Device for determining the angular position of a rotor of a rotary electric machine

Abstract

The invention relates to a method for reducing the bias in a periodic signal having a theoretically zero mean value and provided by a sensor associated with a rotary electric machine, in particular by a sensor for determining the angular position of a rotor of a rotary electric machine, the method comprising: - receiving the signal provided by the sensor; - multiplying this signal by a signal proportional to the rotational speed of the rotor of the rotary electric machine in order to obtain a first signal; - at least one integration over half a period of the first signal in order to obtain a second signal; - applying, to the second signal, a finite impulse response filter, in particular a comb filter, so as to add, to the second signal, a delayed version of this second signal in order to obtain a third signal representative of the bias, in particular equal to the bias of the signal provided by the sensor; and - subtracting the third signal from the signal provided by the sensor, so as to reduce the bias in this signal provided by the sensor.

Description

[0001] Description

[0002] Title of the invention: Device for determining the angular position of a rotating electrical machine rotor

[0003] The present invention relates to a method for reducing bias (or "offset"), optionally combined with normalization, of a periodic signal with a theoretically zero mean value provided by a sensor associated with a rotating electrical machine. This sensor is, in particular, a sensor for determining the angular position of a rotating electrical machine rotor. The present invention also relates to an assembly comprising a rotating electrical machine for the propulsion of a hybrid or electric vehicle in which the aforementioned method is implemented. Such a periodic signal with a theoretically zero mean value, as defined by the present invention, is, for example, in the context above: a sinusoidal signal measuring current, voltage, or position. It may also be a square wave, triangular wave, or other signal whose mean value is theoretically zero.The invention can also, where appropriate, be applied to periodic signals whose average value is not zero when the speed varies.

[0004] An electrical machine is, for example, an alternator or a starter-alternator powered by a nominal voltage of 12V or 48V, or even higher. An electrical machine can also be a propulsion machine powered by a nominal voltage of 12V or 48V, or even higher, for example, a voltage greater than 300V, such as 400V or 800V.

[0005] This electric machine can be integrated into a hybrid or purely electric vehicle, such as a car. More broadly, for the purposes of this application, "vehicle" encompasses any form of mobility powered by pure electric, hybrid, internal combustion, or other means. "Vehicle" thus includes any machine that travels on land with four, three, two, or any other number of wheels, or any machine that travels in the air, on water, or even in space.

[0006] Controlling this electric machine requires knowledge of the rotor's angular position. This angular position is obtained, for example, using a control loop that implements a phase-locked algorithm, processing signals from sensors such as Hall effect sensors, inductive sensors, or resolvers. These sensors can be located at the shaft end if necessary.

[0007] To obtain accurate knowledge of the angular position of the electric machine's rotor, it is useful to address bias and normalize the signals provided by the sensors. In the case of a rotating electric machine providing, at least sequentially, purely electric propulsion for an electric or hybrid vehicle, with:

[0008] - a small number of poles, for example one, two or four pairs of poles, and

[0009] - A low reduction ratio for the gearbox receiving the motor torque makes the need to effectively address bias and normalize effectively even more critical. Indeed:

[0010] - In this case, the vehicle can move at a low speed, on the order of a few km / h for example, and

[0011] - in the event of bias not properly treated in particular, first-order harmonics may appear in the signal representing torque or speed or position, and in the event of insufficient normalization, i.e., processing to bring the amplitude of the fundamental of the signal between -1 and 1, second-order harmonics may appear in the signal representing position.

[0012] Such erroneous processing is sometimes further amplified by biases in the sensor signals varying dynamically, due to the variability of physical electronic components in the electronic acquisition chain, for example sensor(s), operational amplifier(s), analog / digital converter..., or for example due to poor mechanical positioning of the chips of these sensors.

[0013] The effect of temperature and sensor aging can also impact the processing of sensor signal bias and / or their normalization.

[0014] To avoid such problems, it is known to perform a calibration operation at the end of the production line, just before the electric machine is shipped to the customer. However, such an end-of-line calibration operation is complex and time-consuming, relying, for example, on the iterative extraction of Fourier series performed at a constant speed, which then undergoes complex processing. Furthermore, such an end-of-line calibration operation is not possible when the electric machine is supplied by an equipment manufacturer independently of its power and control electronics. Finally, such an end-of-line calibration may be ineffective in some or all of the aforementioned circumstances, notably failing to address cases of dynamically varying bias.

