JPH08294300A - Method for estimating speed of induction machine and controller for induction machine - Google Patents

Abstract

PURPOSE: To provide a speed estimating method of estimating the true speed of an induction machine, and to provide a control device for the induction machine which provides characteristic equivalent to a DC machine by applying this speed estimating method. CONSTITUTION: In a voltage/current detector 31, the primary voltage and the primary current of an induction machine 16 are sampled. In a neuro speed estimating part 32, a relational expression among the primary voltage, primary current and secondary magnetic flux of the induction machine, not including a term depending on the speed of the induction machine in a magnetic flux arithmetic part 33, is used to calculate the secondary magnetic flux corresponding to the sampled primary voltage/current, also to calculate secondary magnetic flux corresponding to the sampled primary current by using a state equation related to the magnetic flux including terms respectively depending on the primary current and the rotational speed. In a speed estimating part 34, the two secondary magnetic fluxes are compared at each sampling, to learn in a direction where both the magnetic fluxes agree with each other a weight coefficient of the term depending upon the rotational speed in the state equation, and the speed of the induction machine is determined correspondingly to the weight coefficient when the secondary magnetic fluxes agree with each other.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for vector-controlling an induction machine without a speed sensor, and more particularly to a speed estimation method and control device using a neural network.

[0002]

2. Description of the Related Art In vector controlled induction machines,
The magnitudes of the magnetizing current and the torque current and the phases thereof are calculated by using the motor parameters, and the corresponding phase current command is given to the current control type power converter. Here, the phase is the phase of the current given to the primary winding and is obtained by integrating the primary frequency. As is well known, the primary frequency is obtained by obtaining the slip angular frequency from the magnetic flux command, the torque current and the motor parameter, and then taking the sum of the slip angular frequency and the rotation angular frequency. Detecting the rotational angular frequency of the electric motor is necessary when forming a speed control loop, but is also essential when determining the phase of the current when performing vector control.

On the other hand, vector-controlled induction machines have come to be used not only in general industry but also in a wide range of fields such as railway vehicles, and the environment in which they are installed is various. In particular, there is no space for mounting the speed sensor or the environment is not good, and there are great expectations for a control device that can perform vector control equivalent to a DC machine without the speed sensor.

FIG. 7 is a block diagram showing the configuration of the most typical velocity sensorless vector control device that has been studied so far. In FIG. 7, a hysteresis comparator 12 attached with a comparator 11, a power converter 13, and
Current source inverter 15 with current transformer 14 for detecting phase current
And the current source inverter 15 is used as the induction motor 1.
The current of 6 is controlled. In this case, the current type inverter 15 uses the primary current commands _iu ^* , _iv ^* , _iw.
^* And, measured primary current by current transformer _{14 i u, i v, i} w
And are compared by the hysteresis comparator 12,
When the difference exceeds an arbitrary set value Δi, the switching element of the power converter 13 is switched to the opposite side to invert the actual current.

Instead of detecting the speed of the induction motor 16 with a speed sensor, the current transformer 14 causes the primary currents i _u , i _v ,
i _w (abbreviated as i in the figure) is detected by the voltage detector 17 as primary voltages v _u , v _v , v _w (abbreviated as v in the figure), and each detected signal is a torque current calculator. Guided to 18. In the torque current calculator 18, the secondary magnetic flux Φ _r is obtained by integrating the voltage equation, and then,
The measured current i is converted into the torque current i ₂ using the result and the phase angle θ of the current. This torque current i ₂ is applied to the slip angular frequency calculator 19 and the comparator 23 installed in the preceding stage of the PI controller 24. Incidentally, symbols subjected to such i ₂ "^" indicates calculated value or the estimated value.

The slip angular frequency calculator 19 determines the torque current i
Upon receipt of ₂ , the slip angular frequency ω _s is obtained and added to the comparator 20. The comparator 20 subtracts the slip angular frequency ω _s from the primary angular frequency ω ₁ which is the output of the PI controller 24, outputs the rotational angular velocity ω _m of the electric motor, and adds it to the comparator 21.
The comparator 21 subtracts the rotational angular velocity ω _m from the set rotational angular velocity command ω _m ^* and adds it to the PI controller 22. The PI controller 22 performs a proportional and integral calculation with respect to the rotational angular velocity difference, outputs a torque current command i ₂ ^*, and applies the torque current command i ₂ ^* to the comparator 23 and the magnetizing current calculator 26. The comparator 23 uses the torque current command i ₂ ^* to calculate the torque current i
_The value obtained by subtracting ₂ is added to the PI controller 24. PI
The controller 24 performs proportional and integral calculations on this difference to obtain the primary angular frequency ω ₁ . Further, the integrator 25 integrates the primary angular frequency ω ₁ and outputs the phase angle θ of the current which changes with the passage of time.

