CN110630448A - Soft measurement method for wind shear effect - Google Patents

Soft measurement method for wind shear effect Download PDF

Info

Publication number
CN110630448A
CN110630448A CN201910857475.5A CN201910857475A CN110630448A CN 110630448 A CN110630448 A CN 110630448A CN 201910857475 A CN201910857475 A CN 201910857475A CN 110630448 A CN110630448 A CN 110630448A
Authority
CN
China
Prior art keywords
wind
torque
soft measurement
shear effect
wind shear
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910857475.5A
Other languages
Chinese (zh)
Inventor
李志勇
王欣
陈有根
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Priority to CN201910857475.5A priority Critical patent/CN110630448A/en
Publication of CN110630448A publication Critical patent/CN110630448A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03D—WIND MOTORS
    • F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
    • 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
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/30—Control parameters, e.g. input parameters
    • F05B2270/304—Spool rotational speed
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/30—Control parameters, e.g. input parameters
    • F05B2270/337—Electrical grid status parameters, e.g. voltage, frequency or power demand
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/70—Type of control algorithm
    • F05B2270/709—Type of control algorithm with neural networks

Abstract

The patent provides a soft measurement method for a wind shear effect, and belongs to the technical field of wind power generation. The invention consists of a phase-locked frequency multiplier (10), an electromagnetic torque soft measurement (20), a torque difference soft measurement (30) and a constraint calculation (40). Firstly (10) outputting frequency multiplication sine and cosine components to (20) and (30); then (20) and (30) adopt an online learning neural network based on the self-adaptive noise cancellation principle to respectively pass through the output current i of the wind driven generatorsAnd the actually measured rotating speed omega of the wind wheel obtains an electromagnetic torque and a torque difference estimated value, and finally, soft measurement of mechanical torque is realized through (40). Because no filtering link exists, the method has good real-time performance, and can be used for not only accurately measuring the wind shear effect torque ripple, but also performing online compensation control on the wind shear effect torque ripple.

