JP4675264B2 - Rotor frequency estimation device and rotor frequency estimation method - Google Patents

Description

  The present invention relates to a rotor frequency estimation apparatus and method for estimating a rotor frequency of an induction motor driven by an inverter that performs a control operation according to an output command value.

  In vector control of an induction motor that does not require a speed sensor, the rotor speed (rotor frequency) is estimated by obtaining electrical information via the rotor flux linkage induced voltage on the stator side (primary side). Yes. Therefore, in an extremely low speed region where the induced voltage is low, it is difficult to estimate the rotor frequency only with electrical information such as current and voltage appearing on the stator side. Therefore, various devices have been devised for estimating the rotor frequency in the extremely low speed range. For example, Non-Patent Document 1 describes a speed sensorless vector control method for an induction motor that drives a railway vehicle, and a method for estimating a rotor frequency based on a motion equation of the vehicle.

A block diagram related to the rotor frequency estimation disclosed in Non-Patent Document 1 is shown in FIG.
FIG. 8 is a diagram showing a relationship between an actual motor system 10 including an induction motor and an inverter that drives the induction motor, and a machine simulator unit 20 that estimates a rotor frequency according to a machine simulator mathematical model. The actual machine system 10 drives an induction motor by supplying an induction motor with a three-phase voltage orthogonally transformed according to voltage output command values V _1d ^* and V _1q ^{* of the} primary side (stator side) magnetic flux and torque, respectively. System. The voltage output command values V _1d ^* and V _1q ^* are obtained in a process in which the actual current value output from the inverter is controlled to follow the current command values I _1d ^* and I _1q ^* . Therefore, it can be said that the inverter drives the motor according to the current command values I _1d ^* and I _1q ^* instead of the voltage output commands V _1d ^* and V _1q ^* .

The machine simulator mathematical model is a mathematical model based on the rotation equation of the wheel in the extremely low speed region where the inverter frequency is about several Hz, and the content is as shown in the following equation (1).

Here, f _rest is the estimated rotor frequency, I _1d is the stator d-axis (magnetic flux component) current, I _1q is the stator q-axis (torque component) current, P _m is the number of pole pairs, G _r is the gear ratio, and M _mset is A mutual inductance setting value, L _2set is a secondary (rotor) self-inductance setting value, J _wset is an equivalent inertia mass setting value, r _w is a wheel radius, and R _tset is a running resistance setting value.

  According to the machine simulator mathematical model according to the mathematical formula (1), the mechanical behavior in the extremely low speed range is modeled by the state equation. Therefore, the stator side current vector that is the actual input is given to the mechanical simulator mathematical model. The rotor frequency can be estimated.

A slip frequency f _s proportional to the ratio of torque to magnetic flux is calculated and added by a predetermined calculator to the estimated rotor frequency f _rest and fed back to the actual system 10 as an inverter frequency f _1. .

As for the magnetic flux component I _1d and the torque component I _1q of the primary current vector value, the current detector detects the three-phase output currents of the inverters u, v, and w, and the detected three-phase current is the dq axis. It is obtained by coordinate conversion to the coordinate system. Since this coordinate transformation is a known technique, a detailed description thereof is omitted.

Kondo, Yuki, “Examination of Application of Induction Motor Speed Sensorless Vector Control to Railway Vehicle Drive”, IEEJ Transactions D, Vol.125-D, January 2005, p.1-8

However, the method of Non-Patent Document 1 estimates the rotor frequency from the magnetic flux component I _1d and the torque component I _{1q of the} primary current based on a mechanical simulator mathematical model that formulates the mechanical state of the induction motor. For this reason, the estimated value of the rotor frequency in the extremely low speed range is determined based on the gradient and load conditions (for example, the number of passengers) with respect to errors and fluctuations of the parameter values of the machine simulator mathematical model and in the case of an induction motor for driving a railway vehicle. ) Will fluctuate greatly with respect to changes. Since the drive control of the motor is performed based on the estimated rotor frequency, the control based on the changed rotor frequency changes the slip frequency, and as a result, the magnetic flux vector changes, resulting in a large torque fluctuation. There was a case. In some cases, there was a possibility of step-out.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to more accurately estimate the rotor frequency in the extremely low speed region.