On the other hand, the magnetizing current calculator 26 calculates the combined value of the torque current command i ₂ ^* and the set magnetizing current i ₀ ^*, and also calculates the phase γ thereof, which is viewed in the rotating magnetic flux coordinate system. The direct current component current i ₁ ^* = i ₁ ε ^{j γ} is calculated. next,
The DC component current i ₁ ^* and the phase angle θ of the current described above are added to the coordinate system conversion unit 27, and the sum of the phase γ and the phase angle θ is the phase of the primary current seen in the stationary coordinate system. Therefore, each of them is distributed to each phase of the induction motor and the primary current commands _iu ^* , _iv ^* , _iw ^* are added to the current source inverter 15. In this way, a system for vector-controlling an induction motor without mounting a speed sensor is realized.

[0008]

In the above vector controller, the PI controller 24 is proportional to the deviation between the torque current command i ₂ ^* and the calculated torque current i ₂ .
Although it is converted into the primary angular frequency ω ₁ by executing the integral calculation, this idea is not strictly true. That is, when there is a deviation of the torque current, the slip frequency should be adjusted considering the original characteristics of the induction motor.However, in the above vector control device, the slip frequency is determined in another block as a result. Will be done. Therefore, the induction motor, which is different from the original characteristic of the vector control, is operated without the speed sensor. As a result, the speed control accuracy was lowered, and it could not be said that a motor drive system equivalent to a DC machine was realized.

The present invention has been made to solve the above problems, and a speed estimation method for estimating the true speed of an induction machine, and by applying this method, characteristics equivalent to a DC machine can be obtained. An object is to provide a control device for an induction machine.

[0010]

An induction machine speed estimation method according to claim 1, wherein the induction machine primary voltage and primary current are sampled and detected, and the induction machine speed is not included. Using the relational expressions between the primary voltage and the primary current and the secondary magnetic flux, the secondary magnetic flux corresponding to the sampled and detected primary voltage and primary current is calculated, and the terms dependent on the primary current and the rotational speed are respectively calculated. The secondary flux corresponding to the primary current detected by sampling is calculated using the equation of state regarding the magnetic flux including, and the two secondary fluxes are compared for each sampling, and the rotational speed in the equation of state in the direction in which both agree It is characterized in that the weighting coefficient of the term depending on is learned, and the speed of the induction machine is determined in association with the weighting coefficient with which the secondary magnetic flux matches.

According to a second aspect of the present invention, there is provided a speed estimation method for an induction machine, wherein the power supply system includes first power conversion means for converting alternating current into direct current to control a direct current output voltage, and the first power conversion means. And a second power conversion means for converting a direct current output from the device into an alternating current to control an alternating current output current,
Sampling and detecting the primary voltage and primary current of the induction machine,
Using the relational expression between the primary voltage and primary current of the induction machine and the secondary magnetic flux, which does not include the term depending on the speed of the induction machine, the secondary magnetic flux corresponding to the sampled primary voltage and primary current is calculated. The secondary magnetic flux corresponding to the sampled and detected primary current is calculated by using the equation of state relating to the magnetic flux including the terms that respectively depend on the primary current and the rotation speed, and the voltage control deviation of the first power conversion means is calculated. Is within a preset value, the two secondary magnetic fluxes are compared for each sampling, and the weighting factor of the term depending on the rotation speed in the equation of state is learned in the direction in which the two agree. Is characterized in that the speed of the induction machine is determined in association with the matching weighting coefficient.

A control device for an induction machine according to a third aspect of the present invention obtains a torque current command according to a difference between a speed command of the induction machine and an actual speed, and based on the torque current command and the magnetic flux command, a primary current command and When calculating the slip frequency command and controlling the current of the induction machine according to the calculated primary current command and slip frequency command, it depends on the detection means for sampling and detecting the primary voltage and primary current of the induction machine and the speed of the induction machine. A first magnetic flux calculation for calculating a secondary magnetic flux corresponding to the sampled detected primary voltage and primary current by using a relational expression between the primary voltage and primary current of the induction machine and the secondary magnetic flux that does not include a term. Second magnetic flux for calculating the secondary magnetic flux corresponding to the sampled and detected primary current by using the means and the equation of state relating to the magnetic flux including terms that respectively depend on the primary current and the rotation speed. The calculation means and two secondary magnetic fluxes are compared for each sampling, and the weighting factor of the term depending on the rotation speed in the equation of state is learned in the direction in which the two agree with each other, and the secondary magnetic flux is associated with the matching weighting factor. And a speed estimating means for determining the speed of the induction machine.

