CN103414359B - Control method when a kind of matrix rectifier input is uneven - Google Patents

Abstract

The present invention be the input of a kind of matrix rectifier uneven time control method, the method real-time monitoring matrix rectifier DC output voltage, and with normal output voltage V _refdo difference and obtain deviation signal, when deviation signal is larger, adopts and control output dc voltage is reached near desired value fast than row P.Then fuzzy control is adopted when deviation signal is less, the deviation variation rate of the deviation of input and input is carried out obfuscation, fuzzy reasoning and decision-making by fuzzy controller, obtain output variable, this output variable obtains space vector modulation coefficient through output quantization factor treatment, in conjunction with input phase current space vector sector angle, the action time of each switch state vector can be obtained, and reach PWM generation unit.The action time of switch state vector is sent to the switch circulation controller of matrix rectifier by PWM generation unit with pwm pulse form, and is determined the turn-on and turn-off signal of each bidirectional switch by this controller, realizes the modulation to matrix rectifier.

Description

Control method when a kind of matrix rectifier input is uneven

Technical field

The present invention relates to electric and electronic technical field, being specifically related to fuzzy than row control method when inputting uneven of matrix rectifier.

Background technology

Matrix rectifier is that a kind of general voltage-dropping type three-phase AC-DC converter can realize real four quadrant running and can produce that to carry out the direct voltage source range of application that amplitude is adjustable, polarity is adjustable more extensive.How to ensure output voltage stabilization when input voltage is uneven, the impact on load equipment under reduction damage is current urgent problem.Adopt traditional PID control to be based upon on the basis of the accurate Mathematical Modeling of object, but the effectiveness comparison of disturbance suppression and the factor such as non-linear is poor, for containing a large amount of non-linear time model very difficult foundation, be difficult to reach control overflow by classical control device.Fuzzy control does not need the accurate Mathematical Modeling of object, can overcome some nonlinear impacts, and some uncertainties and inexactness problem obtain handlability and robustness.Based on this, the present invention proposes control method during a kind of matrix rectifier input imbalance.

Summary of the invention

The object of the present invention is to provide control method when being that the input of a kind of matrix rectifier is uneven, adopt and fuzzyly can keep than row control method that VD is stable improves output waveform quality when matrix rectifier three-phase input voltage damage simultaneously.

The present invention is achieved through the following technical solutions.

Control method when matrix rectifier input is uneven, specifically: monitoring matrix rectifier DC output voltage V in real time ₀, pass through V ₀with matrix rectifier output voltage V during line voltage nominal situation _refdifference obtain deviation signal e, when namely deviation signal e is greater than set point comparatively greatly, adopts and control to make output dc voltage reach near desired value or desired value than row P; When deviation signal is less be namely less than set point time then adopt fuzzy control, the deviation signal e obtained is quantized factor K e process through deviation and obtains fuzzy controller input variable deviation E; Simultaneously by deviation signal differentiate, and obtain fuzzy controller input variable deviation variation rate EC through deviation variation rate factor K ec process; The deviation variation rate EC of the deviation E of input and input is carried out obfuscation, fuzzy reasoning and decision-making by described fuzzy controller, obtain output variable u, this output variable u obtains space vector modulation Coefficient m through the process of output quantization factor Ku, in conjunction with the input phase current space vector sector angle θ calculated, obtain the action time of each switch state vector, and reach PWM generation unit; The action time of switch state vector is sent to the switch circulation controller of matrix rectifier by PWM generation unit with pwm pulse form, and is determined the turn-on and turn-off signal of each bidirectional switch by this switch circulation controller, realizes the modulation to matrix rectifier.

In above-mentioned control method, input phase current space vector sector angle θ is by detecting three-phase input phase voltage V _iphvector, carries out CLARKE conversion (C to phase voltage _3S/2Sconversion), obtain V under rest frame _{i α}, V _{i β}pARK conversion is carried out to it thus obtains phase current space vector sector angle θ.

In above-mentioned control method, under output error varies in size situation, carry out mode selection:

Set an initial modulation and compare m ₀value=expect voltage/466.

When | during e|>=36, modulation ratio m=m ₀, adopting and control than row P, is namely m than row coefficient ₀, make output dc voltage reach near desired value or desired value fast;

When | during e|<36, modulation ratio m=m ₀+ [(1-m ₀)/6] * fuzzy output value, then adopt fuzzy control.

