JPH0757119B2 - Low noise control method for induction motor - Google Patents

Description

The present invention relates to a method for controlling an induction motor by a PWM inverter, and switching of each switching element forming the PWM inverter regardless of the operating speed of the driven induction motor. The present invention relates to a control method for preventing an increase in noise caused by a harmonic frequency of a current overlapping with various mechanical resonance frequencies of an electric motor or a load system by maintaining the frequency at an optimum substantially constant value.

[Conventional technology]

The torque control method to which the present invention is applied is described in a paper entitled "New High Speed Torque Control Method for Induction Motor Based on Instantaneous Slip Frequency Control" on page 9, No. 1, page 9 of the Journal of Electrical Science.

In this paper, the motor input voltage is detected and integrated in the control circuit to obtain the motor magnetic flux. That is,
This is a so-called magnetic flux calculation type control method, and selects an inverter output voltage that follows a magnetic flux command given the length of the magnetic flux vector and draws a circular locus.

Further, the motor generated torque is calculated as a vector product of the magnetic flux and the motor input current, and the inverter output voltage whose magnitude follows the given torque command is selected. The control is performed so that the instantaneous values of the magnetic flux and the torque are kept within a predetermined error, and the inverter output voltage is updated at high speed every moment.

FIG. 2 is a block diagram of a torque control system adopting the output voltage detection method of the PWM inverter proposed in Japanese Patent Application No. 61-99228 filed by the applicant of the present invention in the control method described in the above paper. , DC bus 1 from positive bus 1a and negative bus 1b
Then, power is supplied to the three-phase induction motor 6 via the three-phase PWM inverter 3. The control circuit 7 processes the command and the detected current and voltage signals, and generates an energization signal for the switching element of the PWM inverter 3.

The PWM inverter 3 is composed of six arms, each of which is formed by connecting a switching element such as a transistor and a diode in antiparallel, and the three switching switches _Su , _Sv ,
It can be represented as S _w .

From each output terminal of PWM inverter 3 to current detector 5 _u , 5 _v , 5 _w
Power is supplied to the three-phase induction motor via and the voltage detector 2 is connected between the DC side positive and negative buses 1a and 1b so that each phase current and each phase voltage can be detected from these detectors and the switch state variable table described later. It has become.

If the primary terminal voltage and current of the three-phase squirrel cage induction motor are ▲ ▼ and ▲ ▼, respectively, and the secondary current is ▲ ▼, the voltage equation is However, the symbols ▲ ▼, ▲ ▼, and ▲ ▼ are vector displays of the quantities that have been converted on the direct and horizontal axes, that is, the d and q axes.
For example, ▲ ▼ is v _{1d for the} d-axis component and v _{1q for the} q-axis component. , And ▲ ▼ and ▲ ▼ are defined similarly. The left side of the equation represents the case where both d and q axis components are 0,
In the case of a squirrel cage rotor, the secondary voltage is as follows.

The constant in the formula is R ₁ ; Primary winding resistance L ₁₁ ; Primary inductance R ₂ ; Secondary winding resistance L ₂₂ ; Secondary inductance M; Mutual inductance m is rotational angular velocity, p is differential operator, j is vector Represents the product.

On the other hand, as the definition of magnetic flux, the primary magnetic flux ▲ ▼ is Expand the first line of the expression Substituting expressions and rearranging If you integrate both sides That is, the primary magnetic flux of the electric motor is obtained by the integral calculation of the equation.

The changeover switches S _u , S _v , and S _w may fall to the positive bus line 1a side or to the negative bus line 1b side, and do not take intermediate positions. If the former is the state 1 and the latter is the state 0, the output states of the inverter can all be expressed by the following switch state variable table.

Here, k is a number indicating the state of the changeover switch, and there are only eight of these. Further, _d and _q are switch state variables expressed by _d and _q axis components, and the actual d and q axis voltages v _1d and v _1q are the same as the voltage V of the DC voltage source 1. Multiply by Can be expressed as

The switch state variable table is shown in FIG.
The parentheses next to v ₁ indicate the states of the changeover switches S _u , S _v , and S _w , and represent the voltage vector that steps in 60 ° clockwise steps as k increases.

