The present invention relates to a technique of reducing a harmonic of a load current in a so-called AC-AC power converter.
An inveter is shown as a circuit which drives a multiphase AC load such as a motor. It is known that this inverter employs a configuration (referred to as a “capacitor-less inverter” below) which includes a matrix converter or include a DC link but not a smoothing circuit such as a larger capacitor or a reactor as a circuit system which achieves high efficiency and miniaturization.
Matrix converters include a direct matrix converter which is not provided with DC link and an indirect matrix converter which is provided with DC link. However, as introduced in Japanese Patent Application Laid-Open No. 2004-222338, even a direct matrix converter can be controlled based on an operation of a configuration in which a virtual AC-DC converter and a virtual DC-AC converter which are coupled through virtual DC link without a smoothing circuit.
Further, Japanese Patent No. 4067021 also introduces the above-described capacitor-less inverter, which has a capacitor in a DC link but the capacity of the capacitor is designed to be smaller than a capacity to function as a smoothing circuit. In this technique, voltage of the DC link is supposed to pulsate.
Hence, a circuit which converts AC power without a substantial smoothing circuit irrespectively of whether or not DC link are formally provided or whether or not a capacitor is provided is referred to as a direct AC-AC power converter in the present application.
Direct AC-AC power converter does not involve an energy buffer, and therefore harmonic components produced by a multiphase AC load are propagated to a power side, and harmonics of a power current increase. Reduction of the harmonics of the power current is demanded to prevent negative influences on surrounding environment. A specific example of this demand is, for example, a rule of IEC61000-3-2. According to this rule, harmonic components up to the 40th order with respect to a power frequency are regulated.
Japanese Patent Application Laid-Open No. 2005-27422 points out that, when, for example, a motor is used for a multiphase AC load and concentrated winding is adopted as a winding method of armature windings of the motor, a current (armature current) flowing in the armature windings includes harmonic components and, more particularly, the 5th order component and the 7th order component due to a differential voltage between a voltage outputted from an inverter and an inductive power of a rotating electrical machine. Such harmonic components (with respect to a fundamental wave component of a voltage outputted from the inverter) cause an increase in harmonics of a power current.
Japanese Patent No. 4488122 introduces a technique of reducing the 5th order component and the 7th order component.
However, the technique introduced in Japanese Patent No. 4488122, without taking only takes into account a phase difference between harmonic components for simplification, has a room for improvement of an effect.
It is therefore an object of the present invention to, when an AC voltage is outputted to an inductive load such as a motor, reduce pulsation of effective power caused by odd number-th order harmonic components of a current flowing in the load. This leads to suppression of pulsation of effective power of DC link in a direct power converter which does not include a smoothing capacitor although the direct power converter has the DC link, and suppression of a power harmonic as a result.
A power converter control method according to the present invention is a method of controlling a direct AC-AC power converter (9) which includes: a rectifier circuit (3) which inputs first AC voltages (Vr, Vs, Vt) and outputs a rectified voltage (Vdc); and a voltage source inverter (4) which receives an input of the rectified voltage, applies three-phase second AC voltages (Vu, Vv, Vw) to a load, and outputs three-phase load currents (Iu, Iv, Iw) to the load.
Further, according to a first aspect of the invention, a modulation factor (k) of the voltage source inverter includes a DC component (k0) and an AC component (k6n·cos(6n·ωL·t+φ6n)) of an angular frequency (6n·ωL) which is a 6n multiple of a base fundamental angular frequency (ωL) of the second AC voltage.
Furthermore, according to a second aspect of the invention, a voltage command (Vd** and Vq**) to the voltage source inverter includes a DC component (Vd* and Vq*) and an AC component ((−Ju6n·Eu1)sin(6nωL+φ6N)(−Ju6n·Eu1)·cos(6nωL+φ6N)) of an angular frequency (6n·ωL) which is a 6n multiple of a base fundamental angular frequency (ωL) of the second AC voltage.
According to both of these first aspect and second aspect of the invention, when a fundamental wave component, a (6n−1)th order component and a (6n+1)th order component of the first load current are Iu1, Iu6n−1 and Iu6n+1, respectively, and phase differences of the fundamental wave component, the (6n−1)th order component and the (6n+1)th order component of the load current from a fundamental wave component of the second AC voltage are φ1, φ6n−1 and φ6n+1, respectively, a ratio (−k6n/k0) of an amplitude of the AC component with respect to the DC component takes a ratio represented by
−[m6n2+Iuh6n2+2·m6n·Iuh6n cos(θ−χ6n)]1/2/[Iu1·cos(φ1)]
[m6n=[Iu6n−12+Iu6n+12+2·Iu6n−1·Iu6n+1·cos(φ6n−1−φ6n+1)]1/2).
A phase difference (φ6n) of the AC component from the fundamental wave component of the second AC voltage takes an angle represented by:
tan−1[{m6n·sin(χ6n)+Iuh6n·sin(χ6n)}/{m6n·cos(χ6n)+Iuh6n·cos(χ6n)}]
(χ6n=tan−1[{Iu6n−1·sin(φ6n−1)+Iu6n+1·sin(φ6n+1)}/{Iu6n−1·cos(φ6n−1)+Iu6n+1·cos(φ6n+1)}]
Here, the angle has relationship of Iuh6n<m6n and θ: arbitrary.