In above-mentioned control method, fuzzy controller comprises defuzzification interface (1), inference machine (2), knowledge base (3) and ambiguity solution interface (4), the input variable deviation E of fuzzy controller and input variable deviation variation rate EC sends to inference machine (2) by defuzzification interface (1) after obfuscation, described knowledge base (3-1) is for storing input and output linguistic variable, described rule base (3-2) comprises fuzzy control rule information, input variable deviation E and the corresponding relation between the higher poor rate of change EC of input variable and fuzzy output is comprised in described fuzzy rule information, this inference machine (2) reads corresponding output language variable according to the input variable E after the obfuscation received and input variable deviation variation rate from database (3-1), read corresponding fuzzy control rule information from rule base (3-2) simultaneously, then according to the input variable deviation E after obfuscation, input variable deviation variation rate EC, input language variable and fuzzy control rule obtain the fuzzy control quantity exported, and this fuzzy control quantity is sent to ambiguity solution interface (4), this fuzzy control quantity exports as fuzzy controller output variable u by described ambiguity solution interface (4).

In above-mentioned control method, the fuzzy set of fuzzy controller input deviation e is that { NB, NS, ZO, PS, PB}, namely { bear large, bear little, zero, be just little, honest }, domain [-6+6]; The fuzzy set of input deviation rate of change ec is also that { fuzzy output value u obtains fuzzy set for { NB, NS, ZO, PS, PB}, domain [-6+6] for NB, NS, ZO, PS, PB}, domain [-6+6]; Wherein NB gets Z-shaped membership function, and PB gets S shape membership function, and all the other get Triangleshape grade of membership function.

In above-mentioned control method, fuzzy control deviation quantizes factor K e=6/36 value 0.166, deviation variation rate factor K ec value 10 ^-13, export than row factor value Ku=0.3/6=0.05.

In above-mentioned control method, fuzzy control rule table reaches for If ... and ... then ..., amount to 16, setting up fuzzy rule base is

(1).If(eisNB)and(ecisZO)then(uisNO)(1)

(2).If(eisNB)and(ecisPS)then(uisNS)(1)

(3).If(eisNB)and(ecisPB)then(uisNB)(1)

(4).If(eisNS)and(ecisZO)then(uisNO)(1)

(5).If(eisNS)and(ecisPS)then(uisNS)(1)

(6).If(eisNS)and(ecisPB)then(uisNB)(1)

(7).If(eisZO)and(ecisNB)then(uisPO)(1)

(8).If(eisZO)and(ecisNS)then(uisPO)(1)

(9).If(eisZO)and(ecisPS)then(uisNO)(1)

(10).If(eisZO)and(ecisPB)then(uisNO)(1)

(11).If(eisPS)and(ecisZO)then(misPO)(1)

(12).If(eisPS)and(ecisPS)then(misPS)(1)

(13).If(eisPS)and(ecisPB)then(uisPB)(1)

(14).If(eisPB)and(ecisZO)then(uisPO)(1)

(15).If(eisPB)and(ecisPS)then(uisPS)(1)

(16).If(eisPB)and(ecisPB)then(uisPB)(1)。

Compared with prior art, tool of the present invention has the following advantages and technique effect: adopt such scheme as matrix rectifier VD V ₀when diminishing because of the power supply of electrical network input voltage imbalance, this control method makes space vector modulation Coefficient m become large, and as VD V ₀when becoming large because of the power supply of electrical network input voltage imbalance, spatial modulation coefficient reduces by fuzzy controller, thus keeps output dc voltage constant.

Accompanying drawing explanation

Fig. 1 is matrix rectifier control method schematic diagram.

Fig. 2 is matrix rectifier topological structure.

Fig. 3 is input phase current three dimensional vector diagram.

Fig. 4 is the synthesis of target current vector.

Fig. 5 is fuzzy control principle block diagram.

Fig. 6 is the structural representation of fuzzy controller.

Fig. 7 is not for adopt feedback average anode current voltage waveform.

Fig. 8 is for adopting feedback average anode current voltage waveform.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation, but enforcement of the present invention and protection are not limited thereto.

Shown in the topological structure composition graphs 2 of matrix rectifier, it is made up of 5 parts: I is net side input voltage; II is input filter; III is the main circuit be made up of 6 bidirectional switchs; IV is output filter; V is load.

Suppose that input voltage is positive phase sequence, any two-phase of matrix rectifier demand fulfillment three-phase input voltage can not short circuit and output current can not the requirement of open circuit, have 9 kinds of switch combinations, as shown in table 1:

Current space vector during table 1 input voltage positive phase sequence

In table 1, first 6 kinds is effective vector, and latter 3 kinds is zero vector.In a PWM switch periods, effective vector act on as requested a period of time for export voltage is provided, the remaining time is zero vector action time, does not have voltage to export, only provides oad freewheels.

