CN104967353B - A kind of off-network type wind power generation inverter - Google Patents
A kind of off-network type wind power generation inverter Download PDFInfo
- Publication number
- CN104967353B CN104967353B CN201510402092.0A CN201510402092A CN104967353B CN 104967353 B CN104967353 B CN 104967353B CN 201510402092 A CN201510402092 A CN 201510402092A CN 104967353 B CN104967353 B CN 104967353B
- Authority
- CN
- China
- Prior art keywords
- moment
- inverter
- input
- output
- output voltage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/76—Power conversion electric or electronic aspects
Abstract
A kind of off-network type wind power generation inverter, rectifier, storage capacitor, anti exciting converter, power frequency inverter bridge and LCL filter that composition includes MPU controllers and is sequentially connected between wind-driven generator and load, the MPU controllers gather storage capacitor voltage and inverter output voltage respectively by first voltage sensor and second voltage sensor, the power frequency inverter bridge side inductive current of LCL filter is gathered by current sensor, and anti exciting converter and power frequency inverter bridge are controlled respectively by two drive modules.The present invention controls the output voltage of inverter using the composite controller that neural network contrary modeling and PI controllers are constituted.In near zero-crossing point based on neural network contrary modeling control, to give full play to the advantage of its fast response time；Near line voltage peak value based on PI controls, the advantage good to give full play to its steady-state behaviour, so as to effectively increase the quality of power supply of off-network type wind generator system.
Description
Technical field
The present invention relates to a kind of off-network type wind power generation inversion device based on complex controll, belong to technical field of power generation.
Background technology
Reserves are big, clean the characteristic such as renewable because having for wind-power electricity generation, it has also become the Main way of current generation of electricity by new energy
One of.Off-network type wind generator system have low cost, it is flexible for installation the characteristics of, be widely used in remote mountain areas, island etc. big
Power network cannot provide the area of electric power, and add the simple structure of inverter using minitype permanent magnetism direct driving motor more.With grid type wind
Unlike force generating system, off-network type wind generator system must implement effective inverter output voltage control, to ensure
The stable operation of system.
Use simple PI controllers more existing off-network type blower fan inverter, there is harmonic wave of output voltage content high, anti-
Disturb the shortcoming of ability.Can to a certain extent suppress harmonic wave by configuring multiple-stage filtering unit, but the method is caused
System output impedance order is too high, is easily caused system unstable.Modern society to quality of power supply requirement more and more higher,
The controlled level of off-network type blower fan inverter must be improved constantly.
The content of the invention
Drawback it is an object of the invention to be directed to prior art, there is provided a kind of off-network type wind power generation inverter, to have
Effect improves the quality of power supply of off-network type wind generator system.
Problem of the present invention is realized with following technical proposals：
A kind of off-network type wind power generation inverter, composition includes MPU controllers and is sequentially connected to wind-driven generator
Rectifier, storage capacitor, anti exciting converter, power frequency inverter bridge and LCL filter between load, the MPU controllers lead to
Cross first voltage sensor and second voltage sensor gathers storage capacitor voltage and inverter output voltage respectively, by electric current
Sensor gathers the power frequency inverter bridge side inductive current of LCL filter, and controls inverse-excitation converting respectively by two drive modules
Device and power frequency inverter bridge, the MPU controllers are operated in the following manner：
1. storage capacitor voltage V is gathered by first voltage sensord, inverter is gathered by second voltage sensor defeated
Go out voltage Vo, the power frequency inverter bridge side inductive current i of LCL filter is gathered by current sensorL；Using inverter as object,
Corresponding neural network contrary modeling is set up and trains, it is specific as follows：
A. the inversion model that three layers of BP neural network set up system is chosen, wherein, input layer nodes are 9, hidden
The number of neuron node containing layer is 10, and output layer neuron node number is 1, and hidden layer neuron transfer function is using hyperbolic just