The control device for an induction machine according to a fourth aspect is different from that according to the third aspect in that it further includes a primary voltage and a primary current of the induction machine which do not include a term depending on the speed of the induction machine. Using the relational expression between the secondary magnetic flux, the magnetic flux calculating means for calculating the secondary magnetic flux and the torque current corresponding to the sampled primary voltage and primary current, the magnetic flux command and the torque current command, and the calculated A vector calculation means for alternately inputting the secondary magnetic flux and the torque current and learning the primary weight command and the slip frequency command that match each other while learning using a primary expression with a variable weighting coefficient, and the learned weighting coefficient. Data storage means for storing the electric motor parameter corresponding to, and the speed estimation means determines the speed of the induction machine using the electric motor parameter stored in the data storage means.

According to a fifth aspect of the present invention, there is provided an induction machine control apparatus, wherein the induction machine power supply system includes first power conversion means for converting alternating current into direct current to control a direct current output voltage, and the first power conversion means. A second electric power converting means for converting a direct current outputted from the converting means into an alternating current to control an alternating current output current, and calculating a primary current instruction and a slip frequency instruction based on a magnetic flux instruction and a torque current instruction of the induction machine. Then, in controlling the second power conversion means in accordance with the calculated primary current command and slip frequency command, a detecting means for sampling and detecting the primary voltage and the primary current of the induction machine and a term dependent on the speed of the induction machine are included. A first magnetic flux calculating means for calculating a secondary magnetic flux corresponding to the sampled primary voltage and primary current using the relational expression between the primary voltage and primary current of the induction machine and the secondary magnetic flux; The second magnetic flux calculating means for calculating the secondary magnetic flux corresponding to the sampled and detected primary current using the state equation relating to the magnetic flux including the terms respectively dependent on the secondary current and the rotation speed, and the first power conversion means. Steady state determination means for detecting that the voltage control deviation is within a preset value, and when the voltage control deviation is within the preset value, two secondary magnetic fluxes are compared for each sampling, and the two match. The speed estimation means for learning the weighting coefficient of the term depending on the rotational speed in the state equation in the direction, and determining the speed of the induction machine in association with the weighting coefficient with which the secondary magnetic flux coincides. There is.

The control device for an induction machine according to a sixth aspect is different from that according to the fifth aspect in that the primary voltage and the primary current of the induction machine do not include a term depending on the speed of the induction machine. Using the relational expression between the secondary magnetic flux, the magnetic flux calculating means for calculating the secondary magnetic flux and the torque current corresponding to the sampled primary voltage and primary current, the magnetic flux command and the torque current command, and the calculated A vector calculation means for alternately inputting the secondary magnetic flux and the torque current and learning the primary weight command and the slip frequency command that match each other while learning using a primary expression with a variable weighting coefficient, and the learned weighting coefficient. Data storage means for storing the electric motor parameter corresponding to, and the speed estimation means determines the speed of the induction machine using the electric motor parameter stored in the data storage means.

[0016]

In the speed estimating method for an induction machine according to claim 1, a relational expression between the primary voltage and the primary current and the secondary magnetic flux of the induction machine, which does not include a term depending on the speed of the induction machine, is used. The term that depends on the rotational speed in the equation of state in the direction in which the secondary magnetic flux calculated by the above and the secondary magnetic flux calculated by using the equation of state regarding the magnetic flux including the terms that respectively depend on the primary current and the rotational speed match. Since the weighting factor of the induction machine is learned and the speed of the induction machine is determined in association with this weighting coefficient, the true speed of the induction machine can be estimated.

In the control device for the induction machine according to the second aspect of the invention, the voltage control deviation of the first power conversion means for converting the alternating current into the direct current and controlling the direct current output voltage is within a preset set value. Since the weighting coefficient of the term that depends on the rotation speed in the state equation is learned under this condition, it is possible to estimate the speed different from the actual speed when the voltage control operation for the load change cannot be done in time. Can be prevented.

In the control device for an induction machine according to a third aspect, the method of the first aspect is adopted to estimate the speed of the induction machine, and the induction machine is vector-controlled according to this speed. The true speed can be estimated, and at the same time, the characteristics equivalent to a DC machine can be obtained.

