RU2771794C1 - Method for identifying the time constant of the rotor and the mutual inductance of the stator and rotor of an asynchronous motor in a vector field-oriented control system - Google Patents

Abstract

FIELD: electric motors controlling methods.

SUBSTANCE: invention relates to a method for controlling electric motors using vector control. The method for identifying the time constant T_r of the rotor of an induction motor in the system of vector field-oriented control consists in setting the initial values ​​of the time constant T_r of the rotor and current I_d, then at least two times different values ​​of the current I_q are set, at which the torque on the shaft of the asynchronous motor is measured, fix at a given moment of time, the values ​​of the angular frequency of rotation of the rotor, the magnetic flux and the load on the rotor, and taking into account these parameters, the change in the torque-generating current component I_q is fixed, which will correspond to the new value of I_q1, as well as the value of the electromagnetic torque M1. Further, when the load on the shaft of the asynchronous motor changes, a new value of I_q2 is also recorded, as well as a new value of the electromagnetic torque M₂. Next, the ratios of currents KI and the ratios of moments K_M are calculated according to the equations, where K_I=I_q2/I_q1 and K_M=M₂/M₁, the condition is checked under which K_I=K_M. The asynchronous motor is controlled at the value of the time constant Tr of the rotor, at which the condition K_I=K_M is fulfilled.

EFFECT: improving the field orientation accuracy of the vector control system for an asynchronous motor.

1 cl, 2 tbl, 3 dwg

Description

The present invention relates to the field of electrical engineering, in particular to the control of an electric drive based on an asynchronous motor with a squirrel-cage rotor, and can be used in control systems for asynchronous electric drives of vehicles.

Today, electric drives built on the basis of an asynchronous machine (AM) using vector control with orientation along the rotor flux vector (in foreign literature - "Field-oriented Control") have received a fairly wide application in various electromechanical installations with high requirements for control indicators in static and dynamics.

The vector control algorithm assumes the division of the stator current vector I into two components - the flux-forming Id and the torque-forming Iq, which are projections of the current vector on the axes of the rotating orthogonal coordinate system d-q, oriented along the vector of the rotor flux linkage ψr. It should be noted that it is this orientation principle that makes it possible to separately control the components of the stator current AM, in other words, to separately control the magnetic flux and the electromagnetic moment of the machine. The principle of vector control today is described in detail both in domestic and foreign scientific and professional literature.

At the same time, an asynchronous machine, as a control object, has such properties as non-linearity and non-stationarity.

The non-linearity of AM is due to the fact that the differential equations that mathematically describe an electric machine include terms proportional to the product of the state variables (currents, flux links and rotational speed). In the literature, such components are usually referred to as crosslinks.

As a rule, it is possible to significantly reduce the influence of cross-links by introducing a block for compensating cross-links into the structure of the control system (CS).

The non-stationarity of AM is due to the variation of the parameters of the windings due to the influence of various factors. So, for example, the active resistances of the stator and rotor of the machine change both due to heating of the windings and due to the surface effect (skin effect). Mutual inductance is most dependent on machine flux, but can also be correlated with RPM (typically to a much lesser degree). Such parameters of an electrical machine (EM), as the leakage inductance of the stator and rotor, are quite small in relation to the magnitude of the mutual inductance, so they can be taken as constants.

At the same time, for the qualitative regulation of the components of the stator current along the longitudinal and transverse axes, the control system must have reliable information about the current parameters of AM during operation.

This is explained by the fact that the control system, among others, includes a block for determining (identifying or calculating) the parameters of the rotor flux vector, namely, its module and angle of rotation. This block performs one of the key functions in the structure of the vector control system, since the signal of the angle of rotation of the flux linkage vector is involved in the transformation of coordinates (transition from three-phase stationary to two-phase rotating), and the signal of the module of this vector is involved in the control system to form a given electromagnetic torque, as well as for feedback of the flux linkage control loop.

At present, the above quantities are determined indirectly (sensorless) rather than directly (sensor). In turn, the structure of the block for determining the parameters of the flux vector is built on the basis of a part of the mathematical description AM. Most often, the equations of the EM rotor circuit are taken as a basis for its mathematical description in the orthogonal coordinate system d-q. These equations contain, in particular, such EM parameters as the mutual inductance of the stator and rotor Lm, as well as the time constant of the rotor Tr that depends on it.