Function is cut, output layer neuron transfer function uses S type functions；
B. service data is gathered on inverter；
So that blower fan is respectively at rated wind speed, 80% rated wind speed, 60% rated wind speed, 40% rated wind speed and 20%
Under conditions of rated wind speed, cause that inverter works in rated power, 80% respectively by changing load under every kind of wind friction velocity
The operating mode of rated power, 60% rated power, 40% rated power and 20% rated power, altogether 25 kinds of operating modes, every kind of operating mode
Under carry out inverter output voltage control, and continuous acquisition N group storage capacitor voltages V using single loop PI controllersd, inversion
Device output voltage Vo, anti exciting converter pwm signal dutycycle D, the power frequency inverter bridge side inductive current i of LCL filterL, constitute altogether
25N group service datas；
C. using the service data construction sample of collection；
For the k moment, the sample for being constructed is { Vo(k+1), Vo(k), Vo(k-1), Vo(k-2), iL(k), iL(k-1), D
(k-1), D (k-2), Vd(k), D (k) }, wherein, Vo(k+1) it is the inverter output voltage at k+1 moment, VoK () is the k moment
Inverter output voltage, Vo(k-1) it is the inverter output voltage at k-1 moment, Vo(k-2) it is the inverter output electricity at k-2 moment
Pressure, iLK () is the power frequency inverter bridge side inductive current of the LCL filter at k moment, iL(k-1) it is the LCL filter at k-1 moment
Power frequency inverter bridge side inductive current, D (k-1) is the anti exciting converter pwm signal dutycycle at k-1 moment, and D (k-2) is the k-2 moment
Anti exciting converter pwm signal dutycycle, VdK () is the storage capacitor voltage at k moment, D (k) is the anti exciting converter at k moment
Pwm signal dutycycle, 25N groups service data constitutes 25N sample altogether；
D. 20N in 25N sample data is randomly selected as training sample, and remaining 5N used as test samples, instruction
Practice the three layers of BP neural network set up；In training process, for the sample at k moment, first input of BP neural network is taken
It is the inverter output voltage V at k momentoK (), second input is the inverter output voltage V at k-1 momento(k-1) the, the 3rd
Individual input is the inverter output voltage V at k-2 momento(k-2), the 4th input is the power frequency of the LCL filter at k moment
Inverter bridge side inductive current iLK (), the 5th input is the power frequency inverter bridge side inductive current i of the LCL filter at k-1 momentL
(k-1), the 6th input is the inverter output voltage V at k+1 momento(k+1), the 7th input is the energy storage at k moment
Capacitance voltage VdK (), the 8th input is anti exciting converter pwm signal dutycycle D (k-1) at k-1 moment, the 9th input
Anti exciting converter pwm signal dutycycle D (k-2) for the k-2 moment is held, BP neural network is output as the inverse-excitation converting at k moment
Device pwm signal dutycycle D (k)；
2. inverter is controlled using the BP neural network inversion model and PI controllers that train, is concretely comprised the following steps：
A. the calculating of anti exciting converter pwm signal dutycycle is carried out using the BP neural network inversion model for training, for
At the current k moment, take the inverter output voltage V that first input is the k momentoK (), second input is the k-1 moment
Inverter output voltage Vo(k-1), the 3rd input is the inverter output voltage V at k-2 momento(k-2), the 4th input
Hold the power frequency inverter bridge side inductive current i for the LCL filter at k momentLK (), the 5th input is filtered for the LCL at k-1 moment
The power frequency inverter bridge side inductive current i of ripple deviceL(k-1), the 6th input is the inverter output voltage setting value at k+1 moment
V* o(k+1), the 7th input is the storage capacitor voltage V at k momentdK (), the 8th input becomes for the flyback at k-1 moment
Parallel operation pwm signal dutycycle D (k-1), the 9th input is anti exciting converter pwm signal dutycycle D (k-2) at k-2 moment,
Show that BP neural network is output as d (k)；
B. by subsequent time inverter output voltage setting value V* o(k+1) with current k moment inverter output voltage Vo(k)
Deviation feeding PI controllers, obtain closed-loop control amount dC(k), i.e.,
Wherein：E (k)=V* o(k+1)-Vo(k), KpRepresent proportionality coefficient, KiRepresent integral coefficient；
C. neural network contrary modeling and PI controllers are constituted into composite controller, it is output as the defeated of neural network contrary modeling
Go out the output d of d (k) and PI controllersC(k) proportionally K:(1-K) is superimposed, and is accounted for as final anti exciting converter pwm signal
It is empty than D (k), i.e.,：
D (k)=Kd (k)+(1-K) dC(k)
Wherein, the value of K is：
θ is the phase of inverter output voltage setting value in formula.
