WO2015062834A1

WO2015062834A1 - Method for operating a multiphase electric motor

Info

Publication number: WO2015062834A1
Application number: PCT/EP2014/071653
Authority: WO
Inventors: Peter Stauder; Tom Kaufmann; Matthias SCHAAF; Sighard SCHRÄBLER
Original assignee: Continental Teves Ag & Co. Ohg
Priority date: 2013-10-31
Filing date: 2014-10-09
Publication date: 2015-05-07
Also published as: DE102013222166A1; DE102013222166B4

Abstract

A method for operating a multiphase electric motor comprising a stator with at least two, in particular three, phases (a, b, c) and a rotor, wherein the respective phase (a, b, c) is controlled in each case by a high-side power switch and a low-side power switch, which should distribute the occurring power losses to the power switch involved and by this means prevent hot spots. For this purpose, the following steps are executed: a) Determination of the power losses (PvH, PvL) of the power switches, b) Determination of the maximum power loss of the high-side power switch (PvH_max) and the maximum power loss of the low-side power switch (PvL_max), c) Determination of an offset (O) in the power-on time of the phases, which is set identically for all phases (a, b, c), in such a way that the maximum power loss is reduced with respect to the previously occurring maximum power loss.

Description

description

The invention relates to a method for operating a polyphase electric motor having a stator with at least two, in particular three, phases and a rotor, wherein the respective phase is controlled in each case by a high-side power switch and a low-side power switch. It further relates to a corresponding device.

When operating electric motors, in particular permanent-magnet synchronous machines, in linear actuators in motor vehicles, temporary thermal overloads of the motor output stage can occur, which can lead to the fact that the motor output stage is severely impaired in its functonality or completely fails. Is actively built up pressure in the wheel brake cylinders by, for example, in a Druckbe ^¬ riding provision means of a operated in the 'brake-by-wire "mode, the electro-hydraulic braking system has to be switched in case of failure of the Mo ^¬ torendstufe in the hydraulic auxiliary level.

The motor output stage usually includes a high-side circuit breaker and a low-side circuit breaker for the control or current regulation of each phase. The power switches are designed, for example, as MOSFETs. Their control is preferably carried out with the aid of a pulse width modulation

(PWM) method. The power loss within a final stage of a polyphase, in particular three-phase, electric motor is determined mainly by switching and conduction losses of the circuit breaker. The lead losses result from the resistance of the Leis- tion switch, the tracks, the current in the ent ^¬ speaking phase and the duty cycle of the corresponding half-bridge. High power losses and especially high peaks therein, so-called hot spots, lead to a thermal load on the motor output stage and can lead to the above thermal overloads, on the one hand, the functioning of the motor output stage and thus the motor control is disturbed and on the other hand, a final failure of the Motor control can be caused.

The invention is therefore based on the task of distributing the power losses that occur to the circuit breakers involved and thereby avoiding hot spots. Furthermore, a corresponding device should be specified.

With regard to the method, this object is achieved according to the invention in that the following steps are carried out: a) Determining the power losses of the circuit breakers, b) Determining the maximum power loss of

High-side circuit-breakers and the maximum power dissipation of lowside circuit-breakers,

c) determining an offset in the on-time of the phases, which is set equal for all phases, such that the maximum occurring power loss compared to the previously occurring maximum power loss is reduced.

Advantageous embodiments of the invention are the subject of the dependent claims.

The invention is based on the consideration that in order to avoid a thermal overload of a motor tion switches, via which the phases of the motor are controlled by PWM modulation, the individual maximum power losses should be reduced as far as possible, since individual tips load the system thermally more heavily than more uniformly occurring power losses.

As has now been recognized, a reduction of these peaks or peaks is possible by determining an offset in the turn-on times of the individual phases, whereby the power losses between the low-side and high-side circuit breakers can be shifted. By a suitable choice of the offset, so to speak, the total power loss is redistributed, whereby this redistribution also achieves a more even distribution, in which the maximum peaks can be reduced. The total power loss remains largely unchanged in this way, by the above-described avoidance of

Hot spots can slightly reduce overall losses. Advantageously, the offset, which is computationally reflected as a term in the formulas for calculating the individual voltages, is limited by the available distance of the highest phase voltage to the maximum available DC voltage and the lowest phase voltage to 0 V.

For the determination of the offset is advantageously determines a Dif ^¬ ferenz the time determined in step b) the maximum power loss multiplied by a differential factor and added to the previous value of the offset. Multiplying the difference by a difference factor defines the change in the current offset. The smaller the difference, the less the offset is changed, resulting in an iterative Approach may indicate convergence of the process.

