JPH0820480B2 - Induction motor leakage inductance measuring device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention includes a leakage inductance of an induction motor (also simply referred to as an induction machine) driven through a power converter in an induction machine primary voltage and current. The present invention relates to a measuring device for measuring (identifying) using a high frequency component.

[Conventional technology]

As a measurement method of this kind, conventionally, for example, as shown in FIG. 11, the rotor of the induction machine 2 is constrained and the power converter (or AC power supply) 1 is connected to connect the primary voltage, the primary current vector V ₁ , It is known that the leakage inductance L is obtained from the relation of I ₁ .

[Problems to be Solved by the Invention]

However, such a method has a drawback that the leakage inductance cannot be automatically measured because the rotor needs to be restrained.

Therefore, an object of the present invention is to provide a circuit capable of automatically measuring leakage inductance in a variable speed control device for an induction motor driven via a power converter without restraining the rotor of the induction machine. And

[Means for solving the problem]

An arithmetic means for estimating the high frequency voltage appearing in the leakage inductance from the primary current to obtain an estimation error from the actual value included in the primary voltage, and the parameter (leakage inductance setting value) is slightly changed in the vicinity of the setting value, Adjusting means for adjusting the parameter setting error appearing in the estimated error signal to a minimum.

[Action]

Focusing on the fact that the high frequency component of the voltage generated by switching in the power converter that drives the induction machine is mostly affected by the leakage inductance, and the calculation means for calculating the leakage inductance voltage drop, and the parameter that is changed slightly The leakage inductance is identified (measured) by the adjusting means that converts the setting error of the above into the change amount of the high frequency estimation error signal and adjusts the parameters optimally. However, if the stray load loss cannot be ignored, the high frequency loss The same identification is performed for the above to identify the leakage inductance that is a combination of the leakage inductance and the high frequency loss.

〔Example〕

FIG. 1 is a block diagram showing a first embodiment of the present invention. That is, for the induction machine 2 driven by the pulse width modulation (PWM) inverter 1A, the leakage inductance identifier 3
Is added. The leakage inductance identifier 3 is a coordinate converter 31A, 31B for converting the primary voltage V ₁ of the induction machine and the current I ₁ into voltages and currents of predetermined biaxial components, respectively, and the voltage drop of the leakage inductance from the current components. From the differentiator 32 to be operated, the adder 33 that takes out the deviation between the output of this differentiator 32 and the output of the coordinate converter 31B, the oscillator 34, the high frequency component effective value detector 35, the synchronous detector 36, the parameter adjuster 37, etc. Composed.

FIG. 2 (a) shows a T-type fundamental wave equivalent circuit of the induction machine.
Here, R ₁ and R ₂ ′ are primary and secondary resistances, L ₁ and L ₂ ′ are primary and secondary inductances, M is an exciting inductance, and S is a slip. However, "'" is added to represent the primary conversion value. It is known that a fundamental wave equivalent circuit (b) can be obtained by modifying this circuit in order to obtain the leakage inductance seen from the primary side of this circuit. The leakage inductance L in this equivalent circuit is L = L ₁ −M ² / L ₂ ′. At high frequencies, the induction machine can be approximated by a high frequency equivalent circuit (C).
This leakage inductance is identified here as follows. At high frequencies, the impedance of the leakage inductance becomes large, so the voltage applied to the resistance component in the high frequency equivalent circuit is small, and the phase is 90 degrees behind. First
When the L set value L ^* of the differentiator 32 in the figure is set to the actual value, the input of the high frequency effective value detector 35 is equal to the estimated voltage drop value of the resistance components R ₁ and R ₂ ′. When L ^{* of the} differentiator 32 is changed, the PWM obtained from the high frequency effective value detector 35
The effective value of the waveform component is a curve that takes the minimum value when L ^* is equal to the actual value, as shown in FIG. Therefore, when the output of the oscillator 34, which generates a sine wave having a frequency sufficiently lower than the PWM frequency as shown in FIG. 3B, is added to L ^* and modulated, the effective value detector 35 outputs the signal shown in FIG. As described above, when the set value is larger than the actual value, a signal in phase with the oscillator output is obtained, and when the set value is small, an opposite phase signal is obtained, and when the set value is equal to the actual value, the frequency component of the oscillator becomes zero. When a signal of the same frequency component as this oscillator is detected by the synchronous detector 36, a positive value is obtained in the in-phase and a negative value is obtained in the opposite phase.
When L ^* is adjusted to 37 and L ^* automatically converges to the actual value, the value of L can be identified.