Above-mentioned off-network type wind power generation inverter, the training to BP neural network is calculated using the error-duration model for becoming learning rate
Method, specially：The weight coefficient and threshold value in each layer neuron are initialized as the random quantity between (- 1 ,+1) first, then will be defeated
Enter input of 9 neuron output value of layer as each neuron of hidden layer, 10 neuron output values of hidden layer are used as defeated
Go out the input of the single neuron of layer, the output valve of the single neuron of output layer is network final output, for k-th training sample
This, calculates network reality output after the t times trainingWith desired outputError：
And the error of sum square of 20N training sample is：
Often calculate one time, compared EtWith anticipation error EoIf, Et<Eo, then train end condition to meet, training terminates；It is no
Then, by EtBackpropagation is carried out along access path, and adjusts the weight coefficient and threshold value of each layer one by one, until Et<EoUntill；Then,
Model is tested using 5N test samples, sample standard deviation side error delta MSE is calculated, if there is Δ MSE<(Tr is a certain solid to Tr
Determine threshold value), then it is assumed that the BP neural network model trained meets required precision, now record each neuron weight coefficient and
Threshold value；Otherwise, BP neural network model is trained again, untill it meets required precision.
Above-mentioned off-network type wind power generation inverter, the anti exciting converter include transformer, the first power tube, diode and
Second electric capacity, the primary coil of the transformer connects storage capacitor voltage through the first power tube, and its secondary coil is given through diode
Power frequency inverter bridge is powered, and the second electric capacity is connected to the direct-flow input end of power frequency inverter bridge, and the grid of the first power tube connects the first drive
The output end of dynamic model block.
Above-mentioned off-network type wind power generation inverter, the power frequency inverter bridge is the fully controlled bridge being connected into by four power tubes, four
The grid of individual power tube connects the output end of the second drive module.
Above-mentioned off-network type wind power generation inverter, the LCL filter include power frequency inverter bridge side inductance, the 3rd electric capacity and
Load-side inductance, power frequency inverter bridge side inductance one terminates the first ac output end of power frequency inverter bridge, and other end connects electric current biography
The positive pole of sensor, the positive pole of the 3rd electric capacity connects the negative pole of current sensor, and the negative pole of the 3rd electric capacity connects the second of power frequency inverter bridge
Ac output end, the measurement signal output end of current sensor is connected with MPU controllers, and load-side inductance one terminates the 3rd electric capacity
Positive pole, other end connects the positive pole of load.
The present invention controls the output of inverter using the composite controller that neural network contrary modeling and PI controllers are constituted
Voltage.In near zero-crossing point, the set point change rate of inverter output voltage is maximum, is now controlled to neural network contrary modeling
It is main, to give full play to the advantage of its fast response time；Near line voltage peak value, the set point change of inverter output voltage
Rate is minimum, now based on PI controls, the advantage good to give full play to its steady-state behaviour, so as to effectively increase off-network type wind-force
The quality of power supply of electricity generation system.
Brief description of the drawings
The invention will be further described below in conjunction with the accompanying drawings.
Fig. 1 is the structured flowchart of wind power generation inverter device；
Fig. 2 is neural network contrary modeling schematic diagram；
Fig. 3 is the structural representation of single neuron；
Fig. 4 is wind power generation inverter output voltage curve.
Each list of reference numerals is in figure：GS, blower fan, UT1, first voltage sensor, UT2, second voltage sensor, CT, electricity
Flow sensor, D1, diode, Q1~Q5, the first power tube~the 5th power tube, C1, storage capacitor, C2, the second electric capacity, C3,
3rd electric capacity, L1, power frequency inverter bridge side inductance, L2, load-side inductance, TX1, transformer.
Each symbol inventory is in text：VdIt is storage capacitor voltage, VoIt is inverter output voltage, iLIt is LCL filter power frequency
Inverter bridge side inductive current, D is anti exciting converter pwm signal dutycycle, VdK () is the storage capacitor voltage at k moment, VoK () is
The inverter output voltage at k moment, Vo(k-1) it is the inverter output voltage at k-1 moment, VO(k-2) it is the inversion at k-2 moment
Device output voltage, iLK () is the LCL filter power frequency inverter bridge side inductive current at k moment, iL(k-1) for the LCL at k-1 moment is filtered
Ripple device power frequency inverter bridge side inductive current, Vo(k+1) it is the inverter output voltage at k+1 moment, V* o(k+1) it is the k+1 moment
Inverter output voltage setting value, D (k-1) is the anti exciting converter pwm signal dutycycle at k-1 moment, and D (k-2) is the k-2 moment
Anti exciting converter pwm signal dutycycle, D (k) for the k moment anti exciting converter pwm signal dutycycle,After t training
Network reality output,It is desired output, E after t trainingoIt is anticipation error, EtIt is 20N training sample after t training
Error of sum square.