The difference factor used here is advantageously in the range of 0.05 and 0.2, in particular 0.1. The value of the difference factor depends strongly on the actual application. The larger the currents or the resistances, the larger the differences in the power losses of the circuit breaker, so the factor should then be chosen smaller. If the operating voltage is higher and the currents and resistances are the same, the factor should be larger.

The larger the difference factor, the faster the offset between two iterations changes. If it is too big, there is a risk of overshooting and lack of convergence. Instead of a constant difference factor, alternatively, a difference factor that changes during the course of the iterations can be used, which becomes smaller and smaller during the course of the iterations, for example.

The determination of the offset is preferably carried out iteratively. That is, the calculation of the voltages, the power losses, the determination of the maximum power losses and the calculation of the offset described above are carried out in a loop-like manner. Through the iterative adaptation or determination of the offset, the quality of peak avoidance or redistribution of power losses can be optimized. This is based on the fact that in each case a difference of the two maximum power losses is formed for the determination of the offset, which indicates the change or modification of the offset from its size and its sign. Multiplying the difference by a difference factor effectively defines the increment by which the offset changes. There This step size may not be chosen too large (otherwise convergence problems arise) experience shows that several steps are necessary to achieve an optimized solution. As an alternative to the iterative procedure, in a further preferred embodiment, an analytical calculation of the offset is carried out, whereby the highest loaded high side circuit breaker and the highest loaded low side circuit breaker are determined and the offset is calculated by equating the power loss of these two highest loaded circuit breakers, so that the two previously highly loaded circuit breaker are charged equally.

To an improvement of the analytical solution, in the case that now a different circuit breaker having a higher power dissipation, process steps carried out at least one further passage of the ge ^¬ called, whereby the result of redistribution can be further improved. Preferably, the method is - in both embodiments - only executed when the speed falls below a predetermined speed threshold, especially if it is very low or zero. Especially in these situations, power peaks occur, while at higher speeds the power losses are distributed more homogeneously.

Alternatively or in combination with this, the method can only be carried out if the temperature exceeds a predetermined temperature threshold, so that it is only used if a thermal overload due to the auftre ^¬ border temperatures directly or indirectly threatens. With respect to the device, the above object is achieved according to the invention with at least one hardware and / or software implemented module for performing a method according to any one of the preceding claims. The method is preferably realized in the form of a computer program which, in particular on a microcontroller, runs in a control unit or a control and regulation unit of the motor or the motor output stage. The advantages of the invention are, in particular, that the individual power losses of the power switch of the motor output stage are redistributed in such a manner by the iterative or analytical determination of an appropriate offsets that Leis ^¬ tung tops, which provide a major contribution to the thermal overload of the motor output stage is reduced, become.

An embodiment of the invention will be described with reference to a

Drawing explained in more detail. In it show in a highly schematic representation:

FIG. 1 is a flow chart of a method for operating the final stage of an electric motor in a first preferred embodiment, FIG. 2 is a flowchart of a method for operating the output stage of an electric motor in a second preferred embodiment,

FIG. 3 voltage profiles of three phases of an electric motor as a function of an electrical rotation angle,

FIG. 4 power losses in six circuit breakers as a function of the angle of rotation without carrying out the method, FIG. 5 power losses in six circuit breakers as a function of the angle of rotation when carrying out the method in a preferred embodiment, FIG. FIG. 6 shows the power losses when carrying out the method in the first half of the angular range. FIG.

FIG. 7 shows the voltage waveforms of the three phases at ^¬ By carrying out the method in the first half of the angular range,

FIG. 8 of the voltages according to FIG. 7 corresponding offset,

FIG. 9 the offset in the case of a higher line-line output voltage, and

FIG. 10 to FIG. 9 corresponding phase voltages.

Identical parts are provided with the same reference numerals in all figures.

An in FIG. 1 method shown in the form of a flow chart is used to redistribute the power losses of high-side and low-side circuit breakers of an output stage of an electric motor with three phases a, b, c. The power switches, which are designed as MOSFETs are, by

Pulse width modulation operated. The method is advantageously implemented by software in an engine control unit or a control and regulation unit.

