker post unit in the preset time according to S-N curves corresponding to materials of the blade root, the tower drum top and the tower drum bottom of the marker post unit and the rain flow counting results;
s3 training fatigue damage value neural network model
The method comprises the steps that a high-performance server presets a period, a preset time statistic value of operation data of the benchmarking unit in the preset period is used as input, fatigue damage values of a blade root, the top of a tower barrel and the bottom of the tower barrel in the preset time of the benchmarking unit in the preset period are used as output, and a fatigue damage value optimization neural network model is trained;
s4, calculating the fatigue damage value of each unit in the wind power plant
The high-performance server utilizes the fatigue damage value neural network model and the preset time statistic of all the unit operation data to calculate the preset time fatigue damage value vector D of all the units in the wind power plant on linei,jAnd accumulating the fatigue damage value vectors of all the units at preset time intervals to obtain the accumulated fatigue damage value D of all the unitssum,i,j,Dsum,i,j=Dsum,i,j-1+Di,j;
S5, calculating the predicted fatigue damage values of all units in the wind power plant
For each unit in the wind power plant, a preset time statistic value vector A of the current operation data of each uniti,jReplacing the average power value in the process with a future preset time power set value P of each unitset,i,jTo obtain Ai',j(ii) a A is to bei',jAs the input of the fatigue damage value neural network model, calculating the predicted fatigue damage value vector D of the future preset time of all the unitspred,i,Dpred,i=fANN(Ai',j) Is provided with DpredSum,iCumulative fatigue damage values after a predetermined time for all units, DpredSum,i=Dsum,i,j+Dpred,i;
S6 optimizing power set value combination of all units in wind power plant
Optimizing the future preset time power set value combinations of all the units in the wind power plant by utilizing a particle swarm optimization algorithm to obtain the optimized future preset time power set value combinations of all the units, and outputting a global optimal solution xgbest,KVector quantity;
s7, issuing set value of optimal power of whole plant set
Solving the global optimum xgbest,KThe value of each element of the vector is taken as the power set value, i.e. P, of all the units in the wind farmset=xgbest,KAnd the power is transmitted to each unit through the high-performance server, so that the power in the wind power plant is optimized and scheduled.
2. The method for optimized scheduling of wind farm power based on fatigue damage value estimation according to claim 1, characterized in that in step S1 the operational data comprises generator power, generator speed, pitch angle of three blades, nacelle fore-aft acceleration and nacelle lateral acceleration; the preset time statistic of all the unit operation data comprises the maximum value, the minimum value, the variation range, the average value and the standard deviation of each operation data.
3. The wind farm power optimization scheduling method based on fatigue damage value estimation according to claim 1, characterized by comprising the following steps in step S2:
s201, regarding the bending moment loads of the blade root, the tower drum top and the tower drum bottom of the benchmark unit, the rain flow counting result is a load circulation amplitude S in a preset time load time sequence of each bending moment loadpCorresponding number N of rain flow cyclesrainflow,p,p=1,2,…,128;
S202, regarding materials of the blade root, the tower drum top and the tower drum bottom of the benchmark set, the abscissa of the S-N curve is different load cycle amplitude values SpThe vertical coordinate is the cyclic amplitude S of different loads at the root, the top and the bottom of the tower of the marker post unitpMaximum number of cycles N that can be tolerated before fatigue failure occursp(ii) a The S-N curve is calculated by the following formula:
Sum is the ultimate load of the component, and is related to the material of the component;
s203, calculating fatigue damage values D of the blade root, the tower drum top and the tower drum bottom of the benchmark set within preset time according to S-N curves and rain flow counting results corresponding to materials of the blade root, the tower drum top and the tower drum bottom of the benchmark set, wherein the fatigue damage values D are obtained by calculation according to a linear accumulated damage theory:
according to the linear accumulated damage theory, when the fatigue damage value of the blade root, the top of the tower drum or the bottom of the tower drum of the benchmark set is more than or equal to 1, the fatigue life is reached, and the fatigue damage occurs.
