JPH0767318B2 - Vector control method of induction motor - Google Patents

Description

The present invention relates to a vector control method for an induction motor using a PWM inverter.

B. Overview of the invention The present invention is a vector control method for controlling the output of a PWM inverter so that the magnetic flux and the secondary current of an induction motor are orthogonal to each other, depending on the angular frequency of the PWM inverter or the number of PWM pulses during low speed operation. By correcting the magnetic flux command value and adding a first-order delay to this correction, the torque pulsation and rotational unevenness due to the dead time of the switching element of the PWM inverter are reduced, and the followability is improved.

C. Conventional technology For the variable flux control of the induction motor, a slip frequency control method with good response and accuracy is known, but recently, the primary current of the motor is divided into an exciting current and a secondary current to control it. A vector control method that obtains a response equivalent to that of a DC machine by always making the secondary magnetic flux and the secondary current vector orthogonal to each other has become popular. Furthermore, to further improve responsiveness and accuracy,
A non-interference vector control method (for example, Japanese Patent Laid-Open No. 59-165982) in which the interference of the exciting current with the secondary current and the interference of the secondary current with the magnetic flux are mutually corrected has also been implemented.

D. Problems to be solved by the invention In a conventional vector control method such as a non-interference vector control method, when the inverter is a PWM method, it is difficult to perform high-precision control over a wide range due to an error in the operation of the PWM inverter. Come on. The operation error of the PWM inverter is largely due to the dead time for shifting the on / off time of both transistors in order to prevent a short circuit between the high-voltage side and low-voltage side main switch transistors of the inverter. The presence of the dead time deteriorates the speed following performance, for example, the occurrence of torque pulsation and rotation unevenness during low speed operation.

E. Means for solving the problems The present invention has been made in view of the above problems, as the angular frequency of the PWM inverter becomes smaller during low-speed operation,
Alternatively, the magnetic flux correction that lowers the magnetic flux command value as the number of PWM pulses becomes larger is performed with a first-order lag characteristic.

F. Function A harmonic component due to the dead time of the switching element of the PWM inverter appears as a harmonic current in the secondary current of the induction motor, and the torque pulsation due to the product of this harmonic current and the fundamental wave magnetic flux is corrected by reducing the magnetic flux. However, the magnetic flux correction amount is adjusted according to the angular frequency or the number of pulses based on the fact that the harmonic component is related to the inverter angular frequency and the number of PWM pulses, and the sudden change of the magnetic flux when the angular frequency or the number of pulses is changed. A first-order lag of a second-order time constant or more is added to the correction amount adjustment so as to prevent torque pulsation.

G. Embodiment FIG. 1 is a circuit diagram showing an embodiment of the present invention, which is applied to a non-interference vector control device using a PWM inverter.

Supplying a primary voltage to the induction motor 1 from the transistor-type inverter main circuit 2 _1. Each transistor Tr ₁ to Tr ₆ of the inverter main circuit 2 ₁ switching-control is made by Yotsute PWM waveform to the PWM waveform generation circuit 2 ₂ and the gate circuit 2 _3, the output voltage and frequency is controlled. Primary voltage supplied from the inverter main circuit 2 ₁ to the electric motor 1 is controlled so that the magnetic flux and the secondary current are orthogonal to each other in the electric motor 1. For this control, the direction of the magnetic flux is the α-axis, the direction of the secondary current is the β-axis orthogonal to the α-axis, and the α-phase primary current i ₁ α
^{* And} β-phase primary current i ₁ β ^{* respectively} from α-phase primary voltage e ₁ α,
In order to obtain the two-phase voltage signal of the β-phase primary voltage e ₁ β, the correction operation circuit 3 causes the β-phase primary current i ₁ β of the electric motor 1 to interfere with the magnetic flux and the α-phase primary current i ₁ α I try to remove the interference with the electric current. For this correction computation circuit 3 through the coefficient unit 3 ₁ of the primary resistance r ₁ value, beta-phase primary current i ₁ beta
^* Is multiplied by the power source angular frequency ω ₀ by the multiplier 3 ₂ , and the value obtained by passing through the coefficient unit 3 ₃ having the equivalent leakage inductance Lσ as a coefficient is subtracted from the multiplication result. The coefficient with respect to through the coefficient multiplier 3 ₄ value, the power supply angular frequency omega ₀ is multiplied by the multiplier ₃₅ to the alpha-phase primary current i ₁ alpha ^*, the primary inductance L ₁ as a coefficient to the multiplication result The values passed through the unit 3 ₆ are added.

