Method and device for determining a rotor temperature, computer program, and computer program product

Abstract

The invention relates to a method for determining a rotor temperature of an electric motor, in particular an asynchronous motor, the rotor temperature being determined at least in accordance with reactive powers and/or in accordance with losses of the electric motor. A first rotor temperature is determined in accordance with the reactive powers and, depending on an operating range of the electric motor, plausibility-checked or replaced by a second rotor temperature, which is determined in accordance with the losses.

Claims

1. A computer-implemented method for determining a rotor temperature of an electric machine, the method comprising: exciting the electric machine with an alternating current; determining, with an electronic processor and a current sensor, a first rotor temperature as a function of reactive powers, a first reactive power calculated based on stator phase voltage and stator current measured by a current sensor and a second reactive power calculated based on a flux model, having motor parameters, stator frequency, and current as input variables, determining, with the electronic processor, a second rotor temperature as a function of losses of the electrical machine, the losses being losses in power, depending on an operating range of the electric machine, checking, with the electronic processor, a plausibility of the first rotor temperature by ascertaining a temperature change as a function of the losses and using the temperature change as a limit value in the determination of the first rotor temperature, and controlling, via the electronic processor, torque of the electric machine based on at least one selected from the consisting group of the first temperature and the second temperature.

2. The method as claimed in claim 1, wherein when the operating range of the electric machine is a part-load operating range the first rotor temperature is replaced by the second rotor temperature.

3. The method as claimed in claim 1, wherein system-dependent errors are ascertained and stored in a characteristic map as a function of at least one reactive power for the determination of the first rotor temperature.

4. The method as claimed in claim 1, wherein the second rotor temperature is plausibility-checked by the first rotor temperature.

5. The method as claimed in claim 3, wherein the system-dependent errors are measured before the start-up of the electric machine.

6. The method as claimed in claim 1, wherein the first rotor temperature is replaced by the second rotor temperature when the electric machine is stopped, wherein the electric machine is excited with the alternating current in order to determine an electric resistance of the electric machine, and the second rotor temperature is determined as a function of the electric resistance.

7. A device for operating an electric machine, the device configured to: excite the electric machine with an alternating current, determine, with an electronic processor, a first rotor temperature as a function of reactive powers, a first reactive power calculated based on a stator phase voltage and a stator current measured by a current sensor and a second reactive power calculated based on a flux model, having motor parameters, stator frequency, and current as input variables, determine, with the electronic processor, a second rotor temperature as a function of losses of the electrical machine, the losses being losses in power, depending on an operating range of the electric machine, check, with the electronic processor, a plausibility of the first rotor temperature by ascertaining a temperature change as a function of the losses and using the temperature change as a limit value in the determination of the first rotor temperature, and control, via the electronic processor, torque of the electric machine based on at least one selected from the consisting group of the first temperature and the second temperature.

8. A non-transitory, computer readable medium connected to an electronic processor and comprising instructions that when run with the electronic processor cause the electronic processor to excite an electric machine with an alternating current, determine a first rotor temperature as a function of reactive powers, a first reactive power calculated based on a stator phase voltage and a stator current measured by a current sensor and a second reactive power calculated based on a flux model, having motor parameters, stator frequency, and current as input variables, determine a second rotor temperature as a function of losses of the electrical machine, the losses being losses in power, depending on an operating range of the electric machine, check a plausibility of the first rotor temperature by ascertaining a temperature change as a function of the losses and using the temperature change as a limit value in the determination of the first rotor temperature, and control, via the electronic processor, torque of the electric machine based on at least one selected from the consisting group of the first temperature and the second temperature.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention is explained in greater detail in the following with reference to an exemplary embodiment. Wherein

(2) FIG. 1 shows a method for determining the rotor temperature of an electric machine in different operating ranges,

(3) FIG. 2 shows the ascertainment of a first rotor temperature as a function of reactive powers,

(4) FIG. 3 shows an adaptation of the ascertainment of the first rotor temperature,

(5) FIG. 4 shows a method for ascertaining a second rotor temperature as a function of losses, and

(6) FIG. 5 shows a method for determining the second rotor temperature at rest.

DETAILED DESCRIPTION

(7) FIG. 1 shows an advantageous method for operating an electric asynchronous machine, by means of which a rotor temperature of the asynchronous machine is intended to be determined. The method includes three method parts, i.e., I, II and III, of which at least the method parts I and II are carried out simultaneously. The three method parts I, II and III are controlled and evaluated by a control unit S, in order to determine a rotor temperature {circumflex over ()}.sub.r,F.sub.u.