According to a third aspect of the power converter control method according to the present invention is based on the first aspect or the second aspect of the invention, the ratio of the AC component with respect to the DC component is calculated in advance as a function of a plurality of operation states of the load before the load is actually operated, and the direct AC-AC power converter (9) is controlled based on the function upon the actual operation.
According to a fourth aspect of the power converter control method according to the present invention is based on the third aspect, the operation states include a plurality of power states consumed by the load, and the ratio corresponding to the power states upon the actual operation is taken.
A fifth aspect of the power converter control method according to the present invention is based on any one of the first to fourth aspects of the invention, an amplitude of the AC component is increased in a predetermined range of the base fundamental angular frequency (ωL) of the second AC voltage according to a sensitization amount which increases when the base fundamental angular frequency increases.
For example, Iuh6n=0 is true in all n.
According to the first aspect or the second aspect of the power converter control method according to the present invention, the (6n−1)th order component and the (6n+1)th order component of the load are reduced, and then propagation of the harmonic to a supply source of the first AC voltage is reduced.
According to the third aspect of the power converter control method according to the present invention, the ratio of the AC component with respect to the DC component does not need to be calculated based on the second AC voltage and the load current upon the actual operation, and the burden applied to perform control upon the actual operation is reduced.
According to the fourth aspect of the power converter control method according to the present invention, the dependency on other operating statuses other than power states is small, so that it is possible to provide an effect provided by the first aspect and the second aspect only by calculating a power based on the second AC voltage and the load current upon the actual operation.
According to the fifth aspect of the power converter control method according to the present invention, when the base fundamental angular frequency of the second AC voltage increases, the angular frequency which is a 6n multiple of the base fundamental angular frequency becomes close to the frequency at which the power converter is controlled, so that contribution of the AC component decreases as a whole even when the AC component is reflected at a timing at which the power converter is controlled. Consequently, by increasing the amplitude of the AC component in the predetermined range when the base fundamental angular frequency increases, this decrease is compensated.
An object, features, aspects and advantages of the present invention will be made obvious by the following detailed description and the accompanying drawings.
The converter 3 functions as a rectifier circuit, rectifies three phase (an R phase, an S phase and a T phase, in this case) AC voltages Vr, Vs and Vt (also referred to as “first AC voltages” below) obtained from an AC power source 1, and outputs a rectified voltage Vdc to a pair of DC power lines L1 and L2.
The converter 3 is, for example, a current source rectifier, and operates by pulse width modulation. The converter 3 includes a plurality of current paths which is mutually connected in parallel between the DC power lines L1 and L2. One of the current paths of the converter 3 corresponding to the R phase includes a pair of switching elements Srp and Sm connected in series between the DC power lines L1 and L2. The voltage Vr is applied to a connection point between the switching elements Srp and Sm. One of the current paths of the converter 3 corresponding to the S phase includes a pair of switching elements Ssp and Ssn connected in series between the DC power lines L1 and L2. The voltage Vs is applied to a connection point between the switching elements Ssp and Ssn. One of the current paths of the converter 3 corresponding to the T phase includes a pair of switching elements Stp and Stn connected in series between the DC power lines L1 and L2. The voltage Vt is applied to a connection point between the switching elements Stp and Stn.
The switching elements Srp, Ssp and Stp are connected to the DC power line L1 side, and the switching elements Sm, Ssn and Stn are connected to the DC power line L2 side.
The inverter 4 is, for example, a voltage source inverter, and is operated by pulse width modulation according to, for example, instantaneous space vector control (referred to simply as “vector control” below). The inverter 4 outputs three phase (the U phase, the V phase and the W phase in this case) AC voltages Vu, Vv and Vw (referred to as “second AC voltages” below).
The inverter 4 has a plurality of current paths connected in parallel between the DC power lines L1 and L2.
One of the current paths of the inverter 4 which corresponds to the U phase includes a pair of switching elements Sup and Sun connected in series between the DC power lines L1 and L2. An output voltage Vu is obtained from a connection point of the switching elements Sup and Sun. One of the current paths of the inverter 4 which corresponds to the V phase includes a pair of switching elements Svp and Svn connected in series between the DC power lines L1 and L2. An output voltage Vv is obtained from a connection point between the switching elements Svp and Svn. One of the current paths of the inverter 4 which corresponds to the W phase includes a pair of switching elements Swp and Swn connected in series between the DC power lines L1 and L2. An output voltage Vw is obtained from a connection point between the switching elements Swp and Swn.
The switching elements Sup, Svp and Swp are connected to the DC power line L1 side. These switching elements are regarded as switching elements of an upper arm side below. The switching elements Sun, Svn and Swn are connected to the DC power line L2 side. These switching elements are regarded as switching elements of a lower arm side below. That is, a potential of the DC power line L1 is higher than a potential of the DC power line L2.