The process of obtaining information about the above AM parameters is called identification (or determination, evaluation). There are different ways to identify the parameters of the windings of electrical machines, in particular, asynchronous ones, which can be conditionally divided into two types:

1) online identification (during the normal operation of the electromechanical installation);

2) preliminary identification.

The first type of identification (in online mode), as a rule, is associated with the use of dynamically changing search algorithms directly in the process of operation (operation) of the electric drive. The main advantage of such algorithms is the absence of the need for any preliminary procedures, and the disadvantages are the complexity of their synthesis and the need to allocate additional energy and computing resources of the control controller.

The second type, as the name implies, is based on the procedure for preliminary identification of parameters using, as a rule, subsequent processing of the results and their inclusion in the control algorithm in the form of lookup tables (the term "Lookup Table" is common in foreign literature) or in the form of power polynomial dependencies, for further use in the process of regular work. The main advantage of this method is a single identification procedure with subsequent application of its results.

So, for example, a method for identifying the time constant for the rotor of an induction motor is known from the prior art (see patent document CN 107896079, publ. 04/10/2018). This method belongs to the second type of identification. The known method is based on the construction of a rotor flux formation model and has a rather complex algorithm for determining the time constant.

The present application also describes a method related to the second type of identification. This method assumes the presence of a certain set of software and hardware, and also requires the execution of a certain sequence of actions (algorithm), which is described in this application. This algorithm is implemented using hardware with appropriate software (software) that implements its implementation.

The technical result to be achieved by the proposed technical solution is to increase the accuracy of field orientation of the vector control system of an asynchronous motor by determining the time constant of the rotor with a high degree of accuracy.

The claimed technical result is achieved by creating a method for identifying the time constant T _r of the rotor of an induction motor in a vector field-oriented control system, which consists in performing the following steps:

a) set the initial value of the time constant T _r of the rotor in the block for evaluating the rotor flux linkage and the initial value of the current I _d , then at least two times different values of the current I _q are set, at which the torque on the shaft of the asynchronous motor is measured;

b) fix at a given moment of time the values of the angular frequency of rotation of the rotor, the magnetic flux and the load on the rotor, and taking into account these parameters, fix the change in the torque-generating current component I _q , which will correspond to the new value Iq ₁ , as well as the value of the electromagnetic torque M, which will be correspond to the new value M ₁ , at the same time it is established that the parameters Iq ₁ and M ₁ correspond to the parameters of the operation of the asynchronous motor at the first operating point, then when the load on the shaft of the asynchronous motor changes, the change in I _q is also fixed, which will correspond to the new value Iq ₂ , and also the value of the electromagnetic moment M, which will correspond to the new value of M ₂ , taking into account these parameters, the transition of the operation of the induction motor to a new operating point is fixed;

c) then start the process of calculating the ratio of currents K _I and the ratio of moments K _M at different operating points according to the equations, where K _I =Ig ₂ /Ig ₁ and K _M =M ₂ /M ₁ ;

d) further check the fulfillment of the condition under which the ratio of currents must be equal to the ratio of moments, namely K _I =K _M ;

e) if the condition in step d) is not met, the selected value of the time constant T _r is adjusted by increasing or decreasing it and repeating steps a)-d) until the condition K _I =K _M is met;

f) carry out control of the asynchronous motor at such a value of the time constant T _r of the rotor, which led to the fulfillment of the condition K _I =K _M .

The combination of all the above features is aimed at improving the accuracy of field orientation of the vector control system of an asynchronous motor by determining the time constant of the rotor with a high degree of accuracy.

Brief description of the drawings

Fig. 1. Functional diagram of the installation for carrying out the procedure for automatic identification of parameters.

Fig.2. Basic structure of the control system.

Fig. 3. Scheme of the parameter identification algorithm.

Description of the idea underlying the proposed method for identifying the rotor time constant.