In a control device for an induction machine according to a fourth aspect, a magnetic flux command and a torque current command, and a calculated secondary magnetic flux and a torque current are alternately input, and a linear expression with a variable weighting coefficient is used. While learning by calculating the primary current command and the slip frequency command that match each other and determining the speed of the induction machine using the motor parameter corresponding to the learned weighting factor, induction with large fluctuations in the motor parameter The characteristics equivalent to a DC machine can be obtained even when targeting an electric motor.

In the control device for the induction machine according to the fifth aspect, since the speed of the induction machine is estimated by using the method according to the second aspect, it is possible to estimate a speed different from the actual speed. Can be prevented in advance, and vector control according to this estimated speed becomes possible.

In the control device for an induction machine according to a sixth aspect of the present invention, the magnetic flux command and the torque current command and the calculated secondary magnetic flux and the torque current are alternately input, and the weighting coefficient is a primary expression. While learning using, the primary current command and the slip frequency command that both agree with each other are calculated, and the speed of the induction machine is determined using the motor parameter corresponding to the weighting coefficient learned, so Even when targeting a large induction motor, the characteristics equivalent to a DC motor can be obtained.

[0022]

The present invention will be described in detail below with reference to the embodiments shown in the drawings. FIG. 1 is a block diagram showing the configuration of the first embodiment of the present invention. In the figure, the same elements as those of FIG. 7 showing a conventional apparatus are designated by the same reference numerals and the description thereof will be omitted. FIG.
In the above, although not shown in detail, the vector control unit 30 is composed of a magnetization current calculation unit, a slip frequency calculation unit, a combined current calculation unit, a coordinate conversion unit, and the like. The voltage / current detector 31 detects the primary voltage v and the primary current i supplied to the induction motor 16, and the detected values are added to the neuro speed estimation unit 32 including a neural network. The neuro velocity estimation unit 32 can be represented by a magnetic flux calculation unit 33 and a velocity estimation unit 34. Of these, the magnetic flux calculation unit 33 calculates the torque current i _qs and the secondary magnetic flux λ _r and adds them to the vector control unit 30, and adds the secondary magnetic flux λ _r to the speed estimation unit 34. And
The speed estimation unit 34 calculates the rotational angular speed ω _r . This rotational angular velocity ω _r is added to the comparator 21 and further to the vector control unit 30. The comparator 21 subtracts the rotational angular velocity ω _r from the set angular velocity command ω _r ^* and adds it to the PI controller 22. The PI controller 22 executes a proportional and integral calculation with respect to the rotational angular velocity difference, outputs a torque current command i _qs ^*, and _applies it to the vector control unit 30. The vector control unit 30 and the magnetization current command i _ds ^* matching the flux command lambda _dr ^*, calculates the slip frequency command omega _s ^* matching the torque current command i _qs ^* by the following expression. i _ds ^* = (L _r / R _r M) pλ _dr ^* + (1 / M) λ _dr ^* … (1) ω _s ^* = (R _r Mi _qs ^* / L _r λ _dr ^* … (2) where L _r : Secondary self-inductance R _r : Secondary resistance M: Mutual inductance p: Differential operator The coordinate system conversion unit 27 uses the magnetizing current command i _ds ^* and the magnetizing current command i _ds ^* and the above-mentioned formulas (1) and (2). Slip frequency command ω _s
^* , The torque current command i _qs ^*, and the estimated rotational speed ω _r described below are used to generate the primary current commands i _u ^* , _iv ^* , i
Calculates _w ^* .

FIG. 2 is a functional explanatory diagram of the neuro velocity estimation unit 32 including a neural network.
A magnetic flux calculator 33 that calculates the secondary magnetic flux Φ _r based on the primary voltage v _s and the primary current i _s supplied to the induction motor 16.
A and the secondary magnetic flux λ using the equation of state regarding the magnetic flux including terms that depend on the primary current i _s and the rotation speed ω _r , respectively.
comparator 34 for comparing the neuro-model 33B in the state equation for calculating the _r, these two secondary flux [Phi _r and lambda _r
A and an error calculation unit 34B that repeats the operation of adding a small amount Δω _r to the rotation speed ω _r until the difference ε of the secondary magnetic flux becomes zero and correcting the weighting coefficient depending on the rotation speed ω _r. It Here, the magnetic flux calculator 33A samples and detects the primary voltage v _s and the primary current i _s supplied to the induction motor 16 to calculate the secondary magnetic flux. By the way, the calculation formula for the d-axis component is as follows.

