CN104454346B - Maximum power tracking control method for small permanent-magnet direct-drive wind power generation system - Google Patents
Maximum power tracking control method for small permanent-magnet direct-drive wind power generation system Download PDFInfo
- Publication number
- CN104454346B CN104454346B CN201410622989.XA CN201410622989A CN104454346B CN 104454346 B CN104454346 B CN 104454346B CN 201410622989 A CN201410622989 A CN 201410622989A CN 104454346 B CN104454346 B CN 104454346B
- Authority
- CN
- China
- Prior art keywords
- wind
- speed
- lampyridea
- model
- support vector
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 81
- 238000010248 power generation Methods 0.000 title claims abstract description 16
- 238000012706 support-vector machine Methods 0.000 claims abstract description 7
- 239000013598 vector Substances 0.000 claims description 79
- 239000000523 sample Substances 0.000 claims description 25
- GNBHRKFJIUUOQI-UHFFFAOYSA-N fluorescein Chemical compound O1C(=O)C2=CC=CC=C2C21C1=CC=C(O)C=C1OC1=CC(O)=CC=C21 GNBHRKFJIUUOQI-UHFFFAOYSA-N 0.000 claims description 21
- 238000012360 testing method Methods 0.000 claims description 15
- 241000131894 Lampyris noctiluca Species 0.000 claims description 14
- 230000005389 magnetism Effects 0.000 claims description 12
- 230000008859 change Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 239000003623 enhancer Substances 0.000 claims description 3
- 241000254158 Lampyridae Species 0.000 abstract 1
- 230000008569 process Effects 0.000 description 11
- 230000005611 electricity Effects 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 238000012417 linear regression Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 241000288673 Chiroptera Species 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 230000001373 regressive effect Effects 0.000 description 2
- IGXWBGJHJZYPQS-SSDOTTSWSA-N D-Luciferin Chemical compound OC(=O)[C@H]1CSC(C=2SC3=CC=C(O)C=C3N=2)=N1 IGXWBGJHJZYPQS-SSDOTTSWSA-N 0.000 description 1
- CYCGRDQQIOGCKX-UHFFFAOYSA-N Dehydro-luciferin Natural products OC(=O)C1=CSC(C=2SC3=CC(O)=CC=C3N=2)=N1 CYCGRDQQIOGCKX-UHFFFAOYSA-N 0.000 description 1
- BJGNCJDXODQBOB-UHFFFAOYSA-N Fivefly Luciferin Natural products OC(=O)C1CSC(C=2SC3=CC(O)=CC=C3N=2)=N1 BJGNCJDXODQBOB-UHFFFAOYSA-N 0.000 description 1
- 241001165575 Hylotelephium telephium subsp. maximum Species 0.000 description 1
- DDWFXDSYGUXRAY-UHFFFAOYSA-N Luciferin Natural products CCc1c(C)c(CC2NC(=O)C(=C2C=C)C)[nH]c1Cc3[nH]c4C(=C5/NC(CC(=O)O)C(C)C5CC(=O)O)CC(=O)c4c3C DDWFXDSYGUXRAY-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000005276 aerator Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
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
- F03D7/00—Controlling wind motors
-
- 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
- Y02E10/72—Wind turbines with rotation axis in wind direction
Abstract
The invention relates to a maximum power tracking control method for a small permanent-magnet direct-drive wind power generation system. According to the method, on the basis of collecting a lot of actual samples of wind speed vector-rotating speed-power, a firefly support vector machine regression prediction model of the wind speed is established, and wind speed estimation is performed by utilizing the model; next, the optimal rotating speed, corresponding to the maximum power point, of a draught fan is predicted through the optimum tip speed ratio method; afterwards, the rotating speed of the draught fan is adjusted to the predicted optimum rotating speed of the draught fan, and the maximum power of the draught fan is tracked based on the perturbation and observation method at a set perturbation step length with the rotating speed as an initial value. According to the maximum power tracking control method, wind speed estimation without sensors is achieved, the control cost of a power generation system is greatly reduced, the speed and accuracy for looking for the maximum power point are enhanced, and power loss during perturbation is lowered.
Description
Technical field
The present invention relates to a kind of method for controlling minitype permanent magnetism directly-driving wind power generation system, belong to generation technology neck
Domain.
Background technology
With the continuous aggravation of global energy crisis, regenerative resource is increasingly subject to people's attention.Wind-power electricity generation because
Having many advantages has become one of Main way of current generation of electricity by new energy.Miniature wind power generation system（Less than 15kW）Because of it
Cost is relatively low, cannot provide the place of electric power support great competitiveness in the bulk power grid such as remote mountain area and island.Therefore, right
Miniature wind power generation system carries out correlational study, has important practical significance.
The output of wind-driven generator is relevant with external environment and loading condition, and certain external environment and negative
The maximum power point of existence anduniquess in the case of load（MPP）.For improve generating efficiency, need between wind-driven generator and load
Setting maximal power tracing（MPPT）Circuit.Wind generator system maximum power tracking and controlling method has many kinds, the most at present
There are tip speed ratio control methods, perturbation observation method（P&O）, neural network（ANN）Deng.The pluses and minuses of these methods are as follows：
Tip speed ratio control methods：Under the wind speed of a certain fixation, as long as guarantee wind energy conversion system rotating speed is a certain numerical value, that is, keep
Tip speed ratio is optimum tip-speed ratio, can realize maximal power tracing.The method is relatively simple, however it is necessary that to effective wind
Speed measures.Because the area of wind wheel is larger, and wind speed has certain undulatory property and randomness, is entered by air velocity transducer
The accurate measurement of row effective wind speed is more difficult.In addition, the introducing of air velocity transducer not only increases being manufactured into of electricity generation system
This, and the reliability of whole system can be reduced.