4. The wind farm power optimization scheduling method based on fatigue damage value estimation according to claim 1, wherein in step S3, the method comprises:
the high-performance server takes the preset time statistic value of the operation data of the marker post unit as input and records the preset time statistic value as A1,jRepresenting the statistic value vector of the operation data of the benchmark unit within the jth preset time, wherein the vector dimension is mx 1, namely, the statistic value of the operation data within m preset times is contained, the fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the benchmark unit are used as output and are recorded as D1,jRepresenting a fatigue damage value vector of the benchmark set within the jth preset time, marking the benchmark set as 1, and training a fatigue damage value neural network model, wherein the vector dimension is n multiplied by 1, namely the fatigue damage value vector contains n preset time fatigue damage values; the fatigue damage value neural network model is a system for presetting time of fatigue damage values of all parts relative to operation data of all units within preset time of all unitsNon-linear function of the evaluation value, denoted Di,j=fANN(Ai,j) And i is the number of the unit, and j represents the jth preset time since all the units are put into operation in a grid-connected mode.
5. The wind farm power optimization scheduling method based on fatigue damage value estimation according to claim 4, characterized in that:
the preset time statistic value of the operation data comprises a maximum value of wind speed in preset time, a variation range of wind speed in preset time, an average value of power in preset time, a minimum value of rotating speed of a generator in preset time, a maximum value of a pitch angle in preset time, a variation range of the pitch angle in preset time, a standard deviation of front and rear accelerations of the cabin in preset time and a standard deviation of lateral acceleration of the cabin in preset time.
6. The wind farm power optimization scheduling method based on fatigue damage value estimation according to claim 4, characterized in that:
the fatigue damage values of the blade root, the tower drum top and the tower drum bottom of the benchmark set comprise an in-plane direction fatigue damage value of a first blade root of the benchmark set, an out-plane direction fatigue damage value of the first blade root, an in-plane direction fatigue damage value of a second blade root, an out-plane direction fatigue damage value of the second blade root, an in-plane direction fatigue damage value of a third blade root, an out-plane direction fatigue damage value of the third blade root, a pitching direction fatigue damage value of the tower drum top, a rolling direction fatigue damage value of the tower drum top, a pitching direction fatigue damage value of the tower drum bottom and a rolling direction fatigue damage value of the tower drum bottom.
7. The wind farm power optimization scheduling method based on fatigue damage value estimation according to claim 4, characterized in that:
the hidden layer of the fatigue damage value neural network model is selected from 1 layer, the number of units of the hidden layer is 50, the activation function of the fatigue damage value neural network model is selected as a linear rectification function, and the weight optimization adopts an L-BFGS algorithm.
8. The wind farm power optimization scheduling method based on fatigue damage value estimation according to claim 1, characterized by comprising the following steps in step S6:
s601, P for combining future preset time power set values of all the unitssetVector designation, Pset=[Pset,1,Pset,2,…,Pset,N]T,Pset,NPresetting a power set value of the future preset time for the Nth set of the wind power plant;
s602, taking accumulated fatigue damage value vector D of each unit after preset timepredSum,iMaximum value of, i.e. max (D)predSum,i) Recording the equivalent accumulated fatigue damage value after the preset time of the unit; calculating the average value of the equivalent accumulated fatigue damage values of all the units after the preset timeCalculating the standard deviation of the equivalent accumulated fatigue damage values of all the units after the preset timeK1A weight factor that is the standard deviation of fatigue damage; the standard deviation index of fatigue damage is
S603, calculating the sum of the future preset time power set values of all the units, namelyCalculating a total power tracking indexK2A weight factor that is a total power trace; the total power tracking index is
S604, according to the fatigue damage standard deviation index and the total power tracking index, solving a target function adopted by the particle swarm optimization algorithm as follows:
wherein, Pset,farmPresetting a total power set value of the whole time field in the future;
the particle swarm optimization algorithm adopts the following constraint conditions:
Pset,i<Ppred,i
Pset,i<Pboost,i
wherein, the fatigue damage value of a single unit is restricted to be that the accumulated fatigue damage value after the preset time of each unit is smaller than the expected fatigue damage value, namelytjTime elapsed since the wind farm was put into operation for grid connection, TlifeThe expected life of the wind turbine generator; the predicted power set value of a single unit is restricted to be the power set value P of each unitset,iLess than the predicted power P of a single unitpred,iI.e. Pset,i<Ppred,i(ii) a Single tableThe set over-power setting value of the set is restricted to be the power setting value P of each setset,iOver-power P less than maximum single unitboost,iI.e. Pset,i<Pboost,i(ii) a The full field power set value is restricted to be the sum of the power set values of all the units in the full fieldLess than the power set value P dispatched by the power grid dispatchingset,farmI.e. by
S605, iterating the particle swarm optimization algorithm, wherein the dimension of a position vector and a velocity vector of each particle of the particle swarm optimization algorithm is Nx 1, and N is the number of wind power plant sets;
the iterative formula of the particle swarm optimization algorithm is as follows:
vm,k=ωvm,k-1+C1r1(xpbest,m,k-1-xi-1)+C2r2(xgbest,k-1-xm,k-1)
xm,k=xm,k-1+vm,k
k=k+1
wherein K is the maximum iteration number, K is the current iteration round, omega is an inertia weight factor, the value of the inertia weight factor is linearly reduced along with the increase of the iteration round, and omega isiniIs an initial inertial weight factor, ωendTo end the inertial weight factor, vm,kVelocity vector, x, for the m-th particle at the k-th iterationm,kFor the position vector of the m-th particle at the k-th iteration, vm,k-1Velocity vector, x, for the m-th particle at the k-1 th iterationm,k-1Is the position vector, C, of the m-th particle at the (k-1) th iteration1And C2Is a constant of a learning factor, r1And r2Is Nx 1Dimensional random vectors, i.e. each vector element is random data between (0,1), xpbest,m,k-1For the local optimal solution, x, of the m-th particle at the k-1 iterationgbest,k-1The global optimal solution of the k-1 iteration is obtained; when the iteration number K is equal to the maximum iteration number K, the particle swarm optimization calculation is finished, and a global optimal solution x is outputgbest,KAnd (5) vector quantity.