Note that k = 1 and k = 7 are called zero vectors and coincide with the origin in the figure. k = 0 and = 7 are changeover switches of FIG. 2 which determine the output of the inverter.
S _u , S _v , S _w all fall to the side of positive bus 1a or negative bus
The line voltage of the induction motor 6 is 0, although there is a difference whether it falls to the 1b side, which is the three-phase short-circuit mode. Also,
The reference axes of the u, v, and w phases are k =
It corresponds to the direction of 1, k = 3, k = 5.

The instantaneous torque T is the primary magnetic flux ▲ ▼ and the primary current ▲ in the formula.
It is calculated by the formula as the vector product of ▼.

Here, φ _1d , φ _1q and i _1d , i _1q are respectively the first magnetic flux ▲
▼ Each component when the primary current ▲ ▼ is decomposed into the d and q2 axes.

Blocks 701 and 703b are blocks for calculating the primary terminal voltage ▲ ▼ from the states of the changeover switches S _u , S _v , S _w and the voltage V of the DC voltage source 1 detected by the voltage detector 2, and are switch state variables. Calculated from the table and formula.

Block 702 is 3 detected by current detectors 5 _u , 5 _v , 5 _w
This is a block that converts the phase currents i _u , i _v , and i _w into the d and q biaxial components by the following equation.

This primary current ▲ ▼ is multiplied by the primary winding resistance R ₁ in block 703a, and the primary terminal voltage ▲ in block 704.
▼ ₁ from primary winding resistance R ₁ and subtracting the primary current ▲ ▼ product of.

A block 705 is a block for performing integral calculation of the magnetic flux according to the equation, and the primary magnetic flux ▲ ▼ ₁ both d and q axis components φ _1d , φ _1q
Is calculated, and the magnetic flux vector length φ ₁ is calculated by the following equation in block 706.

Further, in block 710, as shown in the magnetic flux state diagram of FIG. 4, the clockwise rotation angle θ with respect to the d axis of the primary magnetic flux vector is 30 °, 90 °, 150 ° as a boundary line. ,twenty one
The control tag fθ is generated in the following manner depending on which region is divided every 60 ° of 0 °, 270 ° and 330 °.

−30 ° ≦ θ <30 °; fθ = I 30 ° ≦ θ <90 °; fθ = II 90 ° ≦ θ <150 °; fθ = III 150 ° ≦ θ <210 °; fθ = IV 210 ° ≦ θ < 270 DEG; f [theta] = V 270 ° ≦ theta <330 deg; state control diagram f [theta] = VI Figure 6 is a hysteresis comparator, the flux vector length phi ₁ is the magnetic flux command value ▲ φ ^* ₁ ▼ respect, error limit (hysteresis Value) using Δφ A control flag fφ for controlling so that That is, the magnetic flux vector length φ ₁ increases and is the upper limit. Control flag fφ = 0 is generated when the magnetic flux vector reaches φ, and the magnetic flux vector length φ ₁ decreases to the lower limit. Control flag fφ = 1 for instructing the magnetization is generated.

Thus, the magnetic flux vector length φ ₁ is controlled so as to draw a limit cycle in the direction of the arrow shown in FIG. 6, but in reality, the magnetic flux vector length φ ₁ calculated by the equation in block 706. Is subtracted from the magnetic flux command value ▲ φ ^* ₁ ▼ in block 708, and a control flag fφ = 1,0 is generated in block 711 according to the state control diagram of FIG.

The magnetic flux limit cycle shown in FIG. 6 corresponds to the fact that the head of the vector of the primary magnetic flux ▲ ▼ is controlled so as to always exist in the illustrated annular portion, as for FIG. .

A combination of the control flag fφ according to FIG. 6 and the control flag fθ described with reference to FIG. 4, for example, fφ = 1, fθ = I
If the control flag of is set, the area is −30 ° ≦ θ
Since it means the magnetization mode at <30 °, the primary magnetic flux ▲
▼ Primary voltage to be integrated into vector ▲ ▼ _One vector has an outward component of a circle, and there is a possibility that only k = 1, 2, or 6 is selected from FIG.