Description

Soft measurement method for wind shear effect
Technical Field
A wind shear effect soft measurement method relates to mechanical torque soft measurement and belongs to the technical field of wind power generation.
Background
Wind power generation utilizes wind power to drive blades to rotate, and mechanical energy of the generator is converted into electric energy. At high altitudes of the wind farm, the wind speed will vary with vertical height:
Figure BDA0002198753940000011
wherein
Figure BDA0002198753940000012
Respectively is height Zn，Z0The average wind speed, α, is related to the surface structure and the atmosphere stability. Therefore, when the blades rotate, the wind speeds of all points of the blades are different, a wind shear effect is generated, and the blades are subjected to micro-metaanalysis by using a phyllotactic theory method to obtain the mechanical torque of the n blades of the wind turbine:
Figure BDA0002198753940000013
wherein A, B is the correlation constant. From the above formula, it can be seen that if the wind changes with the height, the mechanical torque generated by the wind turbine is composed of a stable component and a ripple component, which are expressed as
Figure BDA0002198753940000014
Wherein, the stable component is derived from the average wind speed, the ripple component is derived from the wind shear effect, the pulsation frequency is n omega, and omega is the actually measured rotating speed of the wind turbine. This phenomenon not only reduces the performance of the wind turbine and the tower life, but also produces power pulsations in the power generation.
Therefore, the wind shear effect torque ripple can be accurately measured, the damage of the wind shear effect torque ripple can be scientifically evaluated, and a basis can be provided for compensation control of the wind shear effect torque ripple. However, at present, no effective scheme exists for measuring the periodic mechanical torque ripple, and the periodic mechanical torque ripple cannot be directly measured by additionally arranging a sensor; mechanical torque pulsation can not be obtained from the above formula by measuring wind speed changes at different heights and combining a phyllotactic theory; and the ripple component frequency is low, and the general filtering method has large error and time delay.
Disclosure of Invention
In order to measure the mechanical torque pulsation and further ensure the safe and reliable operation of the wind power generation device, the invention provides a soft measurement method for the wind shear effect.
The principle of the wind shear effect soft measurement method provided by the invention is that an artificial neural network is learned on line based on the self-adaptive noise cancellation principle, and iterative learning is carried out by combining a reasonable artificial neural network topological structure and parameters, so that the online real-time measurement of mechanical torque pulsation is realized. As shown in fig. 1, the soft measurement method of the present invention comprises: the phase-locked frequency multiplication circuit outputs frequency multiplication sine and cosine components as the input of an artificial neural network, outputs the optimal weight value by combining a reference item and performing linear combination after iterative learning, and the whole system has strong self-adaptive capacity, is easy to realize and has high efficiency.
The invention relates to a wind shear soft measurement process, which comprises the following steps:
the fan rotating speed sensor equipment inputs the real-time rotating speed omega of the wind turbine to the phase-locked frequency multiplier (10);
as shown in the attached figure 2, the input signal passes through a phase-locked circuit (11) to obtain a sine signal sin ω t, and the sine signal sin ω t and a cosine signal cosn ω t are obtained after frequency multiplication (12) and phase shift (13), wherein n is the number of wind wheel blades.
An artificial neural network (21) in the electromagnetic torque soft measurement (20) inputs { sin n ω t, cosn ω t, 1} with a corresponding weight of { w }, as shown in fig. 3(a)1s，w1c，w1Reference item is the output current i of the wind driven generatorsOn-line learning error amount
Figure BDA0002198753940000015
And reference item isSatisfies the formula:wherein
Figure BDA0002198753940000017
Respectively output current i for wind power generatorsA stable component estimation term and a pulsating component estimation term; the learning of the neural network adopts a least mean square algorithm, the weight is adjusted by using error quantity, and the formula is as follows:
Figure BDA0002198753940000018
wherein eta is the learning rate (eta is more than 0 and less than or equal to 1). Over a number of iterations, the weights approach the optimum value and are employed by a linear combination (22). As shown in FIG. 3(b), (22) the optimal weights in the input set { sinn ω t, cosn ω t, 1}, and (21) and the scaling factor K in FIG. 5p4Linearly combined and output
Figure BDA0002198753940000021
The formula is as follows:
Figure BDA0002198753940000022
wherein the content of the first and second substances,
Figure BDA0002198753940000023
an electromagnetic torque estimation term,
Figure BDA0002198753940000024
Is an electromagnetic torque ripple estimation term based on
Figure BDA0002198753940000025
In the above formula
Figure BDA0002198753940000026
Wherein
Figure BDA0002198753940000027
Is a permanent magnet flux linkage, npThe number of pole pairs of the generator.
The torque soft measurement (30) inputs { sinn ω t, cosn ω t, 1} with a weight of { w }2s，w2c，w2}, output
Figure BDA0002198753940000028
As shown in fig. 4; the reference term is torque difference delta T, and error amount e is learned onlineΔTSatisfies the formula with the reference term Δ T:
Figure BDA0002198753940000029
Figure BDA00021987539400000210
is a torque difference estimation term; according to
Figure BDA00021987539400000211
And FIG. 5, determining the parameter T3WhereinAnd J is moment of inertia. The learning of the neural network adopts a least mean square algorithm, the weight is adjusted by using error quantity, and the formula is as follows:
Figure BDA00021987539400000212
wherein eta is the learning rate (eta is more than 0 and less than or equal to 1), and each weight approaches the optimal value after a plurality of iterations. By the formula:
Figure BDA00021987539400000213
performing linear calculation to estimate the output torque difference and its pulsation
Figure BDA00021987539400000214
As shown in fig. 4, linear combination (40) input to (20) output
Figure BDA00021987539400000215
And (30) output
Figure BDA00021987539400000216
Performing constraint calculation and output
Figure BDA00021987539400000217
And obtaining the optimal soft measurement value of the mechanical torque and the pulsation thereof, and measuring the wind shear effect.
Figure BDA00021987539400000218
Drawings
The patent is described in further detail below with reference to the figures and the detailed description.
FIG. 1 is an overall flow chart of soft measurement of wind shear effect torque ripple.
Fig. 2 is a flow chart of the internal phase-locked frequency multiplier of fig. 1.
Fig. 3(a) is a schematic diagram of an electromagnetic torque soft measurement artificial neural network.
Fig. 3(b) is a flow chart of the linear combination of the electromagnetic torque soft measurement.
FIG. 4 is a schematic diagram of a soft torque difference measurement
Fig. 5 is a block diagram of fan speed control.
Detailed Description
The invention discloses a soft measurement method for a wind shear effect, and introduces the basic composition and the measurement process of the method. The basic components comprise a standard phase-locked frequency multiplier, electromagnetic torque soft measurement, torque difference soft measurement and constraint calculation, all the parts are connected as shown in figure 1, and a direct-drive permanent magnet synchronous power generation system is taken as an example to introduce a specific flow of wind shear effect soft measurement of the system.
The output of the actually measured rotating speed sensor of the wind power generation system is connected to the phase-locked frequency multiplier and the torque difference soft measurement, and the output of the stator current sensor is connected to the electromagnetic torque soft measurement. And giving an initial weight value, and accessing sine and cosine components output by the phase-locked frequency multiplier into electromagnetic torque soft measurement and torque difference soft measurement. And the online learning neural network in the electromagnetic torque soft measurement and the torque difference soft measurement outputs the optimal estimated values of the electromagnetic torque and the torque difference after a plurality of times of iterative learning.
The output of the electromagnetic torque soft measurement and the torque difference soft measurement is accessed to the constraint calculation, the equation calculation is carried out, and the optimal estimated value of the mechanical torque is output, the pulsation of the optimal estimated value consists of sine and cosine components, and the weight values respectively correspond to the peak values of the sine component and the cosine component.
The practical significance of the invention is as follows: the wind shear effect soft measurement method can be applied to a wind power generation system to realize real-time mechanical torque pulsation measurement, an artificial neural network learning algorithm is applied to the technical method, the topological structure is simple, complex and redundant lead lines are not needed, and high-accuracy and high-efficiency online soft measurement can be realized without using excessive equipment. The method has important significance for ensuring the safe and reliable operation of the wind power generation system.