The β-phase primary current command i ₁ β ^* is taken out as the output of the speed controller 5 by matching the speed setting value V _S ^* with the detection value ω _r of the speed detector 4 of the electric motor, and the power source angular frequency ω ₀ is the angular frequency. It is obtained from the calculated value of the slip angular frequency ω _S and the detected speed value ω _{r by the} calculation circuit 6. The angular frequency calculation circuit 6 sets the set value i ₁ α
^There is a divider 6 ₁ that calculates ^* and i ₁ β, and a coefficient unit 6 ₂ that multiplies the division result i ₁ β ^* / i ₁ α ^* by the coefficient 1 / τ ₂ with this divider 6 _1. Then, the slip angular frequency ω _S is calculated. Where τ ₂
Is the secondary time constant of the electric motor 1 and is the ratio of the secondary resistance r ₂ to the secondary inductance L ₂ .

Phase voltage calculation circuit 7 is two-phase voltage signals e ₁ alpha, three-phase voltage from e ₁ beta inverter command signals e _a ^*, e _b ^*, is to obtain e _c ^*, a sine wave signals required for this transformation The sinω ₀ t and the cosine wave signal cosω ₀ t are obtained from the triangular relation generating circuit 8 by using the power source angular frequency ω ₀ . Further, PWM waveform generation circuit 2 ₂ three-phase voltage signals e _a ^*, e _b ^*, is to obtain a Yotsute pulse width modulated waveform to a level comparison between e _c ^* and the triangular wave signal (carrier), and thus The triangular wave signal Tri of is the power source angular frequency ω from the circuit 6.
It is obtained from the triangular wave generating circuit 9 which uses ₀ to synchronize with the frequency. 10 is a rectifier for supplying DC power to the inverter main circuit 2 _1.

In this way, when the non-interference vector control of the primary voltage of the motor is performed by the PWM inverter, the inverter main circuit 2 ₁
In the transistor Tr ₄ to Tr ₆ transistors Tr ₁ to Tr ₃ and the lower arm of the upper arm PWM waveform so that at least one of which always conducts the opposing arm in its on period (for example, a transistor Tr ₁ and Tr ₄₎ Due to the dead time controlled and secured at the switching point, an error occurs in the inverter output voltage with respect to the control voltages e _a ^* , e _b ^* , e _c ^* . This dead time causes a low-order harmonic voltage in the output voltage of the inverter, which causes torque pulsation and uneven rotation during low-speed operation. Further, the error in the inverter output voltage deteriorates the tracking performance during acceleration.

Here, in order to reduce the torque pulsation and improve the followability, the α-phase primary voltage command e ₁ α of the correction arithmetic circuit 3 is corrected according to the inverter (power) angular frequency ω ₀ or the frequency of the triangular wave signal Tri. A magnetic flux correction circuit 11 is provided. This magnetic flux correction circuit 11 corrects so that the primary voltage command e ₁ α is decreased to lower the magnetic flux as the inverter angular frequency ω ₀ decreases (low speed), or the frequency of the triangular wave signal Tri increases (low speed). Follow the steps to correct the magnetic flux.

Further, in order to prevent the magnetic flux from suddenly changing due to a sudden change due to switching of the inverter angular frequency ω ₀ or the triangular wave signal Tri for the magnetic flux correction, a primary delay circuit 12 is provided at the subsequent stage of the magnetic flux correcting circuit 11. The primary voltage command e ₁ α is input to the phase calculation circuit 7 through the primary delay circuit 12.