[0024]

[Equation 1] However, L _r : secondary self-inductance M: mutual inductance L _s : primary self-inductance R _s : primary resistance σ: leakage coefficient.

The above equation (3) is characterized in that the secondary magnetic flux can be detected regardless of the rotation speed. Another method is to observe the secondary magnetic flux based on the equation of state. That is, similarly to the above, the expression of the d-axis component is as follows.

[0026]

[Equation 2] Where T _r is a quadratic time constant.

The above equation (4) is characterized in that it includes a term depending on the rotational angular velocity ω _r . When this equation (4) is displayed in a discrete system, it can be summarized as the following equation. λ _dr (k) = W ₁ λ _dr (k-1) + W ₂ λ _dr (k-1) + W ₃ i _ds (k-1) (5) The neuro model 33B of the state equation uses the equation (5). To calculate the secondary magnetic flux λ _r . The comparator 34A compares the secondary magnetic flux Φ _r obtained by the equation (3) with the λ _r obtained by the equation (5), but both of them should originally match.
If there is a deviation between the two, the weighting factors W ₁ , W ₂ , W ₃ used in Eq. (5) will be incorrect. (3),
If the parameters of the electric motor used in the equation (4) are correct, since the weighting factors W ₁ and W ₃ are true values,
In this case, it is considered that the weight coefficient W ₂ depending on the rotation speed contains an error. Therefore, the error calculation unit 34B corrects the rotational angular velocity ω _r included in the weighting coefficient W ₂ in the state equation by adding a minute amount Δω _r , and again calculates the equation (5) for comparison. By repeating this calculation and comparison, the secondary magnetic flux Φ _r
And W _r are matched, the correct W ₂ is learned and the true rotational angular velocity ω _r is estimated.

According to this embodiment, since the detection signal for enabling the estimation of the rotational angular velocity ω _r is the voltage and current supplied to the induction motor, there is no particular difficulty. Therefore,
The speed estimation unit composed of software can be easily realized. Further, according to the present embodiment, the neuro velocity estimation unit 32 can be used as an alternative means of the velocity sensor, and there is also an effect that it can be realized without essentially changing the configuration of the original system.

By the way, the above embodiment is constructed on the premise that the motor parameters are correctly grasped and there is no fluctuation during operation. If the motor parameters are not correctly grasped or if they fluctuate during operation, the weighting factors W ₁ and W _{3 in} Eq. (5) are not correct values, so the secondary magnetic fluxes Φ _dr and λ _dr. The error between and is the rotational angular velocity ω _r
It's not just a fault. That is, the weighting factors W ₁ , W ₃
It is necessary to learn the true value of and to estimate the weighting coefficient W _2, that is, ω _r .

FIG. 3 is a block diagram showing the structure of an embodiment which makes this possible. In the figure, the same elements as those of FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted. In this embodiment, the learning of the electric motor parameter is learned by the neuro vector calculation unit 40, the electric motor parameter obtained by the learning is stored in the data buffer 41, and the neuro speed estimation unit 32 uses the stored electric motor parameter to calculate the rotational angular velocity. It is an estimate. The magnetic flux function 42 is a function of “magnetic flux and secondary magnetic flux command” which is set when it is necessary to strengthen or weaken the magnetic flux according to the speed.
_{The r} and the secondary magnetic flux command Φ _r ^* corresponding thereto are added to the neuro vector calculation unit 40.

The embodiment shown in FIG. 3 is characterized in that a neuro vector operation unit 40 is provided, and the other structure is substantially the same as that described with reference to FIG. 1 or 2. Therefore, vector control using a neural network will be described with reference to FIG.

In FIG. 4, the current source inverter 15 controls the current of the induction motor 16 according to the primary current command output from the coordinate system converter 27. A pulse generator 35 is coupled to the induction motor 16, and the output pulse is added by an adder 36 to rotate the rotational angular velocity ω _r.
Is added to the PI controller 22 and the integrator 25. The PI controller 22 determines the torque current command i _qs based on the rotational angular velocity ω _r and a rotational angular velocity command (not shown).
^* Is added to the neuro vector operation unit 40. Further, the rotational angular velocity ω _r and the slip angular velocity command ω _s ^* calculated by the neuro vector calculation unit 40 are added to the integrator 25, and the integrator 25 adds both of them and integrates them to obtain the phase angle θ. It is added to the coordinate system conversion unit 27. The other coordinate system conversion unit 28 converts the detected primary current i into three-phase / two-phase and adds the primary torque current i _qs and the primary excitation current i _ds to the coordinate system conversion unit 27. On the other hand, the magnetic flux calculation unit 33 calculates, based on the primary voltage v and the primary current i supplied to the induction motor,
The secondary magnetic flux λ _dr and the torque current i _qs are calculated and added to the neuro vector calculation unit 40.