Let us consider the Park-Gorev equations of the rotor circuit AM with vector field-oriented control (1) without taking into account losses in steel, which are taken as the basis for estimating the modulus and the angle of rotation of the rotor flux vector:

where I _d , I _q - components of the stator current along the axes d and q, respectively (or flux-forming and torque-forming currents); ψ - the module of the flux linkage vector (hereinafter referred to as the flux linkage) of the rotor; L _m - mutual inductance of the stator and rotor; T _r =L _r /R _r - time constant of the rotor; L _r =L _rσ +L _m is the total inductance of the rotor; L _rσ - leakage inductance of the rotor; R _r - active resistance of the rotor; z _p - number of pairs of poles; ω - mechanical angular frequency of rotation of the rotor.

It is also known that the electromagnetic torque on the shaft AM with vector control is calculated by formula (2):

In the static mode of operation of AM, the derivative of the flux linkage can be considered equal to zero. Then the flux linkage will be proportional to the flux-forming current through the mutual inductance parameter. With this in mind, the expression for the rotational speed of the rotor flux vector (or d-q coordinate system) in static mode is also transformed. The system of equations (1) for AM operation in statics will take the form (3):

In the second equation of system (3), the second term is called the slip frequency of the rotor flux vector.

In this case, the electromagnetic torque in the static mode, based on (2) and taking into account the first equation (3), is calculated according to equation (4):

Based on equations (3) and (4), the idea of identifying the time constant of the rotor is formed, which is as follows. At a certain fixed value of the machine flux (in static mode, it is equivalent to the flux-generating current I _d ), a change in the torque-generating current component I _q will lead to a proportional change in the moment on the machine shaft if the axis d of the coordinate system dq is oriented exactly along the vector of the rotor flux linkage. In turn, the exact orientation of the SC dq can be achieved by determining the time constant of the rotor with a high degree of accuracy, as well as accurately measuring the rotor speed (for example, using a sensor).

In this case, one should proceed from the condition that in the static mode of operation there is no mutual influence of the components of the stator current (I _d and I _q ).

.The range of currents I _d and I _q is selected as follows. Typically, a machine is given a range of stator current changes in which it must operate (taking into account overload). It is known that the total stator current is

.

In this case, the proportion of the current I _d is approximately 1/3, and the current I _q - 2/3. Although this ratio is purely indicative, it still allows you to choose the right values for identification.

.So, knowing the maximum possible stator current, you can calculate the maximum values of the currents I _d , I _q :

.

The value of the minimum current Id is recommended to be chosen as 10% of the maximum value, or to choose based on the minimum possible current I _d if known. The step or increment with which the current I _d will be incremented during the identification procedure is also recommended to be chosen as 10% of the maximum, which will provide 10 points of the obtained tabular dependences of the time constant and mutual inductance.

The set time for working off the current I _q should be approximately equal to (3 ... 5) T _{r, beginning} .

The set of current values I _q is selected on the basis of those considerations in order to measure the moment of different magnitudes. The goal is to achieve proportionality between I _q and the moment in a wide range of their change.

It is recommended to choose a set of current values I _q as 1/6, 1/3, ½, 2/3, 5/6 and 1 from the maximum current I _q , if we set six values.

The choice of parameters such as the minimum value and the increment of currents is not something strict, and the above is purely advisory. So, for example, the number of specified points in the current range Id, determined by the selected increment, will affect the "detail" of the curves Tr(Id), Lm(Id), i.e. the more points are set - the more detailed curves will be obtained.

In turn, an increase in the number of set values of current Iq will increase the accuracy of parameter identification. The presence of such factors as noise, interference, the accuracy class of measuring sensors, the non-linearity of the electric machine itself, will cause one or another degree of inaccuracy in measuring current and torque. At the same time, too much increase in the number of set I _q values will increase the duration of the identification procedure. For this reason, as a compromise, it is proposed to use 6 values of I _q , at which the moment on the machine shaft is measured.

Let's consider an abstract example. Let's say AM is at some operating point in motor mode with fixed values of angular frequency, magnetic flux and load. For this operating point, denote I _q =I _q1 , M=M ₁ . At some point in time, the load changes and is set at a different level, there is a transition to the operating point. For this point we denote I _q =I _q2 , M=M ₂ .

The ratio of currents I _q and moments M at the second point to those at the first point will be denoted as k ₁ and k _M (5):

Then, with the correct orientation of the SC d-q, the ratio of currents will be equal to the ratio of moments, i.e.