Perturbation observation method：Its principle is to apply periodic disturbance to rotation speed of fan, and the change further according to output determines
The perturbation direction of next step is to realize maximal power tracing.The method implements relatively easily, but what the method searched out
Operating point can only oscillate around operation in MPP, leads to wind energy can not be fully utilized.In addition, the initial value of rotation speed of fan and disturbing
Dynamic step-length has large effect to the accuracy and speed followed the tracks of, and sometimes it also occur that misjudgment phenomenon, have impact on the control of this method
Effect processed.
Neural network ANN：The method predicts maximum power point by training BP neural network model（MPP）, permissible
Avoid in perturbation observation method the power loss caused by disturbance repeatedly.But, occurred by sample limitation and training process
Cross learn or owe study impact, there is certain error between the maximum power point predictive value of model and actual value unavoidably.
In sum, the shortcomings of existing wind generator system control method has high cost, follows the tracks of poor accuracy, limits
The development of wind-power electricity generation is it is therefore necessary to improved.
Content of the invention
Present invention aims to the drawback of prior art, provide a kind of control low cost, tracking accuracy high
Minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method.
Problem of the present invention is to be realized with following technical proposals：
A kind of minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method, methods described is gathering a large amount of wind
On the basis of the actual sample of fast vector-rotating speed-power, set up Lampyridea Support vector regression forecast model the profit of wind speed
Carry out wind estimation with this model, then predict the optimum blower fan corresponding to maximum power point by optimum tip-speed ratio method and turn
Speed, then by the optimum rotation speed of fan of the rotational speed regulation of blower fan to prediction, and with this rotating speed as initial value, using perturbation observation method
Follow the tracks of the peak power of blower fan with the disturbance step-length setting.
Above-mentioned minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method, methods described includes following step
Suddenly：
A. set up Lampyridea Support vector regression forecast model：
1. willnIndividual air velocity transducer is respectively arranged on front side of draught fan impeller, the circular flat inswept with impeller coaxial and
Circle centre position in equal-sized circular flat and apart from the center of circle1/(n-1)At radius,2/(n-1)At radius ..., 1 sesquialter
At footpath, record wind velocity vectorV=[V 1 ,V 2 ...,V n ]T；
2. remember that the wind velocity vector under a certain wind speed environments isV(i)= [V 1 (i),V 2 (i)...,V n (i)]T, with this wind speed to
The DC voltage measuring rectifier output end in corresponding wind generator system isV oc (i), DC current isI oc (i), rotation speed of fan
Forω oc (i), using recordV oc (i)WithI oc (i)Calculate output this momentP oc (i)= V oc (i)• I oc (i), obtain
One group of sample (V(i),P oc (i),ω oc (i))；By collecting the sample group under different wind speed environments, formed sample set (V(i),P oc (i),ω oc (i))}；
3. utilize sample set (V(i),P oc (i),ω oc (i)) training Lampyridea Support vector regression forecast model, its
Middle mode input beP oc (i),ω oc (i), model is output asV(i)；
B. carry out wind estimation using Lampyridea Support vector regression forecast model：
At a time, it is utilized respectively voltage sensor, current sensor gathers commutator output in wind generator system
DC voltageV oc With DC currentI oc , gather rotation speed of fan using speed probeω oc , and calculate output work this moment
RateP oc = V oc • I oc ；Willω oc ,P oc As Lampyridea Support vector regression forecast model input, model be output as work as
The estimated value of front wind velocity vector, and calculate the estimated value of effective wind speed according to the following formula：
（1）
C. utilize optimum tip-speed ratio method, calculate the estimated value of effective wind speedCorresponding optimal rotation speed of fan prediction
Valueω ref ：
（2）
Wherein,λFor tip speed ratio；RFor impeller radius；
D. passing ratio integral control method adjusts the rotating speed of blower fan, makes up to optimal rotation speed of fan predictive valueω ref ；
Then withω ref For initial value, the disturbance step delta adopting perturbation observation method to setωFollow the tracks of the peak power of blower fan；
If the power difference before and after the disturbance that e. perturbation observation method is tried to achieve is more than or equal to given thresholdT r , illustrate that wind speed occurs
Mutation, repeat step b is to step e；Otherwise, the disturbance step-length that continuation adopts perturbation observation method to set follows the tracks of the maximum of blower fan
Power.