Block 707 is a block which calculates the calculated instantaneous torque T a vector product between the output of the block 702, 705 by the equation, by subtracting the instantaneous torque T from the torque command T ^* at block 709, determined by the torque command T ^* and the formula A control flag fτ is generated in block 712 according to the state control diagram of FIG. 7 so that the difference from the instantaneous torque T can be suppressed within a predetermined error limit.

FIG. 7 is a state control diagram of the three-value hysteresis comparator. When the motor is running, the torque deviation T ^* -T is the upper limit value ΔT.
_{When 1} (ΔT ₁ > 0) is reached, the acceleration mode control flag f
Generate τ = 1. When the motor is accelerated and the torque deviation reaches the lower limit value −ΔT ₂ (ΔT ₂ > 0), a zero vector mode control flag fτ = 0 is generated, the torque is gradually decreased, and the deviation is increased again, and the upper limit value ΔT ₁ When it reaches
The limit cycle in which the hysteresis loop in the upper half part of FIG. 7 orbits in the direction of the arrow is drawn.

When this is expressed in the time domain, the instantaneous torque T fluctuates significantly as shown in the torque waveform diagram of FIG. 5, and the allowable error (hysteresis, which is the sum ΔT ₁ + ΔT ₂ of the upper and lower deviations across the torque command T ^*). It reciprocates within the band of (value) ΔT as the maximum value T ₁ and the minimum value T ₂ .

Next, when the motor is performing regenerative braking, a hysteresis loop in the lower half of FIG. 7 is drawn, and when the torque deviation reaches the negative lower limit value ΔT ₁ (ΔT ₁ > 0), the deceleration mode A control flag fτ = -1 is generated. After that, repeat the limit sills in the direction of the arrow as in the case of powering. Thus, block 712 outputs the control flag fτ = 1,0, −1.

A block 713 is a block for determining the most suitable inverter output voltage for each combination of the three control flags fθ, fφ, fτ output from the blocks 710, 711, 712, and the vector length of the primary magnetic flux ▲ ▼ described in FIG. The rotation direction and speed are controlled by these three control flags fθ, fφ, fτ.

For example, when the control flags fφ = 1 and fθ = I as described above, if the voltage vector is represented by ▲ ▼ (k) in accordance with k in the switch state variable table, k may be selected as the voltage vector. However, if the control flag fτ = 1 at this time, the vector k = 2 having the component rotating clockwise, that is, the output voltage vector ▲ ▼ (2) is selected. If fτ = -1, it is (6), and if fτ = 0, it is a zero vector.
▼ (0) or ▲ ▼ (7) is selected.

The following switching table shows three control flags f
The values of the number k of the output voltage vector are shown for all combinations of φ, fθ, and fτ.

By referring to this switching table in block 713 for each operation cycle, a switching signal is sent to the inverter 3 and instantaneous control of magnetic flux and torque is performed.

The inverter frequency can be considered as the rotational speed of the primary magnetic flux vector in FIG. 4, but this is not given from the outside but occurs as a result of integrating the voltage vector by the formula.

[Problems to be solved by the invention]

Generally, the switching frequency of the PWM inverter rises with the speed of the motor, but when the frequency of harmonics of the motor current matches the resonance frequency of the structure including the magnetic circuit of the motor or the load system during the frequency fluctuation, a resonance state occurs. There was a drawback that the noise was significantly increased.

Also in the above-mentioned inverter in the new high speed torque control method, the average switching frequency changes with speed, becomes low at low speed and becomes high as the motor accelerates, and becomes low again at high speed region. The normal switching frequency is the value per switching element that constitutes the inverter, and the high frequency of the motor current is 2, 4, 6 of the switching frequency.
…… It is composed of sideband groups centered on 2n times frequency. The harmonic of the double frequency among them will be called the fundamental wave pulsation frequency.

FIG. 8 shows a current harmonic spectrum generated during steady operation in the above control method. In the case of instantaneous value control like this method, the switching cycle changes every moment, so
It is hard to show a standing spectrum, and the averaged frequencies appear during the analysis period, so the spectrum is dispersed before and after the peak, and the noise is perceived as lower than the conventional PWM inverter.