The reason why the influence of the dead time can be reduced by the magnetic flux correction and that the first-order delay is effective for the magnetic flux correction will be described in detail with the simulation results.

First, the harmonic component generated in the inverter output voltage depending on the dead time will be described. The equivalent voltage due to the dead time td is generated at a period θ _c and a width θ _d of the carrier (triangular wave) frequency f _c with respect to the period of the inverter frequency f ₀ (a of the same figure), as shown in b in FIG. Have a relationship.

In addition, the number of pulses q of the generated voltage in a half cycle is PW
If the number of M pulses is p Is. Now, considering the low speed range where the control rate is small, θ _c can be regarded as substantially constant.

From these facts, when the equivalent voltage waveform due to dead time is Fourier expanded, the Fourier coefficient a _o = 0, b _n = 0 and a _n is In this equation (3), if n is an odd number, When n is an even number, an = 0.

From the relationship shown in FIG. Substituting in, the following equation (5) is obtained.

The above equation (5) holds when the number of pulses q is an even number,
When q is an odd number, it is calculated in the same manner and the following equation (6) is obtained.

In the above equations (5) and (6), As for the above, when the dead time td is about 20 μs and the inverter frequency f ₀ is operated at a low speed of 2 Hz or less, n is sufficient for the torque pulsation when the 19th order or less is considered, and the following equation holds.

From this equation (7), the above equation (5) becomes Becomes Of these, the sum S (n) of the series Σ parts
The following table shows the value obtained from the number of PWM pulses.

In this table, the ratio to the fundamental wave is shown in parentheses. As is clear from this ratio, the magnitude of each order component is the fundamental wave /
It becomes the order. Also, if 2π · S ₍₁₎ is obtained, which PWM
The number of pulses is also 2π · S ₍₁₎ = p. Substituting these relationships into the above equation (5A), Becomes Therefore, the harmonic voltage components for low-order harmonics (30th and lower) can be calculated. This equation corresponds to a square wave having an amplitude Eeq equal to the average value of the PWM waveform shown in FIG.

As described above, the dead time causes the harmonic voltage shown in the formula (8) or (9) to be generated, and the harmonic voltage causes the harmonic current to flow on the secondary side. Torque pulsation occurs between the magnetic flux and the wave magnetic flux, and further rotation fluctuation occurs.

In order to deal with this rotational fluctuation quantitatively, FMT (Figure Merit fo
r Torque is an abbreviation for details. Meiden Jikki No.4.Vol 165.82.P39-4
Consider 0) as the evaluation function. First, the secondary circuit impedance of the induction motor 1 at low speed operation will be considered for two motors of 0.75 kw and 55 kw, which are the specifications shown in the following table.

The leakage reactance x _H (= x ₁ + x ₂ ) for the 5th harmonic during low frequency operation is calculated for the _two EUTs as shown in the table below.

As is clear from this table, the resistance component (r ₁ + r ₂ ) is dominant, and the tendency is stronger for smaller capacity aircraft.

Therefore, the nth harmonic current is almost determined by the resistance component. The torque pulsation due to the harmonic current is n for the order n.
The ± 1st orders have mutually opposite polarities (for example, as shown in Table 1, the 5th order has a positive polarity and the 7th order has a negative polarity), and the magnitude of the torque pulsation depends on the fundamental wave magnetic flux Φ and the harmonic current I _n. The FMT for the nth harmonic can be expressed by the following equation.

Then, from the relationship of the above equation (8), the FMT during low frequency operation becomes the following equation.

From the equation (11), the rotation unevenness due to the torque pulsation changes in proportion to the product of the magnetic flux Φ and the PWM pulse number p if the dead time td is constant. Therefore, at low speed, the magnetic flux Φ may be changed in inverse proportion to the PWM pulse p.
Reference numeral 11 reduces the influence of dead time by correcting the magnetic flux command e ₁ α in inverse proportion to the number of output pulses of the triangular wave generator 9. Further, the PWM pulse number p is increased as the frequency becomes lower and is related to the inverter angular frequency ω _0. Therefore, the magnetic flux correction circuit 11 lowers the magnetic flux command e ₁ α as the inverter angular frequency ω ₀ becomes smaller. Torque pulsation is reduced and uneven rotation is reduced.

