CN110492803A - Permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO - Google Patents
Permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO Download PDFInfo
- Publication number
- CN110492803A CN110492803A CN201910493539.8A CN201910493539A CN110492803A CN 110492803 A CN110492803 A CN 110492803A CN 201910493539 A CN201910493539 A CN 201910493539A CN 110492803 A CN110492803 A CN 110492803A
- Authority
- CN
- China
- Prior art keywords
- parameter
- axis
- follows
- formula
- permanent magnet
- 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
Links
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
- H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
- H02P21/0003—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
- H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
- H02P21/14—Estimation or adaptation of machine parameters, e.g. flux, current or voltage
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
- H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
- H02P21/22—Current control, e.g. using a current control loop
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Control Of Eletrric Generators (AREA)
Abstract
The present invention relates to a kind of permanent-magnetic wind driven generator parameter identification technique fields, it is a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO, include the following steps S1: the required parameter of sampling identification wind-driven generator establishes permanent magnet direct-driving aerogenerator stator voltage current model；S2: establishing and obtains permanent magnet direct-drive wind-force power generation identification model: S3: establishing error target function J (θ)；S4: error target function J (θ) is solved using MDPSO algorithm optimization, obtains aq,bq,cq,dqOptimal solution；S5: wind-driven generator the parameter R, L to be recognized are calculatedd,Lq,ψ；S6: according to MrminCriterion adjusts the PI dynamic state of parameters of generator-side converter wear current inner loop control, obtains new PI parameter value.The present invention proposes the improvement particle swarm algorithm MDPSO of adaptive space locating vector and average optimal location variable, improved MDPSO algorithm has stronger robustness, higher convergence precision and faster convergence rate, and overcomes the problem of general particle swarm algorithm is easily trapped into locally optimal solution.
Description
Technical field
It is that a kind of permanent magnetism based on MDPSO is straight the present invention relates to a kind of permanent-magnetic wind driven generator parameter identification technique field
Driving type aerogenerator parameter identification method.
Background technique
Now, there are 2 kinds by widely applied speed-variable frequency-constant wind-driven generator group: double-fed induction wind driven generator group
(DFIG) and permanent magnet direct-driving aerogenerator group (D-PMSG).It compares for DFIG, the structure of D-PMSG is simple, Generation Rate
The feature high, noise is small and maintenance workload is small, receives more and more attention in wind power generation field.In general,
PMSG generator-side converter wear vector control system is usually revolving speed outer ring, (usually inside and outside ring controller is mainly PI control to current inner loop
Device processed), and the superiority and inferiority of the also controlled device parameter value of vector controlled performance influences.The PI parameter value of controller is generally according to generator
The intrinsic parameter tuning such as resistance, inductance, magnetic linkage obtain, while temperature, magnetic flux saturation, skin effect etc. also influence generator
Intrinsic parameter.In operational process, the intrinsic parameter value of wind-driven generator is variation, and the PI parameter of current transformer control is not moved therewith
State adjusting, will affect the control effect of wind-driven generator.Therefore, it is necessary to which the intrinsic parameter value of accurate recognition generator, improves control
Performance processed improves the wind energy utilization of PMSG.
Traditional discrimination method mainly has: Extended Kalman filter method (EKF), model reference adaptive identification method
(MARS), least square method of recursion (RLS) etc..Permanent magnet direct-driving aerogenerator control system has nonlinear time-varying feature, quasi-
Really the multiparameter problem of identification wind-driven generator is extremely difficult, when identification algorithm need to weigh such as complexity, convergence and calculate
Between etc. factors, often can not find the optimal value of system using traditional discrimination method.With the rapid development of computer technology and wide
General application, researcher recognize the multi-parameter that intelligent optimization algorithm is applied to motor.These algorithms specifically include that the shoal of fish is calculated
Method, genetic algorithm, particle swarm algorithm etc..Genetic algorithm compares other intelligent algorithms, requires height to identification initial value, computation-intensive
And it is time-consuming；Fish-swarm algorithm and particle swarm algorithm are easily ensnared into locally optimal solution although relatively easy.In view of particle swarm algorithm
The problem of being easily trapped into locally optimal solution makes particle not determine track when initializing, after improvement if introducing average desired positions
Algorithm not only increase convergence rate and also enhance ability of searching optimum, but convergence precision is not significantly improved；If
The parameters such as the resistance of PMSM and torque are recognized using conventional particle group algorithm, identification precision is higher, but the algorithm can not
Magnetic linkage and inductance are recognized, is haveed the defects that certain；If having received artificial immune system mechanism and complete learning-oriented particle swarm algorithm
Advantage, design a kind of complete learning-oriented population Evolutionary Computation Model frame based on immune mechanism, which has relatively strong
Convergence and complicated diversity, but calculating process is relative complex, and convergence rate is slow；If introducing adaptive space search
Vector, to help particle to flee from local optimum problem, but convergence rate is slower.
Summary of the invention
The present invention provides a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO, overcomes above-mentioned
The deficiency of the prior art can effectively solve the difficult problem of permanent magnet direct-driving aerogenerator multi-parameter identification, further solve
That there are convergence precisions is poor for traditional parameter identification method, convergence rate is slow, influences the control effect of wind-driven generator, so that forever
The low problem of the wind energy utilization of magnetic direct wind-driven generator group.