The neuro vector calculation unit 40 inputs the magnetic flux command λ _dr ^* and the torque current command i _qs ^*, and calculates the magnetizing current command i _ds ^* and the slip frequency command ω _s ^* by the following equations. i _ds ^* = W ₁₁ λ _dr ^* + W ₁₂ pλ _dr ^* (6) ω _s ^* = W ₂₃ i _qs ^* / λ _dr ^* (7) (6), (7) W ₁₁ and W _{12 in the} equation , W ₂₃ are weighting factors related to the electric motor parameters, and are learned until the secondary magnetic flux λ _dr calculated and detected by the magnetic flux calculator 33 and the actual torque current i _qs become true values. As described above, when detecting the magnetic flux, the use of the above equation (3) does not affect the speed.

The details of the neuro-vector operation unit 40 are described in Japanese Patent Application Laid-Open No. 5-161382 filed and published by the same applicant as the present application, so refer to that publication. Further, the vector control system shown in FIG. 4 is illustrated with a speed detector for ease of understanding, but in the vector control system without a speed sensor, the pulse oscillator 35 and the adder 36 are removed. However, instead, a speed estimator 43 that estimates the speed of the induction motor 16 may be provided, and the rotational angular speed ω _r may be output from this speed estimator.

Therefore, the neuro vector operation unit 40 shown in FIG. 3 learns the parameters of the motor, particularly the mutual inductance M and the secondary resistance R _r , by using the neural network in the process of performing the vector operation. Value is grasped and the result is stored in the data buffer 41. Based on the result, the neuro-velocity estimating unit 32 estimates the velocity in the same manner as described with reference to FIG. 1, so that the velocity estimation using the true motor parameter is performed. In the present embodiment, the neural network of the neuro-vector operation unit 40 and the neural network of the neuro-velocity estimation unit 32 are learned without interfering with each other, and in this process, vector control and its sensorless implementation are simultaneously solved. .

According to this embodiment, the neuro velocity estimation unit 3
2 can be used as an alternative to the speed sensor, and there is also an effect that can be realized without essentially changing the configuration of the original system. In addition, according to the present embodiment, vector control calculation, motor parameter estimation, and motor speed estimation can be performed simultaneously and without interfering with each other, thereby targeting induction motors with large variations in motor parameters. In this case, the characteristic equivalent to that of the DC machine can be obtained.

By the way, in order to improve the control accuracy in controlling the speed of the induction motor, a converter for converting AC into DC and controlling the DC voltage to a set value is provided in the preceding stage of the current source inverter 15. However, the voltage control operation may not be in time for the load variation. It takes a considerable amount of time to learn the weighting coefficient described above, and if the speed changes due to load fluctuation during that period, the estimation result becomes unreliable. That is, the speed estimation operation must be in a steady state regardless of whether the induction motor is in the motor mode or the regenerative braking mode.

FIG. 5 is a block diagram showing the configuration of an embodiment capable of coping with the load fluctuation and the like. In the figure, the same elements as those of FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted. In this embodiment, a converter 51 is connected in front of the current source inverter 15. This converter 51 inputs alternating current,
It outputs a controlled direct current, is provided with a voltage control unit 52, and has a smoothing capacitor 53 connected to the output side. Here, the voltage controller 52 controls the reference voltage V _c.
A deviation ε _v between ^* and the DC voltage V _c detected by a voltage detector (not shown) is obtained, and the converter 51 is controlled so that the deviation ε _v becomes zero. The detailed configurations of the converter 51 and the voltage control unit 52 are publicly known, and thus the description thereof is omitted.

Even if the converter 51 is connected, the control may not be in time for the fluctuation of the load. Even if the speed is estimated in this state, the error becomes large. The neuro-velocity estimation unit 32A estimates the velocity in consideration of this, and has a configuration in which the steady-state determination unit 37 is added to the neuro-velocity estimation unit 32 shown in FIG. The steady state determination unit 37 determines the voltage deviation of the voltage control unit 52, that is,
It is determined whether or not the control deviation ε _v is within a preset reference value (for example, 1%), and when it is within the reference value, a signal for enabling speed estimation is given to the speed estimation unit 34. Speed estimation unit 34
Estimates the speed on the condition that it receives the signal, and if the input signal disappears during the estimation, the estimation results up to that point are invalidated and the speed is estimated again. According to this embodiment, when the voltage control operation is not in time due to the load fluctuation or the like, it is possible to prevent from estimating the speed different from the actual speed.

