CN102307004B - L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method - Google Patents

L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method Download PDF

Info

Publication number
CN102307004B
CN102307004B CN 201110242200 CN201110242200A CN102307004B CN 102307004 B CN102307004 B CN 102307004B CN 201110242200 CN201110242200 CN 201110242200 CN 201110242200 A CN201110242200 A CN 201110242200A CN 102307004 B CN102307004 B CN 102307004B
Authority
CN
China
Prior art keywords
current
voltage
lcl
electrical network
value
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
Application number
CN 201110242200
Other languages
Chinese (zh)
Other versions
CN102307004A (en
Inventor
伍小杰
王颖杰
宗伟林
戴鹏
夏帅
王贵峰
吴玮
周书颖
陆平
袁庆庆
吴迪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
XUZHOU BAODI ELECTRIC CO Ltd
China University of Mining and Technology CUMT
Original Assignee
XUZHOU BAODI ELECTRIC CO Ltd
China University of Mining and Technology CUMT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by XUZHOU BAODI ELECTRIC CO Ltd, China University of Mining and Technology CUMT filed Critical XUZHOU BAODI ELECTRIC CO Ltd
Priority to CN 201110242200 priority Critical patent/CN102307004B/en
Publication of CN102307004A publication Critical patent/CN102307004A/en
Application granted granted Critical
Publication of CN102307004B publication Critical patent/CN102307004B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention discloses an L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method, which belongs to a parameter identification method suitable for LCL filtering. The LCL-filtering-based controlled rectifier parameter identification method comprises the five following steps of: 1, acquiring short-circuit current; 2, performing genetic-algorithm-based parameter identification; 3, acquiring an initial phase angle of a power grid in starting; 4, estimating a virtual power grid flux linkage value in real time under LCL filtering condition; and 5, acquiring an orientation angle of a system under the LCL filtering condition. The method has the advantages that: 1, parameters of the power grid are identified by adopting a genetic algorithm, so an initial virtual flux linkage value and phase of the power grid can be more accurately observed, and the steady starting of a rectifier is realized; 2, the resonance frequency of an LCL filter is obtained in combination with inductance in the power grid, and a reliable basis is provided for the resonance suppression of the LCL filter; and 3, accurate virtual power grid flux linkage orientation is provided for the sensorless running of the LCL-filtering-based controlled rectifier, and the stable running of the system can be controlled.