Here, the correction amount by the magnetic flux correction circuit 11 is determined according to the angular frequency ω ₀ or the triangular wave Tri, but a sudden change due to the switching of the triangular wave Tri due to the change of the angular frequency ω ₀ or the operating frequency causes a sudden change in the correction amount and a transition. The phenomenon occurs. Regarding this transient phenomenon, FIGS. 3 (A) and 3 (B) show the results of simulation performed by changing the response of the rotation speed to the cushion start by changing the pre-excitation.

Fig. 3 (A) shows the case without primary delay, and Fig. 3 (B) shows the case with primary delay, respectively, after 50% pre-excitation for 5 seconds and then accelerating to the rated speed by cushion start. In the figure, ia ^* is the armature current waveform, ie is the excitation current waveform, | λ | is the magnetic flux waveform, and ω _n is the motor angular frequency waveform. In addition, simulation went only by the fundamental wave.

As is clear from these waveform diagrams, when there is no first-order lag, the magnetic flux | λ | cannot follow the abrupt change of the exciting current ie, and vibration is seen at the angular frequency ω _n , whereas the first-order lag When there is, the followability of the magnetic flux is good, and the angular frequency ω
Vibration of _n is reduced.

H. Effects of the Invention As described above, according to the present invention, the magnetic flux command value is corrected according to the angular frequency of the PWM inverter or the number of PWM pulses during low-speed operation, and the magnetic flux correction has a first-order lag.
There is an effect that torque pulsation and rotation unevenness caused by dead time provided in the switching element of the M inverter can be improved by magnetic flux correction to reduce transient characteristics, and speed followability can be improved accordingly.

[Brief description of drawings]

FIG. 1 is a circuit diagram showing an embodiment of the present invention, FIG. 2 is a voltage waveform diagram for explaining a mode of a pulse voltage generated by dead time, FIGS. 3 (A) and 3 (B). 4) is a waveform diagram of the result of simulation based on the present invention. 1 …… Induction motor, 2 ₁ …… Inverter main circuit, 2 ₂ …… PW
M waveform generation circuit, 3 ... correction circuit, 6 ... angular frequency calculation circuit, 7 ... phase voltage calculation circuit, 9 ... triangular wave generation circuit,
11 …… Magnetic flux correction circuit, 12 …… First-order delay circuit.

Claims (2)

[Claims] 1. A vector control method for an induction motor, wherein a PWM inverter is connected as a drive power source, and the output of the PWM inverter is controlled so that the magnetic flux and the secondary current of the induction motor are orthogonal to each other. With the dead time set to be constant, the evaluation function FMT of the following equation that gives torque pulsation to rotation unevenness, k ′: Proportional constant φ: Magnetic flux n: Harmonic order E _d : Inverter DC voltage t _d : Dead time p: According to PWM pulse number, as the angular frequency of the PWM inverter becomes smaller during the low speed operation, the PWM pulse number p becomes smaller. A vector control method for an induction motor, characterized in that a magnetic flux correction is made to be higher and a command value of the magnetic flux φ is lowered, and the magnetic flux correction is provided with a first-order lag characteristic. 2. A vector control method for an induction motor, wherein a PWM inverter is connected as a drive power source, and the output of the PWM inverter is controlled so that the magnetic flux and the secondary current of the induction motor are orthogonal to each other. With the dead time set to be constant, the evaluation function FMT of the following equation that gives torque pulsation to rotation unevenness, k ′: Proportional constant φ: Magnetic flux n: Harmonic order E _d : Inverter DC voltage t _d : Dead time p: According to PWM pulse number, the magnetic flux is inversely proportional to the PWM pulse number p of the PWM inverter during the low speed operation. A vector control method for an induction motor, characterized in that magnetic flux correction for lowering a φ command value is performed, and the magnetic flux correction is provided with a first-order lag characteristic.