Configurations of the above switching elements Sip, Ssp, Stp, Sm, Ssn, Stn, Sup, Svp, Swp, Sun, Svn and Swn are known, and are also disclosed in, for example, in Lixiang Wei, Thomas. A Lipo, “A Novel Matrix Converter Topology With Simple Commutation”, IEEE IAS2001, vol. 3, 2001, pp 1749 to 1754.
The load 2 is an inductive load, and is connected to the inverter 4. More specifically, the load 2 is a motor which has three-phase coils which are connected by way of Y connection and to which the voltages Vu, Vv and Vw are applied. Resistor components of the respective three-phase coils in a circuit diagram are illustrated as resistors connected to the coils in series. Currents iu, iv and iw flow in the coils of these coils corresponding to the U phase, the V phase and the W phase. These currents are monitored by a current sensor (not illustrated).
First, a harmonic of power consumption of the load 2 will be described. When the second AC voltages Vu, Vv and Vw to be applied to the load 2 and the flowing currents (referred to as “load currents” below) iu, iv and iw are discussed, even if the voltage Vu and the load current iu of one phase (U phase) are converted into mathematical expressions like following Expression (1) and Expression (2), universality is not lost. Assuming that the three phases are in stationary states while keeping the balance is sufficient. Hereinafter, only the 5th order harmonic and the 7th order harmonic of the load current iu will be taken into account. In view of Japanese Patent No. 4488122, a coefficient of a fundamental wave component of the second AC voltage Vu includes not only a DC component but also an AC component. Only a harmonic component which is a 6 multiple of the fundamental wave component of the second AC voltage Vu will be first taken into account as the AC component.
[Expression 1]
Vu=21/2[Eu1−Eu6·cos(6ωLt+φ6)] cos(ωLt) (1)
[Expression 2]
iu=21/2[Iu1·cos(χLt+φ1)+Iu5·cos(5ωLt+φ5)+Iu7·cos(7ωLt+φ7)] (2)
Provide, an angular frequency ωL of the fundamental wave component of the current and the voltage outputted from the inverter 4 and a time t are introduced. Further, effective values of the fundamental component, the 5th order component and the 7th order component of the load current iu are Iu1, Iu5 and Iu7, respectively, and phase differences of the fundamental wave component, the 5th order component and the 7th order component of the load current iu from the fundamental wave component of the second AC voltage Vu are φ1, φ5 and φ7, respectively. Furthermore, effective values of the DC component and the AC component of the amplitude of the fundamental wave component of the second AC voltage Vu are Eu1 and Eu6, and the phase difference of this AC component from the fundamental wave component of the voltage Vu is φ6.
An instantaneous power Pu of the U phase is expressed by Expression (3).
A first term on the right side of Expression (3) represents an active power and an inactive power, and second and subsequent terms represent harmonic powers. Harmonic powers other than those of 3n-th order components are 0 in the three-phase circuits, a power represented by a product of the harmonic components is relatively low and, when these powers are ignored, a harmonic power Puh can be approximated by following Expression (4).
Hence, it is only necessary to calculate Eu6 and φ6 which satisfy following Expression (5) and output, from the inverter 4, the voltage Vu defined by Expression (2) using these values to make the harmonic power Puh zero.
[Expression 5]
Eu
1
[Iu
5·cos(6ωLt+φ5)+Iu7·cos(6ωLt+φ7)]=Eu6·Iu1·cos(φ1)·cos(6ωLt+φ6) (5)
Now, setting Pu5=Eu1·Iu5, Pu7=Eu1·Iu7 and Pu6=Eu6·Iu1·cos(φ1), Pu5, Pu6 and Pu7 are allowed to be used as constants. And, Expression (5) is deformed as the following expression.
[Expression 6]
Pu
5·cos(6ωLt+φ5)+Pu7·cos(6ωLt+φ7)=Pu6·cos(6ωLt+φ6) (6)
Further, Expression (6) can be deformed as following Expression (7).
[Expression 7]
[Pu5·cos(φ5)+Pu7·cos(φ7] cos(6ωLt)−[Pu5·sin(φ5)+Pu7·sin(6ωLt)] sin(6ωLt)=Pu6·cos(φ6)·cos(6ωLt)−Pu6·sin(6ωLt)·sin(6ωLt) (7)
Relationships which satisfy Expression (7) are Expressions (8) and (9) according to orthogonality of a sine function and a cosine function.
[Expression 8]
Pu
5·cos(Ψ5)+Pu7·cos(φ7)=Pu6·cos(φ6) (8)
[Expression 9]
Pu
5·sin(φ5)+Pu7·sin(φ7)=Pu6·sin(φ6) (9)
The effective value Eu6 and the phase difference φ6 are calculated according to Expressions (10) and (11) based on Expressions (8) and (9).