In FIG. 8, f ₁ is the inverter frequency, f ₂ is the basic pulsation frequency, and f ₃ is one of the frequencies, which includes frequencies dispersed before and after the peak. f _a and
f _b represents the resonance frequency of the structure.

FIG. 9 is a graph showing how the basic pulsation frequency f ₂ of the motor current changes with respect to the speed. The hysteresis value ΔT of the torque three-value hysteresis comparator, that is,
The sum of the lower deviations ΔT ₁ + ΔT ₂ is used as a parameter. A case is shown in which the hysteresis value ΔT is smaller in the order of A, B, and C from the upper side. The curve A has the smallest hysteresis value ΔT, and the curve C has the largest hysteresis value ΔT.

As a result of the experiment, substantially the same basic pulsation frequency characteristic was obtained even when the allowable error limit Δφ of the magnetic flux hysteresis comparator was used as a parameter.

The reason why the basic pulsation frequency f ₂ increases in the intermediate speed range of the motor is interpreted as follows. That is, the primary current ▲ ▼ is given by the following equation.

When voltage is applied, that is, when current is increased In 3-phase short-circuit mode, that is, when the current decreases Where l ₁ ; primary winding leakage inductance R ₁ ; primary winding resistance ▲ ▼; primary terminal voltage ▲ ▼; motor induced voltage.

Since the primary magnetic flux ▲ ▼ vector draws a substantially circular locus, and its value does not change suddenly, the value of the instantaneous torque T given by the equation is almost equal to the behavior of the primary current ▲ ▼ by
The increase and decrease are repeated exponentially to make a round trip between the maximum value T ₁ and the minimum value T ₂ .

FIG. 10 shows the torque waveform chart of FIG. 5 in more detail. (A) is an instantaneous torque waveform chart at low speed, (b) is at medium speed, and (c) is at instantaneous speed. Absolute value of primary voltage |
Since ▲ ▼ | is always constant, the absolute value of the motor-induced voltage | ▲ ▼ | is small. At low speeds, the torque increase is steep and the damping is slow, while at high speeds, the absolute value of the motor-induced voltage is absolute. Since the value | ▲ ▼ | becomes large, the torque increases slowly and the damping becomes steep.

Further, when the torque is increased and decreased at the medium speed, the situation is almost intermediate between the low speed and the high speed, and therefore the average switching frequency becomes higher than that at the low speed and the medium speed.

As described in detail above, according to the conventional instantaneous magnetic flux and torque control method, the basic pulsation frequency as the PWM inverter also fluctuates as the motor speed changes, and the harmonic current frequency also fluctuates. It was unavoidable that when the mechanical resonance frequency coincided with the resonance frequency, the resonance state was reached and the noise increased.

That is, as shown in FIG. 9, in the case of the A curve having a small torque tolerance (hysteresis value) ΔT, the resonance frequency fa of the structure does not resonate within the normal speed range, but Regarding the resonance frequency fb, resonance occurs within the normal speed range.

In addition, the torque tolerance C (hysteresis value) is large.
In the case of a curve, the fundamental pulsation frequency does not reach the resonance frequency fb of the structure, but the resonance frequency fa of the structure resonates within the normal speed range.

As described above, by appropriately selecting the torque allowable error ΔT, it is unavoidable that a state of resonating with the resonance frequency of the structure including the magnetic circuit of the electric motor occurs.

[Means for solving problems]

A low noise control method for an induction motor according to the present invention, in the conventional instantaneous magnetic flux and instantaneous torque control method, detects the switching frequency of the inverter and instantaneously sets the error signal with a desired average switching frequency value to zero. By variably controlling at least one of the allowable error (hysteresis value) for the command value of the magnetic flux and the instantaneous torque, it is possible to control so that the average switching frequency of the inverter becomes constant from the low speed region to the high speed region of the motor speed. It is a feature.

That is, in FIG. 9, the basic pulsation frequency f ₂ of the PWM inverter is limited to the frequency band f ₂ ′, f ₂ ″ or f ₂ which is above or below the mechanical resonance frequencies fa and fb of the motor structure. If used, noise reduction can be realized without causing a resonance state.