The first invention for solving the above problems is:
A rotor frequency estimation device for estimating a rotor frequency of an induction motor driven by an inverter that performs a control operation according to an output command value,
A machine that estimates a rotor frequency from a primary current vector value of the induction motor and outputs it as a frequency estimated value according to a machine simulator mathematical model (for example, Expression (1)) determined in advance based on the rotation equation of the induction motor. Simulator means (for example, the machine simulator unit 20 in FIG. 1);
In accordance with an induction machine simulator mathematical model (for example, Expression (2)) determined in advance based on the voltage equation of the induction motor, the first order is derived from the output command value, the estimated frequency value, and the slip frequency of the induction motor. Induction machine simulator means for estimating the torque current and outputting it as a primary torque current estimation value (for example, induction machine simulator unit 30 in FIG. 1);
And a rotor frequency estimation device that compensates for an estimation error of the mechanical simulator means based on the estimated value of the primary torque current.

Moreover, as another invention, a rotor frequency estimation method for estimating a rotor frequency of an induction motor driven by an inverter that performs a control operation according to an output command value,
A machine simulator estimation step for estimating a rotor frequency from a primary current vector value of the induction motor according to a machine simulator mathematical model (for example, Expression (1)) determined in advance based on the rotation equation of the induction motor (for example, FIG. 1 machine simulator section 20),
According to the induction machine simulator mathematical model (for example, Expression (2)) determined in advance based on the voltage equation of the induction motor, the output command value, the rotor frequency estimated in the machine simulator estimation step, and the induction motor An induction machine simulator estimation step (for example, induction machine simulator unit 30 in FIG. 1) for estimating the primary torque current from the slip frequency of
A compensation step for compensating for an estimation error in the machine simulator estimation step based on the primary torque current estimated in the induction machine simulator estimation step;
A rotor frequency estimation method including

  According to the first aspect of the invention, the induction machine simulator means estimates the primary torque current from the output command value, the estimated frequency value estimated by the mechanical simulator means, and the slip frequency of the induction motor. That is, if the estimated frequency value is different from the actual rotor frequency, an error may occur in the slip frequency, and the slip frequency error affects the primary torque current. Therefore, the primary torque current is estimated from the output command value, the frequency estimated value, and the slip frequency, and the rotor frequency is appropriately adjusted by compensating for the estimation error of the rotor frequency based on the estimated primary torque current. Can be estimated.

Specifically, for example, as a second invention, the rotor frequency estimation device of the first invention,
Compensation value calculation means for calculating a compensation value by performing a predetermined proportional-integral calculation based on the difference between the primary torque current of the induction motor detected by the predetermined detection means and the estimated value of the primary torque current ( For example, the proportional integrator 40) of FIG.
Compensation value addition means (for example, addition point 50 in FIG. 1) for adding the compensation value calculated by the compensation value calculation means to the frequency estimation value to compensate for the estimation error of the machine simulator means;
With
The induction machine simulator means estimates a primary torque current from the output command value, a frequency estimation value obtained by adding a compensation value by the compensation value addition means, and a slip frequency of the induction motor.
A rotor frequency estimation device may be configured.

Moreover, as 3rd invention, it is the rotor frequency estimation apparatus of 1st invention,
The machine simulator mathematical model is a mathematical model (for example, Equation (4)) that balances the force obtained by subtracting the resistance from the product of the inertia force of the rotor multiplied by the torque and a given torque coefficient.
Torque coefficient changing means for increasing the torque coefficient as the difference between the estimated value of the primary torque current and the primary torque current command value by the inverter is larger, and decreasing the torque coefficient as the difference is smaller (for example, FIG. 7 computing units 80),
A rotor frequency estimation device may be configured.

  According to the present invention, it is possible to appropriately estimate the rotor frequency in the extremely low speed region.

  Hereinafter, an embodiment of a rotor frequency estimation device incorporated in a control device that controls an induction motor that drives a railway vehicle by speed sensorless vector control will be described with reference to the drawings.