The technical scheme is that realized by following measures: a kind of permanent magnet direct-drive wind-force hair based on MDPSO
Parameter of electric machine discrimination method, comprising the following steps:
S1: the required parameter of sampling identification wind-driven generator chooses motor outlet end and measures motor speed and motor output
Three-phase current, use the direct-axis voltage u of current controllerdWith quadrature-axis voltage uq, establish permanent magnet direct-driving aerogenerator stator electricity
Piezoelectricity flow model:
Wherein: ud、uqIt is the d axis and q axis component of stator voltage respectively；id、iqIt is the d axis and q axis point of stator current respectively
Amount；R is the resistance of stator；ωrIt is angular rate；Ld、LqIt is d axis and q axle inductance component；ψ represents permanent magnet flux linkage；P=d/
dt；
S2: carrying out pade to (1) formula and approach and carry out discretization, establishes and obtains permanent magnet direct-drive wind-force power generation identification mould
Type:
In formula: ad、bd、cd、aq、bq、cq、dqRespectively model coefficient, k are discrete point；
A can be obtained according to formula (1)d、bd、cd、aq、bq、cq、dqIt is as follows:
Wherein, TsFor the sampling period；
S3: according to the d axis of stator current and q axis component isd、isqWith identification modelTo obtain identifier and true
The error minimum value of real value establishes error target function J (θ), and formula expression is as follows:
S4: it is solved and is missed using the MDPSO algorithm optimization based on adaptive space locating vector and average optimal location variable
Poor objective function J (θ), obtains aq,bq,cq,dqOptimal solution；
S5: according to a found outq,bq,cq,dqOptimal solution, calculate wind-driven generator the parameter R, L to be recognizedd,Lq,ψ
Are as follows:
Wherein, TsFor the sampling period, R is the resistance of stator, aq,bq,cq,dqThe respectively coefficient of model, Ld、LqIt is d axis
With q axle inductance component；ψ is permanent magnet flux linkage；
S6: the parameter R, L after identification are utilizedd,Lq, ψ, according to MrmiCriterion joins the PI of generator-side converter wear current inner loop control
Number dynamic tuning, obtains new PI parameter value.
Here is the further optimization and/or improvements to invention technology described above scheme:
It is above-mentioned in S1, by direct-axis voltage usdWith quadrature-axis voltage usqDecouple and resolve into two components: one-component is
The u ' exported by PI current controllersdWith u 'sq, expression formula is as follows:
Another component is the Δ u that decoupling obtainssdWith Δ usq, expression formula is as follows:
In formula: isdref、isqrefFor d axis, q shaft current reference value；Kp、KiThe respectively parameter of PI controller；S is multifrequency
Rate.
It is above-mentioned in S4, the calculating of the MDPSO algorithm based on adaptive space locating vector and average optimal location variable
Process is as follows:
S41: the adaptive space search radius R (t) of the DPSO algorithm based on adaptive space locating vector, expression are calculated
Formula are as follows:
Wherein, λ is auto-adaptive parameter (usual λ >=2), and t is the number of iterations, and u is the uniform random number in [0,1] range,It is maximum value, the minimum value of particle position respectively；
S42: adaptive space locating vector (R (t)-x is appliedid(t)) rate equation of dynamic corrections more new particle, obtains
It arrives:
In formula: w, r1、r2、r3For inertia weight, wherein r1、r2Value be (0,1) section random number, r3Be [0,
1] uniform random number in range, c1、c2、c3It is nonnegative constant；VidFor the current flight speed of i-th of particle；pidIt is i-th
The optimal location of particle；xidFor the current location of i-th of particle；gdThe optimal location found by particles all in population；
S43: introducing mean place Pmd in DPSO algorithm, and mean place Pmd is that D ties up m particle individual desired positions
The expression formula of average value, mean place Pmd is as follows:
Then:
It is above-mentioned in S6, the PI parameter of dynamic corrections controller be achieve the purpose that raising system stationarity selection system ginseng
Number, using MrminCriterion keeps its closed loop amplitude-frequency characteristic peak value minimum, and makeover process is as follows:
S61: it is controlled using PI, each PI parameter is set as:
Then as shown in Fig. 5-a, open-loop transfer function may be expressed as: d shaft current inner loop control
In formula: KpwmIt is the pulsewidth modulation gain of SVPWM, Ts is the sampling time；
S62:d axis open-loop transfer function Gid(S) having a pole is-R/Ld, if Ki/Kp=R/Ld, with PI controller
Zero point goes the pole of compensation open-loop transfer function, then after zero pole point compensation that d axis open-loop transfer function is equivalent are as follows:
Wherein, optimum damping ratio is 0.707, current inner loop d axis PI controller parameter are as follows:
S63: according to the calculation method of S62, current inner loop q axis PI controller parameter is similarly calculated are as follows:
S64: speed control outer ring transmission function indicates are as follows:
In formula: J is the rotary inertia of generator, and p is the number of pole-pairs of generator；
S65: according to MrminCriterion design keeps its closed loop amplitude-frequency characteristic peak value minimum, expression formula are as follows:
H=5 is enabled, is obtained:
According to above-mentioned formula, it is concluded that, either the PI parameter of current inner loop or revolving speed outer ring is all consolidated with generator
There is relating to parameters.