[Action]

An example of variably controlling the torque allowable error ΔT will be described with reference to FIG. FIG. 11 is a graph showing the relationship between the motor speed and the average switching frequency, which is essentially the same as in FIG. 9, but the horizontal axis represents the allowable range of torque (hysteresis value) and the vertical axis represents the average switching frequency. And the motor speeds n ₁ , n ₂ , and n ₃ are redrawn as parameters.

That is, the average switching frequency becomes higher as the torque tolerance ΔT becomes smaller, and becomes lower when the motor is at high speed n ₁ or low speed n ₃ and becomes higher when it is at medium speed n ₂ .

Therefore, the detected switching frequency is the switching frequency command Torque tolerance (hysteresis value) Δ
By increasing T, switching frequency command When it is lower, the switching frequency average value sw is set to the switching frequency command by reducing the torque tolerance ΔT. The inverter can be operated at a stable point that coincides with

Although the case where the allowable error ΔT of the torque is variably controlled has been described in detail above, substantially the same operation is performed when the error limit Δφ of the magnetic flux is variably controlled. In this case, the horizontal axis of Fig. 11 is the magnetic flux error limit (hysteresis value) Δφ, and the average switching frequency at a constant motor speed.
When sw is measured, almost the same graph can be obtained.

That is, the average switching frequency is the magnetic flux error limit Δ
It becomes higher as φ becomes smaller, and becomes lower as the magnetic flux error limit Δφ becomes larger.

Therefore, the average value sw of the detected switching frequency is the switching frequency command. Error limit (hysteresis value) Δφ
The switching frequency average value sw is set to the switching frequency command by increasing the error frequency and decreasing the error limit (hysteresis value) Δφ of the magnetic flux when it is low. It is possible to operate the inverter at a stable point corresponding to

Of course, it is apparent that a system with even better response can be obtained by variably controlling both the torque allowable error ΔT and the magnetic flux error limit Δφ as described above.

〔Example〕

FIG. 1 is a block diagram of an embodiment of a low noise control method for an induction motor according to the present invention, and the same reference numerals as those in FIG. 2 indicate parts having the same functions. The difference from FIG. 2 of this embodiment is that blocks 714 to 718 are added to the control circuit 7 and a switching frequency command is issued. Is added as an input.

A block 714 shows that the control flags fτ and fφ change,
It is a switching frequency detector that detects the number of times of switching per unit time. Block 715 is a fast cycle block 7
This is a block that finds the average value sw of the switching frequency by averaging the fluctuations of the switching frequency that is the output of 14.

In block 718, the switching frequency command is the average value sw of the switching frequency output from block 715. Subtract from. Block 716 is the error amplifier, block 7
Invert and amplify 18 outputs, When it becomes, negative output, When it becomes, it produces a positive output.

A block 717 is a block for setting torque and magnetic flux tolerances. The signals from the block 716 give tolerances (hysteresis values) Δφ and ΔT to the hysteresis comparators of the blocks 711 and 712. The block 717 generates standard hysteresis values Δφ and ΔT when the signal from the block 716 is 0.
When the signal from 716 has a positive value, Δφ and ΔT are increased, and when it has a negative value, Δφ and ΔT are decreased.

If the switching frequency drops at low or high speeds, Therefore, the output of block 716, which is the error amplifier, is the absolute value of the difference. It becomes a negative value according to
A small value is commanded for T and Δφ, and the average value sw of the switching frequency operates so as to increase. Also, the electric motor becomes medium speed Then, it operates in the opposite direction and reduces the average value sw of the switching frequency.

In this embodiment, the case where both the allowable error (hysteresis value) Δφ of the magnetic flux hysteresis comparator of block 711 and the allowable error (hysteresis value) ΔT of the three-valued hysteresis comparator of block 712 are variably controlled. However, it is obvious that the same operation can be performed by variably controlling only one of them.

Further, in the above embodiment, the switching frequency command Was controlled to be constant, but the effect can be improved by setting it as variable as necessary.