[0015] There is a need to enable a reduction of bias in a periodic signal and theoretically zero average value provided by a sensor associated with a rotating electrical machine, in particular by a sensor for determining the angular position of a rotating electrical machine rotor, especially in the case of a rotating electrical machine providing purely electric propulsion for an electric or hybrid vehicle, with a small number of poles, for example four pole pairs, and with a low reduction ratio for the reducer receiving the motor torque.

[0016] This need may also be accompanied by, or be a variant of, a need to enable the normalization of such a periodic signal with a theoretically zero average value provided by a sensor associated with a rotating electrical machine. "Theoretically zero average value" means that the presence of a bias in the measurement of this signal is considered an error.

[0017] The invention aims to meet this need and achieves this, according to one of its aspects, by means of a method for reducing bias in a periodic signal with a theoretically zero average value provided by a sensor associated with a rotating electrical machine, in particular by a sensor for determining the angular position of a rotating electrical machine rotor, the method comprising:

[0018] - the reception of the signal provided by the sensor,

[0019] - multiplying this signal by a signal proportional to the rotational speed of the rotor of the rotating electrical machine to obtain a first signal,

[0020] - at least one integration over half a period of the first signal to obtain a second signal,

[0021] - applying a finite impulse response filter, for example a comb filter, to the second signal to add a delayed version of that second signal in order to obtain a third signal representative of the bias of the signal provided by the sensor, the third signal being in particular equal to said bias, and

[0022] - the subtraction of the third signal from the signal provided by the sensor, in order to reduce the bias, in particular to eliminate the bias, of this signal provided by the sensor.

[0023] In all of the above, the half-period over which the integration is performed is the half-period of the period of the signal provided by the sensor.

[0024] The invention thus proposes a method for dynamically estimating the bias of a signal provided by a sensor, particularly in relation to the detection of angular periods, for example, the transition at 0° and 180° of a signal representing the angular position of the rotor of a rotating electrical machine, and in relation to the rotor's rotational speed. The signal representing the angular position of the rotor of the rotating electrical machine is, for example, provided by a circuit implementing a control loop for the rotor's angular position, using as input one or more signals whose bias has been reduced or even eliminated. The rotor's rotational speed corresponds, for example, to the derivative of said signal representing the angular position of the rotor of the rotating electrical machine.The solution according to the invention can also allow, as will be seen later, the estimation of the amplitude of the first harmonic of the signal provided by the sensor, without needing additional processor resources.

[0025] The aforementioned method can be used to estimate the bias for a periodic signal with a theoretically zero average value, even for very low values ​​of machine rotation speed, and therefore of the vehicle.

[0026] This method, which can be described as online calibration, is particularly applicable when the electrical machine and the sensor are supplied by an equipment manufacturer independently of its power and control electronics.

[0027] This method also makes it possible to overcome bias drifts due to temperature changes or aging.

[0028] Since the method according to the invention does not implement a peak detection technique or a bias estimator filtering technique involving Fourier transform processing, curve fitting based on pseudoinverse Fourier transforms for example or a learning technique, it is not sensitive to defects occurring at low and unstable frequencies.

[0029] The fact that the method according to the invention is devoid of the aforementioned techniques is also advantageous in terms of processor resource requirements.

[0030] More specifically, the method according to the invention provides the following advantages:

[0031] - accurate estimation of bias with a short response time, since it implements half-period integration, rather than full-period integration,

[0032] - independence of the dynamics of the signal provided by the sensor, such as its frequency and acceleration, given that the signal provided by the sensor is multiplied by a signal proportional to the rotational speed of the rotor of the rotating electrical machine to obtain the first signal,

[0033] - robustness, given the low sensitivity to noise present in the signal provided by the sensor, its harmonics, or variations in the parameters of this signal provided by the sensor.

[0034] - absence of the use of complex trigonometric functions or adaptive frequency filters,

[0035] - absence of digital overshoot, even at very low frequencies

[0036] - low CPU resource usage.