[Expression 10]
Eu
6
=Eu
1
·[Iu
5
2
+Iu
7
2+2·Iu5·Iu7 cos(φ5−φ7)]1/2/[Iu1·cos(φ1)] (10)
[Expression 11]
φ6=tan−1[{Iu5·sin(φ5)+Iu7·sin(φ7)}/{Iu5·cos(φ5)+Iu7·cos(φ7)}] (11)
Although the 5th order harmonic and the 7th order harmonic have been described above, the other 6n±1th harmonics (n is a positive integer) can also be discussed likewise. Further, even when the AC component of the amplitude of the fundamental wave component of the voltage Vu includes a plurality of 6n-th angular frequencies, approximation is allowed similar to approximation from Expression (3) to Expression (4). Consequently, it is possible to reduce the harmonic power based on the 6n±1th order harmonic of the load current iu by adopting following Expressions (12), (13) and (14) as the voltage Vu. Provide, a summation sign Σ refers to a sum of n.
As is clear from a form of Expression (12), the voltage Vu takes a value obtained by modulating 21/2Eu1·cos ωLt by [1−ΣJu6n (6nωLt+φ6n). Normally, a sinusoidal voltage is outputted as the voltage Vu, and a so-called modulation factor for obtaining a voltage command corresponds to k=(k0−Σk6n·cos(6nωLt+φ6n)). That is, by taking such a modulation factor k, it is possible to obtain a voltage command corresponding to the voltage Vu. Here, Ju6n=k6n/k0 is true, and is a ratio of the modulation factor with respect to the DC component of the 6n-th order component.
Although the U phase has been discussed, the V phase and the W phase can be both analyzed likewise, and Expressions (13) and (14) can be commonly adopted for both of the phases. In other words, the values calculated according to Expressions (13) and (14) are common to every phase.
Hence, it is only necessary to multiply (1−ΣJu6n·cos(6nωLt+φ6n)) on or add −k0ΣJu6n·cos(6nωLt+φ6n) to a modulation factor k0 before reduction of the 6n±1th order harmonic is taken into account.
Eu6 and φ6 which satisfy Expression (5) are calculated above. However, Expression (5) is used to make the harmonic power Puh zero. A value of Expression (4) only needs to be made small instead of actually making the harmonic power Puh zero. More specifically, an absolute value on the left side of Expression (5) only needs to be reduced.
Now, it is assumed that 6th order harmonic power Puh6 after suppression is expressed by following Expression (15). Provide, a phase φuh6 can be arbitrarily set; a view point of allowing any phase difference as long as the absolute value of the harmonic power Puh6 can be suppressed. Provide, by learning the phase φuh6 as a phase difference of the fundamental wave component of the load current iu from the fundamental wave component of the AC voltage Vu, and setting a desired value in the following expression, it is possible to set to a desired value the phase of the harmonic power Puh6 which is left without being completely suppressed.
[Expression 15]
Pu
h6
=Eu
1
·Iu
h6·cos(6ωLt+φuh6) (15)
Further, the current Iuh6 is desired to be smaller than a value in a square bracket on the left side (corresponding to the harmonic power before suppression) of Expression (5) from a view point of suppressing the harmonic power. More specifically, an amplitude Pu57 and a phase difference φu57 of the harmonic power before suppression are calculated according to following Expressions (16) and (17), respectively. The current Iuh6 is desired to be smaller than a value Iu57.
Eu6 and φ6 for correcting the amplitude Pu57 of the harmonic power before suppression are calculated as described below. Provide, the amplitude Pu6 is an amplitude of a power (also referred to as a “correction power” below) obtained by subtracting the right side (corresponding to a harmonic power after suppression) of Expression (15) from the left side (corresponding to the harmonic power before suppression) of Expression (6).
By this means, Expressions (21) and (22) are obtained similar to Expressions (13) and (14).
[Expression 21]
Ju
6n
={Iu
(6n−1)
2
+Iu
(6n+1)
2+2·Iu(6n−1)·Iu(6n+1)·cos(φ(6n−1)−φ(6n+1))+Iuh(6n)2+2·{Iu(6n−1)2+Iu(6n+1)2+2·Iu(6n−1)·Iu(6n+1)·cos(φ(6n−1)−φ(6n+1))}1/2×Iuh(6n)·cos(tan−1[(Iu(6n−1)·sin(φ(6n−1))+Iu(6n+1)·sin(φ(6n+1)))/Iu(6n−1)·cos(φ(6n−1))+Iu(6n+1)·cos(φ(6n+1)))]−(φuh(6n)+π))}1/2/(Iu1)·cos(φ1)) (21)
[Expression 22]
φ6n=tan−1[({Iu(6n−1)2+Iu(6n+1)2+2·Iu(6n−1)·Iu(6n+1)·cos(φ(6n−1)−φ(6n+1))]1/2×sin(tan−1{(Iu(6n−1)·sin(φ(6n−1))+Iu(6n+1)·sin(φ(6n+1)))/(Iu(6n−1)·cos(φ(6n−1))+Iu(6n+1)·cos(φ(6n+1)))})+Iuh(6n)·sin(φuh(6n)+π))/({Iu(6n−1)2+Iu(6n+1)2+2·Iu(6n−1)·Iu(6n+1)·cos(φ(6n−1)−φ(6n+1))}1/2×cos(tan−1{(Iu(6n−1)·sin(φ(6n−1))+Iu(6n+1)·sin(φuh(6n+1)))/(Iu(6n−1)·cos(φ(6n−1))+Iu(6n+1)·cos(φ(6n+1)))})Iuh(6n)·cos(φuh(6n)+π))] (22)
As described above, the phase φuh6 can be arbitrarily set, so that θ=φuh6+π may be used and θ may be arbitrarily set. Further, when the current Iu57 is redescribed as m6, this way of description is applied to the 6n-th order likewise and can be described as follows.