In a block 2, sizes required for the method are initialized. In this case, an iteration number it _max , a voltage advance angle δ, a current advance angle γ, a Peak value of the voltage U _peak and a _peak value of the current I _pe a _k set. In the present embodiment, the following applies:

it _max = 20, δ = 20 °, γ = 20 °, U _p ea _k = 6 V, I _peak = 100 A. The DC voltage or supply voltage U _D c is 12 V, that the motor vehicles in the battery corresponds to the electrical system übli ^¬ cherweise present value. An electrical resistance R of the power switches is assumed to be equal for all six Leis ^¬ tung switch and is 0.003 ohms. The peak value of the voltage U _peak is bounded above by the value U _D c / ^ 3.

After initialization, the method performs a loop 4 through angular values of 0 ° to 360 ° of electrical angle Flip Cellphone ^¬ hung with a step size of 0.1 °. In a block 8, an offset O, which is determined iteratively in the following, is first set to the value 0. The offset O is a voltage offset, which is added later in the method to all voltages of the phases. In block 8, furthermore, the voltages U _a , U, U _c and the currents I _a , I, I _{c of} the phases a, b and c are calculated by

U _x = U _peak * sin ((+ δ + ε _χ ) / 180 * n) + U _DC / 2,

I x = Ipeak * sin ((α + γ + ε _χ ) / 180 * π), where x = a, bc and s _a = 0 °, s _b = 120 ° and s _c = 240 °.

Either currents I _a , I _b and I _{c are used} from the result of a calculation by means of a motor model or the actual currents measured / determined in the last control cycle.

Due to the application at preferably low engine speeds, the changes are negligible. In a block 14, the zero vectors are now evenly distributed. For this purpose, a distance downwards, a _{u is} calculated as the minimum of the voltages U _a , U _b , and U _c . A distance a ₀ above is calculated as the difference of the DC voltage U _D c and the maximum of the voltages U _a , U, U _c . A distance in the middle of these two distances a _m is calculated according to a _m = (a ₀ -a _u ) / 2. A possible or limiting distance a _g , which serves as a limit for the value range of the offset 0, then results in

In a loop 20, the offset 0 is now determined iteratively, wherein a number of loop iterations or iterations it _{max is} performed. Alternatively or in combination with this, a termination criterion can also be selected which is met, for example, if the absolute change and / or the relative change in the offset between two subsequent iterations falls below a predefined threshold value.

In a block 26, the phase voltages U _a , U _b and U _{c are} calculated according to

U _x = U _peak * sin ((+ δ + ε _χ ) / 180 * n) + U _DC / 2 + a _m + 0, again with x = a, bc and s _a = 0 °, Sb = 120 ° and s _c = 240 °. In the first iteration, the offset 0 is still 0, so that in the first iteration the voltages U _a , U _b , U _c are calculated as in block 8, wherein in each case the mean distance a _{m is} added to these voltages. In the following iterations, if the offset 0 assumes values other than 0, then the values of the voltages U _a , U _b , U _c additionally change thereby.

In a block 32 then for each phase a, b, c, the power losses of each high-side circuit breaker ₁

PvH and the respective low-side circuit breaker PvL calculated. They arise too

PvH _x = I _x ² * R * U _X / U _DC ,

PvL _x = I _x ² * R * (1-U _X / UDC) ·

An H designates the highside, an L the lowside component, and x = a, b, c.

In a block 38, the maximum high-side power loss PvH _max and the maximum low-power power loss PvL _{max are} determined (in each case from the PvH _x or PvL _x ). In addition, a difference D of these two maximum power losses is formed according to D = PvH _max - PvL _max .

In a block 44 the offset is 0 then assigned a new value which is given by the previous value of the offset 0, minus the product of the difference D by a Dif ^¬ ferenz factor f _d, which in the present case has a value of 0.1, ie O _n is assigned O _n -i - D * f _d , where n is the current iteration and n-1 is the previous iteration.

The offset 0 is now limited in magnitude to the value of the possible distance a _g , ie 0> a _g 0 is set to a _g ; if 0 <-a _g , 0 is set to -a _g . Both checks are necessary because offset 0 can take both positive and negative values.

In the next following iteration, the voltages are again calculated in block 26, which differ from those of the previous iteration due to the changed value of 0. Due to the changed tensions, new values for the power losses are again arising. The procedure changes iteratively, the value of the offset 0 such that the maximum power losses shift so that their difference D is reduced in the course of the iterations. As a result, peaks in the power losses, the so-called hot spots are avoided, so that the risk of thermal overload of the motor output stage or the circuit breaker installed therein is significantly reduced.