1. Configuration and Operation The rotor frequency estimation device 1 is realized by a computer including a CPU, a ROM storing a program, a RAM, and the like. For example, the rotor frequency estimation device 1 is mounted on a control device of an induction motor as a control board. Moreover, it can be integrally configured as an inverter device including an inverter for supplying drive power to the induction motor.

  FIG. 1 is a block diagram showing the control contents of the rotor frequency estimation device 1. The rotor frequency estimation device 1 includes an induction machine simulator unit 30 and a proportional integrator 40 in addition to the same machine simulator unit 20 as described above as the background art.

The induction machine simulator unit 30 follows the induction machine simulator mathematical model, and based on the inverter frequency f ₁ and the voltage output command values V _1d ^* and V _1q ^* , the primary d-axis current I _1d and the primary q-axis current I _1q. Is a control block for estimating. As described in the background art, the current output command values I _1d ^* and I _1q ^* may be used instead of the voltage output command values V _1d ^* and V _1q ^* .

The content of the calculation by the induction machine simulator mathematical model is based on the following formula (2).

Here, suffix n is the nominal value, suffix est is the estimated value, suffix Ref is the command value, R ₁ is the primary (stator) winding resistance, R ₂ is the secondary (stator) winding resistance, L ₁ represents a primary self-inductance, L ₂ represents a secondary self-inductance, M _m represents a mutual inductance, V _1dRef represents a d-axis voltage command value, and V _1qRef represents a q-axis voltage command value.

Also, R _12n = R _1n + (M _mn / L _2n ) ² R _2n , ω _rest = 2πP _m f _rest , ω _S = R _2n I _1qRef / (L _2n I _2dRef ), ω ₁ = ω _rest + ω _S is there.

A difference obtained by subtracting the actually measured primary q-axis current from the primary q-axis current estimated value I _1quest estimated by the induction machine simulator 30 is input to the proportional integrator 40.

The proportional integrator 40 is an arithmetic unit that proportionally integrates an input value and outputs it as a frequency compensation value Δf _rest . The gain curve of the proportional integrator 40 is shown in FIG. 2, and the transfer function is shown in the following equation (3).

Each gain in equation (3) is adjusted and set as follows. That is, first, the integral gain K _MechI is gradually increased while the proportional gain K _MechP is _kept at “0”. In the process, vibration due to the vibration frequency component unique to the system appears. The vibration angular frequency at this time is set to ω _sys, and the corner frequency 1 / T _Mech of the proportional integrator 40 is set to be smaller than ω _sys . After setting the _break frequency 1 / T _Mech , the gain K _Mech is gradually increased again. At this time, the frequency compensation value Δf _rest is adjusted to be eliminated.

Then, the estimated rotor frequency value output from the machine simulator unit 20 is the pre-compensation estimated value f _rest0, and the frequency compensation value Δf _rest output from the proportional integrator 40 is added and compensated for, so that the rotor frequency The estimated value f _rest is output from the rotor frequency estimation device 1.

Further, the slip frequency f _s is added to the compensated rotor frequency estimated value f _rest to obtain the inverter frequency f ₁ , which is fed back to the actual system 10.

The principle that the above-described rotor frequency estimation device 1 compensates for the rotor frequency f _rest0 estimated by the machine simulator unit 20 is as follows. That is, when the error between the actual rotor frequency f _r and the rotor frequency estimate f _Rest0 uncompensated obtained by mechanical simulator 20 occurs, an error occurs in the slip frequency f _s. This is reflected in the induction machine simulator unit 30 as the primary angular frequency ω ₁ (inverter frequency f ₁ ). The primary voltage vectors V _1d and V _1q reflect the result of constant current control for the actual system 10, but reflect the estimated current vectors I _1dest and I _1quest that are the output current vectors of the induction machine simulator unit 30. Not. Therefore, the error of the slip frequency f _s reflected through the primary angular frequency ω ₁ is reflected in the estimated current vectors I _1dest and I _1quest . Of these, focusing on the q-axis current that is directly proportional to the torque, the deviation between the actual value and the estimated value is corrected by the output Δf _rest of the proportional integrator 40 to correct the rotor frequency estimated value f _rest0 . Thereby, the estimation error of the rotor frequency by the machine simulator unit 20 caused by the error of each parameter value of the machine simulator mathematical model is compensated.