FIG. 6 is a block diagram showing the configuration of another embodiment capable of coping with load fluctuations and the like. In the figure, the same elements as those of FIG. 3 or 5 are designated by the same reference numerals, and detailed description of their configurations will be omitted. This embodiment is similar to that shown in FIG.
The converter 51, the voltage control unit 52, and the smoothing capacitor 53 described above are additionally provided, and the neuro speed estimation unit 32A described in FIG. 5 is provided instead of the neuro speed estimation unit 32. Here, the steady state determination unit 37 determines whether or not the voltage deviation of the voltage control unit 52, that is, the control deviation ε _v is within a preset reference value, and when it is within the reference value, a signal that enables speed estimation. Is given to the speed estimation unit 34.
The speed estimation unit 34 estimates the speed on the condition that the signal is received, and if the input signal disappears during the estimation, the estimation result up to that point is invalidated and the speed estimation is performed again. Thus, according to this embodiment, when the motor parameters are correctly estimated and the motor speed is estimated based on the estimation result, it is possible to prevent the estimation of the speed different from the actual speed due to the load fluctuation or the like. be able to.

[0041]

As is apparent from the above description, according to the present invention, it is possible to estimate the true speed of the induction machine, and use the estimated speed to determine the primary current command and the slip frequency command. A characteristic equivalent to that of a DC machine can be obtained by calculation. Further, while performing learning using a primary expression with a variable weighting coefficient, vector calculation for matching the magnetic flux command and the torque current command with the calculated secondary magnetic flux and torque current is executed, and the corresponding motor parameter is used. By estimating the speed of the induction machine by using the induction machine, characteristics equivalent to those of the DC machine can be obtained even when the induction machine has large fluctuations in the motor parameters.

Further, the speed estimation operation is performed on condition that the voltage control deviation of the voltage control system is within a preset set value, so that the actual speed can be reduced when the voltage control operation with respect to the load fluctuation is not in time. It is possible to prevent the estimation of a speed different from the above.

[Brief description of drawings]

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

FIG. 2 is a functional explanatory diagram of a main part of the embodiment shown in FIG.

FIG. 3 is a block diagram showing the configuration of a second embodiment of the present invention.

4 is a diagram showing a vector control system with a speed detector for explaining the operation of the main elements of the embodiment shown in FIG.

FIG. 5 is a block diagram showing the configuration of a third embodiment of the present invention.

FIG. 6 is a block diagram showing the configuration of a fourth embodiment of the present invention.

FIG. 7 is a block diagram showing a configuration of a conventional most typical vector controller without a speed sensor.

[Explanation of symbols]

 15 current source inverter 16 induction motor 27, 28 coordinate system conversion unit 30 vector control unit 31 voltage / current detector 32, 32A neuro speed estimation unit 33 magnetic flux calculation unit 34 speed estimation unit 37 steady state determination unit 40 neuro vector calculation unit 41 Data buffer 51 Converter 52 Voltage control unit

Claims (6)