The present invention proposes the improvement particle swarm algorithm MDPSO of adaptive space locating vector and average optimal location variable,
Improved MDPSO algorithm has stronger robustness, higher convergence precision and faster convergence rate, and overcomes general
Particle swarm algorithm is easily trapped into the problem of locally optimal solution；Using MDPSO algorithm to permanent magnet direct-driving aerogenerator stator winding
Resistance, d axle inductance, q axle inductance and magnetic linkage recognized；Using the wind-driven generator parameter after identification to generator-side converter wear PI
Control parameter carries out dynamic tuning, using revised PI control parameter, can be improved permanent magnet direct-drive wind turbine group wind energy utilization,
Increase active power output, the output of smooth active power.
Detailed description of the invention
Attached drawing 1 is the permanent magnet direct-driving aerogenerator group structure chart of the embodiment of the present invention 1.
Attached drawing 2 is PMSG parameter identification schematic diagram under the vector controlled of the embodiment of the present invention 1.
Attached drawing 3 is the schematic diagram of the discovery mechanism based on adaptive space locating vector DPSO of the embodiment of the present invention 1.
Attached drawing 4 is the embodiment of the present invention 1 based on MDPSO algorithm parameter discrimination method figure.
Attached drawing 5-a is the d shaft current inner loop control block diagram of the embodiment of the present invention 1.
Attached drawing 5-b is the q shaft current inner ring and revolving speed outer loop control block diagram of the embodiment of the present invention 1.
Attached drawing 6 is the target function value contrast schematic diagram of the embodiment of the present invention 2.
Attached drawing 7 is the stator resistance value contrast schematic diagram of the embodiment of the present invention 2.
Attached drawing 8 is the d axle inductance value contrast schematic diagram of the embodiment of the present invention 2.
Attached drawing 9 is the q axle inductance value contrast schematic diagram of the embodiment of the present invention 2.
Attached drawing 10 is the permanent magnet flux linkage contrast schematic diagram of the embodiment of the present invention 2.
The active power that attached drawing 11 exports for Wind turbines under constant load before and after the PI parameters revision of the embodiment of the present invention 2
Contrast schematic diagram.
Specific embodiment
The present invention is not limited by the following examples, can determine according to the technique and scheme of the present invention with actual conditions specific
Embodiment.
Below with reference to examples and drawings, the invention will be further described:
Embodiment 1:, should the permanent magnet direct-driving aerogenerator parameter identification side based on MDPSO as shown in attached drawing 1,2,3,4
Method, comprising the following steps:
S1: the required parameter of sampling identification wind-driven generator chooses motor outlet end and measures motor speed and motor output
Three-phase current, use the direct-axis voltage u of current controllerdWith quadrature-axis voltage uq, establish permanent magnet direct-driving aerogenerator stator electricity
Piezoelectricity flow model:
Wherein: ud、uqIt is the d axis and q axis component of stator voltage respectively；id、iqIt is the d axis and q axis point of stator current respectively
Amount；R is the resistance of stator；ωrIt is angular rate；Ld、LqIt is d axis and q axle inductance component；ψ represents permanent magnet flux linkage；P=d/
dt；
S2: carrying out pade to (1) formula and approach and carry out discretization, establishes and obtains permanent magnet direct-drive wind-force power generation identification mould
Type:
In formula: ad、bd、cd、aq、bq、cq、dqRespectively model coefficient, k are discrete point；
A can be obtained according to formula (1)d、bd、cd、aq、bq、cq、dqIt is as follows:
Wherein, TsFor the sampling period；
S3: according to the d axis of stator current and q axis component isd、isqWith identification modelTo obtain identifier and true
The error minimum value of real value establishes error target function J (θ), and formula expression is as follows:
S4: it is solved and is missed using the MDPSO algorithm optimization based on adaptive space locating vector and average optimal location variable
Poor objective function J (θ), obtains aq,bq,cq,dqOptimal solution；
S5: a found out according to S2q,bq,cq,dqValue, calculates wind-driven generator the parameter R, L to be recognizedd,Lq, ψ are as follows:
Wherein, TsFor the sampling period, R is the resistance of stator, aq,bq,cq,dqThe respectively coefficient of model, Ld、LqIt is d axis
With q axle inductance component；ψ is permanent magnet flux linkage；
S6: the parameter R, L after identification are utilizedd,Lq, ψ, according to MrminPI of the criterion to generator-side converter wear current inner loop control
Dynamic state of parameters adjusting, obtains new PI parameter value.
The above-mentioned permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO can be made into one according to actual needs
Step optimization or/and improvement:
As shown in attached drawing 1,2, in S1, by direct-axis voltage usdWith quadrature-axis voltage usqIt carries out decoupling and resolves into two components:
One-component is the u ' exported by PI current controllersdWith u 'sq, expression formula is as follows:
Another component is the Δ u that decoupling obtainssdWith Δ usq, expression formula is as follows:
In formula: isdref、isqrefFor d axis, q shaft current reference value；Kp、KiThe respectively parameter of PI controller；S is multifrequency
Rate.