Above-mentioned minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method, described Lampyridea supporting vector
Machine regressive prediction model is to carry out the supporting vector machine model of parameter optimization using glowworm swarm algorithm, and specific optimization step is：
I. the parameter of initialization glowworm swarm algorithm：Initialization Lampyridea numbern, fluorescein volatility coefficient, fluorescein increasing
The strong factorγ, sensing ranger s （I.e. penalty factor to be optimized in Support vector regression forecast modelCWith nuclear parameterChange
Change scope）, neighborhood rate of changeβ, neighbours' threshold valuen t , moving step lengths, initial fluorescence elementl 0 With maximum iteration time；
II. penalty factor to be optimized in Support vector regression forecast modelCWith nuclear parameterSpan in,
As the individual initial position of Lampyridea, fitness function is using Lampyridea present position to random one group of numerical value of generation
（A certain penalty factorCWith nuclear parameter）Average phase on test samples for the Support vector regression forecast model trained
Negative value to error, average relative errorExpression formula is：
；
Wherein,N 1For test samples number,For the true value of wind velocity vector,Wind velocity vector for regression model output is estimated
Evaluation；
III. calculate the fluorescein concentration of the fluorescein concentration, decision domain scope and neighbours of every Lampyridea, by neighbours'
Fluorescein concentration determines the moving direction of each Lampyridea and moves forward；
IV. judging whether glowworm swarm algorithm reaches end condition, if reaching, choosing fitness highest Lampyridea position
For the parameter of supporting vector machine model, and record the supporting vector in corresponding regression model in case using；Otherwise, go to step III.
Above-mentioned minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method, the termination of described glowworm swarm algorithm
Condition is average relative error on test samples for the Support vector regression forecast model 3% or iterationses surpass
Cross maximum iteration time.
The beneficial effects of the present invention is：
（1）The present invention realizes no sensor wind estimation, greatlys save the control cost of electricity generation system, improves searching
The speed of maximum power point and accuracy；
（2）When external environment changes, can directly working speed be adjusted to optimal wind by means of forecast model
Machine rotor speed forecast valueω ref Vicinity, it is to avoid the process that perturbation observation method is progressively soundd out, thus improve tracking velocity；
（3）With optimal rotation speed of fan predictive valueω ref When carrying out disturbance observation for initial value, due to the prediction of optimal rotation speed of fan
Valueω ref It has been sufficiently close to actual optimal rotation speed of fan, therefore less disturbance step-length can be set, thus effectively reduce disturbing
The power loss of dynamic process；
（4）After the characteristic of blower fan changes, new forecast model can be trained to ensure by again collecting sample
Precision of prediction.
Brief description
The invention will be further described below in conjunction with the accompanying drawings.
Fig. 1 is hardware structure diagram of the present invention；
Fig. 2 is the schematic diagram carrying out wind velocity vector estimation using Lampyridea Support vector regression forecast model.
Fig. 3 is tip speed ratio and power coefficient graph of a relation；
Fig. 4 is perturbation observation method（P&O）Algorithm flow chart；
The flow chart of the maximum power tracking method that Fig. 5 provides for the present invention；
In figure is respectively numbered：1- air velocity transducer；2- permanent magnet direct-driving aerogenerator；3- speed probe；4- commutator；
5- voltage sensor；6- first electric capacity；7- second electric capacity；8- current sensor.
In in figure and literary composition, each symbol is：V oc (i)For the DC voltage of rectifier output end in wind generator system,I oc (i)For the DC current of rectifier output end in wind generator system,ω oc (i)For rotation speed of fan,P oc (i)For blower fan output
Power,N 1For test samples number,For the true value of wind velocity vector,For the estimated value of wind velocity vector,Estimating for effective wind speed
Evaluation,ω ref For optimal rotation speed of fan predictive value,λFor tip speed ratio,RFor impeller radius, ΔωFor disturbance step-length,nFor initial
Change Lampyridea number,For fluorescein volatility coefficient,γFor fluorescein enhancer,r s For sensing range,CFor penalty factor,For nuclear parameter,βFor neighborhood rate of change,n t For neighbours' threshold value,sFor the moving step length of Lampyridea,l 0 It is plain for initial fluorescence,For average relative error on test samples for the Support vector regression forecast model,K（x i ,x）For kernel function,For
Lagrange coefficient,y i For linear regression function output vector,It is non-linear from the input space to high-dimensional feature space
Mapping matrix,x i For input vector,WFor weight matrix,bFor bias vector,J(x)For fitness function,l i (t)FortWhen
Carve theiThe fluorescein concentration of Lampyridea.
Specific embodiment
Below in conjunction with the accompanying drawings, preferred embodiment is elaborated.It should be emphasized that the description below is merely exemplary
, rather than in order to limit the scope of the present invention and its application.
Fig. 1 is the hardware structure diagram of the embodiment of the present invention, wherein, the major parameter of minitype permanent magnetism direct wind-driven generator 2
For：Rotor diameter is 1.2m, and rated power is 300W, and rated voltage is 24V, rated speed 800r/min, threshold wind velocity 1m/s,
Rated wind speed 10m/s, survival wind speed 25m/s；The major parameter of maximum power tracking device is：MPPT controller adopts
DsPIC33FJ06GS101 single-chip microcomputer, DC-DC converter adopts Boost circuit, and drive module selects MCP14E3, voltage sensor
LV28-P selected by device 5, and LA25-NP selected by current sensor 8, speed probe 3 adopt voltage zero-cross detect formula cymometer, first
Electric capacity C1=10uF, the second electric capacity C2=100uF.
This minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method comprises the following steps：
Step 1：Set up Lampyridea Support vector regression forecast model：
Step 101：WillnIndividual air velocity transducer is respectively arranged on front side of draught fan impeller, coaxial with impeller circular flat, parallel
And the circle centre position in equal-sized circular flat and apart from the center of circle1/(n-1)At radius,2/(n-1)At radius ..., 1 times
At radius, then can record wind velocity vectorV=[V 1 ,V 2 ...,V n ]T；In embodiment, take n=4, therefore wind velocity vector can be recordedV=[V 1 ,V 2 ...,V 4 ]T；
Step 102：Remember that the wind velocity vector under a certain wind speed environments isV(i)= [V 1 (i),V 2 (i)...,V n (i)]T, now
In wind generator system, the DC voltage of rectifier output end isV oc (i), DC current isI oc (i), rotation speed of fan isω oc (i), calculate output this momentP oc (i)= V oc (i)• I oc (i), then may make up one group of sample (V(i),P oc (i),ω oc (i))；By collecting the sample group under different wind speed environments, formed sample set (V(i),P oc (i),ω oc (i))}；Embodiment
In, constitute small test wind-tunnel using aerator, converter and straight length, control drum by adjusting the frequency setting value of converter
Blower fan is exerted oneself, and then creates different wind speed environments.Set frequency converter frequency from the beginning of 10Hz, risen to for interval with 2Hz
60Hz, it is possible to provide 31 kinds of wind speed environments, that is,V(i)= [V 1 (i),V 2 (i),V 3 (i),V 4 (i)]T,i=1, …, 31}.Every
Under a kind of wind speed environments, by adjusting the dutycycle of DC-DC converter pwm signal, control rotation speed of fan from the beginning of 300r/min,
1500r/min is raised to for interval with 30r/min, totally 41 kinds of rotating speeds, record the straight of corresponding rectifier output end under each rotating speed
Stream voltageV oc With DC currentI oc , and calculate corresponding outputP oc = V oc • I oc , then add up to obtain 1271 groups of samples
This.
Step 103：Using sample set (V(i),P oc (i),ω oc (i)) training Lampyridea Support vector regression prediction
Model, wherein mode input be,ω oc (i)), model is output asV(i)；In embodiment, randomly draw 1271 groups of samples
900 groups of training samples as model in this, remaining 371 groups of test samples as model, for test model precision.
The input space is mapped to the feature space of higher-dimension by Support vector regression SVR by nonlinear mapping, utilizes one
Individual linear function collection is carrying out regression estimates.Lampyridea support vector machine of the present invention are multi output support vector machine, give
Sample set(For input vector,For corresponding output vector, N is number of samples, and n is input
Vector dimension, m is output vector dimension), used by multi output support vector machine SVR, linear regression function is：
(3)
Wherein：
y i For linear regression function output vector；
It is the nonlinear mapping matrix from the input space to high-dimensional feature space；
x i For input vector；
WFor weight matrix；
bFor bias vector.
Weight matrixWAnd bias vectorbTo be calculated by minimizing following formula：
(4)
In formula：
CFor penalty factor；
For the slack variable vector introducing；
Set up Lagrange's equation according to (4), can solve linear regression function is：
(5)
In formula：
K（x i ,x）For kernel function,；
For Lagrange coefficient, it is not zeroCorresponding vectorx i Referred to as supporting vector.
The kernel function selecting multi-form can generate different support vector machine, and conventional kernel function has：Multinomial letter
Number, Gaussian function, Sigmoid function etc..It is kernel function that the present invention chooses Gaussian function, that is,：
(6)
Wherein,Width parameter for gaussian kernel function.
The precision of described Support vector regression forecast model depends primarily on two major parametersC、, penalty factorC
Compromise for Controlling model complexity and approximate error；Nuclear parameterThe distribution or the scope that reflect training sample data are special
Property.In order to obtain high-precision regressive prediction model, the present invention is optimized to this two parameters using glowworm swarm algorithm.
Glowworm swarm algorithm is a kind of new gunz that India scholar Krishnanand and Ghose proposed in 2005
Can optimized algorithm.Glowworm swarm algorithm simulates nature Lampyridea by the luminous behavior attracting companion to seek a spouse or look for food, and has very
Good ability of searching optimum, has been successfully applied to the problems such as function optimization, sensor configuration.
In GSO algorithm, every Lampyridea is distributed in the definition space of object function, the carried luciferin of their own
Brightness is relevant with the object function fitness value on oneself position, and brighter Lampyridea represents that the position that it is located is better.
Assumex i (t)RepresenttMomentiThe position of Lampyridea,J(x)For fitness function,l i (t)RepresenttMomenti
The fluorescein concentration of Lampyridea, then
(7)
In formula (7)：For fluorescein volatility coefficient；For fluorescein enhancer, i.e. fitness withdrawal ratio.
Lampyridea finds neighborhood in the decision region of region, and brighter Lampyridea has higher captivation and goes to attract
The Lampyridea of surrounding moves toward this direction.The size of region decision region can be affected by neighbours' quantity, when neighbours' density relatively
Low, the decision-making radius of Lampyridea can increase and be beneficial to the more neighbours of searching；Conversely, decision-making radius reduces.Decision domain scope is more
New formula is
(8)
In formula (8)：Fort+1 momentiThe decision region of individual Lampyridea；r s For sensing range；Become for neighborhood
Rate；n t For neighbours' threshold value；N i （t）FortMomentiThe neighborhood of Lampyridea.
(9)
In formula (9),RepresentxNorm.Lampyridea in motor process, according to Lampyridea each in its neighborhood
Fluorescein concentration is determining its moving direction.P ij （t）RepresenttMomentiLampyridea is in its neighborhoodjFirefly
The probability of fireworm movement, then
(10)
Select the with wheel disc method according to movement probabilityiThe direction of Lampyridea movement, theniLampyridea existst+ 1 moment
Position be
(11)
In formula (11),Moving step length for Lampyridea.