Description

A kind of controlled rectifier parameter identification method based on LCL filtering
Technical field
The present invention relates to the parameter identification method of a kind of suitable inductor-capacitor-inductance (LCL) filtering, particularly a kind of controlled rectifier parameter identification method based on LCL filtering.
Background technology
Three-phase grid-connected converter has the output current sine, and harmonic content is low, and power factor is adjustable and the advantage of energy in bidirectional flow, and its volume is little and lightweight, is real " green energy converter ".It is in field extensive uses such as generating electricity by way of merging two or more grid systems of the regenerative resources such as active power filtering (APF), unified trend control (UPFC), superconducting energy storage (SMES), high voltage direct current transmission (HVDC), Electric Drive and solar energy, wind energy.Along with country's " 12 " realization of plan, the energy-saving and emission-reduction dynamics continues to increase and the green energy resource demand is constantly increased, and the application of grid-connected converter will be more extensive.
But the devices switch frequency of current transformer is about 2~15kHz, the harmonic frequency of the switching frequency that produces or switching frequency integral multiple lower, after this harmonic wave enters electrical network, can produce certain influence to the electrical network load of humorous wave interference sensitivity, when serious even the generation damage.When adopting LCL filtering, harmonic suppression effect is good, can reduce inductance value, reduces switching frequency.But control this current transformer, not only need detection of grid voltage, electric current and DC bus-bar voltage, if adopt the active damping method to suppress resonance, generally also need detect ac filter capacitive branch voltage or electric current.This not only increases the cost of rectifier, and it is many unfavorable to bring also for the design and installation of whole device.Simultaneously the line voltage transducer detect actual be the voltage that comprises the impact of electrical network induction reactance because the induction reactance of LCL filter own is less, the electrical network induction reactance particularly impact of transformer leakage inductance etc. can't be ignored, thereby has influence on the stable operation of rectifier.
The stator circuit of Three-Phase PWM Rectifier electrical network circuit and three phase alternating current motor has very large similitude.Employing is similar to the method for alternating current machine flux observation, utilize Three-Phase PWM Rectifier DC bus-bar voltage and on off state to construct virtual electrical network flux linkage vector as the directional vector of PWM rectifier, can save the line voltage transducer, and strengthen system's Immunity Performance.But there is the observation initial-value problem in virtual electrical network magnetic linkage, when initial estimate inaccurate may cause rectifier to start, and power network current and DC bus-bar voltage big ups and downs, even can't start.Existing initial estimate method is based on all that the rectifier of pure inductance filtering proposes, because the complexity of LCL three rank filters, the simply above method of reference.Again since electrical network self with induction reactance, particularly the leakage inductance that carries of transformer also can change the LCL filter resonance frequency, might have a strong impact on the electrical network flux observation simultaneously.
Summary of the invention
The objective of the invention is to provide a kind of controlled rectifier parameter identification method based on LCL filtering, it is inaccurate and adopt startup problem out of control after the Converter Without Voltage Sensor control to have solved, system oriented too much based on the controlled rectifier transducer of LCL filtering.
The object of the present invention is achieved like this: described a kind of controlled rectifier parameter identification method based on LCL filtering comprises five steps: 1, the acquisition process of middle short circuit current; 2, based on the parameter identification of genetic algorithm; The initial phase angle of electrical network obtains when 3, starting; 4, the real-time estimation of the virtual magnetic linkage value of electrical network under the LCL filtering condition; 5, the system oriented angle under the LCL filtering condition obtains; Concrete steps are as follows:
The acquisition process of step 1, short circuit current is:
Step 1.1, sampling DC bus-bar voltage, power device is not worked in the controlled rectifier at this moment, is DC bus-bar voltage by the Power Diode Pumped in parallel with power device with the grid ac voltage rectification, is uncontrollable rectifier and gets voltage;
Step 1.2, send the zero vector signal (0 that the duration is T to 6 power devices of three brachium pontis rectification circuits, O, O) or zero vector signal (1,1,1), be equivalent to rectification circuit brachium pontis or lower bridgc arm short on all, the circuit that forms of electrical network and controlled rectifier at this moment is equivalent to only have the three-phase symmetrical circuit of LCL filter composition in electrical network and the controlled rectifier;
Step 1.3, by Hall current sensor certain phase current of sampling, take T/4 as a sampling period, sample 4 times, record 4 sampled values;
Step 2, based on the parameter identification of genetic algorithm:
Step 2.1, set up the genetic algorithm fitness function, with 4 current sampling datas as actual value, will be by short circuit current calculated with mathematical model electric current as estimated value; The fitness that calculates by fitness function is as the choice criteria of identified parameters;