[Claims] 1. A relational expression between the primary voltage and the primary current of the induction machine and the secondary magnetic flux, which is obtained by sampling and detecting the primary voltage and the primary current of the induction machine and does not include a term depending on the speed of the induction machine. Is used to calculate the secondary magnetic flux corresponding to the sampled and detected primary voltage and primary current, and is also sampled and detected using the equation of state regarding the magnetic flux including terms that respectively depend on the primary current and the rotation speed. The secondary magnetic flux corresponding to the primary current is calculated, the two secondary magnetic fluxes are compared for each sampling, and the weighting factor of the term depending on the rotation speed in the equation of state is learned in a direction in which they coincide with each other. The speed estimation method of the induction machine is characterized in that the speed of the induction machine is determined in association with the weighting coefficient with which the secondary magnetic flux coincides. 2. A first power converting means for converting an alternating current into a direct current and controlling a direct current output voltage in a power supply system, and the first power converting means.
And a second electric power converting means for converting a direct current output from the electric power converting means to an alternating current to control an alternating current output current, wherein the primary voltage and the primary current of the induction machine are sampled. Detected, does not include a term depending on the speed of the induction machine, using the relational expression between the primary voltage and primary current of the induction machine and the secondary magnetic flux, to the sampled detected primary voltage and primary current A corresponding secondary magnetic flux is calculated, and a secondary magnetic flux corresponding to the sampled and detected primary current is calculated by using a state equation relating to the magnetic flux including a term that respectively depends on the primary current and the rotation speed, When the voltage control deviation of the power conversion means 1 is within a preset set value, the two secondary magnetic fluxes are compared for each sampling, and the state is set in a direction in which the two secondary magnetic fluxes coincide with each other. Extent learns the weighting coefficient of the term which depends on the rotational speed of the formula, to determine the speed of the induction machine in association with the weighting factor secondary flux matches the speed estimation method of the induction machine, characterized in that. 3. A torque current command is obtained according to the difference between the speed command of the induction machine and the actual speed, and a primary current command and a slip frequency command are calculated based on the torque current command and the magnetic flux command, and the calculated. In an induction machine control device for controlling the current of the induction machine according to the primary current command and the slip frequency command, a detection means for sampling and detecting the primary voltage and the primary current of the induction machine, and a term depending on the speed of the induction machine. A first magnetic flux for calculating a secondary magnetic flux corresponding to the sampled primary voltage and primary current using a relational expression between the primary voltage and primary current of the induction machine and a secondary magnetic flux, which does not include The secondary magnetic flux corresponding to the sampled and detected primary current is calculated by using an arithmetic means and a state equation relating to the magnetic flux including terms that respectively depend on the primary current and the rotation speed. The second magnetic flux calculating means for comparing the two secondary magnetic fluxes for each sampling, and learning the weighting factor of the term depending on the rotation speed in the equation of state in a direction in which the two agree with each other. And a speed estimating unit that determines the speed of the induction machine in association with the weighting coefficient that coincides with the control unit of the induction machine. 4. The primary voltage and the primary sampled and detected by using a relational expression between the primary voltage and the primary current of the induction machine and the secondary magnetic flux, which does not include a term depending on the speed of the induction machine. A magnetic flux calculating means for calculating a secondary magnetic flux and a torque current corresponding to a current, the magnetic flux command and the torque current command, and the calculated secondary magnetic flux and the torque current are alternately input, and a primary expression with a variable weighting factor. While learning by using a vector calculation means for calculating a primary current command and a slip frequency command that match each other and a data storage means for storing a motor parameter corresponding to the learned weighting coefficient, the speed estimation means 4. The speed of the induction machine is determined by using a motor parameter stored in the data storage means.
The control device for the induction machine as described in. 5. An electric power supply system for an induction machine, comprising first power conversion means for converting an alternating current into a direct current and controlling a direct current output voltage,
Second electric power conversion means for converting direct current output from the first electric power conversion means to alternating current to control alternating current output current, and a primary current instruction based on a magnetic flux instruction and a torque current instruction of the induction machine. And a controller for an induction machine that calculates a slip frequency command, and controls the second power conversion means in accordance with the calculated primary current command and slip frequency command, and detects the primary voltage and the primary current of the induction machine by sampling. Detecting means, including the term dependent on the speed of the induction machine, using the relational expression between the primary voltage and primary current of the induction machine and the secondary magnetic flux, the primary voltage and the primary current sampled and detected By using a first magnetic flux calculating means for calculating a secondary magnetic flux corresponding to the above, and a state equation relating to the magnetic flux including terms that respectively depend on the primary current and the rotation speed. Second magnetic flux calculating means for calculating a secondary magnetic flux corresponding to the detected primary current, and a steady state for detecting that the voltage control deviation of the first power converting means is within a preset set value. State determination means, when the voltage control deviation is within a set value, the two secondary magnetic fluxes are compared for each sampling, and the weights of the terms depending on the rotation speed in the state equation in a direction in which they coincide with each other. A controller for an induction machine, comprising: speed estimation means that learns a coefficient and determines the speed of the induction machine in association with the weighting coefficient with which the secondary magnetic flux coincides. 6. The primary voltage and the primary detected by sampling using a relational expression between the primary voltage and the primary current of the induction machine and the secondary magnetic flux, which does not include a term depending on the speed of the induction machine. A magnetic flux calculating means for calculating a secondary magnetic flux and a torque current corresponding to a current, the magnetic flux command and the torque current command, and the calculated secondary magnetic flux and the torque current are alternately input, and a primary expression with a variable weighting factor. While learning by using a vector calculation means for calculating a primary current command and a slip frequency command that match each other and a data storage means for storing a motor parameter corresponding to the learned weighting coefficient, the speed estimation means 6. The speed of the induction machine is determined by using a motor parameter stored in the data storage means.
The control device for the induction machine as described in.