Specific embodiment
Wind power generation inverter of the present invention includes：Rectifier, storage capacitor C1, anti exciting converter are (by transformer
TX1, the first power tube Q1, diode D1 and the second electric capacity C2 are constituted), power frequency inverter bridge is (by the second power tube Q2~the 5th work(
Rate pipe Q5 constitute), LCL filter (being made up of the 3rd electric capacity C3, the first inductance L1, the second inductance L2), the first drive module (i.e.
Drive module 1 in Fig. 1), the second drive module (i.e. drive module 2 in Fig. 1), MPU controllers, first voltage sensor
UT1, second voltage sensor UT2 and current sensor CT.
The three-phase input end of the rectifier is connected with the three-phase output end of blower fan, single-phase output anode and the storage of rectifier
The positive pole connection of energy electric capacity, the single-phase output negativing ending grounding of rectifier；Storage capacitor negative pole is grounded；First voltage sensor is treated
Survey control source anode to be connected with storage capacitor positive pole, the control source negativing ending grounding to be measured of first voltage sensor, the first electricity
The measurement signal output end of pressure sensor is connected with MPU controllers；The input positive terminal of anti exciting converter connects with storage capacitor positive pole
Connect, the input negative terminal ground connection of anti exciting converter；The output plus terminal of anti exciting converter is connected with the input of power frequency inverter bridge one, flyback
The output negative terminal of converter is connected with another input of power frequency inverter bridge；One output end of power frequency inverter bridge and LCL filter one
Input is connected, and another output end of power frequency inverter bridge is connected with another input of LCL filter；One output of LCL filter
End is connected with the control source anode to be measured of second voltage sensor, and another output end and the second voltage of LCL filter are sensed
The control source negative terminal to be measured connection of device；The positive pole of current sensor is connected with power frequency inverter bridge side inductance other end, electric current
The negative pole of sensor is connected with the positive pole of the 3rd electric capacity, and the measurement signal output end of current sensor is connected with MPU controllers；The
The input of one drive module is connected with MPU controllers, the first work(in the output end and anti exciting converter of the first drive module
Rate tube grid is connected；The input of the second drive module is connected with MPU controllers, an output end and power frequency of the second drive module
The grid connection of the second power tube, the 5th power tube in inverter bridge, another output end and the power frequency inversion of the second drive module
The grid connection of the 3rd power tube, the 4th power tube in bridge；MPU controllers are sensed with first voltage sensor, second voltage
Device, current sensor measurement signal output end connection, MPU controllers also with the first drive module, the second drive module it is defeated
Enter end connection；Load positive pole connects load-side inductance other end, and load negative pole connects the negative pole of the 3rd electric capacity.
The present invention is further illustrated below in conjunction with the drawings and specific embodiments.
In the present embodiment, blower fan GS is minitype permanent magnetism direct-drive aerogenerator, and rotor diameter is 1.2m, rated power
It is 300W, rated voltage is 24V, rated speed 800r/min, threshold wind velocity 1m/s, rated wind speed 10m/s.
Inverter structure block diagram for the wind generator design is as shown in Figure 1.The device mainly includes：Rectifier,
Storage capacitor C1, anti exciting converter, power frequency inverter bridge, LCL filter, the first drive module, the second drive module, MPU controls
Device, first voltage sensor UT1, second voltage sensor UT2, current sensor CT；Wherein MPU controllers are from TI companies
High performance float-point digital signal processor TMS320F28335, the digital signal processor be integrated with I/O modules, ADC,
The functional modules such as PWM module, CAN module, UART and SPI modules；Storage capacitor C1 selects 2200uF electrochemical capacitors；Inverse-excitation converting
Transformer TX1 uses NA5814-AL, the first power tube Q1 to select TK50X15J1, diode D1 to use C2D05120E in device, the
Two electric capacity C2 select 0.47uF polarity free capacitors；Second power tube Q2, the 3rd power tube Q3, the 4th power tube Q4, the 5th power
Pipe Q5 selects IPB60R190C6, and the first drive module is MCP14E4 with the second drive module；Power frequency inversion in LCL filter
Bridge side inductance L1 selects 0.5mH inductance, load-side inductance L2 to select 0.1mH inductance, and the 3rd electric capacity C3 is nonpolarity from 0.47uf
Electric capacity, first voltage sensor UT1 uses MCP6022 with second voltage sensor；Current sensor CT uses Hall current
Sensors A CS712ELCTR-058-1.