The above-mentioned d axis because of generator unit stator, there is coupled relations in q shaft-like state variable, in order to design independent d axis, q axis
Electric current isdAnd isqController, it is necessary to d axis, q shaft current isdAnd isqCarry out decoupling control；
As shown in attached drawing 3,4, in S4, the MDPSO based on adaptive space locating vector and average optimal location variable
Algorithm calculating process is as follows:
S41: the adaptive space search radius R (t) of the DPSO algorithm based on adaptive space locating vector, expression are calculated
Formula are as follows:
Wherein, λ is auto-adaptive parameter (usual λ >=2), and t is the number of iterations, and u is the uniform random number in [0,1] range,It is maximum value, the minimum value of particle position respectively；
In order to improve the shortcomings that PSO is easily trapped into local optimum, present invention introduces adaptive space locating vectors, make particle
Initial trace can adaptively be changed, explore new search space, expand ability of searching optimum, obtain globally optimal solution, change
The convergence precision of kind PSO.
S42: adaptive space locating vector (R (t)-x is appliedid(t)) rate equation of dynamic corrections more new particle, obtains
It arrives:
In formula: w, r1、r2、r3For inertia weight, r1、r2Value be (0,1) section random number, r3It is [0,1] range
Interior uniform random number, c1、c2、c3It is nonnegative constant；VidFor the current flight speed of i-th of particle；pidFor i-th of particle
Optimal location；xidFor the current location of i-th of particle；gdThe optimal location found by particles all in population.
Discovery mechanism based on adaptive space locating vector DPSO is as shown in figure 3, unmodified PSO algorithm explores speed
Radius is in vector spaceRegion, improve DPSO algorithm rate equation increase space exploration vector (R (t)-xid
(t)) each particle, can be made to explore radius in velocity vector space to existExtremelyThe region of range.R (t) expression formula (9) can
Know, as the value of u is different, R (t) value is different, searches for Long-term change trend, and when R (t) is larger, particle is explored radius and increased, induction
Particle leaves their current region and pushes particle Xiang Geng great range searching optimal solution；When R (t) is smaller, before utilizing
The optimal solution of secondary iteration particle determines that current iteration particle explores range；No matter how R (t) value, which changes, can all improve convergence essence
Degree.Obviously, by the speed update equation of introducing adaptive space locating vector, solution space as big as possible can be explored, from
And acquire globally optimal solution.
Although improved DPSO algorithm improves convergence precision, since it expands search space range, make to iterate to calculate
Amount increases, and algorithm the convergence speed is slow.
S43: introducing mean place Pmd in DPSO algorithm, and mean place Pmd is that D ties up m particle individual desired positions
The expression formula of average value, mean place Pmd is as follows:
Then:
The present invention is improved on the basis of basic particle group algorithm (PSO), particle swarm optimization algorithm (Particle
Swarm Optimization, PSO) it is proposed in nineteen ninety-five by doctor Kennedy and doctor Eberhart, it is a kind of based on group
Evolution algorithm.PSO algorithm by the solution of each optimization problem as " particle " in search space, and each grain
Son flies at a given speed in D dimension search space, it is assumed that the current location X of particle i (i=1,2 ..., N)i=(xi1,
xi1,…,xid), current flying speed Vi=(vi1,vi2,…,vid), pidFor the current optimal location of particle i, gdFor in population
The optimal location that all particles are found.Its speed, which updates iterative equation, to be indicated are as follows:
vid(t+1)=wvid(t)+c1r1[pid-xid(t)] +c2r2[gd-xid(t)] (7)
Location updating iterative equation are as follows:
xid(t+1)=xid(t)+vid(t+1) (8)
Wherein, w, r1、r2For inertia weight, r1、r2Value be (0,1) section random number, c1、c2It is nonnegative constant.
During initialization because of particle, speed and position are a random numbers to basic particle group algorithm, in iteration
Its position is easier to fall into Local Extremum during update, therefore algorithm is easily trapped into locally optimal solution.
It is above-mentioned particle rapidity is updated to formula (12) by formula (10) after, particle obtain information content it is more, average bit
Set pmdNot only with reference to the posterior infromation of other particles, and further include itself optimum position information, can use particle
More information carry out decision factum.The signal as used in improved MDPSO algorithm is directly to detect
State variable, therefore reduce other disturbing factors, to also improve the accuracy and rapidity of identification；Therefore after improving
MDPSO algorithm improve convergence rate and optimizing ability, solve the problems, such as that PSO algorithm is easily trapped into locally optimal solution.
The generator-side converter wear control block diagram as shown in attached drawing 5-a, 5-b, in S6, the PI parameter of dynamic corrections controller
It is to achieve the purpose that raising system stationarity selects system parameter, keeps its closed loop amplitude-frequency characteristic peak value minimum using Mrmin criterion,
Makeover process is as follows:
S61: it is controlled using PI, each PI parameter is set as:
Then as shown in Fig. 5-a, open-loop transfer function may be expressed as: d shaft current inner loop control
In formula: KpwmIt is the pulsewidth modulation gain of SVPWM, Ts is the sampling time；
S62:d axis open-loop transfer function Gid(S) having a pole is-R/Ld, Ki/Kp=R/Ld is arranged, with PI controller
Zero point go compensation open-loop transfer function pole, then zero pole point compensation after d axis open-loop transfer function is equivalent are as follows:
Wherein, optimum damping ratio is 0.707, current inner loop d axis PI controller parameter are as follows:
In above-mentioned S62, in order to obtain the optimal tracking performance of current inner loop, therefore Ki/Kp=R/Ld is set.