Claims (5)

1. A wind shear effect soft measurement method is characterized in that: comprises a phase-locked frequency multiplier (10), an electromagnetic torque soft measurement (20), a torque difference soft measurement (30) and a constraint calculation (40);(10) outputting frequency multiplication sine signal and cosine signal components to (20), (30); (20) and (30) respectively passing through the output current i of the wind driven generatorsAnd the electromagnetic torque and the torque difference are calculated with the actually measured rotating speed omega of the wind wheel, and then the soft measurement of the wind shear effect is realized through (40).
2. The device (10) of claim 1, wherein: comprises a phase lock (11), a frequency multiplication (12) and a phase shift (13); inputting actually measured wind wheel rotating speed omega, obtaining sin omega t through phase locking (11), obtaining sin n omega t through frequency doubling (12), and finally obtaining cosn omega t through phase shifting (13), wherein n is the number of wind wheel blades.
3. The method as set forth in claim 1 (20), characterized in that: the method comprises an online learning neural network (21) based on the self-adaptive noise cancellation principle and a linear combination (22); (21) the input set is { sin n ω t, cos n ω t, 1} and the corresponding weight is { w }1s，w1c，w1On-line learning of error amount }
Figure FDA0002198753930000011
And reference item isSatisfy the formula
Figure FDA0002198753930000012
Wherein
Figure FDA0002198753930000013
Respectively output current i for wind power generatorsStable component estimation item, pulsation component estimation item, optimal weight { w after iterative learning1s，w1c，w1Participating in linear calculation; (22) the input set is { sin n ω t, cos n ω t, 1}, and the output set is
Figure FDA0002198753930000014
Satisfies the formula:
wherein the content of the first and second substances,
Figure FDA0002198753930000016
an electromagnetic torque estimation term,
Figure FDA0002198753930000017
Is an electromagnetic torque ripple estimation term that is,
Figure FDA0002198753930000018
is a permanent magnet flux linkage, npThe number of pole pairs of the generator.
4. The (30) of claim 1, wherein: the online learning neural network based on the self-adaptive noise cancellation principle has an input set of { sin n ω t, cos n ω t, 1} and a corresponding weight of { w }2s，w2c，w2Put the set of outputs as
Figure FDA0002198753930000019
The on-line learning error amount is obtained by subtracting the output amount from the parameter item delta T
Figure FDA00021987539300000110
Wherein Δ T and ω satisfy the formula:j is moment of inertia.
5. The method (40) of claim 1, wherein: the input amount of which is as described in claim 3
Figure FDA00021987539300000112
And as recited in claim 4
Figure FDA00021987539300000113
Figure FDA00021987539300000114
Warp beam
Figure FDA00021987539300000115
Obtaining an estimate of the mechanical torque and its ripple
Figure FDA00021987539300000116
The wind shear effect is measured.
CN201910857475.5A 2019-09-11 2019-09-11 Soft measurement method for wind shear effect Pending CN110630448A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910857475.5A CN110630448A (en) 2019-09-11 2019-09-11 Soft measurement method for wind shear effect

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910857475.5A CN110630448A (en) 2019-09-11 2019-09-11 Soft measurement method for wind shear effect