The uncontrollable rectification DC bus-bar voltage of step 2.2, sampling gained, its size fluctuates between 1/2 to 1 times of line voltage peak, what the phase voltage amplitude should be for DC bus-bar voltage so Extremely Doubly; Because phase voltage is cosine function, its cycle is 2 π, thus with voltage starting phase angle value at 2 π in the cycle, as between 0 to 2 π, power frequency angular frequency 100 π rad/s, the deviation of getting power frequency ± 1% is the hunting zone; Foundation , L rBe the electrical network inductance, L gBe the net side inductance of LCL filter, LBe LCL filter transform device side inductance, the electrical network inductance ( L r) add affiliation and make resonance angular frequency ω xDull decline, its extreme value are the original resonance frequency of filter Doubly, then the resonance angular frequency hunting zone in the original resonance frequency of filter to the original resonance frequency of filter Doubly between this extreme value;
Step 2.3, according to phase voltage amplitude, voltage starting phase angle, power frequency angular frequency, four parameter search scopes of resonance angular frequency, produce at random the initial parameter population; With the individual calculating current estimation value that is used for of every group of parameter, relatively the error of sampled value and estimated value obtains the fitness function value; Then population is carried out Selecting operation, and keep elite population, then carry out the crossover and mutation computing, generate colony of new generation; According to the algorithm stop condition, such as evolutionary generation or minimum fitness value, judge to calculate whether stop, otherwise recomputate from " selection " computing;
Step 3, the obtaining of the initial phase angle of electrical network when starting:
The computing time of step 3.1, genetic algorithm is longer, and the initial phase angle of electrical network need to recomputate; Again send zero vector, the duration is 2 T/4, sampling short circuit current twice, sampling interval T/4;
Step 3.2, according to the short circuit current model, with first sample rate current, with the resonance frequency omega of line voltage amplitude U, frequencies omega and the LCL filter of identification gained x, the inverse starting phase angle; It has two solutions, calculates gained starting phase angle, substitution with two , in the formula , , , , in the formula MBe the LCL filter resonance frequency With the product of LCL filter resonance frequency calculating parameter, namely , in the formula, L rBe the electrical network inductance, L gBe the net side inductance of LCL filter, LBe LCL filter transform device side inductance, C fBe the LCL filter capacity), obtain the short-circuit current value of second T/4 sampling instant synchronization in two that calculate gained and the steps 3.1, judge to calculate in two current values of gained and the step 3.1 second T/4 constantly the sample rate current value which is the most approaching, will calculate near the starting phase angle of current value as real starting phase angle;
The real-time estimation of the virtual magnetic linkage value of electrical network under step 4, the LCL filtering condition:
Step 4.1, calculate rectification side three-phase voltage according to three-phase bridge arm switch state and DC bus-bar voltage, be transformed under the two-phase rest frame, with the respectively substitution of gained magnitude of voltage , Middle latter half, namely With In, obtain rectification side magnetic linkage under the two-phase rest frame;
Step 4.2, detection rectification side electric current are transformed under the two-phase rest frame foundation , Middle first half, namely With In, respectively with rectification side electric current Multiply each other with electrical network internal inductance and LCL filter inductance sum, acquired results is inductance magnetic linkage under the two-phase rest frame;
Step 4.3, rectification side magnetic linkage and the addition of inductance magnetic linkage obtain based on the virtual magnetic linkage value of the real-time grid of LCL filtering With
Obtaining of system oriented angle under step 5, the LCL filtering condition:
Under step 5.1, the two-phase rest frame The virtual magnetic linkage value of axle electrical network Divided by The virtual magnetic linkage value of axle electrical network , obtain tangent value; Calculate arctan function, obtain the virtual flux linkage orientation of electrical network angle;
Step 5.2, utilize the virtual flux linkage orientation of electrical network angle, rectification side electric current, capacitance current are rotated coordinate transform, rectification side voltage given is rotated the coordinate inverse transformation, thereby realize the virtual flux linkage orientation control of rectifier.
Beneficial effect is owing to adopted such scheme, the method to comprise that rectifier starts front and latter two part of startup.Before rectifier started, the sampling DC bus-bar voltage was sent the short circuit current of zero vector sampling rectification side to rectifier.According to the short circuit current model, resonance frequency by Identification of Genetic Algorithm line voltage amplitude, phase place, frequency and LCL filter etc.Again send zero vector, the sampling short circuit current, the resonance frequency by identification line voltage amplitude, frequency and LCL filter obtains the line voltage initial phase, the virtual magnetic linkage initial value of estimation electrical network.Rectifier according to reconstruct gained rectification side magnetic linkage and inductance magnetic linkage, obtains the virtual magnetic linkage value of real-time grid after starting, and calculates virtual flux linkage orientation angle, realizes the accurate pointing control of system.It is inaccurate and adopt startup problem out of control after the Converter Without Voltage Sensor control to have solved, system oriented too much based on the controlled rectifier transducer of LCL filtering, has reached purpose of the present invention.
Advantage:
1, adopts genetic algorithm that electrical network parameter is carried out identification, realized that the virtual magnetic linkage initial value of electrical network and phase place comparatively accurately observe, realized the rectifier smooth starting.
2, obtain to consider the LCL filter resonance frequency of electrical network internal inductance, suppress to provide reliable basis for the resonance of LCL filter.
3, for the controlled rectifier sensor-less operation based on LCL filtering provides accurate electrical network virtual flux linkage orientation, be conducive to the stable operation of control system.
Description of drawings
Fig. 1 is based on the controlled rectifier Converter Without Voltage Sensor control block diagram of LCL filtering.