That is, setting
[m6n=[Iu6n−12+Iu6n+12+2·Iu6n−1·Iu6n+1·cos(φ6n−1−φ6n+1)]1/2
χ6n=tan−1[{Iu6n−1·sin(φ6n−1)+Iu6n+1·sin(φ6n+1)}/{Iu6n−1·cos(φ6n−1)+Iu6n+1·cos(φ6n+1)}]
Iuh6n<m6n, and
θ is arbitrary,
Ju
6n
=[m
6n
2
+Iu
h6n
2+2·m6n·Iuh6n cos(θ−χ6n)]1/2/[Iu1·cos(φ1)] is true and
φ6n=tan−1[{m6n·sin(χ6n)+Iuh6n·sin(χh6n)}/{m6n·cos(χ6n)+Iuh6n·cos(χ6n)}] is true.
Similarly, the modulation factor can also be expressed as k6n/k0=[m6n2+Iuh6n2+2·m6n·Iuh6n cos(θ−χ6n)]1/2/[Iu1·cos(φ1)].
Naturally, by making Iuh6n=0 for all n, Expressions (21) and (22) match with Expressions (13) and (14).
(c-1) Correction of Modulation Factor
An amplitude/phase extracting unit 701 inputs the load current iu, and generates a fundamental wave component, the 6n−1th order component and the 6n+1th order component of the load current iu. This computation is realized by performing Fourier transform on the load current iu. By this means, effective values Iu1, Iu6n−1 and Iu6n+1 are obtained. Further, the amplitude/phase extracting unit 701 further inputs the voltage Vu, compares the phase of the voltage Vu and phases of the fundamental wave component, the 6n−1th order component and the 6n+1th order component of the load current iu, and generates and outputs the phase differences φ1, φ6n−1 and φ6n+1.
A 6n-th order component computing unit 702 obtains the effective values Iu1, Iu6n−1 and Iu6n+1 and the phase differences φ1, φ6n−1 and φ6n+1 from the amplitude/phase extracting unit 701, and computes a ratio Ju6n and a phase difference φ6n according to Expressions (13) and (14). For example, n=1, 2 is selected. Alternatively, the 6n-th order component computing unit 702 computes and outputs the ratio Ju6n and the phase difference φ6n according to Expressions (21) and (22) by also using the phase difference χ6n and the current Iuh6n. The phase difference χ6n and the current Iuh6n may be externally inputted to or set inside the 6n-th order component computing unit 702.
A product sum computing unit 713 computes [1−ΣJu6n·cos(6nωLt+φ6n)] from the ratio Ju6n and the phase difference φ6n obtained from the 6n-th order component computing unit 702.
A multiplier 715 multiplies the modulation factor k0 for simply outputting a sinusoidal voltage and [1−ΣJu6n cos(6nωLt+φ6n)], and outputs the multiplication result as a new modulation factor k. By taking the modulation factor k as described above, it is possible to reduce the harmonic power based on the 6n±1th harmonic of the load current iu (the load currents iv and iw likewise).
A PWM modulating unit 714 generates gate signals Sup*, Svp*, Swp*, Sun*, Svn* and Swn* using a d axis voltage command value Vd*, a q axis voltage command value Vq* and the modulation factor k. These gate signals are generated by adopting a known technique, and therefore details of this computation will not be described. In other words, only by substituting the modulation factor k0 with the modulation factor k, it is possible to divert a known technique and reduce the harmonic power based on the 6n±1th order harmonics.
A technique for obtaining the command values Vd* and Vq* is also a known technique, and will be briefly described below without details.
A subtractor 705 calculates a deviation of the angular frequency ωL and the command value ωL* of the angular frequency ωL to input to a PI control unit 706. The PI control unit 706 generates a current command value Ia* by performing known PI control (proportion/integration control). A d axis current command value generating unit 707 and a q axis current command value generating unit 708 inputs the current command value Ia* and the phase control command value β*, and calculates a sine component and a cosine component of the current command value Ia* as a d axis current command value Id* and a q axis current command Iq*, respectively, for (−β*), which has negative sign with the phase control command values β*.
A coordinate system converting unit 704 calculates the d axis current Id and the q axis current Iq based on the load currents iu, iv and iw and the motor angle estimation value θ to output to the subtractors 709 and 710, respectively.
The subtractor 709 outputs a deviation ΔId of the d axis current Id and the command value Id* of the d axis current Id. The subtractor 710 outputs a deviation ΔIq of the q axis current Iq and the command value Iq* of the q axis current Iq.