(8) The first method part I is an electromagnetic model for determining a first rotor temperature {circumflex over ()}.sub.r,Fl. This method is based on a reconciliation between the reactive power calculated on the basis of phase voltage and current and a reactive power calculated on the basis of a flux model, having motor parameters, stator frequency, and current as the input variables. Since the phase voltage frequently is not directly measurable, the voltage setpoint value is preferably used by a current controller of the electric machine as the actual voltage value. The difference between the setpoint voltage value and the actual voltage value is considered to be the voltage error. In addition to the voltage error, system-dependent errors such as motor parameter errors or current sensor errors must be taken into account. The overall error that results is dependent, in turn, on operating points or operating ranges of the asynchronous machine and results in corresponding erroneous values in the determination of the rotor temperature. The determination of the rotor temperature can be inaccurate only as a result of method I, in particular when the phase voltage, for example in the part-load range of the electric machine or in the range in which the rotational speed is very low.

(9) The method part II relates to a thermal model for determining a second rotor temperature {circumflex over ()}.sub.r,Th. The thermal model requires losses or loss values as input variables. The accuracy of the loss calculation has substantial influence on the rotor temperature ascertained therefrom. An accurate analytical calculation or experimental ascertainment of losses of the asynchronous machine is very time-consuming. During operation, an error in the ascertainment of the rotor temperature is cumulative and continuously increases.

(10) The method in FIG. 1 proposes a skillful combination of the two method parts I and II, whereby the rotor temperature {circumflex over ()}.sub.r,Fu can be determined very accurately.

(11) System-dependent errors in the method part I are initially ascertained using a prototype machine of the electric machine, which is equipped with telemetry, and are stored as a characteristic map. For further production machines of the electric machine having an identical electromagnetic design or telemetry, the first rotor temperature {circumflex over ()}.sub.r,Fl can be ascertained by means of an adaptation (Adapt.) and by the reconciliation of the reactive powers.

(12) The thermal model (method part II) runs parallel to the method part I and simultaneously yields an estimated temperature change as the limitation for the adaption (Adapt.) in method part I (Limit). In the operating range in which the estimated value according to method part I is reliable, the ascertained first rotor temperature is plausibility-checked (SetPlausy) by the rotor temperature ascertained in the second method part II, and so a cumulative error remains minimized.

(13) The third method part III is provided for an operating state of the electric machine, in which the rotor of the electric machine is stopped and a load is not applied. In this operating state, the method parts I and II are not functional, since the required input variables are missing. In method part III, it is therefore provided that the electric machine is acted upon, in the manner of a transformer, with an excited current, which generates losses in the electric machine. The rotor temperature {circumflex over ()}.sub.r(Stimult) estimated as a result is then used by method step II (SetInit) as the starting value for the thermal model.

(14) The control unit S decides which rotor temperature from which estimating method is sufficiently accurate or suitable at the particular current operating point or operating range and also carries out the aforementioned plausibility check of the rotor temperatures. As a result of the fusion of information from the different method parts I, II and III in the control unit S and due to their coordination and plausibility checking, a reliable estimation of the rotor temperature of the asynchronous machine can be obtained, in particular for electric vehicles. The first method part I or the flux model based on reactive powers can be explained as follows. For the case in which u.sub.s>u.sub.s,min, the following voltage equations and calculations of the reactive power are specified:

