JPH07298699A - Variable speed driver - Google Patents

Abstract

PURPOSE:To obtain a secondary time constant accurately in an early stage by eliminat ing the influence of noise. CONSTITUTION:A current control system obtains voltage commands V1d, V1q for exciting current command I0* and torque current command IT* on a two-phase rotating orthogonal coordinate system and a vector control is effected by determining a slip frequency based on the exciting current command, the torque current command and a secondary time constant tau2 of an induction motor 5. In such variable speed driver, a sequence processing section 21 starts operation while providing a preliminary excitation interval for nullifying the torque current command and generating only the exciting current command. A differential approximation operating section 23 obtains the difference between current and previous values for (n) samples of the exciting voltage V1d of an induction motor during the preliminary excitation interval as a differential approximation sample data string. Each data strmng is fed to a logarithmic operating section thence fed through an accumulating section 28, a matrix operating section 29 and a divider 30 where a linear approximation formula connecting the logarithmic sample data strings is determined by the method of least squares and a secondary time constant tau2 is determined from the reciprocal of the coefficient of linear approximation formula.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a variable speed drive device for vector control of an induction motor, and more particularly to a secondary time constant measuring device for an induction motor.

[0002]

2. Description of the Related Art FIG. 4 exemplifies a variable speed drive device by vector control. The speed control amplifier 1 compares the speed command N * with the detected speed value ω _r of the rotor of the electric motor, performs proportional / integral calculation of the deviation, and outputs the torque current command I _T *.

The current control amplifiers 2 ₁ and 2 ₂ use an exciting current command I ₀ * and a torque current command I _T * for vector control as commands, and rotate on the orthogonal two-phase rotating coordinate system in synchronization with the power supply frequency. The detected value I _1d of the excitation axis current component and the detected value I _1q of the torque axis current component are compared, the proportional / integral calculation of the deviation is performed, and the voltage commands V _1d and V _1q on the orthogonal two-phase rotation coordinate system are compared. Output as.

The coordinate conversion unit 3 has two-phase voltage commands V _1d and V ₁ .
Voltage V _u of the three-phase fixed coordinate system _1q, V _v, performs two-phase / 3-phase coordinate conversion V _w.

The PWM inverter 4 has voltages V _u , V _v , and V
_{Using w} as a command, a voltage having a PWM waveform corresponding to this is supplied to the induction motor 5.

The coordinate transformation unit 6 transforms the three-phase current in the fixed coordinate system of the induction motor 5 detected by the current detector 7 into the three-phase / two-phase coordinate transformation into the currents I _1d and I _1q in the two-phase rotating coordinate system. .

The reference phase calculator 8 generates a reference phase θ ₁ required for coordinate conversion in the coordinate converters 3 and 6. At this time, the delay phase adjustment unit 8A delays the calculation cycle by one cycle and gives it to the coordinate conversion unit 6.

The slip calculation unit 9 obtains a slip frequency component ω _slip from the exciting current command I ₀ *, the torque current command I _T *, the secondary resistance r ₂ of the induction motor 5 and the exciting inductance M. In the reference phase calculation unit 8, this slip frequency component ω
_The speed detection value ω _r is added to _slip to obtain the power source angular frequency ω ₁ , and this is integrated to obtain the reference phase (excitation axis phase) θ ₁ .

To detect the rotor speed ω _r of the induction motor 5,
Rotary encoder 10 axially coupled to the induction motor 5
Then, a pulse having a frequency proportional to the rotation speed is obtained, and the speed calculation unit 11 calculates the pulse cycle or obtains it from the number of pulses for each constant calculation cycle.

In the above-mentioned configuration, control / calculation of each part may be controlled / calculated in an analog manner. However, in order to improve the precision of control / calculation, it is digitally controlled by a microcomputer. Many.

In such vector control of the induction motor, the slip calculation by the slip calculator 9 requires the secondary resistance r ₂ and the exciting inductance M of the induction motor to obtain the secondary time constant as a constant. Since the secondary resistance fluctuates with temperature, it is necessary to compensate the fluctuation component during operation.