In practice, the task of identifying the rotor time constant using the proposed method can be described by a sequence of actions:

1. Some initial value of the rotor time constant is set in the block for estimating the rotor flux linkage based on equations (1), as well as some value of the current I _d .

2. Several times (two or more) different values of the current I _q are set, at which the moment on the machine shaft is measured.

3. Ratios of currents I _q and moments at different points are calculated according to equations (5).

4. The fulfillment of condition (6) is checked. If the condition is not met, the selected value of the time constant must be corrected and the described actions should be carried out until condition (6) is met.

In other words, the process of identifying T _r is, rather, its determination by the method of successive approximations.

Calculation of the mutual inductance of the stator and rotor.

Let us introduce the concept of equivalent inductance, which is equal to:

At the same time, from equation (4), assuming that the d-q coordinate system is oriented along the rotor flux linkage (i.e., the rotor time constant, in accordance with the above, is defined correctly), we write:

Hence, having the value of the measured moment, as well as the currents along the axes d, q, it is possible to calculate the equivalent, and then the mutual inductance, according to the equation following from (7):

The positive root of this quadratic equation will be the desired one for calculating the mutual inductance parameter (since there can be no negative inductance):

Equation (10) includes the parameter of the rotor leakage inductance L _rσ , which, as a rule, is significantly less than the total rotor inductance (usually about 5% of its value) and is considered a constant, i.e. does not depend (or only slightly depends) on any quantities. This parameter must be known during the identification procedure.

Features of the implementation of the identification algorithm in practice.

It is expedient to estimate the time constant of the rotor using a value that characterizes the ratio of the values k _M and k _I as a percentage, using the following formula proposed in this application:

In practice, this value shows how correctly the rotor time constant is determined by its deviation from 100%.

In other words, when the drive is operating with the yet undefined parameter T _r , it is initially necessary to set some of its base value.

Asynchronous machines have nominal equivalent circuit parameters, which include the rotor resistance R _r , the rotor leakage inductance L _rσ , and the mutual inductance L _m . The first value R _r usually depends on the temperature of the rotor and changes during the heating of the machine, the second value L _rσ is relatively small and, as a rule, is considered constant, the third value L _m varies depending on the magnetic flux of the machine in a wide range (it is also subject to identification) .

The rotor time constant is

The initial (or basic) value of T _r is recommended to be calculated based on R _r for a heated machine, which is usually given, and also on the basis of L _m for the nominal flow of the machine.

Nevertheless, even if the base value T _r for some reason turns out to be far from the true initial value (that is, from that at the initial point I _d =I _d,min ), the identification algorithm itself will select this value properly.

Based on the results of measurements of the moments M and currents I _q at some static points, the values k _I , k _M , k _MI are calculated. In the event that k _MI exceeds 100%, the fitted parameter T _r is higher than the true value, otherwise vice versa.

In addition, the identification of the time constant is proposed to be performed in relative units, i.e., given the base value T _r,b , use the variable coefficient k _T , which will show the deviation of the time constant from T _r,b in accordance with formula (12):

It is also important to note that due to not ideal measurements, the presence of noise and interference, in a real identification algorithm, to improve the estimation accuracy, it may be necessary not two (as described above), but more measurements of currents and torques.

For example, the identification algorithm implemented by the authors of this application involves the use of six measurements of currents I _q and the corresponding torques M at a fixed current I _d . In other words, at a fixed current value I _d , six current values I _q are set in succession, at which the torque on the shaft of the electric machine is measured. As a result, we obtain five (one less than the number of measurements) values k _I , k _M and k _MI . After that, the arithmetic mean value k _MI,av should be calculated from the results of six torque measurements using expression (13):

It is also necessary to set some estimation accuracy Δk _MI,tol (in percent), which will form a confidence interval (100 - Δk _MI,tol )…(100 + Δk _MI,tol ), which is necessary because the measurements are not ideal.

The meaning of the value of Δk _MI,tol lies in the fact that when the value of k _MI,av falls within the specified confidence interval according to the results of measurements, the identified value of the time constant can be considered reliable.

If k _MI,av does not fall within the confidence interval, the procedure should be repeated, changing the coefficient k _T properly with the help of the increment Δk _T .