Composition graphs 2,3 matrix rectifier input current space vector vector modulation is carried out, with angular velocity omega in complex plane by input current _iuniform rotation, its target current vector (I _immould for input current space vector).In Fig. 2, I ~ VI represents 6 input current sectors, and it carries out partitioning of input voltage interval according to the zero crossing of phase voltage, and carries out the selection of current phasor.As Fig. 4, suppose at a time to need the current phasor synthesized be two current phasors that it is adjacent, advanced one of phase place is one of delayed phase is with angle be θ, and θ ∈ [0,60 °].Will with carry out PWM synthesis, target current vector can be obtained

While realizing input current and output current adjustment, also will ensure to export direct current can not be interrupted.Therefore, passing through with while synthesis, also need a zero vector, composite formula is as follows:

$\stackrel{\&RightArrow;}{I_{\mathrm{ref}}}=\frac{T_{\mathrm{\&alpha;}}\left(\mathrm{\&theta;}\right)}{T_s}\stackrel{\&RightArrow;}{I_{\mathrm{\&alpha;}}}+\frac{T_{\mathrm{\&beta;}}\left(\mathrm{\&theta;}\right)}{T_s}\stackrel{\&RightArrow;}{I_{\mathrm{\&beta;}}}+\frac{T_0\left(\mathrm{\&theta;}\right)}{T_s}\stackrel{\&RightArrow;}{I_0}---\left(1\right)$

T in formula _sbe the time of a PWM regulating cycle, T _α(θ), T _β(θ) and T ₀(θ) be respectively in a PWM regulating cycle with the action time of three current phasors, its duty cycle functions is respectively:

$f_{\mathrm{\&beta;}}\left(\mathrm{\&theta;}\right)=\frac{T_{\mathrm{\&beta;}}\left(\mathrm{\&theta;}\right)}{T_s}=\frac{2}{\sqrt{3}}\frac{\left|\stackrel{\&RightArrow;}{I_{\mathrm{ref}}}\right|}{\left|\stackrel{\&RightArrow;}{I_{\mathrm{\&beta;}}}\right|}\sin \left(\mathrm{\&theta;}\right)=m\&times;\sin \left(\mathrm{\&theta;}\right)---\left(3\right)$

$f_0\left(\mathrm{\&theta;}\right)=\frac{T_0\left(\mathrm{\&theta;}\right)}{T_s}=1-f_{\mathrm{\&alpha;}}\left(\mathrm{\&theta;}\right)-f_{\mathrm{\&beta;}}\left(\mathrm{\&theta;}\right)---\left(4\right)$

In above formula, f _α(θ), f _β(θ) and f ₀(θ) be respectively with duty cycle functions.Input current space vector the movement locus on summit by extremely the orthohexagonal inner circle that summit surrounds.Can obtain according to analysis therefore the index of modulation of electric current is made $m=\frac{2}{\sqrt{3}}\frac{\left|\stackrel{\&RightArrow;}{I_{\mathrm{ref}}}\right|}{\left|\stackrel{\&RightArrow;}{I_{\mathrm{\&alpha;}}}\right|}=\frac{2}{\sqrt{3}}\frac{\left|\stackrel{\&RightArrow;}{I_{\mathrm{ref}}}\right|}{\left|\stackrel{\&RightArrow;}{I_{\mathrm{\&beta;}}}\right|},$ m∈[0,1]。Can prove that the mean value of output voltage in a switch periods is:

Wherein V _imfor inputting the amplitude of phase voltage, for the displacement of the relative input voltage of input current.VD is determined by voltage modulated coefficient, input phase voltage amplitude and power-factor angle.From formula (5), in a switch periods, the mean value of output voltage can by regulating index of modulation m and input power factor angle realize.Generally all wish to adopt unity power factor, so mainly by regulating the index of modulation m of Current Control space vector, realizing the control to output dc voltage, reaching the object that output dc voltage is constant.But output dc voltage is unstable when three-phase input voltage imbalance or voltage sag time.If calculate according to formula 5, then carry out by multiple device the value that transformation of electrical energy calculates and also can produce relatively large deviation with measured value.

Shown in composition graphs 1,2 and Fig. 5, feedback embodiment of the present invention: monitoring matrix rectifier DC output voltage V in real time ₀with matrix rectifier output voltage V during line voltage nominal situation _refdo difference and obtain deviation signal e.