9. Wind power plant power optimization scheduling system based on fatigue damage value estimation is characterized by comprising:
the operation data acquisition and statistics module is used for acquiring operation data of each unit in the wind power plant;
the fatigue damage value on-line calculation module is used for calculating the fatigue damage value of the benchmark set according to the load data of the benchmark set in the wind power plant;
the fatigue damage value neural network model module is used for outputting fatigue damage value vectors of all units in the wind power plant within preset time;
the neural network model offline training module is used for performing offline training on the fatigue damage value neural network model according to the input operation data and the fatigue damage value of the benchmark set;
and the full-field power particle swarm optimization module is used for calculating iterative particle positions corresponding to the iterative times, so that the fatigue damage value neural network model is optimized and the power scheduling of all the units in the wind power field is optimized.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110841483.8A CN113659630B (en) | 2021-07-26 | 2021-07-26 | Wind power plant power optimal scheduling method and system based on fatigue damage value estimation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110841483.8A CN113659630B (en) | 2021-07-26 | 2021-07-26 | Wind power plant power optimal scheduling method and system based on fatigue damage value estimation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113659630A true CN113659630A (en) | 2021-11-16 |
CN113659630B CN113659630B (en) | 2024-03-19 |
Family
ID=78478094
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110841483.8A Active CN113659630B (en) | 2021-07-26 | 2021-07-26 | Wind power plant power optimal scheduling method and system based on fatigue damage value estimation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113659630B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113864137A (en) * | 2021-12-06 | 2021-12-31 | 天津发现技术有限公司 | Fatigue life monitoring method and system for whole wind turbine generator |
CN114123357A (en) * | 2021-11-24 | 2022-03-01 | 华能威宁风力发电有限公司 | AGC power optimization control method for wind power plant |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014201849A1 (en) * | 2013-06-18 | 2014-12-24 | 国网辽宁省电力有限公司电力科学研究院 | Method for actively optimizing, adjusting and controlling distributed wind power plant provided with energy-storage power station |
CN108547735A (en) * | 2018-04-17 | 2018-09-18 | 中南大学 | The integrated optimization control method of wind power plant active output and unit fatigue |
CN110120682A (en) * | 2019-05-08 | 2019-08-13 | 新疆大学 | A kind of minimum tower elevator supply Optimization Scheduling for losing abandonment amount |
CN110535174A (en) * | 2019-07-23 | 2019-12-03 | 电子科技大学 | A kind of active power controller method considering wind power plant fatigue load distribution and production capacity |
CN111291514A (en) * | 2020-02-03 | 2020-06-16 | 华北电力大学 | Method for rapidly predicting fatigue life of wind turbine generator |
CN112260325A (en) * | 2020-11-24 | 2021-01-22 | 武汉大学 | Wind power plant active power control method considering fatigue load of wind driven generator |
-
2021
- 2021-07-26 CN CN202110841483.8A patent/CN113659630B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014201849A1 (en) * | 2013-06-18 | 2014-12-24 | 国网辽宁省电力有限公司电力科学研究院 | Method for actively optimizing, adjusting and controlling distributed wind power plant provided with energy-storage power station |
CN108547735A (en) * | 2018-04-17 | 2018-09-18 | 中南大学 | The integrated optimization control method of wind power plant active output and unit fatigue |
CN110120682A (en) * | 2019-05-08 | 2019-08-13 | 新疆大学 | A kind of minimum tower elevator supply Optimization Scheduling for losing abandonment amount |