The criterion for successful selection (identification) of the value of T _r is the fact that the ratio of currents I _q at the current and previous steps is exactly the same as the ratio of moments. This means that the coordinate system dq is correctly oriented, i.e. the value of T _r , which is embedded in the vector control system (ie, in the control controller) coincides with that in the machine. In order to assess how close these ratios are in reality, there are such quantities as the ratio of currents I _q (k _I ) and the ratio of moments (k _M ).

In other words, if k _I =k _M , we assume that at a given point, i.e. at the current given current I _d we have chosen the correct value of T _r , and we can move on to the next point I _d .

However, in practice, k _I and k _M will never be exactly equal, but can be quite close in value. To establish the degree of this proximity, there is a value k _MI =k _M /k _I ⋅ 100%, which shows in percentage how much the ratio of moments is greater or less than the ratio of currents I _q . The confidence interval determines the fact that, for example, the identified rotor time constant is determined correctly with such and such an accuracy.

Let's say we set the accuracy to 1%. Based on the measurement results, we obtained that k _MI =99.2%, i.e. we do not reach 100% by 0.8%, which is less than the established confidence interval. So, everything was chosen correctly. If it turned out that k _MI =98.9% or lower, then this would mean that the current set T _r is lower than the true one, which means that it must be corrected as described above and the measurement repeated.

Let's look at a specific example.

Example 1. Let the confidence interval be (99.5…100.5)%. According to the results of six measurements, the value of k _MI,av was 101.3%. Therefore, the identified value of T _r is greater than the real one, a series of measurements should be repeated, reducing k _T . According to the results of the repeated series, k _MI,av = 100.1% was obtained. Since this value falls within the confidence interval, the measurement at this point I _dr can be considered complete, and the identified value of T _r is reliable.

In addition, the presence of noise and interference, or a too low value of Δk _MI,tol can cause the identification procedure to hang, namely, multiple failures of k _MI,av in the confidence interval.

Let's consider two typical examples, when during the identification procedure in practice, such cases of hanging can occur:

Example 2. Let the confidence interval be (99.5…100.5)%.

Measurement 1. k _MI,av =102.3%, k _T should be reduced, repeat.

Measurement 2. K _MI,av \u003d 98.5%), k _T should be increased, repeat.

Measurement 3. k _{MI, av} = 101.5%, k _T should be reduced, repeated.

Measurement 4. k _MI,av = 99.0%), increase k _T , repeat.

Example 3. Let the confidence interval be (99.1…100.1)%.

Measurement 1. k _{MI, av} = 99.8%), k _T should be reduced, repeated.

Measurement 2. k _MI,av =100.1%, increase k _T , repeat.

Measurement 3. k _{MI, av} = 99.9%), k _T should be reduced, repeated.

Measurement 4. k _MI,av =100.2%, increase k _T , repeat.

The reason for the freeze in example 2 is most likely the increased noisiness of the torque signal or too large increment Δk _T , which can be either positive or negative depending on the deviation of k _MI,av in the positive or negative direction from 100%).

The reason for freezing in example 3 is the too low value of Δk _MI,tol i.e. too "narrow" confidence interval.

In order to avoid freezing, it is necessary, firstly, to smooth the signal of the measured torque (and, if necessary, the stator current), for example, by filtering; secondly, one should properly select the value of the increment Δk _T ; thirdly, it is necessary to choose a compromise value Δk _MI,tol i.e. such that, on the one hand, to ensure good estimation accuracy, and on the other hand, to increase the probability that the value k _MI,av falls within the confidence interval. For example, to identify the parameters of traction machines, the authors of the present application used Δk _MI,tol =0.5%.

The necessary tools for carrying out the procedure for identifying Tr and Lm.

The proposed functional diagram of the installation for carrying out the procedure for automatic identification of parameters is shown in Fig. one.

Below is a description of the components that make up the proposed installation.

Block 1. The investigated machine (IM) is asynchronous with a squirrel-cage rotor with built-in speed and temperature sensors (DS, DT). If necessary, a cooling system for the IM windings is installed.

Block 2. Racing car (GM) with built-in speed and temperature sensors (DS, DT). If necessary, a GM winding cooling system is installed.