S63: according to the calculation method of S62, current inner loop q axis PI controller parameter is similarly calculated are as follows:
Revolving speed controls outer ring block diagram as shown in Fig. 5-b, wherein
S64: speed control outer ring transmission function indicates are as follows:
In formula: J is the rotary inertia of generator, and p is the number of pole-pairs of generator；
S65: according to MrminCriterion design keeps its closed loop amplitude-frequency characteristic peak value minimum, expression formula are as follows:
H=5 is enabled, is obtained:
Can be obtained according to above-mentioned formula, either the PI parameter of current inner loop or revolving speed outer ring all with the intrinsic ginseng of generator
Number is related.
Embodiment 2: as shown in attached drawing 6 to 11 and as shown in table 1,2, the present embodiment is that Simulink is flat in MATLAB
Permanent magnet direct-driving aerogenerator group is built under platform, the setting emulation sampling time is Ts=0.00001s, SVPWM modulating frequency are set as
The parameter of 20KHz, magneto alternator model are as shown in table 1.
The case where it is 0 that permanent magnet direct-driving aerogenerator, which is in d shaft current, constant torque 2Nm, revolving speed is 500r/min
Under, obtain 4000 groups of data, sample frequency 20kHz.Particle group parameters setting are as follows: population dimension is 7, population number 80,
The number of iterations is 30, c1, c2, and c3 is set as 1.59.
The number of iterations for the fitness function value that the present embodiment takes tri- kinds of methods of PSO, DPSO, MDPSO to obtain respectively is such as
Shown in Fig. 6, from convergence rate angle analysis, the number of iterations can be seen that three kinds of convergence speed of the algorithm differences, and MDPSO is repeatedly
It is restrained after generation 9 times, DPSO restrains after iteration 13 times, and PSO restrains after iteration 22 times；From convergence precision angle point
Analysis, it be 0.0001, PSO convergence precision is 0.001 that the convergence precision of MDPSO, which is in the convergence precision of 0.00001, DPSO,.It can be seen that
MDPSO algorithm is better than DPSO and PSO in convergence precision and convergence rate, and MDPSO has jumped out locally optimal solution, avoided
" precocity " phenomenon.
It is stator resistance value, d axle inductance, q axle inductance and permanent magnet flux linkage parameter value respectively as shown in attached drawing 7 to 10
The curve graph changed with the number of iterations.Stator resistance identification initial value is 48 Ω, and by 9 iteration, parameter identification value drops to
0.9587 Ω, the relative error with rated value are 0.07%；It is 10.8H that d axle inductance, which recognizes initial value, by 8 iteration, parameter
Identifier drops to 0.0053H, and the relative error with rated value is 0.9%；It is 20.6H that q axle inductance, which recognizes initial value, by q
Secondary iteration, parameter identification value drop to 0.01225, and the relative error with rated value is 2%；Permanent magnet flux linkage value recognizes initial value
For 0.72Wb, by 10 iteration, parameter identification value drops to 0.1865Wb, and the relative error with rated value is 2.08%.It is comprehensive
It is upper described, each parameter value in identification process, initial value all far from rated value, after iteration 10 times identifier decline
To near rated value, error meets control and requires, identification result is as shown in table 2 less than 2.1%.
The stator resistance of direct-drive permanent magnet wind power generator, dq axle inductance value and permanent magnet flux linkage are obtained by parameter identification
Value.Utilize MrminCriterion carries out dynamic tuning to generator-side converter wear PI control parameter using the wind-driven generator parameter after identification,
Obtain revised PI parameter value, Kpd=0.265, Kid=71.28；Kpq=1.4875, Kiq=71.28, Kwp=4.165Kwi=
833.The PI parameter of wind-driven generator generator-side converter wear current inner loop control uses revised PI parameter, in constant load condition
Under, the output of the active power of directly driven wind-powered unit as shown in figure 11, and does not correct Wind turbines generator-side converter wear PI control parameter
When active power for wind power export comparison, discovery:
1) after parameter identification, using the wind-driven generator parameter after identification, amendment generator-side converter wear PI control is joined again online
Number, so that the active power fluctuation range of directly driven wind-powered unit is smaller, power output is more smooth；
2) generator-side converter wear PI control parameter is corrected again online using the wind-driven generator parameter after identification, so that straight drive
Formula Wind turbines wind energy utilization efficiency increases.
The above technical characteristic constitutes highly preferred embodiment of the present invention, with stronger adaptability and best implementation effect
Fruit can increase and decrease non-essential technical characteristic, according to actual needs to meet the needs of different situations.