A method in a second preferred embodiment is shown in FIG. 2 shown. In this method, an analytical solution for the offset is sought. The method steps in blocks 2, 8 and 14, 26, 32 correspond to those in connection with FIG. 1 described method, with the difference that in block 2 no maximum iteration number must be determined. Here too, a loop is again performed on the 4 angle by ^¬, a loop 20, the offset 0 is determined iteratively in which, however, omitted here.

In a block 50, in each case an index i _{H of} the maximum highside power loss (i _H = a, b, or c) and an index i _{L of} the maximum highside power dissipation (i _L = a, b or c) are determined. With the help of these indices i _H , 1 L, a value for the offset 0 is now determined. The following cases are distinguished:

1. Both indices i _H , 1L are equal, d. h the maximum

Highside power loss and maximum lowside power dissipation belong to the same phase. In this case, the offset 0 is assigned the value 0.5 - U _x / U _D c, where x = i _H = i _L.

2. The indices i _H and i _L are different, ie the

maximum highside power dissipation and the maximum

Lowside power losses correspond to different phases. With i = i _H and j = i _L , the value 0 = (Ij ² * (1-Uj / U _DC ) -I, ² * LVUDci / U ² + I _D ² ) is then assigned to the offset 0. In both cases, the offset 0 is still multiplied by the DC voltage U _D c in order to arrive at the new value.

In a block 56, the offset 0 is now, as described above, limited in amount to the possible distance a _g . Now in a block 62 again, as discussed above in connection with the blocks 26 and 32, again the voltages U _a , b, c _/ the Ver ^¬ loss performance PvH _a , _b , _c and PvL _a , _b , _c and the maximum

Highside power loss and the maximum

Lowside power dissipation calculated.

In a decision 68 it is checked whether the amount of the difference between maximum high-side power loss and maximum low-power power loss is greater than the mean value thereof multiplied by a factor that has been selected to be 0.01 in the present case. If this is not the case, the method is terminated in block 70, since it has succeeded in avoiding power peaks. Otherwise, the method ^¬ steps are performed again from block 50. Although the previously known maximum power losses were reduced by the calculation of the offset - a power loss of another circuit breaker has increased by the choice of the offset but now so that this peak is now to be reduced.

If necessary, this repetition of the method steps can also be repeated until the power loss peaks have been sufficiently reduced. The following figures are taken from a simulation calculation to clarify the method discussed above. In FIG. 3 is plotted on an x-axis 90 an angle or phase angle, while on a y-axis 92, a voltage in the unit volts is applied. Three curves 96, 98, 100 illustrate the course of the voltage of the three phases a, b, c over an electrical revolution with uniform distribution of

Zero vectors, as provided according to the invention at the beginning of the process. In this case, the voltage reserve, ie the voltage range which is available for variation to higher (up to the maximum voltage U _D c of 12 V) and lower voltages (down to 0 V), is the same. In other words, the difference of the lowest voltages in the phases to 0 V is equal to the difference of the highest voltage U _D c to the occurring voltage maxima.

In FIG. 6 are plotted by the curves 104, 106, 108, 110, 112, 114, the corresponding power losses in the three highside and the three low-side circuit breakers on the y-axis 92 in the unit watts (W). The maxima of all loss lines are the same.

In FIG. 5, curves 11, 122, 124, 126, 128, and 130 now show the power losses that result when using the method with the analytical solution for the offset O. It was up to the angle of 180 ° carried out a run to Be ^¬ humor from O, in the range of 180 ° to 360 °, two runs were performed for the determination of O. It is clear that was rounded as compared to only one pass of the profile of the power losses in their tip portions, while the absolute maxima remained approximately two loading ^¬ rich, that is 0 ° to 180 ° and 180 ° to 360 °, the same. Compared to FIG. 4 it becomes clear that, due to the redistribution of the power losses, their maxima are lower, which reduces the thermal load on the motor output stage. When using the analytical method, first the circuit breaker without a Intervention to generate the maximum power losses in low-side and high-side, searched. The distribution of the power loss of a phase on high and low side is determined by the voltage, the total amount of losses from the current. Therefore, at high current and low voltage (relative to U _b at / 2) a small

Power loss can be calculated, which increases greatly when using an offset, because the current is large in this phase.

As a result, it may happen that, after the first pass has been carried out, the power loss of a switching element has increased in such a way that it now represents the dominant power loss. If this improved, but not yet optimal power loss distribution is not accepted, the error can be corrected by re-running.