Block 3. Power converter supplying the windings of the IM (SPI). SPI power modules are controlled by a controller. Phase current sensors (CFT) and voltage sensor ZPT (DN) are usually mounted inside the SPI block.

Block 4. Power converter supplying the windings of the GM (LNG).

Block 5. A source of constant high voltage voltage (HPV) for powering the DC link (DC link).

Block 6. Torque sensor (DKM) on the shaft of the IM-GM twin. The authors of this application use the torque sensor manufactured by "Tensotekhnika".

Block 7. Decoder that converts the measured torque signal from analog type to digital type. The authors of this application use the decoder "T-35", which is included with the torque sensor.

Block 8. Programmer (JTAG emulator) for controller firmware. The present applicants use a JTAG emulator manufactured by Sauris.

Block 9. USB-CAN adapter for data exchange between PC and controller.

Block 10. Personal computer (PC).

Experiment preparation.

1. The test bench is prepared according to the scheme of Fig. 1. In this case, the IM and GM shafts must be rigidly connected to each other.

2. The UDC voltage is supplied from the source of the high-voltage source of the HPVN, supplying the SPI and LNG. The voltage must be such that the maximum value of the three-phase alternating voltage converted by means of the SP is not lower than the nominal value of the supply voltage of the IM stator windings.

3. The GM shaft spins up to the required rotational speed, at which the experiment will be carried out. The GM must operate in the speed maintenance mode and withstand the maximum load that will be thrown by means of the IM, which should be taken care of in advance.

4. If necessary (and available), the IM and GM cooling systems are switched on;

5. The SPI control controller is “flashed” with the necessary software through the programmer, which allows the identification procedure to be carried out. This issue will be discussed in more detail below;

6. The identification program starts on the PC, after which the procedure can be started.

Features of the software of the control controller that implements the vector control of the asynchronous machine under study.

Let's give a description of the basic structure of the control system (shown in figure 2), which is required to carry out the described identification procedure.

The structure of the SU includes:

• regulators of flux-forming and torque-forming stator current (RPT, RMT);

• unit for determining the module and the angle of rotation of the rotor flux linkage vector (BOP);

• Park and Clark coordinate converters for transferring currents from a fixed three-phase coordinate system to an orthogonal rotating SC (A,B,C-d,q) and voltages at the output of regulators in the opposite direction (d,q-A,B,C)

• block for pulse-width modulation (PWM).

Optionally, the control system can be supplemented with other blocks that implement various kinds of protections and restrictions, however, in this application, these issues are not considered, since they relate to the nuances of the synthesis of the control system.

The input signals of the specified currents Idr, Iqr are control actions that are set by the user in the identification program, as described above.

By means of the three-phase current signal IABC received from the DFT, the coordinate converter A,B,C-d,q, as well as the signal of the angle of rotation 0 of the coordinate system d-q, feedback signals of the currents Id, Iq are formed. The error signals of processing are fed to the regulators RPT, RMT, which, in turn, form the signals of the specified voltages Ud, Uq. These signals are converted into signals UA, UB, UC of a three-phase coordinate system, and are fed to the PWM generation unit to control the SPI keys. The BOP block generates a signal of the angle of rotation of the rotor flux linkage vector (d-q coordinate system).

Description of the identification algorithm.

The scheme of the parameter identification algorithm 100 is shown in Fig.3.

Block 101. Entry point to the identification algorithm.

Block 102. Initial data is entered:

• Flux-forming current range and step (increment), which corresponds to the parameters I _drmin , I _drmax , ΔI _dr ;

• Six values of torque-generating current at which measurements will be taken to identify the rotor time constant and one value for calculating the mutual inductance, which corresponds to the parameters I _qrT1 ... I _qrT6 and I _qrL ;

• The initial value of the time constant coefficient and its increment, which corresponds to the values k _T,init , Δk _T ;

• Confidence interval for estimating the time constant of the rotor, corresponding to the value of Δk _MI,tol ;

• Number of pole pairs AM and rotor leakage inductance.

Block 103. After all the variables have been initialized in the program and all preliminary procedures have been completed, it is ready directly for the beginning of the identification procedure. Waiting for the start button to be pressed.