CN110535174A (en) * | 2019-07-23 | 2019-12-03 | 电子科技大学 | A kind of active power controller method considering wind power plant fatigue load distribution and production capacity |
CN111291514A (en) * | 2020-02-03 | 2020-06-16 | 华北电力大学 | Method for rapidly predicting fatigue life of wind turbine generator |
CN112260325A (en) * | 2020-11-24 | 2021-01-22 | 武汉大学 | Wind power plant active power control method considering fatigue load of wind driven generator |
Non-Patent Citations (5)
Title |
---|
庞辉庆等: "风电机组尾流与疲劳载荷关系分析", 农业工程学报, vol. 33, no. 10, pages 239 - 244 * |
张奔;徐伟;冯浩;: "基于疲劳分布的风电场有功控制策略", 人民长江, no. 09 * |
张晋华;吴文静;王卓然;刘永前;田德;: "基于降低风电场损耗的风电场优化调度研究", 太阳能学报, no. 04 * |
薛磊: "大型风电机组载荷监测与延寿评估技术研究", 《中国优秀硕士学位论文全文数据库工程科技II辑》 * |
许帅;霍志红;许昌;韩星星;丁佳煜;: "限功率控制下风电机组叶片疲劳损伤研究", 太阳能学报, no. 01 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114123357A (en) * | 2021-11-24 | 2022-03-01 | 华能威宁风力发电有限公司 | AGC power optimization control method for wind power plant |
CN114123357B (en) * | 2021-11-24 | 2023-08-15 | 华能威宁风力发电有限公司 | Wind farm AGC power optimization control method |
CN113864137A (en) * | 2021-12-06 | 2021-12-31 | 天津发现技术有限公司 | Fatigue life monitoring method and system for whole wind turbine generator |
Also Published As
Publication number | Publication date |
---|---|
CN113659630B (en) | 2024-03-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6906354B2 (en) | Wind turbine generator fatigue damage calculation device, wind power generation system, and wind turbine generator fatigue damage calculation method | |
CN112219028B (en) | Wind turbine control method | |
JP5984791B2 (en) | Wind power generator monitoring system and monitoring method | |
Jamieson et al. | Multi-rotors; a solution to 20 MW and beyond? | |
US10584680B2 (en) | Method for operating a wind turbine generator | |
CN113659630A (en) | Wind power plant power optimization scheduling method and system based on fatigue damage value estimation | |
CN111794909B (en) | Sector regulation-oriented wind farm level yaw dynamic optimization method and system | |
KR101706508B1 (en) | Data Analysis System of Mechanical Load Measurement Data for Wind Turbine | |
CN108953060B (en) | Wind power plant field level yaw control method based on laser radar anemoscope | |
CN106815771B (en) | Long-term assessment method for wind farm load | |
CN102419394B (en) | Wind/solar power prediction method with variable prediction resolution | |
CN110454329A (en) | A kind of Wind turbines award setting method | |
CN103541862A (en) | Method and system to actively pitch to reduce extreme loads on wind turbine | |
CN103061980A (en) | Feed-forward control system and feed-forward control method for wind generating set based on laser wind finding radar | |
GB2560223A (en) | Wind turbine farm level loads management control strategy | |
CN104123456A (en) | Classified statistic method and system of wind curtailment electric quality of wind power plant | |
US11639710B2 (en) | Systems and methods of coordinated yaw control of multiple wind turbines | |
CN115030866A (en) | Wind power plant group control system | |
CN108665102A (en) | Method based on the real-time generated energy of mesoscale data prediction wind power plant | |
CN113283035B (en) | Method, system, equipment and storage medium for constructing double-parameter wind turbine engine room transfer function | |
CN108345996A (en) | A kind of system and method reducing wind power checking energy | |
CN117521282B (en) | Density-dependent fan parameterization method for wind farm climate characteristic simulation | |
EP4372225A1 (en) | Method and system for wind speed determination using rpm and blade pitch angle | |
CN116345435A (en) | Technical improvement unit power increase evaluation method and system based on cluster characteristics of wind turbine generator | |
BARANAUSKAS et al. | The evaluation of the load and wind power forecasting errors |
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 |