[0037] The sensor(s) can be position sensors, or voltage or current sensors.

[0038] The method may include applying low-pass filtering to the third signal.

[0039] This low-pass filtering can be achieved using a variable cutoff frequency, which depends on the rotational speed of the rotating electrical machine. For example, the cutoff frequency may be inversely proportional to the rotational speed of the electrical machine. Alternatively, the cutoff frequency may vary inversely with the rotational speed of the electrical machine, without any proportional relationship. For instance, the cutoff frequency may have a fixed value at low speeds and a lower fixed value at high speeds.

[0040] A cutoff frequency chosen in this way can compensate for undersampling of points at high speeds by the value of the cutoff frequency for these high speeds.

[0041] If necessary, saturation can be applied to the third signal output of the low-pass filter. Saturation involves, for example, comparing the value of the third signal output of the low-pass filter to a maximum value and a minimum value, and:

[0042] - when the value of the third signal is greater than the maximum value, this third signal takes on the maximum value at the end of this saturation.

[0043] - when the value of the third signal is less than the minimum value, this third signal takes on the minimum value at the end of this saturation.

[0044] - when the value of the third signal is between the minimum and maximum values, this third signal retains its value after this saturation.

[0045] The integration of the first signal over half a period to obtain the value of the second signal can be triggered by a synchronization signal that is a function of the signal representing the angular position of the rotating electric machine's rotor. A change of state in this synchronization signal, for example, when the rotating electric machine's rotor's angular position is detected as passing through 0° and 180°, can cause this integration. This change of state can correspond to a transition from "0" to "1" of the synchronization signal when the rotating electric machine's rotor's angular position is detected as passing through 0°, and a transition from "1" to "0" when the rotating electric machine's rotor's angular position is detected as passing through 180°.Alternatively, each detection of the passage of the angular position of the rotor of the rotating electrical machine through 0° or 180° corresponds to a brief passage from "0" to "1" of the synchronization signal which returns to "0" at the end of the pulse.

[0046] For the purposes of this application, "angular position of the rotating electrical machine" and "angular position of the rotor of the rotating electrical machine" are synonymous.

[0047] The synchronization signal, particularly through its change of state, can simultaneously cause:

[0048] - an initialization of the integration of the first signal over half a period to obtain a value for the second signal, and

[0049] - the application of a finite impulse response filter, such as a comb filter, to the second signal obtained at the end of this integration, to obtain the value of the second signal for the previous iteration, also referred to as the "delayed version of the signal" in this application.

[0050] According to a first example of implementation, the synchronization signal can each time control the initialization of a single integration over half a period of the first signal.

[0051] This first implementation example involves iterating over only half a period. The stored value of the second signal thus corresponds to the value of an integration over just half a period. According to this first implementation example, the delayed value of the second signal corresponds to the value of an integration over the immediately preceding half-period.

[0052] According to a second implementation example, the synchronization signal can trigger the initialization of an even number of integrations each time, each integration being performed over half a period of the first signal, with these half-periods occurring consecutively. The stored value of the second signal is obtained by summing the values ​​of these integrations and dividing the sum by the number of half-periods considered. In this second implementation example, the delayed value of the second signal corresponds to the value of the same even number of integrations, each integration being performed over half a period of the first signal, with these half-periods occurring consecutively. The stored value of the second signal is obtained by summing the values ​​of these integrations and dividing the sum by the number of half-periods considered. This delayed version is shifted half a period into the past.

[0053] Throughout the above, the low-pass filter can receive a previously stored third signal value, each storage being performed upon receipt of a change of state of the synchronization signal.

[0054] In all of the above, the method may still include:

[0055] - subtracting the delayed version of the second signal from the second signal to obtain a fourth signal representative of the amplitude of the first harmonic of the signal provided by the sensor, this fourth signal being notably equal to this amplitude, and

[0056] - the division of the signal provided by the sensor whose bias has been reduced by the fourth signal.

[0057] This results in a normalized signal with reduced bias, improving the accuracy of the signal initially provided by the sensor.

[0058] The signal representing the amplitude of the first harmonic can be a signal in which even-numbered harmonics have been suppressed and odd-numbered harmonics above 1 are attenuated, for example, by their order and possibly by their relative phase. This provides a more accurate estimate of the first harmonic than, for example, using a peak detection technique.