Publications (1)

Publication Number Publication Date
CN110630448A true CN110630448A (en) 2019-12-31

Family

ID=68972550

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910857475.5A Pending CN110630448A (en) 2019-09-11 2019-09-11 Soft measurement method for wind shear effect

Country Status (1)

Country Link
CN (1) CN110630448A (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009026930A2 (en) * 2007-08-31 2009-03-05 Vestas Wind Systems A/S Method for controlling at least one adjustment mechanism of a wind turbine, a wind turbine and a wind park
US20140097618A1 (en) * 2012-10-09 2014-04-10 Alstom Renovables Espana, S.L. Method of operating a variable speed wind turbine
US20160115941A1 (en) * 2014-10-27 2016-04-28 General Electric Company System and method for adaptive rotor imbalance control
CN108518305A (en) * 2018-03-27 2018-09-11 沈阳工业大学自控技术有限公司 A kind of Wind turbines control method and system
CN109596255A (en) * 2017-09-25 2019-04-09 森维安有限责任公司 Measure device, wind-driven generator operation method and the wind-driven generator of wind-driven generator torque
CN109936163A (en) * 2019-03-27 2019-06-25 中南大学 A kind of suppressing method of ripple caused by wind shear effect

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009026930A2 (en) * 2007-08-31 2009-03-05 Vestas Wind Systems A/S Method for controlling at least one adjustment mechanism of a wind turbine, a wind turbine and a wind park
US20140097618A1 (en) * 2012-10-09 2014-04-10 Alstom Renovables Espana, S.L. Method of operating a variable speed wind turbine
US20160115941A1 (en) * 2014-10-27 2016-04-28 General Electric Company System and method for adaptive rotor imbalance control
CN109596255A (en) * 2017-09-25 2019-04-09 森维安有限责任公司 Measure device, wind-driven generator operation method and the wind-driven generator of wind-driven generator torque
CN108518305A (en) * 2018-03-27 2018-09-11 沈阳工业大学自控技术有限公司 A kind of Wind turbines control method and system
CN109936163A (en) * 2019-03-27 2019-06-25 中南大学 A kind of suppressing method of ripple caused by wind shear effect

Similar Documents

Publication Publication Date Title
CN103410659B (en) Effective wind speed obtaining method of wind generating set based on High-Gain observer
CN107061158B (en) A kind of prediction of low wind speed leeward power generator and tracking and controlling method
Ciampichetti et al. Sliding mode control of permanent magnet synchronous generators for wind turbines
CN111293686A (en) ARMAX system identification-based real-time evaluation method for inertia of power system
Mishra et al. Comprehensive review on maximum power point tracking techniques: wind energy
CN101106279A (en) Merged network and power adjustment system for direct drive permanent magnetic synchronization generator and its method
CN107341299A (en) The blower fan Reliability Modeling that meter and running environment influence
Azad et al. Parameter estimation of doubly fed induction generator driven by wind turbine
CN104675629A (en) Maximum wind energy capturing method of variable-speed wind generating sets
Sarkar et al. A study of MPPT schemes in PMSG based wind turbine system
Yan et al. Torque estimation and control of PMSM based on deep learning
Afrasiabi et al. Ensemble kalman filter based dynamic state estimation of PMSG-based wind turbine
Corradini et al. Fault-tolerant sensorless control of wind turbines achieving efficiency maximization in the presence of electrical faults
Thiringer et al. Control of a variable-speed pitchregulated wind turbine
CN109599889A (en) DFIG low voltage traversing control method, system under unbalance voltage based on fuzzy active disturbance rejection
Syskakis et al. MPPT for small wind turbines: Zero-oscillation sensorless strategy
Tian et al. A Gaussian RBF network based wind speed estimation algorithm for maximum power point tracking
He et al. Grey prediction pi control of direct drive permanent magnet synchronous wind turbine
CN110630448A (en) Soft measurement method for wind shear effect
Gong et al. Wind speed and rotor position sensorless control for direct-drive pmg wind turbines
CN105958886A (en) Online estimating device and method of observing impeller fatigue service life real-timely based torques
El Aimani Modeling and control structures for variable speed wind turbine
Fu et al. MPPT control based fuzzy for wind energy generating system
Hocine et al. A hybrid sensorless control of PMSG wind-power generator with frequency signal injection method and extended Kalman filter
CN114465280A (en) Dynamic equivalent modeling method for new energy grid-connected system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20191231