2. Experimental Results A numerical simulation of electric vehicle control incorporating the rotor frequency estimation device 1 was performed, and the effect of the rotor frequency estimation device 1 was confirmed. In the numerical simulation, the current was raised 2 seconds after the start of the simulation. This is for simulating the time required for the rising of the power running current after releasing the brake in an actual vehicle, and is a strict condition of 2 seconds longer than the actual time. As a result, the start from the slow forward state is reproduced on the down slope, and the start from the reverse state is reproduced on the up slope.

(1) Empty vehicle / downhill slope In a commuter train or the like, a simulation was performed assuming an activated state with a 35 permil downhill slope in an empty state. At this time, the equivalent load setting value and the running resistance setting value of the induction machine simulator section 30 are completely different values, that is, a case where the commuter train is 200% full (the load is about half of the empty vehicle mass) and is running flat. The assumed value was used.

FIG. 3 shows a simulation result when the induction machine simulator unit 30 is started up without compensation. Since there is no compensation by the induction machine simulator unit 30, control is performed with a configuration substantially similar to the configuration shown in FIG. As a result of the simulation, as shown in FIG. 3A, the generated torque _Tm is about half of the estimated torque _Tmest . For this reason, although acceleration is also accelerated at a downward slope, it is only about half of the planned value (estimated value).

On the other hand, FIG. 4 shows a simulation result in the case where compensation by the induction machine simulator unit 30 is present. For those with compensation by the induction machine simulator 30, FIG. (A) generating torque _T m, any of (b) q-axis torque _{I 1q,} and (c) the rotor frequency _{f r,} the simulation starts after 5 seconds, i.e. After 3 seconds from the start of estimation by the rotor frequency estimation device 1, a result that coincided with each estimated value was obtained. It can be said that appropriate acceleration was performed in the extremely low speed region by compensating for the influence of setting errors of mechanical parameters.

(2) Full occupancy / uphill gradient In a commuter train, etc., a simulation was performed assuming a starting state at a 35 per mil uphill in a 200% full state (load is about half of the empty car mass). At this time, the equivalent load setting value and the traveling resistance setting value of the induction machine simulator unit 30 are completely different values, that is, values assuming a case where the vehicle is running empty and flat.

FIG. 5 shows a simulation result when the induction machine simulator unit 30 is started without compensation. Since there is no compensation by the induction machine simulator unit 30, control is performed with a configuration substantially similar to the configuration shown in FIG. Simulation results, as shown in FIG. 5 (c), deviate rotor frequency estimate f _rest and the rotor frequency f _r is, since it is assumed start of the up steep in this case, leading to loss of synchronism It turns out that it is set back later.

On the other hand, FIG. 6 shows a simulation result in the case where compensation by the induction machine simulator unit 30 is present. FIG (a) generating torque _T m, (b) q-axis torque _{I 1q,} and (c) both of the rotor frequency _{f r,} the simulation starts 5 seconds later, that is, 3 seconds after the start of estimation by the rotor frequency estimating apparatus 1 Corresponds to the respective estimated values, and the control in the extremely low speed region functions well even under full uphill conditions, and can be started appropriately.

3. Modified Example In the above embodiment, the rotor frequency estimation device 1 has been described as intended for an induction motor for a railway vehicle. However, induction motors are also used as power sources for various machine devices such as machine tools, manufacturing equipment, lifts, fans and pumps, in addition to electric vehicles such as railway vehicles and electric vehicles. Therefore, it goes without saying that the rotor frequency estimation device 1 may be applied to the induction motor that is the power source to improve the starting characteristics.