[0059] The method may include applying low-pass filtering to the fourth signal. This low-pass filtering can be performed using a variable cutoff frequency, which is a function of the rotational speed of the rotating electrical machine. For example, the cutoff frequency may be inversely proportional to the rotational speed of the electrical machine. Alternatively, the cutoff frequency may vary inversely with the rotational speed of the electrical machine, without any proportional relationship. For example, the cutoff frequency may have a fixed value at low speeds and a lower fixed value at high speeds.

[0060] A cutoff frequency chosen in this way can compensate for undersampling of points at high speeds by the value of the cutoff frequency for these high speeds.

[0061] The cutoff frequency of the low-pass filter applied to the fourth signal can be identical to that of the low-pass filter applied to the third signal.

[0062] If necessary, saturation as explained previously can be applied to the fourth signal before application of the low-pass filter.

[0063] Throughout the preceding discussion, the signal is, for example, a signal provided by a position sensor, such as a sine or cosine wave. Alternatively, it could be a signal provided by a voltage or current sensor. In the latter case, obtaining the amplitude of the first harmonic of the signal, or a precise representation of this amplitude, via the fourth signal can prove useful for purposes other than normalization, which is then unnecessary. These other purposes include, for example, the control of the electrical machine.

[0064] The invention also relates, according to another aspect, to a method for determining the angular position of a rotating electrical machine rotor based on two sensor signals provided by position sensors, each sensor providing a periodic signal with a theoretically zero average value, a method in which: - the bias of each signal provided by a sensor is reduced using the above method, and

[0065] - on the basis of these two signals whose bias has been reduced the angular position of the rotor, in particular via the realization of a control loop, by supplying a signal representative of the angular position of the rotating electrical machine.

[0066] One of these two signals is, for example, a cosine signal and the other of these two signals is, for example, a sine signal.

[0067] The control loop receives as input, for example, each sensor signal whose bias has been reduced.

[0068] The method may include, for each of these two signals, the aforementioned normalization step.

[0069] The aforementioned method can be implemented for any rotational speed of the rotating electrical machine.

[0070] Alternatively, in all the above, when the method only implements bias reduction and not normalization, the integration of the first signal over half a period to obtain a second signal can be performed for a first speed range, for example, speeds below 600 rpm or 800 rpm. For a second speed range, for example, speeds above 600 rpm or 800 rpm, the integration of the first signal can be performed by period of that signal, or even by groups of periods of that signal. A single threshold, for example, 600 rpm or 800 rpm, separates the first speed range from the second speed range.

[0071] Working in periods, or even groups of periods, for higher speeds can allow for bias filtering.

[0072] The invention also relates, according to another aspect, to an assembly comprising:

[0073] - a rotating electrical machine for the propulsion of a hybrid or electric vehicle, and

[0074] - a control device for this electrical machine, comprising two sensors each providing a periodic signal of theoretically zero average value associated with this rotating electrical machine, to determine the angular position of the rotor of this rotating electrical machine, this control device being adapted to carry out the steps of the method above.

[0075] Everything stated above regarding the method still applies to the whole.

[0076] The invention also relates, according to another of its aspects, to a computer program product comprising instructions which lead the above assembly to execute the steps of the method for determining the angular position of an electric machine rotor above.

[0077] The invention also relates, according to another aspect, to a computer-readable medium on which the above-mentioned computer program product is recorded. The rotating electrical machine is, for example, a synchronous machine, such as a three-phase synchronous machine or a synchronous machine whose stator winding defines a dual three-phase system. The stator winding is, for example, formed by wires or conductive bars connected to one another.

[0078] In all the above, the rotor can be a claw rotor. This rotor then comprises a first and a second interlocking pole wheel, the first pole wheel defining a series of claws of generally trapezoidal shape, each claw extending axially towards the second pole wheel. A permanent magnet can be received between two consecutive claws circumferentially along the rotor.

[0079] Alternatively, the rotor may be something other than a claw rotor, comprising for example a pack of laminations or being a cage rotor.

[0080] Throughout all of the above, the rotor can comprise any number of pole pairs, for example three, four, six, or eight pole pairs.