Fig. 2 is three layers of set up BP neural network inversion model.Input layer nodes are 9, hidden layer nerve
First nodes are 10, and output layer neuron node number is 1.
Fig. 3 is the structural representation of single neuron, x0,x1,…xiIt is the input signal of neuron, wijIt is neuron
Weight coefficient, θjIt is the threshold value of neuron, then neuron is output as：
Hidden layer neuron transfer function uses hyperbolic tangent function, i.e.,
Wherein：
X is input into for transfer function；
F (x) is exported for transfer function.
Output layer neuron transfer function uses S type functions, i.e.,
Wherein：
X is input into for transfer function；
F (x) is exported for transfer function.
So that blower fan is respectively at rated wind speed (10m/s), 80% rated wind speed (8m/s), 60% rated wind speed (6m/
S), under conditions of 40% rated wind speed (4m/s) and 20% rated wind speed (2m/s), loaded by changing under every kind of wind friction velocity
So that inverter works in rated power (300W), 80% rated power (240W), 60% rated power (180W), 40% respectively
The operating mode of rated power (120W) and 20% rated power (60W), 25 kinds of operating modes, utilize single loop PI altogether under every kind of operating mode
Controller carries out inverter output voltage control, with the sample frequency continuous acquisition N=4000 group storage capacitor voltages of 40KHz
Vd, inverter output voltage Vo, anti exciting converter pwm signal dutycycle D, the power frequency inverter bridge side inductive current of LCL filter
iL, 25N=100000 sample is constituted altogether；80000 therein are randomly selected as training sample, remaining 20000 conduct
Test samples.
E is set in embodimentoAfter=4.0, threshold value Tr=0.02, repetitive exercise 15422 times, Et=3.92 meets and terminates
Condition Et<Eo, training terminates.Now, recycle test samples to test model, calculate sample standard deviation side error delta MSE
=0.014, there is Δ MSE<Tr, it is taken as that meeting inversion model required precision, now records the weight coefficient and threshold of each neuron
Value.
The parameter tuning of PI controllers is K in embodimentp=0.02, Ki=1200；Inverter output voltage setting value V* o
It is given according to the change of discrete sine rule by digital signal processor.Software program is write according to above-mentioned steps and be implanted into DSP cores
In piece, you can realize the blower fan inverter output voltage complex controll with PI controllers based on neural network contrary modeling.
Fig. 4 is the output voltage curve using control method rear fan inverter of the present invention, it is seen that output voltage is in zero passage
Point and peak value nearby shake without obvious, and with sine degree higher.
In order to further analyze the effect of control method of the present invention, respectively by itself and conventional PI control side under different operating modes
Method (takes Kp=0.02, Ki=1200) contrasted, table 1 is the inverter output voltage harmonic content contrast knot of two methods
Really, it is known that the inventive method has preferable robustness for the change of operating mode, the inverter output voltage under different operating modes is humorous
Ripple content is significantly less than conventional PI control method.
Table 1：The Contrast on effect of composite control method of the present invention and PI control methods
The above, the only present invention preferably specific embodiment, but protection scope of the present invention should not limit to
In this, any one skilled in the art according to the invention discloses core technology, the change to be made or replace
Change, should all be included within the scope of the present invention.