(15) $\begin{array}{cc}{u}_{\mathrm{sd}}=R_s{i}_{\mathrm{sd}}+\frac{d}{dt}\left(L_s{i}_{\mathrm{sd}}\right)-{}_s*L_s{i}_{\mathrm{sq}}+\frac{L_m}{L_r}*\frac{d}{\mathrm{dt}}\left({}_{rd}\right)& \left(\mathrm{Equation}1\right)\\ u_{\mathrm{sq}}=R_s{i}_{\mathrm{sq}}+\frac{d}{\mathrm{dt}}\left(L_s{i}_{\mathrm{sq}}\right)+{}_s*L_s{i}_{\mathrm{sd}}+{}_s*\frac{L_m}{L_r}{}_{rd}& \left(\mathrm{Equation}2\right)\\ \begin{array}{c}{u}_{\mathrm{sq}}{i}_{\mathrm{sd}}-u_{\mathrm{sd}}{i}_{\mathrm{sq}}{L}_s\left[i_{\mathrm{sd}}*\frac{{\mathrm{di}}_{\mathrm{sq}}}{\mathrm{dt}}-i_{\mathrm{sq}}*\frac{{\mathrm{di}}_{\mathrm{sd}}}{\mathrm{dt}}\right]+\\ {}_s\left[L_s\left(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2\right)+\frac{L_m}{L_r}{}_{rd}{i}_{\mathrm{sd}}\right]+\\ \frac{L_m}{L_r}\frac{d}{\mathrm{dt}}\left({}_{rd}\right)i_{\mathrm{sq}}\\ {}_s\left[L_s\left(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2\right)+\frac{L_m}{L_r}{}_{rd}{i}_{\mathrm{sd}}\right]+\\ \frac{L_m}{L_r}\frac{d}{\mathrm{dt}}\left({}_{rd}\right)i_{\mathrm{sq}}\end{array}& \left(\mathrm{Equation}3\right)\\ u_{\mathrm{sq}}{i}_{\mathrm{sd}}-u_{\mathrm{sd}}{i}_{\mathrm{sq}}-{}_s\left[L_s\left(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2\right)+\frac{L_m}{L_r}{}_{rd}{i}_{\mathrm{sd}}\right]+\frac{L_m}{L_r}\frac{d}{\mathrm{dt}}\left({}_{rd}\right)i_{\mathrm{sq}}\underset{R,\mathrm{correct}}{.fwdarw.}0& \left(\mathrm{Equation}4\right)\\ \mathrm{define}\text{:}d{Q}_{\mathrm{norm}}=\frac{u_{\mathrm{sq}}{i}_{\mathrm{sd}}-u_{\mathrm{sd}}{i}_{\mathrm{sq}}-\frac{L_m}{L_r}\frac{d}{\mathrm{dt}}\left({}_{rd}\right)i_{\mathrm{sq}}}{{}_s}-\left[L_s\left(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2\right)+\frac{L_m}{L_r}{}_{rd}{i}_{\mathrm{sd}}\right]\underset{R,\mathrm{correct}}{.fwdarw.}0& \\ d{Q}_{\mathrm{norm}}\left(\mathrm{steady}\right)=\frac{u_{\mathrm{sq}}{i}_{\mathrm{sd}}-u_{\mathrm{sd}}{i}_{\mathrm{sq}}}{{}_s}-\left[L_s\left(i_{\mathrm{sd}}^2+i_{\mathrm{sq}}^2\right)+\left(1-\right)L_s{i}_{\mathrm{sd}}^2\right]=\frac{u_{\mathrm{sq}}{i}_{\mathrm{sd}}-u_{\mathrm{sd}}{i}_{\mathrm{sq}}}{{}_s}-\left[\left(i_{\mathrm{sd}}^2+*i_{\mathrm{sq}}^2\right)\right]*L_s\underset{R,\mathrm{correct}}{.fwdarw.}0& \left(\mathrm{Equation}5\right)\end{array}$

(16) The left side of Equation 3 relates to the voltage and the current level. The right side of Equation 3 relates to the current, flux, and motor parameters. It is assumed that the rotor resistance is the only parameter that is dependent on the rotor temperature. If the information regarding the rotor temperature is false, for example due to false flux coordinates, the dq components in Equation 4 and in reality no longer match. Equation 4 is then not equal to zero. In this case, the estimated rotor resistance or the first rotor temperature must be adapted until Equation 4 is correct again. For this purpose, in particular the previously determined, system-dependent errors, in particular in the voltage, the current, and the motor parameters, are taken into account and are stored as characteristic map dQ.sub.sy.sub._.sub.serr0. For further production machines having the same electromagnetic design without telemetry, the first rotor temperature {circumflex over ()}.sub.r,Fl is ascertained by means of the adaption (Adapt.) of

(17) $d{Q}_{\mathrm{norm}}\underset{R,\mathrm{correct}}{.fwdarw.}d{Q}_{\mathrm{sys\_err}}0.$
The plausible temperature change is delivered by a thermal model of the second method part II and is used as the limit value for the adaption (Adapt.). The operating principle is shown in FIGS. 2 and 3.

(18) The following symbols and abbreviations apply:

(19) =flux linking

(20) .sub.rd=flux linking on the d-axis (rotor flux coordinate)

(21) L.sub.m=main inductance

(22) L.sub.s=stator leakage inductance

(23) .sub.s=stator dispersion coefficient:

(24) ${}_s=\frac{L_s}{L_m}$

(25) L.sub.x=leakage inductance, x=s: stator, x=r: rotor

(26) .sub.r=rotor dispersion coefficient:

(27) ${}_r=\frac{L_r}{L_m}$

(28) $=1-\frac{1}{\left(1+{}_s\right)\left(1+{}_r\right)}=\mathrm{total}\mathrm{dispersion}\mathrm{coefficient}$

(29) .sub.mr, {circumflex over ()}.sub.mr=rotor flux angle, estimated rotor flux angle

(30) .sub.x, {circumflex over ()}.sub.x, .sub.x=temperature, estimated temperature, temperature difference x=r: rotor, x=s: stator

(31) .sub.S.sub.NTC=measured stator temperature (NTC sensor)

(32) dQ.sub.norm=normalized reactive power difference

(33) dQ.sub.sys.sub._.sub.err=system-dependent error in the calculation of reactive power difference