[0081] Throughout the above, the electrical machine may include a stator cooling circuit through which a fluid such as air or liquid circulates. This liquid may be water or oil.

[0082] The rotor can be cooled by this same cooling circuit or by another cooling circuit in which air, or liquid such as water or oil, circulates.

[0083] The electric machine may have a mechanical rated power between 4 kW and 35 kW, for example 4 kW, 8 kW, 15 kW, 25 kW or 35 kW, or the electric machine may have a mechanical rated power between 40 kW and 400 kW, for example 40 kW, 80 kW, 100 kW, 150 kW, 180 kW, 200 kW, 300 kW or 400 kW.

[0084] This rotating electric machine can be electrically powered from an electrical energy storage unit via an inverter / rectifier of the assembly, this inverter / rectifier allowing, depending on whether the electric machine is operating as a motor or as a generator, to charge an on-board network of the vehicle or to be electrically powered from this network.

[0085] The nominal voltage of the electrical energy storage unit can be 12V, 48V or have another value, for example another value greater than 300V, 400V or 800V.

[0086] The rotating electric machine may also include a pulley or other means of connection to the rest of the vehicle's powertrain. For example, the electric machine is connected, notably via a belt, to the crankshaft of the vehicle's internal combustion engine. Alternatively, the electric machine is connected to other locations within the powertrain, for example, to the input of the gearbox from the perspective of torque transmitted to the vehicle's wheels, to the output of the gearbox from the perspective of torque transmitted to the vehicle's wheels, to the gearbox itself from the perspective of torque transmitted to the vehicle's wheels, or even to the front or rear axle of the powertrain.

[0087] The rotating electrical machine is not necessarily a synchronous machine, it can be an asynchronous machine.

[0088] The invention will be better understood upon reading the following description of a non-limiting example of its implementation and upon examination of the attached drawing in which:

[0089] [Fig. 1] schematically represents, in axial section, an example of a rotating electrical machine to which the invention can be applied.

[0090] [Fig.2] shows in elevation a different type of rotor than that of figure 1

[0091] [Fig.3] schematically represents the device for determining the angular position of the rotor of the machine to which the invention is applied,

[0092] [Fig.4] schematically represents another example of a device for determining the angular position of the rotor of the machine to which the invention is applied,

[0093] [Fig.5] schematically represents an example of a dynamic bias reduction circuit.

[0094] Figure 1 shows a rotating polyphase electrical machine 1, particularly for motor vehicles, to which the invention can be applied.

[0095] This rotating electrical machine can form an alternator or a starter-alternator for the vehicle. This rotating electrical machine can be powered via a power electronic component 9 comprising an inverter / rectifier by a battery with a nominal voltage of 12 V or 48 V or a value greater than 300 V, for example.

[0096] The rotating electrical machine 1 comprises a housing 2. Inside this housing 2, it also includes a shaft 3, a rotor 4 fixed to the shaft 3 for rotation, and a stator 5 surrounding the rotor 4. The rotation of the rotor 4 occurs about an axis X. In this example, the housing 2 includes a front bearing 6 and a rear bearing 7, which are assembled together. These bearings 6 and 7 are hollow and each centrally supports a respective ball bearing 10 and 11 for the rotational mounting of the shaft 3.

[0097] In the example considered, a pulley 12 is fixed to the front end of the shaft 3, at the level of the front bearing 6, for example by means of a nut bearing against the bottom of the pulley's cavity. This pulley 12 transmits the rotational motion to the shaft 3 and can be connected via a belt to the crankshaft of the vehicle's internal combustion engine.

[0098] The rear end of the shaft 3 carries, here, slip rings belonging to a commutator and connected by wire links to the winding. Brushes belonging to a brush holder 8 are arranged so as to rub against the slip rings.

[0099] The front bearing 6 and the rear bearing 7 may also include substantially lateral openings for the passage of air in order to allow the cooling of the rotating electrical machine by air circulation generated by the rotation of a front fan 13 on the front dorsal face of the rotor 4, i.e. at the level of the front bearing 6 and of a rear fan 14 on the rear dorsal face of the rotor, i.e. at the level of the rear bearing 7.