There is a method of obtaining the measured value including the temperature fluctuation by performing the secondary time constant measurement during the pre-excitation period of the induction motor (for example, Japanese Patent Laid-Open No. 2-106190).

In this measuring method, the primary voltage or the primary current when a constant current or a constant voltage is applied to the induction motor is measured at three points, and the measured value is calculated. FIG. 5 shows a primary voltage waveform by applying a constant current. The secondary time constant τ ₂ is calculated from the measured voltages v ₁ , v ₂ , v ₃ at time intervals Δt by the following equation.
Ask for.

[0014]

[Equation 2]

By the vector control using the secondary time constant, the vector control condition is always satisfied even if the induction motor has a temperature change, and high torque accuracy is obtained.

[0016]

In the conventional secondary time constant measuring method, when applied to an actual device, a constant current method that quickly establishes a magnetic flux is often adopted. Also, the measured voltage is inexpensive because a voltage detector is not provided outside the variable speed drive device, but the output voltages V _1d and V _1q of the amplifiers 2 ₁ and 2 ₂ of the current control system are used instead of the measured values. The device can be realized.

However, in order to set the current control characteristic of the current control system to a high-speed response, when the control gain is set high,
The voltages V _1d and V _1q , which are current control outputs, include large ripples.

Therefore, if the secondary resistance is measured from only the voltage measurement data at three points as in the conventional case, the ripple component becomes noise, and the accurate constant cannot be measured.

It is an object of the present invention to provide a variable speed drive device which eliminates the influence of noise and obtains an accurate secondary time constant to improve control accuracy.

Another object of the present invention is to provide a variable speed drive device capable of obtaining the secondary time constant at an early stage of the pre-excitation period.

[0021]

In order to solve the above-mentioned problems, the present invention uses the excitation current command and the torque current command on the orthogonal two-phase rotational coordinate system and the deviation of the detected values of the respective current components to determine the orthogonality 2. A current control system for obtaining voltage commands V _1d and V _1q on the phase rotation coordinate system, and a slip calculation unit for obtaining a slip frequency from the excitation current command, the torque current command, and the secondary time constant τ _{2 of the} induction motor are provided. In the variable speed drive device for an induction motor, a sequence processing unit having a pre-excitation period in which the torque current command is set to zero and only an excitation current command is generated at the start of operation of vector control, and an excitation voltage component of the induction motor during the pre-excitation period. the (V _1d) sampling by n samples, a differential approximation calculation unit for obtaining a difference between the previous value and the current value of each sample value as a derivative approximation sample data string, the derivative approximation sample de A logarithmic operation unit for obtaining the individual log of data columns, the calculated by the least square method linear approximation connecting the logarithmic sample data sequence to obtain the reciprocal of the coefficient of the straight line approximation formula as the time constant tau ₂ the secondary A least squares arithmetic unit,
It is characterized by having.

According to the present invention, the least squares arithmetic unit is

[0023]

[Equation 3]

According to the above, the second-order time constant τ ₂ is obtained.

Further, according to the present invention, the least-squares arithmetic operation section stores a first memory section for storing an intermediate time t _i 'of each sample time up to each sampling time as an auxiliary matrix element, and the auxiliary matrix element. The second memory unit for storing each element m of the inverse matrix of the normal equation, Σt _i ′, Σ (ti ′) ² and Δi, and the logarithmic value F _i ′ of the logarithmic calculation unit are accumulated at each sample time i. added value .SIGMA.F _i ', and the pair numerical and the auxiliary matrix elements t _i' Σt _i 'obtained by cumulative addition is multiplied by
A cumulative addition computing unit that obtains Fi ′, a matrix computing unit that multiplies both outputs of the cumulative addition computing unit and the value of the inverse matrix from the second memory unit, and adds this value with a sign A divider that divides the inverse matrix element Δi of the second memory unit by the output of the matrix calculation unit to obtain the secondary time constant (τ ₂ ) _i at each sample time. To do.