Fig. 2 is the three-phase VSR topology diagram of LCL filtering.
Single-phase LCL circuit diagram when Fig. 3 is short circuit.
Fig. 4 is the genetic algorithm flow chart of power grid parameter identification.
Fig. 5 is short circuit current waveform figure.
Fig. 6 is without initial virtual magnetic linkage value startup figure.
Fig. 7 has accurately initial virtual magnetic linkage value startup figure.
Fig. 8 adopts parameter identification to estimate initial virtual magnetic linkage value startup figure.
Fig. 9 is that the magnetic linkage angle is followed the tracks of analogous diagram.
Figure 10 is busbar voltage step figure.
Embodiment
Embodiment 1: described a kind of controlled rectifier parameter identification method based on LCL filtering comprises five steps: 1, the acquisition process of short circuit current; 2, based on the parameter identification of genetic algorithm; The initial phase angle of electrical network obtains when 3, starting; 4, the real-time estimation of the virtual magnetic linkage value of electrical network under the LCL filtering condition; 5, the system oriented angle under the LCL filtering condition obtains; Concrete steps are as follows:
The acquisition process of step 1, short circuit current is:
Step 1.1, sampling DC bus-bar voltage, power device is not worked in the controlled rectifier at this moment, is DC bus-bar voltage by the Power Diode Pumped in parallel with power device with the grid ac voltage rectification, is uncontrollable rectifier and gets voltage;
Step 1.2, send the zero vector signal (0 that the duration is T to 6 power devices of three brachium pontis rectification circuits, O, O) or zero vector signal (1,1,1), be equivalent to rectification circuit brachium pontis or lower bridgc arm short on all, the circuit that forms of electrical network and controlled rectifier at this moment is equivalent to only have the three-phase symmetrical circuit of LCL filter composition in electrical network and the controlled rectifier;
Step 1.3, by Hall current sensor certain phase current of sampling, take T/4 as a sampling period, sample 4 times, record 4 sampled values;
Step 2, based on the parameter identification of genetic algorithm:
Step 2.1, set up the genetic algorithm fitness function, with 4 current sampling datas as actual value, will be by short circuit current calculated with mathematical model electric current as estimated value; The fitness that calculates by fitness function is as the choice criteria of identified parameters;
The uncontrollable rectification DC bus-bar voltage of step 2.2, sampling gained, its size fluctuates between 1/2 to 1 times of line voltage peak, what the phase voltage amplitude should be for DC bus-bar voltage so Extremely Doubly; Because phase voltage is cosine function, its cycle is 2 π, thus with voltage starting phase angle value at 2 π in the cycle, as between 0 to 2 π, power frequency angular frequency 100 π rad/s, the deviation of getting power frequency ± 1% is the hunting zone; Foundation , L rBe the electrical network inductance, L gBe the net side inductance of LCL filter, LBe LCL filter transform device side inductance, the electrical network inductance ( L r) add affiliation and make resonance angular frequency ω xDull decline, its extreme value are the original resonance frequency of filter Doubly, then the resonance angular frequency hunting zone in the original resonance frequency of filter to the original resonance frequency of filter Doubly between this extreme value;
Step 2.3, according to phase voltage amplitude, voltage starting phase angle, power frequency angular frequency, four parameter search scopes of resonance angular frequency, produce at random the initial parameter population; With the individual calculating current estimation value that is used for of every group of parameter, relatively the error of sampled value and estimated value obtains the fitness function value; Then population is carried out Selecting operation, and keep elite population, then carry out the crossover and mutation computing, generate colony of new generation; According to the algorithm stop condition, such as evolutionary generation or minimum fitness value, judge to calculate whether stop, otherwise recomputate from " selection " computing; This step process is according to specification content shown in Figure 4.
Step 3, the obtaining of the initial phase angle of electrical network when starting:
The computing time of step 3.1, genetic algorithm is longer, and the initial phase angle of electrical network need to recomputate; Again send zero vector, the duration is 2 T/4, sampling short circuit current twice, sampling interval T/4;
Step 3.2, according to the short circuit current model, with first sample rate current, with the resonance frequency omega of line voltage amplitude U, frequencies omega and the LCL filter of identification gained x, the inverse starting phase angle; It has two solutions, calculates gained starting phase angle, substitution with two , in the formula , , , , in the formula MBe the LCL filter resonance frequency With the product of LCL filter resonance frequency calculating parameter, namely (it is shown in Figure 1 such as specification, L rBe the electrical network inductance, L gBe the net side inductance of LCL filter, LBe LCL filter transform device side inductance, C fBe the LCL filter capacity), obtain the short-circuit current value of second T/4 sampling instant synchronization in two that calculate gained and the steps 3.1, judge to calculate in two current values of gained and the step 3.1 second T/4 constantly the sample rate current value which is the most approaching, will calculate near the starting phase angle of current value as real starting phase angle;
The real-time estimation of the virtual magnetic linkage value of electrical network under step 4, the LCL filtering condition:
Step 4.1, calculate rectification side three-phase voltage according to three-phase bridge arm switch state and DC bus-bar voltage, be transformed under the two-phase rest frame, with gained magnitude of voltage substitution specification formula (10) respectively , and (11) Middle latter half With , obtain rectification side magnetic linkage under the two-phase rest frame;