[0100] In this embodiment, the stator 5 comprises a body 15 in the form of a stack of laminations with notches, for example, of the semi-closed or open type, fitted with notch insulation for mounting the stator's polyphase electrical winding. Each phase has a winding 16 passing through the notches of the body 15 and forming, with all the phases, a front and a rear winding on either side of the stator body. The windings 16 are, for example, made from a continuous wire coated with enamel or from bar-shaped conductive elements such as pins connected together. The stator's electrical winding is, for example, three-phase, thus employing a star or delta connection whose outputs are connected to the power electronic component 9.

[0101] The rotor 4 in Figure 1 is a claw rotor. It has two pole wheels 17. The first pole wheel 17 is turned towards the power electronic component 9 while the second pole wheel 17 is turned towards the pulley 12.

[0102] Each of the polar wheels 17 comprises a base 18 extending radially on either side of the X-axis, the wheel defining a series of claws 19 of generally trapezoidal shape. Each claw of a polar wheel 17 extends axially towards the other polar wheel from a base disposed on the radially outer periphery of the base 18.

[0103] The rotor 4 also includes, between the radially inner portions 20 and the claws 19, a coil wound on a coil insulator 22.

[0104] The rotor 4 may also include permanent magnets (not shown) interposed between two adjacent claws 19 on the outer periphery of the rotor. Alternatively, the rotor 4 may be without such permanent magnets.

[0105] The rotor 4 may still be different from that shown in Figure 1, being for example formed by a stack of sheets, as shown in Figure 2.

[0106] The number of pole pairs defined by rotor 4 can be any number of times, for example be equal to three, four, six or eight.

[0107] The machine also includes sensors for measuring the position of rotor 4, for example two Hall effect sensors grouped in the same plastic housing. One of these sensors provides a first sensor signal "S", which is a sinusoidal signal, and the other of these sensors provides a second sensor signal "C", which is a cosine signal.

[0108] These sensors are, for example, positioned at the rear bearing 7 of the machine and they interact with a magnetic target attached to the rotating rotor.

[0109] The sensor signals S and C are used by a control device for the rotating electrical machine 1, which includes a device 100 for determining the angular position of the rotor. The determining device 100 can be a computer, for example a microcontroller, or an integrated circuit, for example an FPGA or an ASIC, capable of receiving and processing electrical signals, including sensor signals.

[0110] The two signals S and C provided by these two sensors are used by the determination device 100 for the determination of the angular position of the rotor 4, which will now be described with reference to figures 3 and following.

[0111] Device 100 includes a circuit 101 for dynamically reducing the bias ("offset" in English) in the sensor signals S and C. In the example considered, the circuit 101 also allows normalization of the sensor signals S and C whose bias has been reduced.

[0112] In the example in Figure 3, we obtain the following output from this circuit 101:

[0113] - the signal S, whose bias has been removed and which has been normalized,

[0114] - the C signal whose bias has been removed and which has been normalized

[0115] These two signals are received as input to a circuit 103 that estimates the rotor's position. This circuit 103 outputs a signal 200 representing the angular position of the rotor 4, via an angle 0 measured relative to a reference position of the rotor. In the example described, this angle corresponds to the electrical angle characterizing the position of the rotor 4, but could alternatively correspond to the mechanical angle characterizing this position. This circuit 103 implements a control loop for the position of the rotor 4.

[0116] This circuit 103 also provides the following outputs:

[0117] - the rotational speed of the electric machine, which is used as input data by the dynamic bias reduction circuit 101, and

[0118] - A synchronization signal 201 changes state when signal 200 crosses 0° and when signal 200 crosses 180°. This change of state in the synchronization signal corresponds, for example, to a pulse from 0 to 1. The synchronization signal 201 can be common to both sensor signals or, according to an alternative (not shown), two synchronization signals 201 can be provided, each assigned to a sensor signal. In the alternative shown in Figure 4, the dynamic bias reduction circuit 101 is not used to process sensor signals S and C. These sensor signals are received directly by circuit 103, which estimates the rotor position. The rotational speed output of circuit 103, as well as the synchronization signal, are received as inputs to the dynamic bias reduction circuit 101, which is assigned to a current sensor signal.The output of this dynamic bias reduction circuit 101 yields an unbiased current signal. Such an unbiased current signal is, for example, a phase current measurement.