That is, when the torque command T ^* is large, the inverter output current is generally large and the switching loss is large. On the contrary, when the torque command T ^* is small, the inverter output current becomes small and the switching loss also becomes small.

Therefore, when the torque command T ^* becomes large, the switching frequency command When the torque command T ^* is small, the switching frequency command Is set to be large, it is possible to prevent the loss due to the current and the loss due to the number of times of switching of the switching elements that constitute the PWM inverter 3 from increasing at the same time.

That is, when it is necessary for the electric motor to generate a large torque, this can be prioritized, and when a small torque is sufficient, the switching frequency can be increased to reduce noise.

Generally, magnetic noise tends to be more annoying compared to mechanical noise when the motor speed is low.Therefore, when the motor speed is low, the switching frequency command To reduce noise and increase switching speed command as speed increases. It is also effective to reduce

〔effect〕

According to the low noise control method for an induction motor of the present invention,
By operating the inverter at a harmonic current frequency that avoids frequencies such as the mechanical resonance frequency fa or fb of the structure, low noise operation that suppresses the generation of annoying electromagnetic noise due to resonance becomes possible.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of a low noise control method for an induction motor according to the present invention, FIG. 2 is a block diagram of an example of a conventional instantaneous magnetic flux and torque control system, and FIG. 3 is a switch state variable table. Inverter output voltage vector diagram, 4th
Fig. 5 is a magnetic flux state diagram showing an instantaneous control method of the primary magnetic flux vector of the electric motor, Fig. 5 is a torque waveform diagram, Fig. 6 is a state control diagram of the magnetic flux hysteresis comparator, and Fig. 7 is a torque three-value hysteresis comparator. State control diagram, FIG. 8 is a current harmonic spectrum generated during steady operation, FIG. 9 is a graph showing changes in the motor current basic pulsation frequency with respect to the motor speed, and FIG. 10 is the torque waveform diagram of FIG. 5 in more detail. (A) at low speed, (b) at medium speed,
(C) is a waveform diagram of the instantaneous torque at high speed.
Figure 11 is a graph showing the relationship between torque tolerance and average switching frequency. 1 ... DC voltage source, 2 ... voltage detector, 3 ... PWM inverter, 5 _u , 5 _v , 5 _w ... current detector, 6 ... induction motor,
7 ... Control circuit.

Claims (4)

[Claims] 1. An instantaneous value of a magnetic flux vector and torque is calculated from a voltage and current vector component signal in a stationary coordinate system, and when the error between the magnitude of the calculated magnetic flux value and its command value exceeds a predetermined allowable range, the magnetic flux increases. Alternatively, a binary hysteresis that generates the first control flag that commands magnetic flux reduction
Second control for instructing to increase / decrease torque or to maintain current state depending on the case where the error between the comparator and the calculated torque value and its command value exceeds a predetermined permissible range and is within the predetermined permissible range A three-value hysteresis comparator for generating a flag, and a magnetic flux position detecting means for generating a third control flag which indicates in which arc area the current circumference is divided from the magnitude and sign of the magnetic flux vector component. In addition, the torque generated by the induction motor is controlled so as to follow the command value by directly designating the switching state of the inverter that generates the voltage vector that optimizes the torque response by combining the three control flag values. In addition to the method of controlling the magnetic flux vector to draw an approximate circular locus, Means for calculating or detecting the switching frequency, means for calculating or calculating the switching frequency from a preset switching frequency command value, means for inverting and amplifying the output signal of the calculating means, and output signal for the inverting amplifying means. Is a positive value, at least one of the magnetic flux or torque allowable error set value is increased, and a negative value is changed to a small value, and the switching frequency is controlled to follow the given command value. Low noise control method for induction motor. 2. The low noise control method for an induction motor according to claim 1, wherein the switching frequency commanded in advance is a constant value that does not depend on the operating state of the motor. 3. The low noise control method for an induction motor according to claim 1, wherein the switching frequency commanded in advance is set to an optimum value according to the torque command value of the motor. 4. The low noise control method for an induction motor according to claim 1, wherein the switching frequency commanded in advance is set to an optimum value according to the speed of the motor.