Step 4.2, detection rectification side electric current are transformed under the two-phase rest frame, according to specification formula (10) , and formula (11) Middle first half With , respectively with rectification side electric current Multiply each other with electrical network internal inductance and LCL filter inductance sum, acquired results is inductance magnetic linkage under the two-phase rest frame;
Step 4.3, rectification side magnetic linkage and the addition of inductance magnetic linkage obtain based on the virtual magnetic linkage value of the real-time grid of LCL filtering With
Obtaining of system oriented angle under step 5, the LCL filtering condition:
Under step 5.1, the two-phase rest frame The virtual magnetic linkage value of axle electrical network Divided by The virtual magnetic linkage value of axle electrical network , obtain tangent value; Calculate arctan function, obtain the virtual flux linkage orientation of electrical network angle;
Step 5.2, utilize the virtual flux linkage orientation of electrical network angle, rectification side electric current, capacitance current are rotated coordinate transform, rectification side voltage given is rotated the coordinate inverse transformation, thereby realize the virtual flux linkage orientation control of rectifier.
In Fig. 2, according to the symmetry of topological structure, ignore the inductance internal resistance, provide the electric current and voltage equation group of single-phase branch road
(1)
(2)
(3)
(4)
In the formula Be respectively line voltage and electric current;
Be respectively capacitance voltage and electric current;
Be respectively rectifier AC electric current and voltage;
Be respectively grid side inductance, rectification side inductance and the electric capacity of LCL filter;
Be respectively DC bus-bar voltage and electric current;
Be respectively bus filter capacitor and load;
Can get rectifier AC voltage and bus current according to each brachium pontis power device on off state, respectively suc as formula shown in (5) and the formula (6).
(5)
(6)
In the formula, Be each brachium pontis power device switch function, upper brachium pontis is open-minded, and lower brachium pontis is 1 when turn-offing, otherwise is 0.
If directly estimate line voltage according to formula (1) and formula (2), the noise that current differential is introduced can cause the line voltage estimation inaccurate, even is unable to estimate.According to virtual magnetic linkage principle, use virtual electrical network magnetic linkage to replace line voltage, can effectively avoid noise jamming.Because the electric capacity effect of LCL filter can be ignored when low frequency, adopt the virtual electrical network flux linkage model of pure inductance filtering, use rectifier side electric current to replace power network current, the virtual magnetic linkage of Three-Phase PWM Rectifier that obtains LCL filtering under the two-phase rest frame is
(7)
(8)
(9)
In the formula
For rectifier AC reconstruct voltage , Component, , Be respectively virtual electrical network magnetic linkage , Component, Be virtual electrical network flux linkage vector angle, it falls behind the line voltage azimuth
Because virtual electrical network magnetic linkage initial value for integral is unknown, estimates its value such as employing formula (7), (8), can in the virtual electrical network magnetic linkage that estimates, introduce the direct current biasing relevant with initial value for integral.General normal employing low pass filter replaces pure integrator, shown in (10), (11).
(10)
(11)
, Low pass filter pair is adopted in representative respectively , Filtered result, so under the two-phase rest frame the virtual magnetic linkage value of the Three-Phase PWM Rectifier of LCL filtering respectively suc as formula shown in (10), (11).But its initial value the unknown can cause the filtering time lengthening equally, causes in the start-up course virtual flux linkage orientation inaccurate, and the electric current and voltage fluctuation is violent.In addition, frequency converter generally was connected to transformer before the access electrical network, play the effect of transformation isolation, but transformer leakage inductance With respect to less LCL inductance, just become an amount of can not ignore, affect the accurate observation of virtual magnetic linkage.More than two amounts must be before starting Accurate Prediction.
Before rectifier starts, send zero vector to rectifier, be equivalent to rectifier side three-phase shortcircuit, single-phase LCL circuit diagram is as shown in Figure 3 during short circuit.At this moment Be zero, be conducive to circuit reduction, but because the filter capacitor effect can produce resonance frequency component, the initial virtual electrical network magnetic linkage estimation equation in the time of therefore can not directly applying mechanically pure inductance filtering in short circuit current.If A phase voltage , set up single-phase branch road rectifier ac-side current according to Fig. 3 Instantaneous equation.
(12)
In the formula
(13)
(14)
(15)
(16)
, MBe respectively the LCL filter resonance frequency of consideration electrical network inductance and amassing of its parameter.According to the relation of the virtual flux linkage vector of electrical network and line voltage vector, get the virtual magnetic linkage formula of electrical network under the two-phase rest frame.
(17)
(18)
Through type (17) and (18) can be converted into the identification of magnetic linkage initial value the identification of A phase line voltage initial value, comprise amplitude, the starting phase angle of line voltage.Through type (12) as can be known, the identification of transformer leakage inductance can be converted into the identification to the LCL filter resonance frequency.The actual electric network frequency has certain fluctuation needs accurate identification equally, needs so identified parameters to have: the resonance angular frequency of the amplitude of line voltage, starting phase angle, power frequency angular frequency and LCL filter.
Sampling A phase current 4 times, set up thus fitness function, shown in (19).
(19)
In the formula, For Sampled value constantly, For through type (12) calculates Moment estimated value.The evaluation of parameter just is converted into fitness function minimizing problem.