An interference compensating unit 711 performs computation for compensating for a mutual interference between the d axis and the q axis according to impedances ωL·Ld and ωL·Lq based on inductances Ld and Lq of a motor. The interference compensating unit 711 inputs the d axis current Id, the q axis current Iq, the deviations ΔId and ΔIq and the angular frequency ωL, and generates the d axis voltage command Vd* and the q axis voltage command Vq*. This computation is a known technique, and will not be described in details.
(c-2) Correction of Voltage Command
Although
In addition, the inverter control circuit 7 illustrated in
Hereinafter, generation of the corrected d axis voltage command Vd** and the q axis voltage command Vq** will be described. First, similar to the inverter control circuit 7 illustrated in
The correction command generating unit 703 inputs the ratio Ju6n and the phase difference φ6n, and generates a d axis voltage command correction value ΔVd* and a q axis voltage command correction value ΔVq*.
The adder 716 adds the d axis voltage command Vd* and the d axis voltage command correction value ΔVd* to output the corrected d axis voltage command Vd**. The adder 717 adds the q axis voltage command Vq* and the q axis voltage command correction value ΔVq* to output the corrected q axis voltage command Vq**.
The correction command generating unit 703 has a multiplier 703a, a product sum computing unit 703b, a sine value calculator 703c, a cosine value calculator 703d, multipliers 703e and 703f and total sum calculators 703g and 703i.
The multiplier 703a inputs the angular frequency ωL, multiplies the angular frequency ωL with a 6n multiple and outputs the 6n-th order angular frequency 6nωL. Here, n is a positive integer, and that n angular frequencies 6nωL are outputted is indicated by adding a diagonal line to an arrow from the multiplier 703a to the product sum computing unit 703b, and “n” is indicated near the arrow. That is, this diagonal line and “n” indicate that n pieces of information are transmitted. The same applies to diagonal lines added to other arrows, and “n”.
The product sum computing unit 703b adds the phase difference φ6n of the same order, to a result obtained by multiplying the angular frequency 6nωL with the time t. That is, the product sum computing unit 703b performs n pairs of multiplication and addition per corresponding angular frequency. By this means, the product sum computing unit 703b feeds the phase (6nωL+φ6n) to both of the sine value calculator 703c and the cosine value calculator 703d. The sine value calculator 703c and the cosine value calculator 703d output a value obtained by multiplying a sine value with a (−1) multiple and a value obtained by multiplying a cosine value with a (−1) multiple as the inputted phases, respectively. More specifically, the sine value calculator 703c and the cosine value calculator 703d output −sin(6nωL+φ6n) and −cos(6nωL+φ6n), respectively.
Each of n values outputted from the sine value calculator 703c is multiplied with the ratio Ju6n corresponding to each value n, the DC component Eu1 of the amplitude of the fundamental wave component of the second AC voltage Vu and a coefficient √(3/2) in the multiplier 703e, and then is outputted as a value √(3/2)(−Ju6n·Eu1)·sin(6nωL+φ6n). Similarly, the multiplier 703f outputs √(3/2)(−Ju6n·Eu1)·cos(6nωL+φ6n).
The DC component Eu1 can be obtained by performing Fourier transform on the second AC voltage Vu in, for example, the amplitude/phase extracting unit 701 illustrated in
The total sum calculator 703g calculates a total sum of values √(3/2)(−Ju6n·Eu1)·sin(6nωL+φ6n) by n to output as the d axis voltage command correction value ΔVd*. The total sum calculator 703i calculates a total sum of values √(3/2)(−Ju6n·Eu1) cos(6nωL+φ6n) by n to output as the q axis voltage command correction value ΔVq*.
As described above, the (corrected) d axis voltage command Vd** is obtained by adding the d axis voltage command correction value ΔVd* to the axis voltage command Vd*, and the (corrected) q axis voltage command Vq** is obtained by adding the q axis voltage command correction value ΔVq* to the q axis voltage command Vq*.
Generally, the modulation factor k is proportional to a ratio of a line voltage of the second AC voltages with respect to the rectified voltage Vdc, and the line voltage is proportional to a square root of a sum of squares of the d axis voltage command and the q axis voltage command.
Hence, performing processing based on the modulation factor k0 and the corrected d axis voltage command Vd** and q axis voltage command Vq** upon generation of the gate signals Sup*, Svp*, Swp*, Sun*, Svn* and Swn* by the PWM modulating unit 714 is equivalent to performing processing based on the modulation factor k, and the d axis voltage command Vd* and the q axis voltage command Vq* which are not corrected.
Further, the d axis voltage command Vd* and the q axis voltage command Vq* can be both learned as phasors which are orthogonal in a rotating coordinate system which rotates at the base fundamental angular frequency ωL of the second AC voltage, and both correspond to √(3/2)Eu1. Hence, in the corrected d axis voltage command Vd**, the d axis voltage command Vd* before correction can be learned as a DC component, and the d axis voltage command correction value ΔVd* can be learned as an AC component including the 6n-th order angular frequency. In the q axis voltage command Vq**, the q axis voltage command Vq* before correction can be learned as a DC component, and the q axis voltage command correction value ΔVq* can be learned as an AC component including the 6n-th order angular frequency.