FIG. 4 is a block diagram showing a second embodiment of the present invention. This is an example when stray load loss cannot be ignored,
Therefore, in addition to the parameter adjuster 37A as a leakage inductance adjuster, a parameter adjuster 37B as a stray load adjuster is provided, and switches S ₁ and S ₂ for selecting them are further added. Further, in order to obtain the voltage drop due to the stray load and the leakage inductance, an incomplete differentiator 32A is provided and the leakage inductance identifier 4
Is composed.

That is, as the equivalent circuit of the induction machine when there is a stray load loss, as shown in the basic equivalent circuit of FIG. 5 (a) and the high frequency equivalent circuit of FIG. 5 (b), the stray load loss equivalent resistance is parallel to the leakage inductance L. Assuming a circuit to which R is added, the leakage inductance at this time is identified as follows.

A) L set value adjustment switches S ₁ and S ₂ are parameter adjusters 37A for adjusting L ^*
Set to the input / output side of. When L ^* given to the incomplete differentiator 32A is changed, the PWM obtained from the high frequency effective value detector 35
As shown in Fig. 6, the effective value of the waveform component is a curve that takes the lowest value near the value where L ^* is equal to the actual value. Therefore,
Similar to the example of FIG. 1, the output of the oscillator 34 that generates a sine wave of a frequency sufficiently lower than the PWM frequency is added to L ^* and modulated to synchronize the same frequency component as the oscillator output included in the high frequency effective value. The wave is detected by the wave detector 36 and given to the parameter adjuster 37A so that L ^* becomes the minimum.

B) R set value adjustment switches S ₁ and S ₂ are parameter adjusters 37B for adjusting R.
Set to the input / output side of. When the R set value R ^* given to the incomplete differentiator 32A is changed, the effective value of the PWM waveform component shows a curve having the lowest value near the value where R ^* is equal to the actual value, as shown in FIG. Therefore, similarly to the L adjustment, the oscillator frequency component included in the high frequency effective value is detected by the synchronous detector 36 and given to the parameter adjuster 37B to adjust R ^* .

By repeatedly executing the above A) and B) alternately, R ^* and L ^* respectively converge to their actual values.

Although as used in the switching oscillator 34 and the synchronous detector 36 by a switch S _1, S ₂ or more, by the oscillator and the synchronous detector of providing two each, omitting switch S _1, S ₂ You can

FIG. 7 is a block diagram showing a third embodiment of the present invention.

In the circuit of the third embodiment shown in FIG. 7, the stray load loss is not negligible as in the circuit of the second embodiment already described with reference to FIG. The first random signal generator 41A for adjusting the minute adjustment parameter and the second random signal generator 41B for adjusting the stray load loss minute adjustment parameter are separately installed, and the leakage inductance / stray load loss identifier 5 Are configured.

Here, the control period of the second random signal output from the second random signal generator 41B is the first random signal generator 41A.
By setting it to be an integral multiple of the control cycle of the first random signal from, the leakage inductance that is easy to adjust and the high-frequency loss that has low detection sensitivity of the adjustment error are both accurate and stable. I am expecting identification.

In the third embodiment shown in FIG. 7, the PWM inverter 1A
The leakage inductance / stray load loss identifier 5 connected to the induction machine 2 that operates with the power source as a power source is a coordinate converter 31B that converts the primary voltage V ₁ into a biaxial component and the primary current I ₁ is a biaxial component. Coordinate converter 31A for converting to, incomplete differentiator 32A for calculating leakage inductance voltage drop, high frequency effective value detector 3
5, first and second random signal generators 41A and 41B, multiplier 42A
And 42B, and parameter adjusters 37A and 37B.

FIG. 8 is an equivalent circuit diagram for one phase when there is a stray load loss in the leakage inductance.

As shown in FIG. 8, the leakage inductance is
This can be approximated by connecting the leakage inductance L and the high frequency loss equivalent resistance R in parallel. Note that M is the exciting inductance, ω is the primary angular frequency, and ω ₂ is the axial angular velocity. Here, the incomplete differentiator 32A estimates the leakage inductance voltage drop V _L using the primary current I ₁ by the following equation (1).
Where L ^* is the leakage inductance estimate and R ^* is the high frequency loss equivalent resistance estimate.