For reducing finding the solution the time of genetic algorithm, and avoid the local convergence of separating as far as possible, need to do further regulation to the scope of finding the solution parameter. Before sending zero vector, the uncontrollable rectification DC bus-bar voltage of sampling gained, its size fluctuates between 1/2 to 1 times of line voltage peak, what the phase voltage amplitude should be for DC bus-bar voltage so Extremely Doubly.The voltage starting phase angle is got 2 π cycles.The span of power frequency angular frequency is got power frequency according to the regulation of the system frequency deviation standard GB/T15945-2008 of State Grid Deviation be the hunting zone.The electrical network inductance adds affiliation makes resonance angular frequency ω xDull decline, its extreme value are the original resonance frequency of filter Doubly, then the resonance angular frequency hunting zone is arrived between this extreme value in the original resonance frequency of filter.
The genetic algorithm flow process of power grid parameter identification as shown in Figure 4.At first determine the hunting zone of each parameter, produce the initial parameter population.With the individual calculating current estimation value that is used for of every group of parameter, relatively the error of sampled value and estimated value obtains the fitness function value.Then population is carried out Selecting operation, and keep elite population, then carry out the crossover and mutation computing, generate colony of new generation.According to the algorithm stop condition, judge to calculate whether stop, otherwise recomputate from Selecting operation.
Because genetic algorithm is long computing time, the starting phase angle of identification gained is not as using when starting.Can utilize other identification parameters obtaineds, add two zero vectors before startup, sampling A phase current by first A commutating phase device side current sampling data, gets two starting phase angles by formula (12).Again with phase angle substitution formula (12), calculate the second current value constantly, with actual sample value relatively, get both near the time the phase angle.
The control method that the present invention proposes is passed through checking in software Matlab and experimental prototype.Emulation and experiment parameter are as follows: the rated power 1KW of system, and Circuit Fault on Secondary Transformer open-circuit line voltage effective value 40V, busbar voltage 150V, switching frequency 5KHz, LCL filter and transformer leakage inductance parameter are respectively rectifier side inductance L=1.41mH, electric capacity C f=4.75 μ F (△ type), the grid side inductance L g=0.82mH (transformer leakage inductance 0.22mH).
Before sending zero vector, the DC bus-bar voltage that the experiment sampling obtains is 50.64V, need to determine the identified parameters hunting zone according to experiment parameter.After sending zero vector, to rectifier side A phase current values sampling 4 times, sampling interval 0.5ms.Electric current and voltage detecting adopt Hall element in the experiment, and adopt 16 quick conversion chip AD7656.Fig. 5 is for sending zero vector A commutating phase side electric current and busbar voltage figure, because the afterflow of three-phase current, and the busbar voltage rising of after stopping the zero vector effect, charging.Call GAs Toolbox, write the fitness function file, population scale 100 is set, the elite leaves over 2, crossover probability 0.8, variation probability 0.01.
Identification parameters obtained input digit controller DSP TMS320F2812 control program is sent out zero vector again, obtains the voltage starting phase angle, and judges actual value.Calculate to get the virtual magnetic linkage initial value of electrical network according to formula (17), (18) again, calculate orientation angle by formula (9), start rectifier.After the startup, the virtual flux linkage calculation of electrical network switches to formula (10), (11).
Fig. 6, Fig. 7 and Fig. 8 be respectively rectifier without initial virtual magnetic linkage value, by detection of grid voltage calculate accurate initial virtual magnetic linkage value and adopt A that parameter identification estimates initial virtual magnetic linkage value current on line side and busbar voltage start oscillogram mutually.Starting current and busbar voltage when accurate initial virtual magnetic linkage value is relatively arranged, when starting without initial virtual magnetic linkage value, busbar voltage overshoot is large, and the vibration of starting current and busbar voltage is violent, needs long stabilization time, and adopt parameter identification to estimate when initial virtual magnetic linkage value starts, the fluctuation of starting current reduces, and busbar voltage overshoot diminishes, but because certain initial virtual magnetic linkage evaluated error is arranged, starting current also has certain fluctuation, and certain prolongation is also arranged stabilization time.When being the rectifier startup, Fig. 9 estimates the magnetic linkage angle tracking analogous diagram of initial virtual magnetic linkage value without initial virtual magnetic linkage value and employing parameter identification.Long dotted line represents the virtual magnetic linkage angle pursuit path without initial virtual magnetic linkage value, and the short dash line representative adopts parameter identification to estimate the virtual magnetic linkage angle pursuit path of initial virtual magnetic linkage value, and solid line represents the virtual magnetic linkage angle of actual electric network track.Adopt parameter identification to estimate that the virtual magnetic linkage angle pursuit path of initial virtual magnetic linkage value overlaps substantially with the virtual magnetic linkage angle of electrical network track, and need to could follow the tracks of correctly through half fundamental frequency cycles without the virtual magnetic linkage angle pursuit path of initial virtual magnetic linkage value.With Fig. 6 figure and 7 figure more as can be known, the actual magnetic linkage without initial virtual magnetic linkage value is followed the tracks of the stable longer time that also needs.Figure 10 is rectifier busbar voltage A phase current on line side oscillogram during from the 150V step to 200V, because this moment, magnetic linkage was followed the tracks of stablely, busbar voltage and A phase current on line side are rapidly stable.