Using concretely comprising the following steps that glowworm swarm algorithm is optimized to support vector machine：
1) initialize the parameter of glowworm swarm algorithm.Initialization Lampyridea numbern, fluorescein volatility coefficient, fluorescein increasing
The strong factorγ, sensing ranger s , neighborhood rate of changeβ, neighbours' threshold valuen t , moving step lengths, initial fluorescence elementl 0 With greatest iteration
Number of times, sensing ranger s For parameter to be optimized in Support vector regression forecast model（Penalty factorCWith nuclear parameter）Change
Change scope.In embodiment, settingn=20,=0.4,β=0.08,n t =5,s=0.03, initial fluorescence elementl 0 =5, maximum changes
Generation number is 1000, penalty factorCSpan be [10-1,104], nuclear parameterSpan be [10-2,104]；
2) in parameter to be optimizedCWithSpan in, random generate one group of numerical value as individual initial of Lampyridea
Position.Fitness function is using Lampyridea present position（A certainCWith）The Support vector regression prediction trained
The negative value of average relative error on test samples for the model.Average relative errorExpression formula is：
(12)
Wherein,N 1For test samples number,For the true value of wind velocity vector,Wind velocity vector for regression model output is estimated
Evaluation.Understand, average relative error on test samples for the model is less, and fitness is higher；
3) according to formula (7)-(11), calculate the fluorescein of the fluorescein concentration, decision domain scope and neighbours of every Lampyridea
Concentration, determines the moving direction of each Lampyridea by the fluorescein concentration of neighbours and moves forward；
4) judge whether glowworm swarm algorithm reaches end condition.If reaching, choose fitness highest Lampyridea position
For the parameter of supporting vector machine model, and record the supporting vector in corresponding regression model in case using；Otherwise, 3 are turned).This reality
Apply in example, the end condition of setting is average relative error on test samples for the Support vector regression forecast model 3% or iterationses exceed maximum iteration time.Through 128 iteration, optimization process terminates, the fitness highest of acquisition
Supporting vector machine model parameter beC=486.2,=69.54, this model includes 89 supporting vectors altogether.
Using the Support vector regression forecast model obtaining, it is estimated that known power and rotating speedP oc ,ω oc ?
Wind velocity vector under operating mode, as shown in Figure 2.
Step 2：Carry out wind estimation using Lampyridea Support vector regression forecast model：P oc (i)= V oc (i)• I oc (i)
At a time, it is utilized respectively voltage sensor, current sensor gathers in wind generator system through commutator
VD afterwardsV oc With output currentI oc , gather rotation speed of fan using speed probeω oc , and calculate this moment
OutputP oc = V oc • I oc ；WillP oc ,ω oc As Lampyridea Support vector regression forecast model input, model is defeated
Go out the estimated value for current wind velocity vector, and draw the estimated value of effective wind speed according to the following formulaV m ：
VD in the present embodiment, after the commutator that a certain moment collectsV oc For 15V, output currentI oc
For 12A, then output this moment is 180W, the rotary speed information that speed probe collectsω oc For 600r/min, incite somebody to action 180W,
600r/min } as Lampyridea Support vector regression forecast model input, output wind speed vector estimated value beV=
[8.92,9.14,8.94,9.02]T, then the estimated value of effective wind speedV m =9.0m/s.
Step 3：According to optimum tip-speed ratio method, draw wind speedV m Corresponding optimal rotation speed of fan predictive valueω ref ：
Wherein,For tip speed ratio；RFor impeller radius；Tip speed ratio be blade tip tangential velocity and wind wheel before wind speed it
Ratio, according to the formula of wind energy utilization, in different wind speed, has fixing tip speed ratio so that wind energy utilization is fixed on
Maximum, as shown in Figure 3.In embodiment, take=6.7, when effective wind speed is 9m/s,RDuring=0.6m, optimal blower fan can be drawn
Rotor speed forecast value is 961r/min.
Step 4：Passing ratio integral control method adjusts the rotating speed of blower fan, makes up to optimal rotation speed of fan predictive valueω ref ；And withω ref For initial value, the peak power of the disturbance step delta ω tracking blower fan adopting perturbation observation method to set；Real
Apply in example, take disturbance step delta ω=5r/min.
Fig. 4 is perturbation observation method（P&O）Flow chart, its principle be periodically disturbance blower fan rotating speed (ω+Δω),
Compare the changed power before and after its disturbance again, if output increases then it represents that perturbation direction is correct, continue in the same direction (+
Δω) disturbance；If output reduces, towards contrary (- Δω) direction disturbance.In embodiment, due to
The A/D module of dsPIC33FJ06GS101 completes once to sample and only needs to 0.5US, the VD of measurement blower fan, direct current
During output current, in order to eliminate the signal jitter that DC-DC converter medium-high frequency copped wave causes, the continuous conversion of A/D module is made to take for 10 times
Meansigma methodss are as measured value.
Step 5：Power difference before and after the disturbance that perturbation observation method is tried to achieve is more than or equal to given thresholdT r When, wind is described
Speed there occurs mutation, and repeat step 2 is to step 5；Otherwise, the disturbance step-length continuing to adopt perturbation observation method to set follows the tracks of blower fan
Peak power, as shown in Figure 5.In embodiment, takeT r =15W.