Claims (5)
1. a kind of off-network type wind power generation inverter, it is characterized in that, composition includes MPU controllers and is sequentially connected to wind-force
Rectifier, storage capacitor (C1), anti exciting converter, power frequency inverter bridge and LCL filter between generator and load, it is described
MPU controllers gather storage capacitor voltage and inverse respectively by first voltage sensor (UT1) and second voltage sensor (UT2)
Become device output voltage, the power frequency inverter bridge side inductive current of LCL filter is gathered by current sensor (CT), and by two
Drive module controls anti exciting converter and power frequency inverter bridge respectively, and the MPU controllers are operated in the following manner：
1. storage capacitor voltage V is gathered by first voltage sensor (UT1)d, gather inverse by second voltage sensor (UT2)
Become device output voltage Vo, the power frequency inverter bridge side inductive current i of LCL filter is gathered by current sensor (CT)L；With inversion
Corresponding neural network contrary modeling is set up and trained to device as object, specific as follows：
A. the inversion model that three layers of BP neural network set up system is chosen, wherein, input layer nodes are 9, hidden layer
Neuron node number is 10, and output layer neuron node number is 1, and hidden layer neuron transfer function uses tanh letter
Number, output layer neuron transfer function uses S type functions；
B. service data is gathered on inverter；
So that blower fan to be respectively at rated wind speed, 80% rated wind speed, 60% rated wind speed, 40% rated wind speed and 20% specified
Under conditions of wind speed, cause that inverter works in rated power, 80% specified respectively by changing load under every kind of wind friction velocity
The operating mode of power, 60% rated power, 40% rated power and 20% rated power, 25 kinds of operating modes, equal under every kind of operating mode altogether
Inverter output voltage control, and continuous acquisition N group storage capacitor voltages V are carried out using single loop PI controllersd, inverter it is defeated
Go out voltage Vo, anti exciting converter pwm signal dutycycle D, the power frequency inverter bridge side inductive current i of LCL filterL, 25N is constituted altogether
Group service data；
C. using the service data construction sample of collection；
For the k moment, the sample for being constructed is { Vo(k+1), Vo(k), Vo(k-1), Vo(k-2), iL(k), iL(k-1), D (k-
1), D (k-2), Vd(k), D (k) }, wherein, Vo(k+1) it is the inverter output voltage at k+1 moment, VoK () is the inversion at k moment
Device output voltage, Vo(k-1) it is the inverter output voltage at k-1 moment, Vo(k-2) it is the inverter output voltage at k-2 moment,
iLK () is the power frequency inverter bridge side inductive current of the LCL filter at k moment, iL(k-1) for the k-1 moment LCL filter work
Frequency inverter bridge side inductive current, D (k-1) is the anti exciting converter pwm signal dutycycle at k-1 moment, and D (k-2) is the k-2 moment
Anti exciting converter pwm signal dutycycle, VdK () is the storage capacitor voltage at k moment, D (k) is the anti exciting converter PWM at k moment
Signal dutyfactor, 25N groups service data constitutes 25N sample altogether；
D. 20N in 25N sample data is randomly selected as training sample, and remaining 5N is individual as test samples, and training is built
Three layers of vertical BP neural network；In training process, for the sample at k moment, when first input for taking BP neural network is k
The inverter output voltage V at quarteroK (), second input is the inverter output voltage V at k-1 momento(k-1), the 3rd it is defeated
It is the inverter output voltage V at k-2 moment to enter endo(k-2), the 4th input is the power frequency inversion of the LCL filter at k moment
Bridge side inductive current iLK (), the 5th input is the power frequency inverter bridge side inductive current i of the LCL filter at k-1 momentL(k-
1), the 6th input is the inverter output voltage V at k+1 momento(k+1), the 7th input is the storage capacitor at k moment
Voltage VdK (), the 8th input is anti exciting converter pwm signal dutycycle D (k-1) at k-1 moment, the 9th input is
Anti exciting converter pwm signal dutycycle D (k-2) at k-2 moment, BP neural network is output as the anti exciting converter PWM at k moment
Signal dutyfactor D (k)；
2. inverter is controlled using the BP neural network inversion model and PI controllers that train, is concretely comprised the following steps：
A. the calculating of anti exciting converter pwm signal dutycycle is carried out using the BP neural network inversion model for training, for current k
At the moment, take the inverter output voltage V that first input is the k momentoK (), second input is the inverter at k-1 moment
Output voltage Vo(k-1), the 3rd input is the inverter output voltage V at k-2 momento(k-2) when, the 4th input is k
The power frequency inverter bridge side inductive current i of the LCL filter at quarterLK (), the 5th input is the work of the LCL filter at k-1 moment
Frequency inverter bridge side inductive current iL(k-1), the 6th input is the inverter output voltage setting value V at k+1 moment* o(k+1),
7th input is the storage capacitor voltage V at k momentdK (), the 8th input is believed for the anti exciting converter PWM at k-1 moment
Number dutycycle D (k-1), the 9th input is anti exciting converter pwm signal dutycycle D (k-2) at k-2 moment, draws BP god
D (k) is output as through network；
B. by subsequent time inverter output voltage setting value V* o(k+1) with current k moment inverter output voltage Vo(k) it is inclined
Difference feeding PI controllers, obtain closed-loop control amount dC(k), i.e.,
Wherein：E (k)=V* o(k+1)-Vo(k), KpRepresent proportionality coefficient, KiRepresent integral coefficient；
C. neural network contrary modeling and PI controllers are constituted into composite controller, its output d for being output as neural network contrary modeling
The output d of (k) and PI controllersC(k) proportionally K:(1-K) is superimposed, used as final anti exciting converter pwm signal dutycycle
D (k), i.e.,：
D (k)=Kd (k)+(1-K) dC(k)
Wherein, the value of K is：
θ is the phase of inverter output voltage setting value in formula.