[0119] In the example that will now be described, the dynamic bias reduction circuit 101 also allows for the normalization of each sensor signal it receives at input. The output thus yields a version of the input signal whose bias has been reduced, in particular eliminated, and which has been normalized by a value representative of the amplitude of its first harmonic, which is considered normalization.

[0120] In other undescribed examples, only a reduction or elimination of bias was performed, with no normalization being carried out.

[0121] We will first describe how the sensor signal in S is processed by circuit 101.

[0122] This sensor signal in S is first multiplied by a weight 202 in 120. This weight

[0123] 202 is for example representative of the rotational speed of the rotating electrical machine, which can be proportional to this rotational speed such as the electrical pulsation of this speed, or even be this rotational speed.

[0124] The weight applied at 120 is, for example, generated by a block 301 which receives the rotational speed as input and provides, as output, in addition to the weight 202, a cutoff frequency 203 whose value can vary dynamically. This cutoff frequency is, for example, inversely proportional to the rotational speed of the electric machine.

[0125] Figure 5 shows that another block 302 is present here. This block 302 generates, based on the angular position of the electrical machine, a synchronization signal 204 for processing by circuit 101 of the sensor signal at S, and a synchronization signal

[0126] 205 for the processing by circuit 101 of the sensor signal in C. Each of these synchronization signals 204 and 205 is, for example, identical, being determined similarly to the synchronization signal 201 above, that is:

[0127] - that the signal 204, 205 receives a pulse at "1" from the value "0" upon detection of the transition of the signal value 200 to 0°, and

[0128] - that this signal 204, 205 receives another pulse at "1" from the value "0" upon detection of the passage of this value through 180°

[0129] Upon a change of state of the synchronization signal 204, the following occurs simultaneously: - the initialization of the integration over half a period of the output signal 120, to obtain a second signal, and

[0130] - the storage of the value of this integration for the previous half-period, that is to say the value of the second signal for the previous half-period.

[0131] A finite impulse response filter, such as a comb filter, is applied in the example considered in 211 to the value of the second signal at each change of state of the synchronizing signal to obtain the value of this second signal for the previous integration, or "delayed version of the second signal".

[0132] In the example in Figure 5, two operations are then performed.

[0133] The second signal and its delayed version are summed in 212 to obtain a third signal representative of the bias of the sensor signal in S, in particular equal to said bias.

[0134] We observe that a gain is applied to the third signal at 213. In the described case, the gain can be equal to 1. In a variant, the reception of the synchronization signal 204 causes an even number of integrations of the signal to which the weight has been applied over respective half-periods, rather than a single integration over a half-period at 210 to obtain a second signal. Each integration occurs over a half-period, so the number of half-periods to consider is the aforementioned even number. In this case, the gain at 213 is equal to the inverse of the number of half-periods.

[0135] Furthermore, the second signal is subtracted from its delayed version at 214 to obtain a fourth signal representing the amplitude of the first harmonic of the signal provided by the sensor. A gain, which can be equal to nJ2, is then applied to the fourth signal at 215. After the gain has been applied at 215, an absolute value calculation can be performed at 240 for the fourth signal.

[0136] A recording of the value obtained after application of the gains can be made in 250, 251 at each change of state of the synchronization signal 204, 205.

[0137] We observe that two respective low-pass filters are then applied at 216 and 217 to each of the signals resulting from the third and fourth signals. In this case, each low-pass filter implements the same cutoff frequency inversely proportional to the rotational speed.

[0138] We then observe that the signal from the third signal and at the output of the low-pass filter is subtracted from the sensor signal at S, at 220.

[0139] In step 221, the signal obtained at the end of step 220 is also divided by the signal from the fourth signal and at the output of the low-pass filter in step 217.

[0140] This yields a version of the signal from the signal at S that is unbiased and normalized. In the example in Figure 5, the sensor signal at C undergoes exactly the same processing.

[0141] In a concrete example, the sensor signal at S has the equation

[0142] [Math] where "o" denotes bias.