Table 1PMSG simulation parameter
| Parameter | Value |
| Stator resistance R/ Ω | 0.958 |
| D axle inductance/mH | 5.25 |
| Q axle inductance/mH | 12 |
| Permanent magnet flux linkage/Wb | 0.1827 |
| Rotary inertia/(Kg*m2) | 0.003 |
| Number of pole-pairs/P | 4 |
2 direct wind-driven generator parameter identification result of table
| Title | R/Ω | Ld/mH | Lq/mH | Ψ/Wb |
| True value | 0.958 | 5.25 | 12 | 0.1827 |
| Identifier | 0.9587 | 5.3 | 12.25 | 0.1865 |
| Relative error | 0.07% | 0.95% | 2.08% | 2.08% |
Claims (5)
1. a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO, it is characterised in that: the following steps are included:
S1: the required parameter of sampling identification wind-driven generator chooses motor outlet end measures motor speed and motor output three
Phase current uses the direct-axis voltage u of current controllerdWith quadrature-axis voltage uq, establish permanent magnet direct-driving aerogenerator stator voltage electricity
Flow model:
Wherein: ud、uqIt is the d axis and q axis component of stator voltage respectively；id、iqIt is the d axis and q axis component of stator current respectively；R
It is the resistance of stator；ωrIt is angular rate；Ld、LqIt is d axis and q axle inductance component；ψ represents permanent magnet flux linkage；P=d/dt；
S2: carrying out pade to (1) formula and approach and carry out discretization, establishes and obtains permanent magnet direct-drive wind-force power generation identification model:
In formula: ad、bd、cd、aq、bq、cq、dqRespectively model coefficient, k are discrete point；
A is obtained according to formula (1)d、bd、cd、aq、bq、cq、dqIt is as follows:
Wherein, TsFor the sampling period；
S3: according to the d axis of stator current and q axis component isd、isqWith identification modelTo obtain identifier and true value
Error minimum value, establish error target function J (θ), formula expression is as follows:
S4: error mesh is solved using the MDPSO algorithm optimization based on adaptive space locating vector and average optimal location variable
Scalar functions J (θ), obtains aq,bq,cq,dqOptimal solution；
S5: according to a found outq,bq,cq,dqOptimal solution, calculate wind-driven generator the parameter R, L to be recognizedd,Lq, ψ are as follows:
Wherein, TsFor the sampling period, R is the resistance of stator, aq,bq,cq,dqThe respectively coefficient of model, Ld、LqIt is d axis and q axis
Inductive component；ψ is permanent magnet flux linkage；
S6: the parameter R, L after identification are utilizedd,Lq, ψ, according to MrminPI parameter of the criterion to generator-side converter wear current inner loop control
Dynamic tuning obtains new PI parameter value.
2. a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO according to claim 1, feature
It is: in S1, by direct-axis voltage usdWith quadrature-axis voltage usqCarry out decoupling and resolve into two components: one-component is by PI electric current
The u ' of controller outputsdWith u 'sq, expression formula is as follows:
Another component is the Δ u that decoupling obtainssdWith Δ usq, expression formula is as follows:
In formula: isdref、isqrefFor d axis, q shaft current reference value；Kp、KiThe respectively parameter of PI controller；S is complex frequency.
3. a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO according to claim 1 or 2,
Be characterized in that: in S4, the calculating process of the MDPSO algorithm based on adaptive space locating vector and average optimal location variable is such as
Under:
S41: the adaptive space search radius R (t) of the DPSO algorithm based on adaptive space locating vector, expression formula are calculated
Are as follows:
Wherein, λ is auto-adaptive parameter (usual λ >=2), and t is the number of iterations, and u is the uniform random number in [0,1] range,It is maximum value, the minimum value of particle position respectively；
S42: adaptive space locating vector (R (t)-x is appliedid(t)) rate equation of dynamic corrections more new particle, obtains:
In formula: w, r1、r2、r3For inertia weight, wherein r1、r2Value be (0,1) section random number, r3It is [0,1] range
Interior uniform random number, c1、c2、c3It is nonnegative constant；VidFor the current flight speed of i-th of particle；pidFor i-th particle
Optimal location；xidFor the current location of i-th of particle；gdThe optimal location found by particles all in population；
S43: introducing mean place Pmd in DPSO algorithm, and mean place Pmd is that D ties up being averaged for m particle individual desired positions
Value, the expression formula of mean place Pmd are as follows:
Then:
4. a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO according to claim 1 or 2,
Be characterized in that: in S6, the PI parameter of dynamic corrections controller is to achieve the purpose that raising system stationarity selects system parameter, is adopted
Use MrminCriterion keeps its closed loop amplitude-frequency characteristic peak value minimum, and makeover process is as follows:
S61: it is controlled using PI, each PI parameter is set as:
Then the open-loop transfer function of d shaft current inner loop control may be expressed as:
In formula: KpwmIt is the pulsewidth modulation gain of SVPWM, Ts is the sampling time；
S62:d axis open-loop transfer function Gid(S) having a pole is-R/Ld, if Ki/Kp=R/Ld, with the zero point of PI controller
Go the pole of compensation open-loop transfer function, then it is after zero pole point compensation that d axis open-loop transfer function is equivalent are as follows:
Wherein, optimum damping ratio is 0.707, current inner loop d axis PI controller parameter are as follows:
S63: according to the calculation method of S62, current inner loop q axis PI controller parameter is similarly calculated are as follows:
S64: speed control outer ring transmission function indicates are as follows:
In formula: J is the rotary inertia of generator, and p is the number of pole-pairs of generator；
S65: according to MrminCriterion design keeps its closed loop amplitude-frequency characteristic peak value minimum, expression formula are as follows:
H=5 is enabled, is obtained:
According to above-mentioned formula it is concluded that, either the PI parameter of current inner loop or revolving speed outer ring all with the intrinsic ginseng of generator
Number is related.