Consequently, irrespectively of the d axis or the q axis, the ratio of the amplitude of the AC component with respect to the DC component is |[√(3/2)(−Ju6n·Eu1)]/[√(3/2)Eu1]|=Ju6n, and is equal to the ratio in case where a modulation factor is corrected.
Meanwhile, a case where an effect to reduce a harmonic is remarkable will be described. That is, a case where a current Iuh6n=0 in all n will be described.
As is seen from
As is understood from above Expression (4), the harmonic power Puh pulsates at an angular frequency which is a 6n multiple of the angular frequency (which is also an angular frequency of the second AC voltage) ωL of the load current iu. Hence, even when the rectified voltage Vdc is controlled to keep constant, a current flowing in the DC link fluctuates at cos(6n·ωLt).
By the way, when a current source rectifier is used as the converter 3, it is possible to make an input current be a sine wave by modulating a current flowing in the DC link by a sine wave taking into account a conduction ratio of each phase (see, for example, Lixiang Wei, Thomas. A Lipo, “A Novel Matrix Converter Topology With Simple Commutation”, IEEE IAS2001, vol. 3, 2001, pp 1749 to 1754. and Japanese Patent Application Laid-Open No. 2007-312589). More specifically, the current is modulated by cos(ωSt) by using the angular frequencies ωS of the voltages Vr, Vs and Vt. Hence, the input current Ir pulsates at cos(ωSt)·cos(6n·ωLt). This pulsation component is deformed as (½)(cos(6n·ωLt−ωSt)+cos(6n·ωLt+ωSt)). Hence, the harmonic components of the input current Ir come to peaks at the frequency 6n·fL±fS (fL=ωL/2π and fS=ωS/2π).
In
Similarly, in
A graph L11 in
Similarly, a graph L21 in
A graph L22 in
To reduce the harmonic components in this way, the adequate ratio Ju6n and phase difference φ6n are set depending on the harmonic power Puh as described in the above B section (Expressions (4) to (14) in particular). In other words, these values depend on power.
As is known from
In view of the above and based on the discussion of the B section, it is predicted that the ratio Ju6 decreases following an increase in the power, and the ratio Ju6 in case where distributed winding is adopted is lower than the ratio Ju6 in case where concentrated winding is adopted. Further, when the power is the same, it is also predicted that there is little dependency on parameters other than a winding method (concentrated winding/distributed winding) of armature windings. Hereinafter, validity of these predictions will be described using experiment results.
As is understood from
Hence, by obtaining dependencies of the ratios Ju6 and Ju12 on powers in a motor which includes armature windings for which distributed winding is adopted as a winding method, it is possible to take the ratios Ju6 and Ju12 commonly between operation states of motors such as torques other than a power or types of motors of other than a winding method of armature windings such as field magnetic fluxes.
More specifically, the ratios Ju6 and Ju12 of a plurality of power states may be calculated before the load 2 is actually operated, and be stored the relationship between a plurality of power states and the ratios Ju6 and Ju12 in a table or as a numerical expression. This table or numerical expression may be stored by the 6n-th order component computing unit 702 or the amplitude/phase extracting unit 701.
Hence, it is desirable to calculate and store in advance the ratios Ju6 and Ju12 per winding method of armature windings or per power before a load is actually operated. By this means, a burden of the calculation amount upon an actual operation is reduced.
Although the above discussion has been made on the ratio Ju6, it goes without saying that this discussion is also validly applied to the d axis voltage command correction value ΔVd* and the q axis voltage command Vq* in view of the principle described in the B section. That is, it is desirable to calculate and store in advance the d axis voltage command correction value ΔVd* and the q axis voltage command Vq* per winding method of armature windings or per power before a load is actually operated. By this means, the calculation amount upon an actual operation is reduced.
Naturally, dependencies of the ratio Ju6n or the d axis voltage command correction value ΔVd* and the q axis voltage command Vq* on not only a power state and a winding method of armature windings but also a plurality of operation states such as a magnitude of a motor torque or a motor rotation speed may be obtained and stored as a table or a function. The power state can also be learned as a type of an operation state.
The PWM modulating unit 714 generates the gate signals Sup*, Svp*, Swp*, Sun*, Svn* and Swn* at temporally discrete timings using the d axis current command value Id*, the q axis current command value Iq* and the modulation factor k or the corrected d axis voltage command Vd** and q axis voltage command Vq** and the modulation factor k0. A cycle of the timings is determined according to a carrier frequency which serves as a basis of comparison when, for example, a carrier comparison method is adopted in the PWM modulating unit 714.
Normally, the carrier frequency is fixed, so that, when a fundamental frequency component of a second AC voltage increases, the number of gate signals to be updated for the fundamental frequency decreases.
Particularly, the modulation factor k and the corrected d axis voltage command Vd** and q axis voltage command Vq** include frequencies which are a 6n multiple of the fundamental frequency, and are not likely to be reflected at adequate timings to update gate signals.