In the in FIG. In the simulation shown in FIG. 6, the method is carried out at the electrical rotation angles of 0-180 ° and is then inactive, whereby the reduction of the power loss maxima is clearly visible.

In FIG. 7, with the curves 96, 98, 100 again as in FIG. 3 shows the voltage curves in the three phases a, b, c, this time in the angular range 0 ° to 180 °, the method described above is operated and then inactive. For this reason, in the angular range 180 ° to 360 °, a voltage curve results again, which corresponds to that of FIG. 3 corresponds. In the angular range 0 ° to 180 °, the reduction of the maximum power losses and the associated determination of the offset 0 has led to the voltage curves covering a larger range of values, in other words: the distance between voltage maxima and voltage minima in the three phases has increased. In FIG. 8 is plotted in the curve 138 of the corresponding offset 0, which is used for all phase voltages and corresponds to an offset in the on-time. Through a curve 140, an upper limit or voltage reserve that is available to the method is indicated for the offset 0, through a curve 142 a corresponding lower limit or lower voltage reserve. The value for the offset 0 must move for every angle in this value range. As shown in FIG. 8, this value range has not been fully utilized in the present cycle. From the angle 180 °, the offset 0 0, since the method was not applied here.

In FIG. 9 are again the offset 0 through the curve 138 and again upper and lower limits for the offset 0 by the curves 138 and 142 in the case of a higher line-line output voltage. The Aussteuergrad here is higher overall, therefore, the selected offset is almost always in one of the upper or lower limits. In between, there are only very short transitions, during which the offset changes quickly from lower to upper limit. Therefore, in FIG. 10 ranges in which the output voltage of the phases reaches the upper or lower limit. As explained above, this case (high modulation) seldom coincides with the thermally significant case (nearly standstill) .The corresponding voltage curves are shown in FIG. , _r

1 b

LIST OF REFERENCE NUMBERS

2 block

4 loop

8 block

14 block

20 loop

20 block

26 block

32 block

38 block

44 block

50 block

56 block

62 block

68 decision

70 block

90 x-axis

92 y-axis

96 curve

98 curve

100 curve

104, 106,

108, 110,

112, 114 curve

120, 122,

124, 126,

128, 130 curve

138, 140,

142 curve

Claims

Method for operating a polyphase electric motor with a stator having at least two, in particular three, phases (a, b, c) and a rotor, wherein the respective phase (a, b, c) in each case by a high-side circuit breaker and a low-side circuit breaker characterized by the execution of the following steps: a) determining the power losses (PvH, PvL) of the circuit breakers,

b) determining the maximum power dissipation of

High-side circuit-breaker (PvH _max ) and the maximum power dissipation of low-side power switches (PvL _max ), c) determining an offset (0) in the on-time of the phases, which is set equal for all phases (a, b, c), such that the maximum occurring power loss (PvH _max , PvL _max ) compared to the previously occurring maximum power loss (PvH _max , PvL _max ) is reduced.

The method of claim 1, wherein the offset (0) is limited by the available distance (a _m ) of the highest phase voltage to the maximum available DC voltage (U _D c) and the lowest phase voltage to 0 V.

Method according to Claim 1 or 2, wherein a difference (D) of the maximum power losses determined in step b) is determined, multiplied by a difference factor (f _d ) and added to the previous value of the offset (0).

The method of claim 3, wherein the difference factor (f _d ) in the range 0.05 and 0.2, in particular at 0.1, is located. Method according to one of claims 1 to 4, wherein the determination of the offset (0) is iterative.

The method of claim 1 or 2, wherein the highest loaded highside circuit breaker and the highest loaded

Lowside circuit breakers are determined and the offset (0) is calculated by equalizing the power loss of these two most heavily loaded circuit breakers, so that the two previously highly loaded circuit breakers are charged equally.

A method according to claim 6, wherein in the case that now another circuit breaker has a higher power loss, at least a further pass of the procedural ^¬ rensschritte according to claim 6 takes place.

Method according to one of claims 1 to 7, which is executed only when the speed falls below a predetermined speed threshold.

Method according to one of claims 1 to 8, which is executed only when the temperature exceeds a predetermined temperature threshold.

Device for operating a polyphase electric motor with a stator having at least two, in particular three, phases (a, b, c) and a stator, wherein the respective phase (a, b, c) in each case by a high-side circuit breaker and a low-side circuit breaker is driven, with at least one hardware and / or software rea ^¬ Lisiert module for performing a method according to any one of the preceding claims.