Carry out mode selection according to error e size, set an initial modulation and compare m ₀value=target voltage/466.

When | during e|>=36, modulation ratio m=m ₀, adopt and control output dc voltage is reached near desired value fast than row P.

When | during e|<36, modulation ratio m=m ₀+ [(1-m ₀)/6] * fuzzy output value, then adopt fuzzy control.

The structural representation of composition graphs 6 fuzzy controller, fuzzy controller comprises defuzzification interface 1, inference machine 2, knowledge base 3 and ambiguity solution interface 4, and described knowledge base comprises database 3-1 Sum fanction storehouse 3-2.

Input variable deviation E and the input variable deviation variation rate EC of described fuzzy controller send to inference machine 2 by defuzzification interface 1 after obfuscation, described knowledge base 3-1 is for storing input and output linguistic variable, described rule base 3-2 comprises fuzzy control rule information, input variable deviation E and the corresponding relation between input variable deviation variation rate EC and fuzzy output is comprised in described fuzzy rule information, this inference machine 2 reads corresponding output language variable according to the input variable E after the obfuscation received and input variable deviation variation rate from database 3-1, read corresponding fuzzy control rule information from rule base 3-2 simultaneously, then according to the input variable deviation E after obfuscation, input variable deviation variation rate EC, input language variable and fuzzy control rule solve fuzzy control relation, obtain the fuzzy control quantity exported, and this fuzzy control quantity is sent to ambiguity solution interface 4, this fuzzy control quantity exports as fuzzy controller output variable u by described ambiguity solution interface 4.

Set the fuzzy set of input deviation e as { NB, NS, ZO, PS, PB}, domain [-6+6]; The fuzzy set of input deviation rate of change ec is also that { fuzzy set of fuzzy output value u is { NB, NS, ZO, PS, PB}, domain [-6+6] for NB, NS, ZO, PS, PB}, domain [-6+6].Wherein NB gets Z-shaped membership function, and PB gets S shape membership function, and all the other get Triangleshape grade of membership function.

Fuzzy control rule table reaches for If ... and ... then ..., amount to 16, setting up fuzzy rule base is

1.If(eisNB)and(ecisZO)then(uisNO)(1)

2.If(eisNB)and(ecisPS)then(uisNS)(1)

3.If(eisNB)and(ecisPB)then(uisNB)(1)

4.If(eisNS)and(ecisZO)then(uisNO)(1)

5.If(eisNS)and(ecisPS)then(uisNS)(1)

6.If(eisNS)and(ecisPB)then(uisNB)(1)

7.If(eisZO)and(ecisNB)then(uisPO)(1)

8.If(eisZO)and(ecisNS)then(uisPO)(1)

9.If(eisZO)and(ecisPS)then(uisNO)(1)

10.If(eisZO)and(ecisPB)then(uisNO)(1)

11.If(eisPS)and(ecisZO)then(misPO)(1)

12.If(eisPS)and(ecisPS)then(misPS)(1)

13.If(eisPS)and(ecisPB)then(uisPB)(1)

14.If(eisPB)and(ecisZO)then(uisPO)(1)

15.If(eisPB)and(ecisPS)then(uisPS)(1)

16.If(eisPB)and(ecisPB)then(uisPB)(1)

Fig. 7 be input voltage uneven time, wherein A phase effective value is 187V, B phase effective value be 253V, C phase effective value is 220V, and the target voltage of setting is 300V, does not adopt feedback matrix rectifier output current voltage simulation waveform.Can be found out by simulation result, VD mean value is 300V, and ripple voltage is comparatively large, easily produces a large amount of harmonic wave, is unfavorable for that load equipment normally runs.

Fig. 8 be input voltage uneven time, wherein A phase effective value is 187V, B phase effective value be 253V, C phase effective value is 220V, and the target voltage of setting is 300V, adopts feedback matrix rectifier output current voltage simulation waveform.Can be found out by simulation result, VD quality is improved well, and ripple voltage is less, improves electric current and voltage stability.

Non-elaborated part of the present invention belongs to the known technology of those skilled in the art.