FIG. 9 is a graph showing a curve in which the high frequency component effective value Y changes depending on the leakage inductance estimated value and the stray load loss equivalent resistance estimated value.

That is, the curved surface Y is a function of L ^* and R ^* , and this curved surface Y
The curve that composes has one point where Y is always a minimum when either L ^* or R ^* is fixed and the other is changed. (The solid line shows the relationship between L ^* and Y when R ^{* is} constant,
The broken line shows the relationship between R ^* and Y when L ^{* is} constant. ) However, it is easy to find a point where ∂Y / ∂L ^* is large and this ∂Y / ∂L ^* is equal to zero, but ∂Y / ∂R ^* is small and the slope detection sensitivity is small. Since R is low, R ^* such that ∂Y / ∂R ^* is equal to zero has the property of greatly changing depending on the value of L ^* .

Then, on the curved surface Y, (∂Y / ∂L ^* ) = 0
Along the curve P drawn by the dotted line, a point Q where (∂Y / ∂R ^* ) = 0, that is, a point where Y = 0 (at this time, the leakage inductance estimated value L ^* matches the actual value L, Moreover, the stray load loss equivalent resistance estimated value R ^* coincides with its actual value R) is obtained as follows.

A) Control cycle The control cycle of the first random signal generator 41A and the parameter adjuster 37A is T ₁ , the control cycle of the second random signal generator 41B and the parameter adjuster 37B is T _2, and these T ₁ and T ₂ Let and be the relationship shown in equation (2). However, n is a positive integer.

T ₂ = N · T ₁ (2) Here, n is set to a sufficiently large value, and as shown in the timing chart of FIG. 10, L ^* is adjusted in the cycle T ₁ (that is, T ₂
R ^* is adjusted in the cycle T ₂ while adjusting n times during the period.

10 is a timing chart showing the operation of each part of the circuit of the third embodiment shown in FIG. 7. FIG. 10 (a) shows the change of the first random signal E ₁ , and FIG. ) Is the change of the second random signal E ₂ , FIG. 10 (c) is the change of the high frequency detection value output from the high frequency effective value detector 35, FIG. 10 (d)
Shows changes in the leakage inductance estimated value L ^* , and FIG. 10 (e) shows changes in the stray load loss equivalent resistance estimated value R ^* .

B) Adjustment of L ^{* in} period T ₁ Sum of R ^* and second random signal E ₂ , ie R ^* (j) +
E ₂ (j) (where j is 0, 1, 2 ...) Is used to identify the actual leakage inductance value L in the control cycle T ₁ during the control cycle T ₂ set in the arithmetic circuit. At this time, E ₂ (j)
Is constant, and E ₁ is + ΔL ^* or − for each cycle of T _1.
The value ΔL ^* will be taken randomly (Fig. 10 (a)).
reference).

In the parameter adjuster 37A, the multiplication result of the high frequency detection value and the first random signal E ₁ is integrated by the gain K ₁ , and the leakage inductance estimated value L ^* is determined by the equation (3), for example. (However, i = 0,1,2,3 ……) In addition, Therefore, If the gain K ₁ is set so that, (∂Y / ∂L ^* ) =
At 0, it converges to L ^* which gives the minimum value of Y. If a low pass filter is combined with the parameter adjuster 37A, the ripple of L ^* can be reduced.

C) Adjustment of R ^{* in} cycle T _{2 When} the sum of R ^* and the second random signal E ₂ changes for each control cycle of T ₂ , whereas L ^* is a minimum value of E is given. , L ^* , and R ^* exist on the curve P where (∂Y / ∂L ^* ) = 0 (see FIG. 9). Therefore, for each control cycle of T ₂ ,
The second random signal E ₂ is randomly changed to + ΔR ^* or −ΔR ^* (see FIG. 10B), and the parameter adjuster 37B
Then, R ^* is adjusted by the following equation (5) based on the calculation result of the high frequency detection value and the second random signal E ₂ . Note that K ₂ is an adjustment gain. (However, j = 0,1,2, ...) In addition, Therefore, set n to a sufficiently large value, If the adjustment gain K ₂ is set so as to satisfy the following condition, R ^* converges to a value that makes the high-frequency component detection value Y a minimum value of zero at the overshoot amount ΔR ^* or less.