[0143] Following the aforementioned step 120, this signal is integrated over half a period as follows:

[0144] [Math2]

[0145] The sum at 212 allows us to obtain the third signal which is here equal to the bias of the sensor signal at S:

[0146] [Math3]

[0147] Subtracting from 214 yields the fourth signal, which is an accurate estimate of the amplitude of the first harmonic of the sensor signal at S:

[0148] [Math4]

[0149] Thus, the mag(n) estimate obtained differs from the amplitude hi(n) of the first harmonic only by the presence of odd-rank harmonics greater than or equal to 3 whose value decreases as the rank of the harmonic increases, and possibly also as a function of the value of their relative phase. We observe that the estimate does not depend on even-rank harmonics and remains relatively insensitive to noise.

[0150] The invention is not limited to the example just described.

[0151] The implementation of the determination of the angular position of the rotor 4 can be carried out by a computer program product, which, when executed by the control device of the rotating electrical machine 1, makes it possible to carry out the process of determining the angular position of the rotor 4 of the rotating electrical machine 1 on the basis of the two sensor signals 20 provided by position sensors, as described with reference to Figures 3 to 5.

[0152] This computer program product can be saved on a storage medium readable by the control device in order to implement the determination of the angular position of rotor 4.

Claims

1. Claims 1. A method for reducing bias in a periodic signal with a theoretically zero average value provided by a sensor associated with a rotating electrical machine, in particular by a sensor for determining the angular position of a rotating electrical machine rotor, the method comprising: - receiving the signal provided by the sensor - multiplying this signal by a signal proportional to the rotational speed of the rotor of the rotating electrical machine to obtain a first signal - at least one integration over half a period of the first signal to obtain a second signal - applying a finite impulse response filter, in particular a comb filter, to the second signal to add a delayed version of that second signal in order to obtain a third signal representative of the bias, in particular equal to the bias of the signal provided by the sensor, and - the subtraction of the third signal from the signal provided by the sensor, in order to reduce the bias of this signal provided by the sensor.

2. Method according to claim 1, comprising applying low-pass filtering to the third signal.

3. Method according to claim 2, the low-pass filtering being carried out using a variable cutoff frequency, a function of the rotational speed of the rotating electrical machine.

4. Method according to any one of the preceding claims, the integration over half a period of the first signal to obtain a value of the second signal being caused by a synchronization signal as a function of a signal representative of the angular position of the rotor of the rotating electrical machine.

5. Method according to claim 4, the synchronization signal each time controlling the initialization of a single integration over half a period of the first signal.

6. The method according to claim 4, wherein the synchronization signal triggers the initialization of an even number of integrations each time, each integration being performed over half a period of the first signal, these half-periods occurring consecutively, and the stored value of the second signal being obtained by summing the values ​​of these integrations and dividing the sum obtained by the number of half-periods considered.

7. A method according to any one of the preceding claims, comprising: - subtracting the delayed version of the second signal from the second signal to obtain a fourth signal representative of the amplitude of the first harmonic of the signal provided by the sensor, and - the division of the sensor signal whose bias has been reduced by the fourth signal.

8. Method according to claim 7, comprising applying low-pass filtering to the fourth signal.

9. Method according to claim 8, the low-pass filtering being carried out using a variable cutoff frequency, a function of the rotational speed of the rotating electrical machine.

10. Method for determining the angular position of a rotating electrical machine rotor based on two sensor signals provided by position sensors, each sensor providing a periodic signal with a theoretically zero average value, method in which: - the bias of each signal provided by a sensor is reduced using the method according to any one of the preceding claims, and - the position of the rotor is estimated on the basis of these two signals whose bias has been reduced, in particular via the realization of a control loop, by providing a signal representative of the angular position of the rotor of the rotating electrical machine.

11. Set comprising: - a rotating electrical machine for the propulsion of a hybrid or electric vehicle, and - a control device for this electrical machine, comprising two sensors each providing a periodic signal of theoretically zero average value associated with this rotating electrical machine, to determine the angular position of the rotor of this rotating electrical machine, this control device being adapted to perform the steps of claim 10.

12. Product computer program comprising instructions which lead the assembly according to claim 11 to execute the steps of the process according to claim 10.

13. Computer-readable medium on which the computer program product according to claim 12 is recorded.

Images