Claims (5)

1. control method when matrix rectifier input is uneven, is characterized in that, real-time monitoring matrix rectifier DC output voltage V ₀, pass through V ₀with matrix rectifier output voltage V during line voltage nominal situation _refdifference obtain deviation signal e, when namely deviation signal e is greater than set point comparatively greatly, adoption rate P controls to make output dc voltage reach near desired value or desired value; When deviation signal is less be namely less than set point time then adopt fuzzy control, the deviation signal e obtained is quantized factor K e process through deviation and obtains fuzzy controller input variable deviation E; Simultaneously by deviation signal differentiate, and obtain fuzzy controller input variable deviation variation rate EC through deviation variation rate factor K ec process; The deviation variation rate EC of the deviation E of input and input is carried out obfuscation, fuzzy reasoning and decision-making by described fuzzy controller, obtain output variable u, this output variable u obtains space vector modulation Coefficient m through the process of output quantization factor Ku, in conjunction with the input phase current space vector sector angle θ calculated, obtain the action time of each switch state vector, and reach PWM generation unit; The action time of switch state vector is sent to the switch circulation controller of matrix rectifier by PWM generation unit with pwm pulse form, and is determined the turn-on and turn-off signal of each bidirectional switch by this switch circulation controller, realizes the modulation to matrix rectifier; Described control method carries out mode selection under output error varies in size situation:

Set an initial modulation and compare m ₀value=expect voltage/466;

When | during e|>=36, modulation ratio m=m ₀, adoption rate P controls, and namely proportionality coefficient is m ₀, make output dc voltage reach near desired value or desired value fast;

When | during e|<36, modulation ratio m=m ₀+ [(1-m ₀)/6] * fuzzy output value, then adopt fuzzy control.

2. control method when a kind of matrix rectifier input according to claims 1 is uneven, it is characterized in that fuzzy controller comprises defuzzification interface (1), inference machine (2), knowledge base (3) and ambiguity solution interface (4), described knowledge base comprises database (3-1) Sum fanction storehouse (3-2), the input variable deviation E of fuzzy controller and input variable deviation variation rate EC sends to inference machine (2) by defuzzification interface (1) after obfuscation, described database (3-1) is for storing input and output linguistic variable, rule base (3-2) comprises fuzzy control rule information, input variable deviation E and the corresponding relation between the higher poor rate of change EC of input variable and fuzzy output is comprised in described fuzzy rule information, this inference machine (2) reads corresponding output language variable according to the input variable E after the obfuscation received and input variable deviation variation rate from database (3-1), read corresponding fuzzy control rule information from rule base (3-2) simultaneously, then according to the input variable deviation E after obfuscation, input variable deviation variation rate EC, input language variable and fuzzy control rule obtain the fuzzy control quantity exported, and this fuzzy control quantity is sent to ambiguity solution interface (4), this fuzzy control quantity exports as fuzzy controller output variable u by described ambiguity solution interface (4).

3. control method when a kind of matrix rectifier according to claim 1 inputs uneven, is characterized in that the fuzzy set of fuzzy controller input deviation E is for { NB, NS, ZO, PS, PB}, namely { bear large, bear little, zero, be just little, honest }, domain [-6+6]; The fuzzy set of input deviation rate of change EC is also that { fuzzy output value u obtains fuzzy set for { NB, NS, ZO, PS, PB}, domain [-6+6] for NB, NS, ZO, PS, PB}, domain [-6+6]; Wherein NB gets Z-shaped membership function, and PB gets S shape membership function, and all the other get Triangleshape grade of membership function.

4. control method when a kind of matrix rectifier input according to claim 1 is uneven, is characterized in that fuzzy control deviation quantizes factor K e=6/36 value 0.166, deviation variation rate factor K ec value 10 ^-13, output-scale-factor value Ku=0.3/6=0.05.

5. control method when a kind of matrix rectifier input according to claim 1 is uneven, is characterized in that fuzzy control rule table reaches for If ... and ... then ..., amount to 16, setting up fuzzy rule base is

If(eisNB)and(ecisZO)then(uisNO)(1)

If(eisNB)and(ecisPS)then(uisNS)(1)

If(eisNB)and(ecisPB)then(uisNB)(1)

If(eisNS)and(ecisZO)then(uisNO)(1)

If(eisNS)and(ecisPS)then(uisNS)(1)

If(eisNS)and(ecisPB)then(uisNB)(1)

If(eisZO)and(ecisNB)then(uisPO)(1)

If(eisZO)and(ecisNS)then(uisPO)(1)

If(eisZO)and(ecisPS)then(uisNO)(1)

If(eisZO)and(ecisPB)then(uisNO)(1)

If(eisPS)and(ecisZO)then(misPO)(1)

If(eisPS)and(ecisPS)then(misPS)(1)

If(eisPS)and(ecisPB)then(uisPB)(1)

If(eisPB)and(ecisZO)then(uisPO)(1)

If(eisPB)and(ecisPS)then(uisPS)(1)

If(eisPB)and(ecisPB)then(uisPB)(1)。