Claims (3)

1. controlled rectifier parameter identification method based on LCL filtering, it is characterized in that: described controlled rectifier parameter identification method comprises five steps: 1, the acquisition process of short circuit current; 2, based on the parameter identification of genetic algorithm; The initial phase angle of electrical network obtains when 3, starting; 4, the real-time estimation of the virtual magnetic linkage value of electrical network under the LCL filtering condition; 5, the system oriented angle under the LCL filtering condition obtains; Concrete steps are as follows:
The acquisition process of step 1, short circuit current is:
Step 1.1, sampling DC bus-bar voltage, power device is not worked in the controlled rectifier at this moment, is DC bus-bar voltage by the Power Diode Pumped in parallel with power device with the grid ac voltage rectification, is uncontrollable rectifier and gets voltage;
Step 1.2, send the zero vector signal (0 that the duration is T to 6 power devices of three brachium pontis rectification circuits, O, O) or zero vector signal (1,1,1), be equivalent to rectification circuit brachium pontis or lower bridgc arm short on all, the circuit that forms of electrical network and controlled rectifier at this moment is equivalent to only have the three-phase symmetrical circuit of LCL filter composition in electrical network and the controlled rectifier;
Step 1.3, by Hall current sensor certain phase current of sampling, take T/4 as a sampling period, sample 4 times, record 4 sampled values;
Step 2, based on the parameter identification of genetic algorithm:
Step 2.1, set up the genetic algorithm fitness function, with 4 current sampling datas as actual value, will be by short circuit current calculated with mathematical model electric current as estimated value; The fitness that calculates by fitness function is as the choice criteria of identified parameters;
The uncontrollable rectification DC bus-bar voltage of step 2.2, sampling gained, its size fluctuates between 1/2 to 1 times of line voltage peak, what the phase voltage amplitude should be for DC bus-bar voltage so Extremely Doubly; Because phase voltage is cosine function, its cycle is 2 π, thus with voltage starting phase angle value at 2 π in the cycle, as between 0 to 2 π, power frequency angular frequency 100 π rad/s, the deviation of getting power frequency ± 1% is the hunting zone; Foundation , L rBe the electrical network inductance, L gBe the net side inductance of LCL filter, LBe LCL filter transform device side inductance, the electrical network inductance ( L r) add affiliation and make resonance angular frequency ω xDull decline, its extreme value are the original resonance frequency of filter Doubly, then the resonance angular frequency hunting zone in the original resonance frequency of filter to the original resonance frequency of filter Doubly between this extreme value;
Step 2.3, according to phase voltage amplitude, voltage starting phase angle, power frequency angular frequency, four parameter search scopes of resonance angular frequency, produce at random the initial parameter population; With the individual calculating current estimation value that is used for of every group of parameter, relatively the error of sampled value and estimated value obtains the fitness function value; Then population is carried out Selecting operation, and keep elite population, then carry out the crossover and mutation computing, generate colony of new generation; According to the algorithm stop condition, such as evolutionary generation or minimum fitness value, judge to calculate whether stop, otherwise recomputate from " selection " computing;
Step 3, the obtaining of the initial phase angle of electrical network when starting:
The computing time of step 3.1, genetic algorithm is longer, and the initial phase angle of electrical network need to recomputate; Again send zero vector, the duration is 2 T/4, sampling short circuit current twice, sampling interval T/4;
Step 3.2, according to the short circuit current model, with first sample rate current, with the resonance frequency omega of line voltage amplitude U, frequencies omega and the LCL filter of identification gained x, the inverse starting phase angle; It has two solutions, calculates gained starting phase angle, substitution with two , in the formula , , , , in the formula MBe the LCL filter resonance frequency With the product of LCL filter resonance frequency calculating parameter, namely , in the formula L rBe the electrical network inductance, L gBe the net side inductance of LCL filter, LBe LCL filter transform device side inductance, C fBe the LCL filter capacity), obtain the short-circuit current value of second T/4 sampling instant synchronization in two that calculate gained and the steps 3.1, judge to calculate in two current values of gained and the step 3.1 second T/4 constantly the sample rate current value which is the most approaching, will calculate near the starting phase angle of current value as real starting phase angle;
The real-time estimation of the virtual magnetic linkage value of electrical network under step 4, the LCL filtering condition:
Step 4.1, calculate rectification side three-phase voltage according to three-phase bridge arm switch state and DC bus-bar voltage, be transformed under the two-phase rest frame, with the respectively substitution of gained magnitude of voltage , Middle latter half With , obtain rectification side magnetic linkage under the two-phase rest frame;
Step 4.2, detection rectification side electric current are transformed under the two-phase rest frame foundation , First half With , respectively with rectification side electric current Multiply each other with electrical network internal inductance and LCL filter inductance sum, acquired results is inductance magnetic linkage under the two-phase rest frame;
Step 4.3, rectification side magnetic linkage and the addition of inductance magnetic linkage obtain based on the virtual magnetic linkage value of the real-time grid of LCL filtering With
Obtaining of system oriented angle under step 5, the LCL filtering condition:
Under step 5.1, the two-phase rest frame The virtual magnetic linkage value of axle electrical network Divided by The virtual magnetic linkage value of axle electrical network , obtain tangent value; Calculate arctan function, obtain the virtual flux linkage orientation of electrical network angle;
Step 5.2, utilize the virtual flux linkage orientation of electrical network angle, rectification side electric current, capacitance current are rotated coordinate transform, rectification side voltage given is rotated the coordinate inverse transformation, thereby realize the virtual flux linkage orientation control of rectifier.
2. a kind of controlled rectifier parameter identification method based on LCL filtering according to claim 1, it is characterized in that, topological structure with the controlled rectification of LCL filter comprises electric network source and electrical network internal reactance (1), LCL filter (2), three brachium pontis rectification circuits (3), DC side filter capacitor C (4) and DC load (5), electric network source links to each other with the three-phase input end of LCL filter respectively with the electrical network internal reactance, the output of LCL filter links to each other with each brachium pontis node in the three brachium pontis rectification circuits respectively, DC side filter capacitor C is in parallel with load, and is connected to two outputs of three brachium pontis rectification circuits.
3. a kind of controlled rectifier parameter identification method based on LCL filtering according to claim 1, it is characterized in that, comprise the detection (6) of rectification side electric current with the controlled rectification of LCL filter without the transducer control structure, the detection (7) of LCL filter capacity electric current, the detection of DC bus-bar voltage (8), parameter identification unit (9), the computing unit of the virtual magnetic linkage of electrical network (10), the computing unit (11) at the virtual flux linkage orientation of electrical network angle, DC voltage control unit (12), current control unit (13), active damping control unit (14) and space vector of voltage generating unit (15); Given direct voltage ( ) with the detected value (U of DC bus-bar voltage Dc) send into DC voltage control unit (12), obtain active current given ( ), the employing orientation angle ( ) with the detected value (i of rectification side electric current A2, i B2) coordinate transform, obtain respectively d, q shaft current (i d, i q); D, q shaft current (i d, i q) and given value of current ( , ) send into current control unit (13), that the current feed-forward item equals power supply internal inductance and LCL filter inductance and multiply by d, q shaft current (i d, i q); Detected value (the i of LCL filter capacity electric current Ca, i Cb), through the coordinate inverse transformation based on orientation angle (θ), send into damping control module, obtain damping control magnitude of voltage, respectively with the addition of current control unit output valve, through the coordinate inverse transformation based on orientation angle (θ), obtain rectification side voltage given ( , ), send into space vector of voltage generating unit (15), obtain on off state (S A, b, c), and with on off state (S A, b, c) feed three brachium pontis rectification circuits (3); On off state (S A, b, c), the detected value (U of DC bus-bar voltage Dc), given value of current ( , ) and parameter identification unit (9) parameters obtained, send into the computing unit (10) of the virtual magnetic linkage of electrical network, obtain the virtual magnetic linkage of electrical network , The virtual magnetic linkage of electrical network , Send into again the computing unit (11) at the virtual flux linkage orientation of electrical network angle, obtain orientation angle (θ); So circulation has realized Converter Without Voltage Sensor control.
CN 201110242200 2011-08-22 2011-08-22 L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method Expired - Fee Related CN102307004B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110242200 CN102307004B (en) 2011-08-22 2011-08-22 L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110242200 CN102307004B (en) 2011-08-22 2011-08-22 L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method