5. a kind of permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO according to claim 3, feature
Be: in S6, the PI parameter of dynamic corrections controller is to achieve the purpose that raising system stationarity selects system parameter, is used
MrminCriterion keeps its closed loop amplitude-frequency characteristic peak value minimum, and makeover process is as follows:
S61: it is controlled using PI, each PI parameter is set as:
Then the open-loop transfer function of d shaft current inner loop control may be expressed as:
In formula: KpwmIt is the pulsewidth modulation gain of SVPWM, Ts is the sampling time；
S62:d axis open-loop transfer function Gid(S) having a pole is-R/Ld, if Ki/Kp=R/Ld, with the zero point of PI controller
Go the pole of compensation open-loop transfer function, then it is after zero pole point compensation that d axis open-loop transfer function is equivalent are as follows:
Wherein, optimum damping ratio is 0.707, current inner loop d axis PI controller parameter are as follows:
S63: according to the calculation method of S62, current inner loop q axis PI controller parameter is similarly calculated are as follows:
S64: speed control outer ring transmission function indicates are as follows:
In formula: J is the rotary inertia of generator, and p is the number of pole-pairs of generator；
S65: according to MrminCriterion design keeps its closed loop amplitude-frequency characteristic peak value minimum, expression formula are as follows:
H=5 is enabled, is obtained:
According to above-mentioned formula (13) to formula (20) it is concluded that, either the PI parameter of current inner loop or revolving speed outer ring all with
The intrinsic relating to parameters of generator.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910493539.8A CN110492803A (en) | 2019-06-06 | 2019-06-06 | Permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910493539.8A CN110492803A (en) | 2019-06-06 | 2019-06-06 | Permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110492803A true CN110492803A (en) | 2019-11-22 |
Family
ID=68545908
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910493539.8A Pending CN110492803A (en) | 2019-06-06 | 2019-06-06 | Permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110492803A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111293693A (en) * | 2020-03-30 | 2020-06-16 | 华北电力大学 | Doubly-fed wind turbine converter control parameter identification method based on extended Kalman filtering |
| CN111725840A (en) * | 2020-06-29 | 2020-09-29 | 浙江大学 | Parameter identification method for direct-drive wind generating set controller |
| CN112000003A (en) * | 2020-08-31 | 2020-11-27 | 新疆大学 | Temperature control method of oxidation tank based on fractional order controller |
| CN112653184A (en) * | 2020-12-15 | 2021-04-13 | 上海交通大学 | Method, device, terminal and medium for identifying black box model of wind power generation equipment |
| CN113315432A (en) * | 2021-03-31 | 2021-08-27 | 新疆金风科技股份有限公司 | Parameter identification method and parameter identification device of wind driven generator |
| CN113990017A (en) * | 2021-11-21 | 2022-01-28 | 特斯联科技集团有限公司 | Forest and grassland fire early warning system and method based on PNN neural network |
| CN114094890A (en) * | 2021-11-02 | 2022-02-25 | 南京航空航天大学 | Optimal dynamic PI parameter setting method for magnetic-regulating and voltage-regulating direct-current generator system |
| CN114104013A (en) * | 2021-11-12 | 2022-03-01 | 中国矿业大学 | Speed control method for inhibiting wheel sliding of mining unmanned electric locomotive |
| CN117353617A (en) * | 2023-12-06 | 2024-01-05 | 小神童创新科技(广州)有限公司 | Closed loop non-inductive control method for DC brushless motor |
| CN117559518A (en) * | 2023-10-16 | 2024-02-13 | 国家电网有限公司华东分部 | Equivalent model construction method and device for wind power generation system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105680746A (en) * | 2015-11-20 | 2016-06-15 | 宋正强 | Method for designing current of permanent-magnet synchronous motor and parameter of speed controller PI by using online particle swarm optimization algorithm |
| RO131666A2 (en) * | 2015-06-02 | 2017-01-30 | Nextrom Industries S.R.L. | Method for estimating parameters of brushless dc motors |
| CN109546912A (en) * | 2019-01-07 | 2019-03-29 | 常州工学院 | Permanent Magnet Synchronous Motor Controller based on modified form particle swarm optimization algorithm |
-
2019
- 2019-06-06 CN CN201910493539.8A patent/CN110492803A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RO131666A2 (en) * | 2015-06-02 | 2017-01-30 | Nextrom Industries S.R.L. | Method for estimating parameters of brushless dc motors |
| CN105680746A (en) * | 2015-11-20 | 2016-06-15 | 宋正强 | Method for designing current of permanent-magnet synchronous motor and parameter of speed controller PI by using online particle swarm optimization algorithm |
| CN109546912A (en) * | 2019-01-07 | 2019-03-29 | 常州工学院 | Permanent Magnet Synchronous Motor Controller based on modified form particle swarm optimization algorithm |
Non-Patent Citations (4)
| Title |
|---|
| ZHAO-HUA LIU等: "Parameter Estimation for VSI-Fed PMSM Basedon a Dynamic PSO With Learning Strategies", 《IEEE TRANSACTIONS ON POWER ELECTRONICS》 * |
| 傅小利等: "基于柯西变异粒子群算法的永磁同步电机参数辨识", 《电工技术学报》 * |
| 姜久春 等: "《电动汽车电机及驱动系统》", 31 March 2018, 北京交通大学出版社 * |
| 沈蛟骁等: "标准粒子群算法在永磁同步电机参数辨识中的应用研究", 《微电机》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111293693A (en) * | 2020-03-30 | 2020-06-16 | 华北电力大学 | Doubly-fed wind turbine converter control parameter identification method based on extended Kalman filtering |