(34) k.sub.P, k.sub.I=P component and I component gain of a PI controller (adaptation law)

(35) .Math.=voltage amplitude (output current controller)

(36) =current amplitude (current sensor)

(37) =phase angle between voltage and current phasor

(38) R.sub.x, {circumflex over (R)}.sub.x=resistance, estimated resistance, x=r: rotor, x=s: stator

(39) P.sub.loss.sub.x=power loss, x=r: rotor, x=s: stator

(40) Z.sub.Th, Z.sub.ij=thermal impedance

(41) n=rotational speed

(42) us=stator voltage

(43) is=Is=stator current

(44) d=d-axis

(45) q=q-axis

(46) Isq-R=stator current q-axis real

(47) Isq-M=stator current q-axis model

(48) Isd-R=stator current d-axis real

(49) Isd-M=stator current d-axis model

(50) The rotor temperature {circumflex over ()}.sub.r,Th is then described by the following relationship:

(51) $\begin{array}{cc}{}_r=\left(\frac{R_r}{R_{r20C.}}-1\right)*\frac{1}{{}_{r20C.}}+20C.& \left(\mathrm{Equation}6\right)\end{array}$

(52) Wherein R.sub.r,20 C. is the rotor resistance at 20 C. and .sub.r,20 C. is the temperature coefficient of resistance of rotor conductors.

(53) The loss-based thermal model according to method part II applies in this case for:

(54) 0<u.sub.s<u.sub.s,min:

(55) In an operation in which little voltage is requested, for example at low rotational speeds in a part-load operating range, the voltage error is relatively great, since only setpoint voltage values can be used for the determination method of the first method part I. The thermal model is therefore used for the second method part II, as depicted in FIG. 4 in a simplified manner. Losses of the electric machine can be ascertained in advance either via analytical calculation or with the aid of measurements and can then be stored as a characteristic map as a function of operating points or operating ranges. The transfer functions Z.sub.ii, Z.sub.22, Z.sub.12 and Z.sub.21 are preferably modeled as a PT1 element, or as a higher-order element for greater accuracy. The parameters of the transfer functions can be measured on the basis of the losses of the measured or simulated temperature progressions. It is also advantageous when a cooling temperature .sub.c is taken into account. If the cooling temperature is not available, it can be calculated by integrating the difference between the most recent rotor temperature {circumflex over ()}.sub.s and the measured stator temperature {circumflex over ()}.sub.s,NTC:
{circumflex over ()}.sub.c=(.sub.S.sub.NTC{circumflex over ()}.sub.S)dt(Equation 7)

(56) The ascertained stator and rotor temperatures are then:
{circumflex over ()}.sub.S=.sub.S+{circumflex over ()}.sub.c(Equation 8)
{circumflex over ()}.sub.r=.sub.r+{circumflex over ()}.sub.c(Equation 9)

(57) In the operating range in which the first method part I yields reliable results, the first rotor temperature is used as a relevant rotor temperature in order to correct the estimation in the thermal model of the second method part II, so that the cumulative error is minimized, as is also shown in FIG. 1. The operating principle is shown in FIG. 4.

(58) The first method part III accounts for the operating state in which the electric machine is stopped. The methods according to I and II do not function in this case. In order to permit the rotor temperature to be ascertained nevertheless, the asynchronous machine is operated as a transformer. For this purpose, a sinusoidal current is excited in the d-axis of the electric machine. At a higher frequency, a relationship can be established between the electric variable and resistances of the electric machine.

(59) $\begin{array}{cc}\cos *\frac{\hat{U}}{\hat{I}}{R}_s+{\left(\frac{1}{1+{}_r}\right)}^2*R_r& \left(\mathrm{Equation}10\right)\\ R_r\left(\cos *\frac{\hat{U}}{\hat{I}}-R_s\right)*{\left(1+{}_r\right)}^2& \left(\mathrm{Equation}11\right)\end{array}$

(60) Finally, the rotor temperature {circumflex over ()}.sub.r,St determined by the method part II according to Equation 6.

(61) FIG. 5 shows a simplified representation of the third method part III. The following abbreviations also apply here:

(62) : current amplitude (current sensor)

(63) .Math.: voltage amplitude (output current controller)

(64) : phase angle between voltage and current phasor

(65) R.sub.s: stator resistance, temperature dependence can be compensated by the measured stator temperature (NTC sensor in the winding) or can be ascertained on-line in the inverter using the following equation:

(66) $R_s\frac{\hat{U}}{\hat{I}}.$

(67) The control unit now selects, as a function of the current operating range of the electric machine, the correct method part I, II or III for determining the rotor temperature {circumflex over ()}.sub.r,Fu for the plausibility checks and combines the results in order to determine a single, plausibility-checked signal for the rotor temperature, as shown in FIG. 1.