Publications (2)

Publication Number Publication Date
CN102307004A CN102307004A (en) 2012-01-04
CN102307004B true CN102307004B (en) 2013-10-30

Family

ID=45380827

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110242200 Expired - Fee Related CN102307004B (en) 2011-08-22 2011-08-22 L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method

Country Status (1)

Country Link
CN (1) CN102307004B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4020782A1 (en) * 2020-12-23 2022-06-29 ABB Schweiz AG Determining parameters of a filter circuit

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103001246A (en) * 2012-11-12 2013-03-27 华中科技大学 Method for controlling a PWM (pulse width modulation) rectifying type energy feeding device on basis of virtual flux linkage
CN103954875B (en) * 2013-12-26 2017-01-18 深圳市英威腾电气股份有限公司 LCL-filtering-based controllable-rectifier-transformer leakage inductance value identification method
CN105720580A (en) * 2014-12-02 2016-06-29 国家电网公司 Method and device for suppressing harmonic effect of inverter LCL filter
CN106921163B (en) * 2015-12-25 2019-08-13 北京计算机技术及应用研究所 Active Power Filter-APF electric network impedance adaptive approach
CN109742963B (en) * 2018-12-12 2020-10-23 西南交通大学 Single-phase pulse rectifier power grid voltage estimation method
CN110456165B (en) * 2019-07-04 2021-09-10 哈尔滨工程大学 Direct current converter inductance and capacitance parameter identification circuit
CN111313733A (en) * 2020-04-01 2020-06-19 南通大学 Low-parasitic-oscillation two-level SiC MOSFET H-bridge main circuit topology
CN113315443A (en) * 2021-05-14 2021-08-27 苏州汇川联合动力系统有限公司 Three-electric-system resonance point identification method, motor controller and storage medium
CN113285481B (en) * 2021-05-25 2022-07-12 山东大学 Grid-connected converter inductance parameter online estimation method, prediction control method and system

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
基于LCL滤波器的电压源型PWM整流器控制策略综述;孙蔚等;《电工技术学报》;20080131;第23卷(第1期);62-67 *
基于虚拟磁链的超导储能装置直接功率控制策略;张辉等;《西安理工大学学报》;20091231;第25卷(第1期);90-96 *
孙蔚等.基于LCL滤波器的电压源型PWM整流器控制策略综述.《电工技术学报》.2008,第23卷(第1期),
张辉等.基于虚拟磁链的超导储能装置直接功率控制策略.《西安理工大学学报》.2009,第25卷(第1期),

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4020782A1 (en) * 2020-12-23 2022-06-29 ABB Schweiz AG Determining parameters of a filter circuit
WO2022136141A1 (en) * 2020-12-23 2022-06-30 Abb Schweiz Ag Determining parameters of a filter circuit

Also Published As

Publication number Publication date
CN102307004A (en) 2012-01-04

Similar Documents

Publication Publication Date Title
CN102307004B (en) L-capacitance-L (LCL)-filtering-based controlled rectifier parameter identification method
CN108616141B (en) Control method for LCL grid-connected inverter power nonlinearity in microgrid
CN106655947B (en) A kind of permanent magnet synchronous motor control algolithm improving low capacity dc-link capacitance voltage transient stability
CN103683319B (en) Based on the control method of grid-connected inverter that stagnant ring is modulated during unbalanced source voltage
CN103326611B (en) A kind of prediction direct Power Control method of three-phase voltage source type PWM converter
CN105375804A (en) Model prediction current control method based on NPC topology grid connected inverter under asymmetric voltage
CN103036462A (en) Model prediction control method of voltage source type rectifier when network voltage is unbalanced
CN106130381B (en) The control method of power feedforward prediction Direct Power based on Virtual shipyard orientation
CN109217698B (en) A kind of double-closed-loop control method based on traditional VSR closed-loop current control
CN110034690B (en) Vienna rectifier model prediction virtual flux linkage control method
CN104167941B (en) Three-phase PWM rectifier control method
CN103944428B (en) A kind of control method of the Three-Phase PWM Rectifier being applicable to waveform distortion of the power supply network
WO2017024596A1 (en) Intelligent cooperative control system and method for multi-unit permanent magnet synchronous motor
CN106877408A (en) Improve the method that T-shaped three level permanent magnet direct-drive wind power system predicts Direct Power
CN102055205A (en) Voltage unbalanced mathematical model-based control strategy for brushless double-feed motor grid-side converter
CN108879775B (en) Power grid unbalanced photovoltaic inverter coordination control method considering current limit value
CN102916596A (en) Input and output power resonance control method of PWM (pulse width modulation) rectifier under voltage unsymmetrical fault
CN103326399A (en) Grid-connected inverter control method under unbalanced and harmonic wave power grids
CN103595279A (en) Photovoltaic inverter fixed-frequency-type model prediction control method in power grid asymmetric fault
CN203339701U (en) Multi-target loss reduction device
Das et al. State of art control techniques for DSTATCOM
CN105743367A (en) Dead beat control method of pulse width modulation (PWM) rectifier under unbalanced power grid voltage
CN107204727B (en) A kind of low capacity thin-film capacitor permanent magnet synchronous motor direct-axis current given controled method
CN104701845A (en) PWM rectifier direct power control method under unbalanced power grid based on virtual magnetic linkage
CN102969716A (en) Method for controlling active power filter of marine electrical network

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20131030

Termination date: 20170822