2. off-network type wind power generation inverter according to claim 1, it is characterized in that, the training to BP neural network is used
Become the error back propagation algorithm of learning rate, specially：The weight coefficient and threshold value in each layer neuron are initialized as first (- 1 ,+
1) random quantity between, then using 9 neuron output values of input layer as each neuron of hidden layer input, the 10 of hidden layer
Individual neuron output value as the single neuron of output layer input, it is final that the output valve of the single neuron of output layer is network
Output, for k-th training sample, calculates network reality output after the t times trainingWith desired outputError：
And the error of sum square of 20N training sample is：
Often calculate one time, compared EtWith anticipation error EoIf, Et<Eo, then train end condition to meet, training terminates；Otherwise, will
EtBackpropagation is carried out along access path, and adjusts the weight coefficient and threshold value of each layer one by one, until Et<EoUntill；Then, utilize
5N test samples are tested model, sample standard deviation side error delta MSE are calculated, if there is Δ MSE<Tr, Tr are a certain fixed threshold
Value, then it is assumed that the BP neural network model trained meets required precision, now records the weight coefficient and threshold value of each neuron；
Otherwise, BP neural network model is trained again, untill it meets required precision.
3. off-network type wind power generation inverter according to claim 1 and 2, it is characterized in that, the anti exciting converter includes
Transformer (TX1), the first power tube (Q1), diode (D1) and the second electric capacity (C2), the primary coil of the transformer (TX1)
Storage capacitor voltage is connect through the first power tube (Q1), its secondary coil is powered through diode (D1) to power frequency inverter bridge, the second electricity
Hold the direct-flow input end that (C2) is connected to power frequency inverter bridge, the grid of the first power tube (Q1) connects the output of the first drive module
End.
4. off-network type wind power generation inverter according to claim 3, it is characterized in that, the power frequency inverter bridge is by four
The fully controlled bridge that power tube is connected into, the grid of four power tubes connects the output end of the second drive module.
5. off-network type wind power generation inverter according to claim 4, it is characterized in that, the LCL filter includes power frequency
Inverter bridge side inductance (L1), the 3rd electric capacity (C3) and load-side inductance (L2), the termination power frequency of power frequency inverter bridge side inductance (L1)
First ac output end of inverter bridge, other end connects the positive pole of current sensor (CT), and the positive pole of the 3rd electric capacity (C3) connects electricity
The negative pole of flow sensor (CT), the negative pole of the 3rd electric capacity (C3) connects the second ac output end of power frequency inverter bridge, current sensor
(CT) measurement signal output end is connected with MPU controllers, and load-side inductance (L2) terminates the positive pole of the 3rd electric capacity (C3), separately
The positive pole of an outer terminating load.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510402092.0A CN104967353B (en) | 2015-07-08 | 2015-07-08 | A kind of off-network type wind power generation inverter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510402092.0A CN104967353B (en) | 2015-07-08 | 2015-07-08 | A kind of off-network type wind power generation inverter |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104967353A CN104967353A (en) | 2015-10-07 |
| CN104967353B true CN104967353B (en) | 2017-06-16 |
Family
ID=54221325
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510402092.0A Expired - Fee Related CN104967353B (en) | 2015-07-08 | 2015-07-08 | A kind of off-network type wind power generation inverter |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104967353B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108075482A (en) * | 2016-11-11 | 2018-05-25 | 中科诺维(北京)科技有限公司 | Permanent magnet direct-drive wind turbine gird-connected inverter device and control method |
| CN108075481A (en) * | 2016-11-11 | 2018-05-25 | 中科诺维(北京)科技有限公司 | Wind-power electricity generation Miniature inverter device and control method |