[0026]

(First Invention) FIG. 2 shows an operation pattern of the sequence processing unit when the induction motor is started from the stopped state. A constant value of the exciting current command I ₀ * is generated at the same time as the operation command given at the time t _{1 in} the figure, and the torque current command I _T * remains zero, and is generated from the time t ₂ .

During this pre-excitation period (t _{1 to} t ₂ ), the secondary magnetic flux λ _2d of the induction motor exponentially increases. At this time, the torque axis component (q-axis voltage component) V _1q of the output voltage remains zero, but the magnetic flux axis component (d-axis voltage component) V _1d exponentially decreases as the secondary magnetic flux increases. . This voltage equation can be expressed by the following equation.

[0028]

[Equation 4]

R ₁ : primary resistance of induction motor r ₂ : secondary resistance of induction motor τ ₂ : secondary time constant of induction motor (M / r ₂ ) M: exciting inductance of induction motor t: time Therefore, in the pre-excitation period, the excitation current I ₀ *,
Assuming that the resistances r ₁ and r ₂ and the exciting inductance M are constant,
Further, assuming that the voltage V _1d contains noise but the basic component is generated according to the characteristic formula, the secondary time constant τ ₂ can be obtained by the calculation for solving the characteristic formula.

In this calculation, since unnecessary terms are first deleted, the following formula (2) is obtained by differentiating the formula (1).

[0031]

[Equation 5]

Further, since the exponential function component is linearly converted, the following equation (3) is obtained by taking the logarithm with e being the base.

[0033]

[Equation 6]

The relationship of these equations (1) to (3) is as shown in FIG. 3, and the waveform of the magnetic flux axis voltage V _1d is shown in (a).
A differential waveform is shown in (b), and a logarithmic waveform is shown in (c). As a result, the secondary time constant τ ₂ is obtained from the reciprocal of the slope of the waveform in (c).

Here, in the differential operation of the voltage V _1d, an approximate operation by the differential approximate operation unit is performed. That is, n points of the excitation axis voltage V _1d (t _i ) at the sample time t _i are taken out, and the differential value is obtained by the difference approximation as in the following equation. However, the time t ′ _i of the approximated differential component is an intermediate time of the difference approximation period.

[0036]

[Equation 7]

In order to linearly convert the exponential function component from this sample data string, the logarithmic calculation unit obtains the logarithm of each sample data. This logarithm is as follows.

[0038]

[Equation 8]

Next, the least-squares method computing section obtains a quadratic time constant τ ₂ by computing the reciprocal of the slope of the straight line connecting the logarithmized sample data strings by the least-squares method.

(Second Invention) The calculation of the quadratic time constant τ ₂ by the least-squares calculation unit is only required to find the slope of a straight line taking the sample data string as a value, and it is found as a coefficient of a linear equation. The calculation procedure for obtaining this coefficient calculation by the least square method will be described in detail below.

As the sample data string, n pieces of data (x ₁ , f ₁ ), (x ₂ , f ₂ ), ... (X _n , f _n ) are given, and an approximate straight line passing near each point is given. In order to obtain the equation by the method of least squares, first, the auxiliary matrix C for making the normal equation is a linear equation, and therefore, it is as follows.

[0042]

[Equation 9]

The coefficient matrix A of the normal equation is as follows.

[0044]

[Equation 10]

The constant matrix b is as follows.

[0046]

[Equation 11]

Then, in order to solve the equation of Aa = b, the inverse matrix of the coefficient matrix A is obtained as follows.

[0048]

[Equation 12]

Therefore, assuming that the constant matrix b is obtained from the measurement result, the coefficient matrix a is as follows.

[0050]

[Equation 13]

From the above, the coefficient matrix a can be obtained by the least squares method in the case of linear function approximation. In the present invention,
In order to obtain only the slope of the approximate straight line,
You only have to ask for the line.

In the second line, the correspondence with the differentially approximated sample data of FIG. 3 is such that xi is replaced by t _i 'and f _i is replaced by F _i ' and the approximate linear function y (x) =
a ₀ + a ₁ x is replaced with V _1d (t) = a ₀ + a ₁ F ′ (t).

Therefore, the secondary time constant τ ₂ can be calculated by the following equation.