Block 104. The next step is the initialization of variables, i.e. assignment to variables k _T , Δk _T .

Block 105. The beginning of the "big" cycle. The variable I _dr is assigned the initial (minimum) value I _dr,min .

Block 106. Since the field of the machine grows for some time (the transient process lasts [3...5]T _r ), the program counts this time and the transition to the next block is performed.

Block 107. Beginning of the inner loop. The n parameter is set to 1.

Block 108. The variable I _qr is assigned one of the user-specified values corresponding to the ordinal number n.

Block 109. Waiting for some time t _tp until the process is steady. After that, the torque on the shaft is measured by means of a sensor several times during the time t _meas , filtered if necessary, then averaged. The result is written to the variable M _T , _av , _n .

Block 110. The condition “n !=1” is checked, after which the transition to block 111 occurs along the “-” branch, and to block 112 along the “+” branch.

Block 111. In the case of being at the first point (first measurement), the parameters k _I,n , k _M,n , k _{MI,n are} assigned zero values.

Block 112. If n=2...6, parameters k _I , _n , k _M,n , k _MI,n are calculated by formulas (6), (11) (block 112).

Block 113. After six measurements of the moment, the internal loop is exited, the value of k _MI,av is calculated (block 13).

Block 114. The hit of k _MI,av in the confidence interval is checked, i.e. the condition |k _MI,av - 100|>Δk _{MI,tol is checked}

Block 115. If the condition of the previous step is met (branch "+"), the deviation in the positive or negative side of the value of k _MI,av relative to 100% is checked.

Block 116. If k _MI,av deviates in the positive direction, the value of k _T receives a positive increment Δk _T , _m , then the algorithm returns to block 107.

Block 117. Otherwise, the value of k _T receives a negative increment - Δk _T , _m , then the algorithm returns to block 107.

Block 118. If k _MI,av falls into the DI, the algorithm proceeds to the calculation of L _m . For this, the variable I _qr is assigned the value I _qrL .

Block 119. Waiting for some time t _tp until the process is steady. After that, the moment is measured several times during the time t _meas , filtered if necessary, then averaged. The result is written to the variable M _L,av,n .

Block 120. The calculation of the equivalent inductance according to the formula (8) and the mutual inductance according to the formula (10) is performed

Block 121. Parameters I _dr , k _T , L _m are entered into the database in the form of a table, as shown below (Table 1):

Block 122. As soon as the identification ends at the last point I _dr =I _drmax , the variables I _qr and I _dr are set to zero.

Block 123. End of identification procedure.

Claims (7)

A method for identifying the time constant T _r of the rotor of an induction motor in a vector field-oriented control system, which consists in performing the following steps: a) set the initial value of the time constant T _r of the rotor in the block for evaluating the rotor flux linkage and the initial value of the current I _d , then at least two times different values of the current I _q are set, at which the torque on the shaft of the asynchronous motor is measured; b) fix at a given moment of time the values of the angular frequency of rotation of the rotor, the magnetic flux and the load on the rotor, and taking into account these parameters, fix the change in the torque-generating current component I _q , which will correspond to the new value Iq ₁ , as well as the value of the electromagnetic torque M, which will be correspond to the new value M ₁ , at the same time it is established that the parameters Iq ₁ and M ₁ correspond to the parameters of the operation of the asynchronous motor at the first operating point, then when the load on the shaft of the asynchronous motor changes, the change in I _q is also fixed, which will correspond to the new value Iq ₂ , and also the value of the electromagnetic moment M, which will correspond to the new value of M ₂ , taking into account these parameters, the transition of the operation of the induction motor to a new operating point is fixed; c) then start the process of calculating the ratio of currents K _I and the ratio of moments K _M at different operating points according to the equations, where K _I = Iq ₂ /Iq ₁ and K _M = M ₂ / M ₁ ; d) further check the fulfillment of the condition under which the ratio of currents must be equal to the ratio of moments, namely K _I = K _M ; e) if the condition in step d) is not met, the selected value of the time constant T _r is adjusted by increasing or decreasing it and repeating steps a)-d) until the condition K _I = K _M is met; f) carry out control of the asynchronous motor at such a value of the time constant T _r of the rotor, which led to the fulfillment of the condition K _I = K _M .

Images