| CN108173279A (en) * | 2016-12-08 | 2018-06-15 | 华能新能源股份有限公司辽宁分公司 | The soft grid-connection control device and control method of permanent magnet direct-drive wind turbine |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5224025A (en) * | 1992-04-21 | 1993-06-29 | Wisconsin Alumni Research Foundation | Forward converter with two active switches and unity power factor capability |
| GB2497275A (en) * | 2011-11-25 | 2013-06-12 | Enecsys Ltd | Modular adjustable power factor renewable energy inverter system |
| CN202524301U (en) * | 2012-05-02 | 2012-11-07 | 浙江日风电气有限公司 | Grid-connected inverter of wind-driven generator |
| CN103887955B (en) * | 2014-04-08 | 2016-08-17 | 盐城工学院 | The combining inverter of suppression fuel cell output low frequency current ripples and control device |
-
2015
- 2015-07-08 CN CN201510402092.0A patent/CN104967353B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN104967353A (en) | 2015-10-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104806450B (en) | A kind of wind power system MPPT control method based on gravitation neutral net | |
| CN104967353B (en) | A kind of off-network type wind power generation inverter | |
| CN103166489B (en) | A kind of control circuit of Three-Phase SVPWM Rectifier | |
| CN104158418B (en) | A kind of photovoltaic network inverter | |
| CN107069807A (en) | Containing uncertain budget adjust without balance nodes microgrid Robust Scheduling method | |
| CN103168403A (en) | Controlling a wind power plant transformer | |
| Yan et al. | General energy filters for power smoothing, tracking and processing using energy storage | |
| CN107425736A (en) | Multi-end flexible direct current transmission system current conversion station control method based on model prediction | |
| CN104333034B (en) | Grid-connected Miniature inverter based on SVMs Adverse control | |
| CN105870974A (en) | Bidirectional measurement device of electric quantity parameters of distributed photovoltaic grid-connected system | |
| CN204103501U (en) | A kind of distributed energy network | |
| CN104242336A (en) | New energy power station equivalent modeling method based on interaction coefficient of power generating units | |
| CN205960662U (en) | Two -way measuring device of distributing type grid -connected PV electric quantity parameter of system | |
| CN106100398A (en) | A kind of micro-capacitance sensor energy storage inverter control method based on gravitation neutral net | |
| CN203151389U (en) | Control circuit of three-phase high power factor rectifier | |
| CN104795842B (en) | Symmetrical fault ride-through control method for hybrid wind farm group containing double-fed wind farm and permanent magnetic direct drive wind farm | |
| CN104377718B (en) | A kind of active parallel-connection type mixing energy storing system and method for work thereof | |
| CN107766596A (en) | A kind of low voltage ride-through capability method of tire based on typical fault operating mode collection | |
| CN115102200A (en) | Energy storage system model identification method and system | |
| CN104485679B (en) | It is applicable to self-balancing control method and the system of the wind storage integration of distribution | |
| CN108075482A (en) | Permanent magnet direct-drive wind turbine gird-connected inverter device and control method | |
| CN104485685A (en) | Solar and storage integrated self-balancing control method and system applicable to distribution network | |
| CN108448595B (en) | Photovoltaic-series compensation system time domain simulation small disturbance analysis method and system | |
| CN206490441U (en) | A kind of inverter energy-saving control device and control system | |
| Samosir et al. | Design of DC to DC converter for solar photovoltaic power plant applications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Haining Inventor after: Liu Weiliang Inventor after: Yang Libin Inventor after: Wang Yinsong Inventor after: Chen Wenying Inventor after: Liu Changliang Inventor after: Lin Yongjun Inventor after: Ma Liangyu Inventor before: Liu Weiliang Inventor before: Chen Wenying Inventor before: Liu Changliang Inventor before: Lin Yongjun Inventor before: Ma Liangyu |
|
| COR | Change of bibliographic data | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20151222 Address after: The victory road 810008 in Qinghai province Xining City West District No. 89 Applicant after: State Grid Qinghai Electric Power Company Applicant after: North China Electric Power University (Baoding) Applicant after: Electric Power Research Institute of State Grid Qinghai Electric Power Company Address before: 071003 Hebei province Baoding Yonghua No. 619 North Street Applicant before: North China Electric Power University (Baoding) |
|
| 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: 20170616 Termination date: 20180708 |