If a low-pass filter is combined with the parameter adjuster 37B, the ripple of R ^* can be reduced as in the case of the item B).

In each of the above-mentioned three types of embodiments, each of the leakage inductance identifiers 3 and 4 or the leakage inductance / stray load loss identifier 5 is composed of hardware that processes an analog quantity, but processes a digital quantity. Of course, it can be replaced by software.

〔The invention's effect〕

According to this invention, since the leakage inductance is identified by using the high frequency component generated in the power converter, it can be measured regardless of the operating state of the induction motor, and the rotor restraint is unnecessary. Measurement can be realized.
Further, since the method of detecting the parameter error using the sine wave is used, the correlation function with the switching waveform becomes zero, and stable identification operation can be obtained. Further, in order to identify the leakage inductance and the high frequency loss, a separate random signal is used for the modulation signal instead of the sine wave, so that the interference with the high frequency due to the switching of the power converter can be removed and the stability is improved. Furthermore, by appropriately setting the relationship between the control periods of both random signals, there is no beat between the adjustment loops of each parameter, and the transient overshoot of the parameter estimation value can be suppressed, and stable identification can be achieved. Have.

[Brief description of drawings]

FIG. 1 is a block diagram showing the first embodiment of the present invention, and FIG.
FIG. 3 is a circuit diagram showing an induction motor equivalent circuit in which stray load loss is ignored, FIG. 3 is a graph showing the relationship between leakage inductance setting value and high frequency effective value, and FIG. 4 is a block diagram showing a second embodiment of the present invention. Fig. 5 is a circuit diagram showing an induction machine equivalent circuit in which stray load loss is taken into consideration, Fig. 6 is a graph showing the relationship between each set value of stray load resistance and leakage inductance, and the high frequency effective value. Fig. 7 is FIG. 8 is a block diagram showing a third embodiment of the present invention, FIG. 8 is an equivalent circuit diagram for one phase when there is a high frequency loss in the leakage inductance, and FIG. 9 is a leakage inductance estimated value and a stray load loss equivalent resistance estimated value. FIG. 10 is a timing chart showing the operation of each part of the circuit of the third embodiment shown in FIG. 7, and FIG. 11 is a conventional example of an induction machine leakage inductance measuring method. In the schematic diagram shown is there. 1 ... Power converter or AC power supply, 1A ... PWM inverter, 2 ... Induction machine, 3,4 ... Leakage inductance identifier, 5 ... Leakage inductance / stray load loss identifier, 31
A, 31B ... Coordinate converter, 32 ... Differentiator, 32A ... Incomplete differentiator, 33 ... Adder, 34 ... Oscillator, 35 ... High frequency effective value detector, 36 ... Synchronous detector, 37,37A, 37B ...... parameter adjuster, 41A ...... first random signal generator, 41B ...... second random signal generator, 42A, 42B ...... multiplier, S _1, S ₂ ...
…switch.

Claims (2)

[Claims] 1. To measure the leakage inductance of an induction motor driven through a power converter including a switching element, a high frequency voltage appearing in the leakage inductance is estimated from a primary current and an actual value included in the primary voltage is estimated. Calculating means for obtaining an estimated error between the calculating means and the adjusting means for adjusting the leak inductance setting value of the calculating means to a minimum in a sinusoidal manner so as to minimize the leak inductance setting error appearing in the estimated error signal. A leakage inductance measuring apparatus for an induction motor, comprising: 2. In order to measure the leakage inductance and the high frequency loss of an induction motor driven through a power converter composed of a switching element, a high frequency voltage appearing in a combined circuit of the leakage inductance and the high frequency loss is converted into a primary current. From the actual value contained in the primary voltage, and the leakage inductance estimated value of the arithmetic means is slightly changed at a predetermined control cycle using the first random signal, First adjusting means for adjusting a leakage inductance estimation error appearing in the estimated value error signal to a minimum, and a high frequency loss estimated value of the computing means is controlled by the first random signal using a second random signal. Secondly, it is adjusted so as to minimize the high frequency loss estimation error appearing in this estimation value error signal by slightly changing it in a control cycle that is an integral multiple of the cycle. Leakage inductance measuring apparatus of an induction motor characterized by comprising providing a node means.