That is, when the fundamental frequency component of the second AC voltage increases, the ratio Ju6n, the d axis voltage command correction value ΔVd* and the q axis voltage command correction value ΔVq* substantially decrease. Further, this tendency becomes remarkable when the order of a harmonic, that is, the value n increases.
Hence, it is desirable to set a sensitization amount to increase the ratio Ju6n, the d axis voltage command correction value ΔVd* and the q axis voltage command correction value ΔVq*, and increase the sensitization amount when the fundamental frequency components of the second AC voltages Vu, Vv and Vw increase. This is to compensate for the decrease.
The rotation speed is indicated with the number of rotations (rps) per second, and, when the number of pairs of poles of a motor is 2, 50×2=100 [Hz] obtained by converting the rotation speed 50 rps into a frequency corresponds to the fundamental frequency of the second AC voltage. Hence, the 12th order component of the harmonic reduction component is 1200 Hz. Now, when a carrier frequency is 6 kHz and when the rotation speed is 50 rps, the number of timings is 6000/1200=5.
In addition, the sensitization amount is zero when the rotation speed is less than 50 rps, the harmonic reduction component is not substantially used (which is indicated as “OFF” in
Further, the sensitization amount is zero when the rotation speed is 88 rps or more, and the harmonic reduction component is not substantially used (which is indicated as “OFF” in
Similar to the ratio Ju6, the d axis voltage command correction value ΔVd* and the q axis voltage command correction value ΔVq*, such a sensitization amount may also be obtained in advance per power state and winding method of armature windings, or further, per magnitude of a motor torque or motor rotation speed, and may be stored as a table or a function.
In addition, it is possible to increase the number of timings with respect to a rotation speed by increasing a carrier frequency and, consequently, raise a range of a rotation speed which requires a sensitization amount. In other words, when the carrier frequency is higher, setting the sensitization amount until a range of a high rotation speed is less necessary. Elements which can increase switching frequencies such as wideband gap elements made of SiC or GaN can be used for the switching elements Srp, Ssp, Stp, Srn, Ssn, Stn, Sup, Svp, Swp, Sun, Svn and Swn to support an increase in a carrier frequency.
Further, a main circuit system to which the present embodiment is applied is not limited to a system provided with the DC link illustrated in
The direct matrix converter MCV has input terminals Pr, Ps and Pt and output terminals Pu, Pv and Pw. The input terminals Pr, Ps and Pt input AC voltages Vr, Vs and Vt, and the output terminals Pu, Pv and Pw output three-phase AC output voltages Vu, Vv and Vw.
The direct matrix converter MCV has switching elements Sur, Sus, Sut, Svr, Svs, Svt, Swr, Sws and Swt. The three switching elements Sur, Sus and Sut are connected between each of the input terminals Pr, Ps and Pt and the output terminal Pu. The three switching elements Svr, Svs and Svt are connected between each of the input terminals Pr, Ps and Pt and the output terminal Pv. The three switching elements Swr, Sws and Swt are connected between each of the input terminals Pr, Ps and Pt and the output terminal Pw.
When a control method according to the present embodiment is applied to the direct matrix converter MCV, virtual AC/DC/AC control is adopted. In virtual AC/DC/AC control, for example, a converter 3 and an inverter 4, which are shown in
In order to switch the converter 3 as a virtual rectifier circuit, similarly to switching of the actual converter 3, referring to, for example, Lixiang Wei, Thomas. A Lipo, “A Novel Matrix Converter Topology With Simple Commutation”, IEEE IAS2001, vol. 3, 2001, pp 1749 to 1754. or Japanese Patent Application Laid-Open No. 2007-312589, gate signals Srp*, Ssp*, Stp*, Srn*, Ssn* and Stn* which control conduction/non-conduction of switching elements Srp, Ssp, Stp, Srn, Ssn and Stn are obtained.
Further, matrix transform is performed according to the following equation to obtain switch signals of the direct matrix converter MCV from the gate signals Srp*, Ssp*, Stp*, Srn*, Ssn*, Stn*, Sup*, Svp*, Swp*, Sun*, Svn* and Swn*.
Switch signals S11, S12, S13, S21, S22, S23, S31, S32 and S33 are switch signals for the switching elements Sur, Sus, Sut, Svr, Sys, Svt, Swr, Sws and Swt, respectively. That this matrix transform is valid is already known from Japanese Patent Application Laid-Open No. 2004-222338.
Further, as introduced in Japanese Patent No. 4067021, application to a conversion circuit which uses a very small capacitor is allowed. Alternatively, an input side of the converter 3 which feeds an output to DC link may be a single-phase input or a multiphase input.
Although the present invention has been described in details, the above description is exemplary in all aspects and the present invention is by no means limited to this. It is understood that an infinite number of modified examples which have not been described are assumed as long as the modified examples do not deviate from the scope of the present invention.
Number | Date | Country | Kind |
---|---|---|---|
2011-209176 | Sep 2011 | JP | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/JP2012/073609 | 9/14/2012 | WO | 00 | 3/25/2014 |