[0054]

[Equation 14]

(Third Invention) In the second invention, after measuring the data at n points, the secondary time constant τ ₂ is obtained. However, the time required to establish the secondary magnetic flux is about the time of the secondary time constant, and in order to predict the establishment of the secondary magnetic flux itself, the secondary time constant of the actual machine is used immediately after the start of pre-excitation. It is necessary to perform an estimation calculation of the secondary magnetic flux.

Therefore, in the present invention, i = 3, 4, 5, ... N
At each measurement point, the secondary time constant (τ ₂ ) _i is calculated using the data measured up to that point.

In this calculation, as shown in the following equation, the auxiliary matrix C is calculated for each of the cases of n = 2, 3, 4, ... I and stored in the first memory section.

[0058]

[Equation 15]

Similarly, the inverse matrix A of the coefficient matrix of the normal equation
^For each element of ^-1 , the coefficient up to each data point is calculated and stored in the second memory unit.

The constant matrix b can also be calculated using the previous data as follows.

[0061]

[Equation 16]

These sampling times i = 2, 3,
Using each data up to 4, ... n, the secondary time constant τ ₂ at each time is obtained by the calculation by the cumulative addition calculation unit, the matrix calculation unit, and the divider for each sample data from the logarithmic calculation unit. This arithmetic expression is as follows.

[0063]

[Equation 17]

[0064] b _i1: constant matrix b first row of elements b of _{i i2:} In the second row elements present invention of constant matrix b _i, from the data to each time i from the start of measurement,
Since the slope is obtained by the method of least squares, the quadratic time constant can be calculated sequentially.

[0065]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram of a secondary time constant calculation showing an embodiment of the present invention.

The sequence processing unit 21 generates an exciting current command I ₀ * and a torque current command I _T * having a pre-excitation period as shown in FIG. 3 at the start of vector control operation.

The sample time generator 22 generates n points of sampling time data set from the pre-excitation start time.

The differential approximating operation unit 23 includes the current control amplifier 2
_The excitation voltage command V _1d that outputs ₁ is sampled at the generation timing of the sample time generation unit 22, and the difference between the previous value and the current value of each sample value is obtained as differential approximation data.

The logarithmic calculation unit 24 individually obtains the logarithm having e as a base for each sample data string from the differential approximation calculation unit 23.

The memory units 25, 26 and 27 store the matrix elements up to each sample time i as table data. Intermediate time t _i 'of the differential approximation is stored in the memory unit 25.
Is stored as an auxiliary matrix element, each element m and Σt _i ′ of the inverse matrix of the normal equation is stored in the memory unit 26, and Σ (t _i ′) ² and (Σ (t _i )) are stored in the memory unit 27. ^The Δi obtained from ² and m is stored.

The cumulative addition computing unit 28 computes the components of the constant matrix b for the sample data Fi 'from the logarithmic computing unit 24. In 28 ₁ , the value corresponding to the first row element of the equation (14) is obtained, and in 28 ₂ and 28 ₃ , the value corresponding to the second row element is obtained.

The matrix calculation unit 29 calculates the constant m and Σt _i 'from the memory unit 26 for each value from the cumulative addition calculation unit 28.
And the difference is obtained to obtain the denominator term of Expression (15).

The divider 30 obtains a quadratic time constant (τ ₂ ) _i by division with Δi from the memory unit 27 as the numerator and the value from the matrix operation unit 29 as the denominator.

According to this embodiment, the secondary time constant can be obtained immediately after the start of pre-excitation. That is, when the number of data points is small, it is affected by noise, so the accuracy is poor, but the secondary time constant can be obtained early. Then, as the number of data points increases, the accuracy improves, and when the pre-excitation period ends and the torque current is output and used for slip calculation, vector control using a highly accurate secondary time constant can be performed.

In the embodiment, the secondary time constant τ ₂ after measuring the data at n points can be obtained from the calculation result of each calculation unit by sampling the final data.

In the embodiment, the exciting voltage may be detected directly from the primary voltage of the induction motor 5. In addition, each unit 22 for calculating the secondary time constant
˜30 can be replaced as software processing provided in the control device having a microcomputer configuration.