Further, the rotor frequency estimation apparatus to which the present invention is applicable is not limited to the configuration shown in FIG. For example, it is good also as a structure like the rotor frequency estimation apparatus 2 shown in FIG. That is, the calculator 90 calculates the ratio between the primary q-axis current estimated value I _1qest estimated by the same induction machine simulator 30 as the rotor frequency estimation device 1 of FIG. 1 and the primary q-axis current command value I _1qRef. Then, the calculator 80 calculates (1-G _tm ) K _gain with the result as G _tm , and the machine simulator unit 25 estimates and calculates the rotor frequency estimated value f _rest .

A machine simulator mathematical model which is a mathematical model of the machine simulator unit 25 is as follows. Each constant and coefficient are the same as in equation (1).

The operating principle of the rotor frequency estimation device 2 is based on the fact that a slip frequency error appears in the primary q-axis current command value I _1qRef . That is, if the difference between the primary q-axis current estimated value I _1quest and the primary q-axis current command value I _1qRef is small, the value of (1-G _tm ) K _gain by the calculator 80 approaches “0”, and the difference Is large, the value of K _gain is effective for the torque term (first term on the right side) of Equation (4).

  Such a rotor frequency estimation device 2 can also improve the estimation of the rotor frequency in the extremely low speed region.

The block diagram of a rotor frequency estimation apparatus. Proportional calculator gain curve. Simulation results with no compensation by the induction machine simulator. Simulation results with compensation by induction machine simulator. Simulation results with no compensation by the induction machine simulator. Simulation results with compensation by induction machine simulator. The block diagram of the rotor frequency estimation apparatus as a modification. The block diagram in connection with the conventional rotor frequency estimation.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1, 2 Rotor frequency estimation apparatus 10 Real system 20 Machine simulator part 30 Induction machine simulator part 40 Proportional integrator

Claims (4)

A rotor frequency estimation device for estimating a rotor frequency of an induction motor driven by an inverter that performs a control operation according to an output command value,
Mechanical simulator means for estimating a rotor frequency from a primary current vector value of the induction motor and outputting it as a frequency estimated value according to a machine simulator formula model determined in advance based on the rotation equation of the induction motor;
A primary torque current is estimated by estimating a primary torque current from the output command value, the estimated frequency value, and the slip frequency of the induction motor according to a predetermined induction machine simulator mathematical model based on the voltage equation of the induction motor. Induction machine simulator means for outputting as an estimated torque current value;
And a rotor frequency estimation device that compensates for an estimation error of the machine simulator means based on the estimated value of the primary torque current. Compensation value calculating means for calculating a compensation value by performing a predetermined proportional-integral calculation based on a difference between the primary torque current of the induction motor detected by the predetermined detecting means and the estimated value of the primary torque current; ,
Compensation value adding means for adding the compensation value calculated by the compensation value calculating means to the frequency estimated value to compensate for the estimation error of the machine simulator means;
With
The induction machine simulator means estimates a primary torque current from the output command value, a frequency estimation value obtained by adding a compensation value by the compensation value addition means, and a slip frequency of the induction motor.
The rotor frequency estimation apparatus according to claim 1. The machine simulator mathematical model is a mathematical model in which the inertial force of the rotor balances the force obtained by subtracting the resistance from the product of the torque and the given torque coefficient,
Torque coefficient changing means for increasing the torque coefficient as the difference between the primary torque current estimated value and the primary torque current command value by the inverter increases, and decreasing the torque coefficient as the difference decreases;
The rotor frequency estimation apparatus according to claim 1. A rotor frequency estimation method for estimating a rotor frequency of an induction motor driven by an inverter that performs a control operation according to an output command value,
A machine simulator estimating step of estimating a rotor frequency from a primary current vector value of the induction motor according to a machine simulator mathematical model predetermined based on a rotation equation of the induction motor;
In accordance with an induction machine simulator mathematical model determined in advance based on the voltage equation of the induction motor, the output command value, the frequency estimation value estimated in the machine simulator estimation step, and the slip frequency of the induction motor are linear. An induction machine simulator estimation step for estimating torque current;
A compensation step for compensating for an estimation error in the machine simulator estimation step based on the primary torque current estimated in the induction machine simulator estimation step;
A rotor frequency estimation method including:

Images