By above-mentioned maximal power tracing（MPPT）Method is passed throughCLISP program LISP writes control chip
DsPIC33FJ06GS101, output pwm signal drives DC-DC converter, you can realize maximal power tracing function.
It is observed on hardware platform of the present invention by the correctness of extracting method in order to verify with conventional disturbance
（P&O）Compare.It is specially：Under same wind speed environments（The frequency converter frequency of setting wind-tunnel is 40Hz）, control blower fan
Initial speed is ω0=600r/min, is respectively compared the tracking velocity of two methods and the mean power of steady-state process.
P＆O method is observed in conventional disturbance（Adjustment cycle T=5s, step delta ω=20r/min）Tracking process be：Through even
After the positive direction disturbance of continuous 17 bats, start in maximum power point MPP（ω opt =950r/min）Left and right vibration, that is, enter stable state mistake
Journey, required time about 85 seconds altogether.
The tracking process of the method for the invention is, is 180W first according to power this moment, rotating speed is 600r/min, leads to
Cross SVR model and estimate wind velocity vector and beV=[8.92,9.14,8.94,9.02]T, then the estimated value of effective wind speedV m =9.0m/
s.Maximum power point rotor speed forecast value ω is obtained according to optimum tip-speed ratio method ref After=961r/min, direct by PI controller
Working speed ω is adjusted to 961r/min, then proceeds by little step-length disturbance and observe P＆O（Adjustment cycle T=5s, step delta
ω=5r/min）, due to ω ref Closely maximum power point MPP itself（ω opt =950r/min）, negative through continuous 3 bats
After the disturbance of direction, enter steady-state process, required time is about 15 seconds altogether.It follows that the tracking of institute of the present invention extracting method
Speed will be apparently higher than conventional perturbation observation method.
After entering steady-state process, calculate the mean power of 60 seconds steady-state process respectively, show that conventional disturbance is observed（P&O）Side
Method is 209 watts, and the method for the invention is 222 watts, and this explanation can effectively reduce power using the method for the invention
Loss.
The above, the only present invention preferably specific embodiment, but protection scope of the present invention is not limited thereto,
Any those familiar with the art the invention discloses core technology in the range of, the change that can guess or replace
Change, all should be included within the scope of the present invention.
Claims (3)
1. a kind of minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method, is characterized in that, methods described is being adopted
On the basis of collecting the actual sample of a large amount of wind velocity vector-rotating speed-power, set up the Lampyridea Support vector regression prediction of wind speed
Model simultaneously carries out wind estimation using this model, is then predicted corresponding to maximum power point by optimum tip-speed ratio method
Excellent rotation speed of fan, then by the optimum rotation speed of fan of the rotational speed regulation of blower fan to prediction, and with this rotating speed as initial value, using disturbing
Dynamic observational method follows the tracks of the peak power of blower fan with the disturbance step-length setting；
Methods described is processed according to the following steps：
A. set up Lampyridea Support vector regression forecast model：
1. n air velocity transducer is respectively arranged on front side of draught fan impeller, the circular flat inswept with impeller be coaxial and size
Circle centre position in equal circular flat and at the center of circle 1/ (n-1) radius, at 2/ (n-1) radius ..., at 1 times of radius,
Record wind velocity vector V=[V1, V2..., Vn]T；
2. remember that the wind velocity vector under a certain wind speed environments is V (i)=[V1(i), V2(i) ..., Vn(i)]T, with this wind velocity vector pair
In the wind generator system answered, the DC voltage of rectifier output end is VocI (), DC current is IocI (), rotation speed of fan is
ωocI (), using the V recordingoc(i) and IocI () calculates output P this momentoc(i)=Voc(i)·IocI (), obtains one
Group sample (V (i), Poc(i), ωoc(i))；By collecting the sample group under different wind speed environments, form sample set { (V (i), Poc
(i), ωoc(i))}；
3. sample set { (V (i), P are utilizedoc(i), ωoc(i)) } training Lampyridea Support vector regression forecast model, its middle mold
Type inputs as { Poc(i), ωoc(i) }, model is output as V (i)；
B. carry out wind estimation using Lampyridea Support vector regression forecast model：
At a time, be utilized respectively voltage sensor, current sensor gather wind generator system in commutator output straight
Stream voltage VocWith DC current Ioc, gather rotation speed of fan ω using speed probeoc, and calculate output P this momentoc
=Voc·Ioc；By { ωoc, PocAs Lampyridea Support vector regression forecast model input, model is output as current wind speed
The estimated value of vectorAnd calculate estimated value V of effective wind speed according to the following formulam：
C. utilize optimum tip-speed ratio method, calculate estimated value V of effective wind speedmCorresponding optimal rotation speed of fan predictive value
ωref：
Wherein, λ is tip speed ratio；R is impeller radius；
D. passing ratio integral control method adjusts the rotating speed of blower fan, makes up to optimal rotation speed of fan predictive value ωref；Then
With ωrefFor initial value, the peak power of the disturbance step delta ω tracking blower fan adopting perturbation observation method to set；
If the power difference before and after the disturbance that e. perturbation observation method is tried to achieve is more than or equal to given threshold Tr, illustrate that wind speed there occurs prominent
Become, repeat step b is to step e；Otherwise, the disturbance step-length that continuation adopts perturbation observation method to set follows the tracks of the maximum work of blower fan
Rate.
2. minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method according to claim 1, its feature
It is that described Lampyridea Support vector regression forecast model is the support vector machine carrying out parameter optimization using glowworm swarm algorithm
Model, specific optimization step is：
I. the parameter of initialization glowworm swarm algorithm：Initialization Lampyridea number n, fluorescein volatility coefficient ρ, fluorescein enhancer
γ, sensing range rs(i.e. penalty factor to be optimized and nuclear parameter δ in Support vector regression forecast model2Excursion),
Neighborhood rate of change β, neighbours' threshold value nt, moving step length s, initial fluorescence element l0With maximum iteration time；
II. penalty factor to be optimized and nuclear parameter δ in Support vector regression forecast model2Span in, give birth at random
Become one group of numerical value as the individual initial position of Lampyridea, fitness function is (a certain to punish using Lampyridea present position
Penalty factor C and nuclear parameter δ2) the Support vector regression forecast model the trained average relative error on test samples
Negative value, average relative error ΔMREExpression formula is：
Wherein, N1For test samples number, ViFor the true value of wind velocity vector,Wind velocity vector estimated value for regression model output；
III. calculate the fluorescein concentration of the fluorescein concentration, decision domain scope and neighbours of every Lampyridea, by the fluorescence of neighbours
Plain concentration determines the moving direction of each Lampyridea and moves forward；
IV. judge whether glowworm swarm algorithm reaches end condition, if reaching, choosing fitness highest Lampyridea position is to prop up
Hold the parameter of vector machine model, and record the supporting vector in corresponding regression model in case using；Otherwise, go to step III.
3. minitype permanent magnetism directly-driving wind power generation system maximum power tracking and controlling method according to claim 2, its feature
It is that the end condition of described glowworm swarm algorithm is average relative error on test samples for the Support vector regression forecast model
ΔMRE≤ 3% or iterationses exceed maximum iteration time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410622989.XA CN104454346B (en) | 2014-11-09 | 2014-11-09 | Maximum power tracking control method for small permanent-magnet direct-drive wind power generation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410622989.XA CN104454346B (en) | 2014-11-09 | 2014-11-09 | Maximum power tracking control method for small permanent-magnet direct-drive wind power generation system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104454346A CN104454346A (en) | 2015-03-25 |
| CN104454346B true CN104454346B (en) | 2017-02-15 |
Family
ID=52900963
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410622989.XA Expired - Fee Related CN104454346B (en) | 2014-11-09 | 2014-11-09 | Maximum power tracking control method for small permanent-magnet direct-drive wind power generation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104454346B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104950982B (en) * | 2015-06-25 | 2017-06-06 | 华南理工大学 | Dual feedback wind power generation system maximum power tracking and controlling method based on many step-lengths |
| CN106194582B (en) * | 2016-09-19 | 2018-09-04 | 华能新能源股份有限公司辽宁分公司 | Wind power system MPPT control device and methods based on measuring wind speed and estimation |
| CN106837704A (en) * | 2017-04-17 | 2017-06-13 | 北京耀能科技有限公司 | A kind of medium-sized low-speed permanent magnetic direct-drive Wind turbines and its Parameter Self-learning control method |
| CN107524563A (en) * | 2017-08-21 | 2017-12-29 | 华南理工大学 | A kind of control method based on slip form extremum search |
| CN108336822A (en) * | 2018-01-30 | 2018-07-27 | 深圳众厉电力科技有限公司 | Transmission line of electricity monitoring and warning system applied to intelligent grid |
| CN108506163B (en) * | 2018-04-25 | 2024-01-30 | 华北电力科学研究院有限责任公司 | Doubly-fed wind power virtual synchronous machine rotating speed recovery method, device and system |
| CN108983863B (en) * | 2018-08-30 | 2019-08-06 | 同济大学 | A kind of photovoltaic maximum power tracking method based on improvement glowworm swarm algorithm |
| CN109667728B (en) * | 2018-12-21 | 2020-09-08 | 北京金风科创风电设备有限公司 | Fault detection method and device for wind generating set rotating speed sensor |
| CN110361963B (en) * | 2019-06-10 | 2022-04-01 | 岭南师范学院 | Method and device for optimizing PI (proportional integral) parameters of permanent magnet fan |
| CN110376895B (en) * | 2019-07-30 | 2022-03-04 | 华能国际电力股份有限公司营口电厂 | Thermal power generating unit coordination control method based on hierarchical limited predictive control |
| CN110985289B (en) * | 2019-12-04 | 2021-01-01 | 浙江大学 | SVR and SMC based MPPT method with preset performance for wind turbine generator |
| CN114510111B (en) * | 2021-12-29 | 2023-09-12 | 北京华能新锐控制技术有限公司 | Global MPPT control method and device for partial shading photovoltaic array |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009091923A (en) * | 2007-10-04 | 2009-04-30 | Univ Of Ryukyus | Maximum power point follow-up control device for wind power generation device |
| CN101793235A (en) * | 2010-04-15 | 2010-08-04 | 哈尔滨工业大学 | Maximum power tracking type wind power generation device with energy predicting function and method thereof |
| CN102291050A (en) * | 2011-08-17 | 2011-12-21 | 华北电力大学（保定） | Maximum power point tracking method and device for photovoltaic power generation system |
| CN102338808A (en) * | 2011-08-26 | 2012-02-01 | 天津理工大学 | Online hybrid forecasting method for short-term wind speed of wind power field |