[0077]

As described above, according to the present invention, the excitation voltage component of the pre-excitation period is sampled by n samples, the differential approximation operation is performed on this sample data string, and the logarithm is calculated. Since the linear approximation formula to be connected is obtained by the method of least squares and the reciprocal of the coefficient is obtained as the quadratic time constant, the following effects are obtained.

(1) Since the noise component is suppressed by a statistical method using the least squares method from the sampling data of multiple points from the excitation voltage waveform in the preliminary excitation period, the measurement accuracy can be improved.

(2) Even if it is obtained from the output command of the current control system in which the measured voltage has a high gain, the influence of noise can be eliminated and the measurement accuracy can be improved.

(3) The coefficients used for calculating the secondary time constant by the least square method can be made into table data if the sampling time is predetermined, and the previous value can be used for the calculation of the measurement data. Therefore, the amount of calculation for each sampling is constant regardless of the number of data points, and an increase in the number of data points does not increase the calculation processing time.

(4) The secondary time constant can be obtained immediately after the start of pre-excitation.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a waveform diagram of a pre-excitation period.

FIG. 3 is a second-order time constant measurement waveform diagram for explaining the present invention in principle.

FIG. 4 is a block diagram of vector control.

FIG. 5 is a voltage measurement mode diagram of a conventional example.

[Explanation of symbols]

1 ... Speed control amplifier 2 ₁ , 2 ₂ ... Current control amplifier 5 ... Induction motor 8 ... Reference phase calculation unit 9 ... Slip calculation unit 21 ... Sequence processing unit 22 ... Sample time generation unit 23 ... Differential approximation calculation unit 24 ... Logarithmic calculation Units 25, 26, 27 ... Memory unit 28 ... Cumulative addition computing unit 29 ... Matrix computing unit 30 ... Divider

Claims (3)

[Claims] 1. A current for obtaining a voltage command V _1d , V _1q on the orthogonal two-phase rotating coordinate system from a deviation between an excitation current command and a torque current command on the orthogonal two-phase rotating coordinate system and detected values of respective current components. In the variable speed drive device of the induction motor, which includes a control system, the excitation current command, the torque current command, and the slip calculation unit that determines the slip frequency from the secondary time constant τ _{2 of} the induction motor, the vector control is started at the start of operation. A sequence processing unit having a pre-excitation period in which the torque current command is zero and only the excitation current command is generated, and the excitation voltage component (V _1d ) of the induction motor is sampled by n samples during the pre-excitation period, and each sample value A differential approximation calculator that obtains the difference between the current value and the current value as a differential approximation sample data string, a logarithm calculator that obtains each logarithm of the differential approximation sample data string, and the logarithm Was determined by the method of least squares linear approximation equation connecting the sample data string, variable speed, characterized in that a least-squares method calculation unit which the inverse of the coefficient of the straight line approximation formula obtained as the secondary time constant tau ₂ Drive. 2. The least-squares method computing unit is defined by the following equation: _2. The variable speed drive device according to claim 1, wherein the secondary time constant τ ₂ is obtained in accordance with the above. 3. The least-squares method computing unit stores, as an auxiliary matrix element, an intermediate time t _i 'of each sample time up to each sampling time, and a normal equation from the auxiliary matrix element. A second memory unit for storing each element m of the inverse matrix, Σt _i ′, Σ (ti ′) ² and Δi, and a value ΣF obtained by cumulatively adding the logarithmic value F _i ′ of the logarithmic operation unit for each sample time i. _i ′, and a cumulative addition computing unit for multiplying the logarithmic value by the auxiliary matrix element t _i ′ and cumulatively adding Σt _i ′ · Fi ′, both outputs of the cumulative addition computing unit, and the second memory A matrix operation unit that multiplies the value of the inverse matrix from the unit, adds the value with a sign, and divides the inverse matrix element Δi of the second memory unit by the output of the matrix operation unit, and every sample time. to wherein said secondary time constant (tau ₂₎ further comprising a divider to obtain a _i in Variable speed drive according to claim 1, wherein.