| CN111725840A (en) * | 2020-06-29 | 2020-09-29 | 浙江大学 | Parameter identification method for direct-drive wind generating set controller |
| CN112000003A (en) * | 2020-08-31 | 2020-11-27 | 新疆大学 | Temperature control method of oxidation tank based on fractional order controller |
| CN112653184B (en) * | 2020-12-15 | 2023-02-28 | 上海交通大学 | Identification method, device, terminal and medium for black box model of wind power generation equipment |
| CN112653184A (en) * | 2020-12-15 | 2021-04-13 | 上海交通大学 | Method, device, terminal and medium for identifying black box model of wind power generation equipment |
| CN113315432A (en) * | 2021-03-31 | 2021-08-27 | 新疆金风科技股份有限公司 | Parameter identification method and parameter identification device of wind driven generator |
| CN113315432B (en) * | 2021-03-31 | 2023-09-12 | 金风科技股份有限公司 | Parameter identification method and parameter identification device of wind driven generator |
| CN114094890A (en) * | 2021-11-02 | 2022-02-25 | 南京航空航天大学 | Optimal dynamic PI parameter setting method for magnetic-regulating and voltage-regulating direct-current generator system |
| CN114094890B (en) * | 2021-11-02 | 2023-12-19 | 南京航空航天大学 | Optimal dynamic PI parameter setting method for magnetic-regulation voltage-regulation direct-current generator system |
| CN114104013A (en) * | 2021-11-12 | 2022-03-01 | 中国矿业大学 | Speed control method for inhibiting wheel sliding of mining unmanned electric locomotive |
| CN114104013B (en) * | 2021-11-12 | 2024-04-09 | 中国矿业大学 | Speed control method for inhibiting wheel slip of mining unmanned electric locomotive |
| CN113990017A (en) * | 2021-11-21 | 2022-01-28 | 特斯联科技集团有限公司 | Forest and grassland fire early warning system and method based on PNN neural network |
| CN113990017B (en) * | 2021-11-21 | 2022-04-29 | 特斯联科技集团有限公司 | Forest and grassland fire early warning system and method based on PNN neural network |
| CN117559518A (en) * | 2023-10-16 | 2024-02-13 | 国家电网有限公司华东分部 | Equivalent model construction method and device for wind power generation system |
| CN117353617A (en) * | 2023-12-06 | 2024-01-05 | 小神童创新科技(广州)有限公司 | Closed loop non-inductive control method for DC brushless motor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110492803A (en) | Permanent magnet direct-driving aerogenerator parameter identification method based on MDPSO | |
| CN109642536B (en) | Control or processing system and method | |
| Yin et al. | Recurrent neural network based adaptive integral sliding mode power maximization control for wind power systems | |
| CN103184972B (en) | Parameter self-turning method for torque/propeller pitch controller of megawatt asynchronous double-feed wind driven generator | |
| CN104836498A (en) | Auto-tuning PID generator control system based on artificial neural network | |
| CN102664580B (en) | Mixed smart control method of power system multi-stabilizer | |
| CN110808610B (en) | Doubly-fed wind turbine optimization method based on proportional-integral-derivative control idea | |
| CN103762921B (en) | A kind of multi objective control method based on DFIG under the unbalanced power grid of particle cluster algorithm | |
| CN106452263B (en) | DFIG is based on the sliding moding structure direct Power Control method for expanding active power under a kind of unbalanced power grid | |
| CN104980069A (en) | Multipurpose optimization method for double closed-loop speed governing system of brushless DC motor | |
| CN111478365B (en) | Optimization method and system for control parameters of virtual synchronizer of direct-drive wind turbine generator | |
| CN116738636A (en) | Multi-machine equivalent method of doubly-fed wind power station considering impedance characteristics and synchronization mechanism | |
| Cheng et al. | Modeling and optimization control for aircraft AC generator brushless excitation system based on improved adaptive PSO | |
| CN106685290A (en) | Method for self-turning parameters of rotary speed controller of permanent magnet synchronous motor based on chaos molecule kinetic theory algorithm | |
| Xie et al. | Optimization research of FOC based on PSO of induction motors | |
| Sun et al. | Design and optimization of control parameters based on direct-drive permanent magnet synchronous generator for wind power system | |
| CN112072693B (en) | High-dimensional multi-fractional order optimization method for high-dimensional multi-fractional order doubly-fed fan control | |
| Chetouani et al. | Maximum power point tracking design using particle swarm optimization algorithm for wind energy conversion system connected to the grid | |
| Yin et al. | Multi-objective high-dimensional multi-fractional-order optimization algorithm for multi-objective high-dimensional multi-fractional-order optimization controller parameters of doubly-fed induction generator-based wind turbines | |
| Xiao et al. | VSCF wind turbine control strategy for maximum power generation | |
| Teng et al. | Recurrent wavelet neural network controller with improved particle swarm optimisation for induction generator system | |
| CN117318553B (en) | Low-wind-speed permanent magnet direct-driven wind turbine control method based on TD3 and Vienna rectifier | |
| CN113067522B (en) | Elevator output voltage control method based on RF-GA-SVM algorithm | |
| CN105932860B (en) | A kind of power converter double-loop control strategy based on particle swarm algorithm | |
| Pucci | Sensors-less neural maximum power point tracking control of induction machines wind generators by growing neural gas and minor component analysis EXIN+ reduced order observer |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191122 |