| CN102664426A (en) * | 2012-05-10 | 2012-09-12 | 华北电力大学 | Anti-normalization interval correction method for improving air speed prediction precision of SVM (Support Vector Machine) |
| CN103437955A (en) * | 2013-08-13 | 2013-12-11 | 华北电力大学（保定） | Maximum power tracking device for mini permanent magnetic direct drive wind power generation system and control method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2358140B1 (en) * | 2009-10-23 | 2012-04-12 | Gamesa Innovation & Technology S.L | METHODS OF AIRCRAFT CONTROL TO IMPROVE ENERGY PRODUCTION. |
-
2014
- 2014-11-09 CN CN201410622989.XA patent/CN104454346B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009091923A (en) * | 2007-10-04 | 2009-04-30 | Univ Of Ryukyus | Maximum power point follow-up control device for wind power generation device |
| CN101793235A (en) * | 2010-04-15 | 2010-08-04 | 哈尔滨工业大学 | Maximum power tracking type wind power generation device with energy predicting function and method thereof |
| CN102291050A (en) * | 2011-08-17 | 2011-12-21 | 华北电力大学（保定） | Maximum power point tracking method and device for photovoltaic power generation system |
| CN102338808A (en) * | 2011-08-26 | 2012-02-01 | 天津理工大学 | Online hybrid forecasting method for short-term wind speed of wind power field |
| CN102664426A (en) * | 2012-05-10 | 2012-09-12 | 华北电力大学 | Anti-normalization interval correction method for improving air speed prediction precision of SVM (Support Vector Machine) |
| CN103437955A (en) * | 2013-08-13 | 2013-12-11 | 华北电力大学（保定） | Maximum power tracking device for mini permanent magnetic direct drive wind power generation system and control method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104454346A (en) | 2015-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104454346B (en) | Maximum power tracking control method for small permanent-magnet direct-drive wind power generation system | |
| CN104806450B (en) | A kind of wind power system MPPT control method based on gravitation neutral net | |
| CN102291050B (en) | Maximum power point tracking method and device for photovoltaic power generation system | |
| CN103437955B (en) | Minitype permanent magnetism directly-driving wind power generation system maximum power tracking device and controlling method | |
| CN106194582B (en) | Wind power system MPPT control device and methods based on measuring wind speed and estimation | |
| CN106452355B (en) | A kind of photovoltaic generating system maximum power tracking method based on Model Distinguish | |
| CN102419394B (en) | Wind/solar power prediction method with variable prediction resolution | |
| CN103885521A (en) | Photovoltaic array MPPT method based on cuckoo search algorithm | |
| CN105590144A (en) | Wind speed prediction method and apparatus based on NARX neural network | |
| CN102938562B (en) | Prediction method of total wind electricity power in area | |
| CN105673322A (en) | Variable parameter nonlinear feedback control method achieving wind turbine MPPT control | |
| Zhu et al. | Research and test of power-loop-based dynamic multi-peak MPPT algorithm | |
| CN108979957B (en) | Obtain the non-linear predication control method of Variable Speed Wind Power Generator maximal wind-energy | |
| CN106786669B (en) | A kind of active power of wind power field change rate control method and system | |
| CN107045574A (en) | The low wind speed section effective wind speed method of estimation of wind power generating set based on SVR | |
| CN108549962A (en) | Wind power forecasting method based on history fragment sequence search and sequential rarefaction | |
| CN105024599A (en) | Wave energy power generation system maximum power tracking device and control method | |
| CN103293950B (en) | The control method that a kind of maximum photovoltaic power point based on LSSVM is followed the tracks of | |
| CN107633368A (en) | Wind power generating set output performance estimating method and device | |
| CN108536212A (en) | A kind of novel variable step photovoltaic maximum power tracking method based on power prediction | |
| Pan et al. | Research on variable pitch control strategy of direct-driven offshore wind turbine using KELM wind speed soft sensor | |
| CN106485603A (en) | A kind of short-term wind speed forecasting method being applied to wind-power electricity generation | |
| CN206386223U (en) | Wind power system MPPT control devices based on measuring wind speed with estimation | |
| CN103995559A (en) | Constant voltage MPPT control method and system based on environmental parameter model | |
| CN108418252A (en) | A kind of the mixed tensor collection device and operation method of wireless sensor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| CB03 | Change of inventor or designer information |
Inventor after: Ma Liangyu Inventor after: Yu Ping Inventor after: Wang Ting Inventor after: Song Shengnan Inventor after: Liu Weiliang Inventor after: Liu Changliang Inventor after: Lin Yongjun Inventor before: Ma Liangyu Inventor before: Song Shengnan Inventor before: Chen Wenying Inventor before: Liu Weiliang Inventor before: Lin Yongjun Inventor before: Liu Changliang Inventor before: Ma Jin Inventor before: Ma Yongguang |
|
| COR | Change of bibliographic data | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20170112 Address after: 100094 Beijing City, North Road, No., building 01310, room 1, room 68 Applicant after: ZHONGKE INNOVATION (BEIJING) TECHNOLOGY Co.,Ltd. Applicant after: NORTH CHINA ELECTRIC POWER University (BAODING) Address before: 071003 Hebei province Baoding Yonghua No. 619 North Street Applicant before: NORTH CHINA ELECTRIC POWER UNIVERSITY (BAODING) |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170215 Termination date: 20211109 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |