Method for determining the angular position of an electric machine rotor

Abstract

The present invention relates to a method for determining the angular position of a rotor of a three-phase synchronous electric machine, in which the following steps are implemented: a) the instantaneous voltages (U) and currents (I) of each phase are measured; b) a first rotational speed (ωest) and a first angular position (θest) of a voltage vector are estimated from at least the voltages measured in step a); c) the angular position of the rotor (θrotor) is determined from the first rotational speed (ωest) and the first angular position (θest) of the voltage vector, the measured currents (I), and an optimization method based on finding the maximum power of the electric machine. The invention also relates to a control method, an electric machine, a computer program product, and a computer-readable storage medium. Figure 1 to be published

Description

Title of the invention: Method for determining the angular position of an electric machine rotor technical field

[0001] The invention relates to determining the angular position of a synchronous electric machine rotor, particularly for the purpose of controlling the electric machine. More specifically, the invention is aimed at applications where the rotational speed of the electric machine is high (up to 200,000 revolutions per minute, in particular). Indeed, the higher the rotational speed, the more problematic the inaccuracy in the rotor's angular position becomes, especially for controlling the electric machine.

[0002] Furthermore, we seek to determine the angular position of the electric machine's rotor without using a position sensor. Indeed, for the targeted high-speed applications, implementing a position sensor proves difficult. Prior art

[0003] Several methods for determining the angular position of an electric machine rotor are known, called "sensorless", i.e. not using a position sensor.

[0004] Some of these methods are based on the calculation of electromotive forces (known by the abbreviation EMF) and their phase, in steady state. The calculation of the angular position can then be deduced, for example, using an observer as described in B. NAHID-MOBARAKEH, "Sensorless vector control of synchronous magnet machines: methods, convergence, robustness, "online" identification of parameters", Doctoral thesis, Institut National Polytechnique de Lorraine, 2001.

[0005] This type of method is simple and easy to implement, but the simplified model of the machine introduces significant position calculation errors.

[0006] Other methods use an observer such as the Kalman filter, for example described in the following publications: - Z. Zheng, Y. Li, M. Fadel, “Sensorless control of PMSM based on extended kalman filter,” 2007 European Conference on Power Electronics and Applications, Sep 2007, Aalborg, Denmark. 8 p., ffl0.1109 / EPE.2007.4417275ff. ffhal-03540901. - J. Dilys, V. Stankevic, K. Luksza, “Implementation of Extended Kalman Filter with Optimized Execution Time for Sensorless Control of a PMSM Using ARM Cortex-M3 Microcontroller”, Energies 2021, 14, 3491. https: / / doi.org / 10.3390 / enl4123491 - W. van Meijl, G. Muisers, A. Borisavljevic, M. Brands and E. Lomonova, "Sensorless observation of a very-high-speed permanent magnet synchronous machine," 2013 IEEE International Symposium on Sensorless Control for Electrical Drives and Predictive Control of Electrical Drives and Power Electronics (SLED / PRECEDE), 2013, pp. 1-8, doi: 10.1109 / SLED-PRECEDE.2013.6684502.

[0007] This type of method has two major drawbacks. The first is the dependence of the system of equations and the modeling solely on Gaussian white noise. The second drawback is the high computation time, requiring the use of high-performance processors. For high rotational speeds, it is generally necessary to increase the switching frequency of the variable speed drives and therefore the computation frequency.

[0008] Methods using a sliding mode observer are also known. The document M. Fadel, R. Ruelland, G. Gateau, JC. Hapiot, P. Brodeau, JP. Carayon, "Sensorless control of embedded actuators," 2004 conference of the EEA club's electrotechnical section; March 2004, Cergy Pontoise, France, DO-10.1051 / bib-j3ea:2005609, describes such a method. The principle of these observers can be divided into two steps. The first step consists of converging the dynamics of a system of order n into a manifold of order nm, where m is the number of measurable outputs. This manifold is then called the "sliding surface." The aim is to determine a portion of the matrix in order to first converge the estimated quantities of the model that are measurable. When this step is completed, a system with nm order dynamics is obtained.The second step consists of imposing the dynamics of this system using the remaining gains. This can be calculated using the Filipov method or a similar method known as the equivalent control method. The goal is to cancel the remaining observation error, which corresponds to the estimated but unmeasurable quantities.

[0009] The drawback of this method is that it is highly discontinuous and generates more or less significant noise on the observed quantities (a phenomenon known as "chattering"). It is therefore often necessary to filter these quantities in order to reconstruct a usable high-resolution position. The use of filters generates delays that are more problematic at very high rotational speeds, where the slightest delay can translate into a significant delay angle and thus destabilize the control loop.

[0010] Some methods also seek to minimize the rotor position error, taking into account iron losses. For example, the document J. Kim, I. Jeong, K. Nam, J. Yang and T. Hwang, "Sensorless Control of PMSM in a High-Speed ​​Region Considering Iron Loss," in IEEE Transactions on Industrial Electronics, vol. 62, no. 10, pp. 6151-6159, Oct. 2015, doi: 10.1109 / TIE.2015.2432104 uses finite element modeling to determine iron losses. However, this method requires accurate modeling of the electrical machine and the iron losses. Indeed, the iron losses Pf consist of hysteresis losses Ph and eddy current losses Pe. Pf = Ph + Pe = + keB2^

[0011] With Æ the Steinmetz constant, kh and ke the hysteresis and eddy current constants, and B the magnetic field and (!) the rotation speed.

[0012] When iron losses are modeled by a resistance, only eddy current losses are modeled, and these are proportional to w2. Thus, the finite element calculation consisting of calculating the iron losses provides an imprecise result.

[0013] Moreover, the models used generally do not take into account the evolution of the stator resistance with temperature, which negatively impacts their reliability. Summary of the invention

[0014] The object of the invention is to propose a method for determining the angular position of the rotor, in a precise manner, in particular for high rotational speed applications, without using a position sensor to be implemented in the electrical machine.

[0015] The method of the invention can in particular be insensitive to parametric variations of a model used in the method, to improve the accuracy of the angular position of the rotor.

[0016] The invention relates to a method for determining the angular position of a rotor of a three-phase synchronous electric machine, in particular a smooth-pole or salient-pole electric machine, according to which at least the following steps are implemented: a) We measure the instantaneous voltages and currents of each phase of the electrical machine; b) We estimate a first rotational speed and a first angular position of a voltage vector of the electrical machine from at least the voltages measured in step a); c) said angular position of the rotor is determined from said first rotational speed and said first angular position of the voltage vector, from the instantaneous currents measured in step a) and from a method for finding the maximum power of the electric machine.

[0017] Preferably, no rotor position sensor or torque sensor is used.

[0018] According to one configuration of the invention, an acquisition filter is used which generates a time delay between the physical voltage or current signal measured in step a and the corresponding digital signal) and the time delay generated by the acquisition filter is compensated.

[0019] Advantageously, in step b), a phase-locked loop method is applied.

[0020] Preferably, in step b), at least the following sub-steps are carried out: 1) A Park transformation is applied to the measured voltages to determine the forward voltage and the quadrature voltage, the voltage vector corresponding to the vector defined by the forward voltage and the quadrature voltage, 2) A proportional-integral controller is used to determine a comparative rotational speed from the forward voltage. 3) then we estimate a comparative position of the tension vector by integrating (Int) said comparative rotational speed over time, 4) Substeps 1) to 3) are repeated, taking into account, in substep 1), the last estimated comparative position of the voltage vector, until the forward voltage is less than a predetermined threshold, 5) We determine said first rotational speed and said first angular position as corresponding respectively to the last comparative rotational speed and the last comparative position of the tension vector.

[0021] Advantageously, in step c), a first estimate of the angular offset between the angular position of the voltage vector and the angular position of the rotor is made, from parameters of the electric machine, then the angular offset of the rotor is modified from the first estimate of the angular offset and the instantaneous electrical power of the electric machine to maximize the electrical power, and the angular position of the rotor is determined from the modified angular offset.

[0022] Preferably, in step c), to modify the angular offset of the rotor, at least the following sub-steps are carried out: i) the instantaneous power of the electrical machine is calculated from the measured currents and voltages and the first estimate of the angular shift; ii) we determine (Var) a time variation of instantaneous power; iii) as a function of the time variation of instantaneous power, a comparative calibration angle is determined; iv) the instantaneous power is recalculated from the currents and voltages taking into account the comparative calibration angle; (v) Preferably, substeps i) to iv) are repeated at least once, replacing in step i) the first estimate of the angular offset with the angle of comparative calibration determined and the calibration angle is determined as the last of the comparative calibration angles determined.

[0023] According to one embodiment of the invention, in substep iii), when the time variation is negative, the comparative calibration angle is determined by the following formula:

[0024] ®cal_comp = 'PW-P

[0025] and when the time variation is positive, the comparative calibration angle is determined by the following formula:

[0026] $ p^ff P

[0027] with Kl, K2, K3 and K4 predefined parameters, s the Laplace operator and Diff_P the time variation of power.

[0028] The invention also relates to a method of controlling a synchronous electric machine in which the angular position of the rotor of said electric machine is determined according to the method of determining the angular position of the rotor of the electric machine according to one of the variants or combinations of variants described above and the electric machine is controlled from the determined angular position of the rotor.

[0029] Preferably, the synchronous electric machine is a variable reluctance machine, a smooth-pole electric machine or a salient-pole electric machine, the electric machine comprising or not permanent magnets.

[0030] Advantageously, when the electrical machine has salient poles, a deflux angle is determined to control the electrical machine.

[0031] The invention also relates to a synchronous electric machine in which the control method described is implemented according to one of the variants or combinations of variants described above or in which the angular position of the rotor is determined from the method of determining the angular position of the rotor of the electric machine according to one of the variants or combinations of variants described above.

[0032] The invention also relates to a computer program product downloadable from a communication network and / or recorded on a medium readable by a computer, such as a computer or calculator, and / or executable by a processor, comprising program code instructions for implementing the method of determining the angular position of the rotor according to one of the variants or combinations of variants described above or the control method according to one of the variants or combinations of variants described above, when said program is executed on a computer.

[0033] The invention also relates to a storage medium readable by a computer means, such as a computer or a calculator, and storing instructions which, when executed by a computer means, imply that the computer means implements the method of determining the angular position of the rotor according to one of the variants or combinations of variants described above or the control method according to one of the variants or combinations of variants described above. List of figures

[0034] Other features and advantages of the methods and systems according to the invention will become apparent from the following description of non-limiting examples of embodiment, with reference to the figures attached and described below. [Fig 1]

[0035] Figure 1 shows a block diagram of the method for determining the angular position of the rotor of a synchronous electric machine according to the invention. [Figure 2]

[0036] The figure represents a representative block diagram of an example of step b) of the method for determining the angular position of the rotor of an electric machine according to the invention. [Fig 3]

[0037] Fig. 3 illustrates the angular offset between the angular position of the voltage vector and the angular position of the rotor, in a Park frame, for the application of the method for determining the angular position of the rotor of an electric machine according to the invention. [Fig 4]

[0038] Fig. 4 represents an example of step c) of the method for determining the angular position of the rotor of an electric machine according to the invention. [Fig 5]

[0039] Fig. 5 represents the comparison between the theoretical angular position of the rotor and the angular position of the rotor obtained from the method of determining the angular position according to the invention. [Fig 6]

[0040] Figure 6 illustrates the comparison between the actual direct current and the direct current from the method of determining the angular position of the rotor of an electric machine according to the invention. [Fig 7]

[0041] Fig. 7 illustrates the comparison between the actual quadrature current and the quadrature current obtained from the method of determining the angular position of the rotor of an electric machine according to the invention. Description of the implementation methods

[0042] The following annotations are used in this description and / or in the figures:

[0043] d: direct direction of a frame of reference (in particular a Park frame of reference)

[0044] q: direction in quadrature of a frame of reference (in particular of a Park frame)

[0045] destimated: direct direction of the estimated Park reference frame of the tension vector

[0046] qestimated: direction in quadrature of the estimated Park frame of the tension vector

[0047] drei: direct direction of the Park reference frame of the actual current vector in the rotor

[0048] qrei: quadrature direction of the Park frame of the real current vector in the rotor

[0049] <hcai_comp : angle de calage comparatif

[0050] dLir angular offset modified when the time variation of power is positive

[0051] 3>cai2: angular offset modified when the time variation of power is negative

[0052] <hcaiage: décalage angulaire entre le repère défini par (dréei, qréei) et le repère défini par (^estimé, Qestimé)

[0053] K1, K2, K3, K4: predefined parameters

[0054] k: filter parameter

[0055] alternating voltages measured at each phase a, b, c respectively of the electrical machine

[0056] Vmax: maximum voltage

[0057] 0V: actual angular position of the voltage vector

[0058] 0est: first estimated angular position of the tension vector

[0059] Oest_rotor* first estimated angular position value of the rotor

[0060] 0rotor: determined angular position of the rotor

[0061] vdq_PLL\ estimated voltage vector composed of vd PLL and vq_PLL

[0062] Vj pll ; forward voltage estimated at step b)

[0063] vq_PLL: quadrature voltage estimated at step b)

[0064] P(-0esti)C23: Park transformation to go from the three-phase frame to the Park frame

[0065] rs: equivalent resistance of one phase of the stator

[0066] U: voltage measurements

[0067] I: current measurements

[0068] Id: forward voltage

[0069] Idreab actual direct voltage

[0070] Iqreai: real quadrature tension

[0071] Idcontrob forward voltage output of optimization

[0072] Iqcontrob quadrature voltage exiting optimization

[0073] Iq: quadrature tension

[0074] Iref: reference voltage

[0075] co: electrical rotation speed

[0076] coest: estimated electrical rotation speed

[0077] <pd : flux magnétique dans la direction directe

[0078] cpq: magnetic flux in the quadrature direction

[0079] Ld: inductance in the forward direction

[0080] Lq: inductance in the quadrature direction

[0081] rpf: permanent magnet flux seen through the stator windings

[0082] vd: forward voltage

[0083] Vq .quadrature voltage

[0084] p: number of pole pairs of the electrical machine

[0085] PDC: continuous input power (from the inverter for example)

[0086] Diff_P: power time variation

[0087] The invention relates to a method for determining the angular position (and its evolution over time) of a rotor of a three-phase synchronous electric machine (i.e. with three phases).

[0088] The electric machine can be a smooth-pole or salient-pole electric machine. It can also be a variable reluctance machine, for example comprising permanent magnets, or made of a material enabling the generation of an electromotive force under no-load conditions.

[0089] An electric machine generally comprises a rotor and a stator. The stator is usually a fixed part, while the rotor rotates about a longitudinal axis relative to the stator. The rotor may contain permanent magnets, but rotors without permanent magnets also exist. Furthermore, the stator comprises one or more pairs of magnetic poles and generally includes windings (also called coils) in each magnetic pole. Thus, when a voltage is applied to the windings of the stator's magnetic poles, the poles become magnetically activated and cause the rotor to rotate (the electric machine operates as a motor). Induced currents are therefore applied to the rotor. Conversely, the electric machine can operate as a generator.In this case, the rotor is driven in rotation (for example by a belt or a set of gears) and the rotation of the rotor induces a current in the windings, which can thus generate electric current.

[0090] The electric machine is said to be "synchronous" when the rotor rotates in sync with the rotating field of the stator.

[0091] For this method, at least the following steps are implemented: a) We measure the instantaneous voltages and currents of each phase of the electrical machine; b) we estimate a first rotational speed and a first angular position of a voltage vector of the electrical machine from at least the said voltages measured in step a); c) the angular position of the rotor is determined from said first rotational speed and said first angular position of the voltage vector, from the instantaneous currents measured in step a) and from a method for finding the maximum power of the electric machine (for example an optimization method based on finding the maximum power of the electric machine, where the aim is to optimize the accuracy of the angular position of the rotor to achieve the maximum power of the electric machine).

[0092] Preferably, no rotor position sensor or torque sensor is used. Therefore, it is not necessary to implement this type of sensor in the electric machine. The construction of the electric machine is thus simple, and the method allows for good accuracy in determining the angular position of the rotor, particularly for controlling the electric machine.

[0093] The method thus allows the calibration of the angular position of the rotor, without prior knowledge of the parameters of the electric machine (and therefore without needing to integrate a drift over time of these parameters).

[0094] This method is particularly suitable for electrical machines with high rotational speeds (i.e. for example rotational speeds greater than 30,000 revolutions per minute).

[0095] Fig. 1 illustrates, schematically and without limitation, a first variant of a method for determining the angular position of the rotor according to the invention.

[0096] The diagram illustrates a first step in measuring Mes the instantaneous voltages U and currents I of each phase of the electrical machine.

[0097] We can then estimate Estl the first rotational speed coest and the first angular position 0est of a voltage vector representative of the measured voltages U. Indeed, the electric machine being a synchronous machine, the voltage vector is directly linked to the rotation of the rotor of the electric machine.

[0098] From these estimates, the angular position of the rotor of the rotor is then determined by means of a search method (in particular optimization based on search) for the maximum power of the electric machine.

[0099] The angular position of the rotor can then be used to control the electric machine. This control step is outlined in dashed lines to represent its optional nature. When this step is implemented, the method can then be considered an electric machine control method.

[0100] The method according to the invention can in particular be implemented by computer means, such as a computer, a server, or a tablet, in particular by a computer means in direct communication (wired or wifi for example) with the electrical machine (in particular with the means of measuring instantaneous voltages and currents of the electrical machine) in order to work in real time. Step a)

[0101] The measurement step a) of the instantaneous voltages and currents of each phase makes it possible to identify, at each instant, the current of each phase and the voltage of each phase.

[0102] According to one embodiment of the invention, an acquisition filter can be used. The acquisition filter generates a time delay between the physical voltage or current signal measured in step a) and the corresponding digital signal, and this time delay can be compensated. Indeed, signal acquisition generally involves a delay between the actual physical signal and the resulting digital signal. This delay can be due, in particular, to the hardware itself and / or the processing of this signal. When the rotational speed is high, even a small delay can have a significant impact on the angular position error and consequently on the control of the electrical machine.When the time delay is known or can be determined with sufficient precision, the time delay generated by the acquisition filter can be compensated to shift the digital signal by the equivalent of this time delay, thus aligning it with the actual time of the real physical signal. Using time delay compensation for the acquisition filter facilitates subsequent steps by reducing an initial inaccuracy. For example, by knowing the measurement chain and its filters, it is possible to calculate this delay, which can be of the form -k*w in the case of first-order cascaded filters. Step b)

[0103] In this step, we seek to determine the angular position of a voltage vector representing the different phases. Indeed, since this is a synchronous electrical machine, the electrical angular frequency of the voltages is equal to the electrical speed of the rotor. As the rotor rotates, the direction of a voltage vector representing these different phases evolves over time and as a function of the

[0104]

[0105] angular position of the rotor. Thus, from the measured voltages, we can determine the angular position of the voltage vector. Step b) is an estimation because it identifies the angular position of the voltage vector, which does not perfectly match the angular position of the rotor. This is because there is a slight time lag between the current signal and the voltage signal. Given the rotational speed, this time lag (generally called the "power factor") induces an angular shift in the rotor, and naturally, the higher the rotational speed, the greater this angular shift. This time lag also depends on the parameters of the electric machine, such as inductance, stator resistance, and magnetic flux. Controlling the electric machine is therefore inaccurate, or even difficult. The measured (and possibly filtered) stator voltages vabc are presented in the following equation where is the angular position of the voltage vector. Vabc — ^max sin(ov-?f)

[0106]

[0107] The three phases of the electrical machine are each shifted by radians, hence the shifts of 2 and in the previous expression. Knowing these currents and / or voltages measured for each phase allows us, for example, to derive, via a Park transformation, a voltage vector and / or a current vector. The Park transformation is a change of reference frame between the fixed stator frame and the rotating frame under consideration (for example, the rotor frame rotating around the rotor axis relative to the stator), called the Park frame. The Park transformation allows a three-phase problem to be modeled as a two-phase model, which is simpler to model and solve. The Park frame (each frame under consideration) is defined by two axes: a direct axis d and a quadrature axis q, orthogonal to the direct axis d. The direct and quadrature axes d and q are orthogonal to a third axis (usually called the homopolar axis "h") around which the Park frame rotates (for example, the rotor axis).The Park transformation of three-phase voltages and / or currents thus makes it possible to obtain respectively a voltage vector composed of the direct voltage (voltage along the direct axis d) and the quadrature voltage (voltage along the quadrature axis q) and / or a current vector composed of the direct current (current along the direct axis d) and the quadrature current (current along the quadrature axis q). Park's frame of reference is directly related to the angular position of the rotating part, for example the rotor (relative to the electric machine and more precisely the stator). Therefore, in step b), we can estimate the first rotational speed and the first angular position of the voltage vector (for example, the vector resulting from the Park transformation of the voltages in each phase) from the measured voltages (and / or measured currents). For example, we can transform the voltages measured in each phase at each instant into a voltage vector composed of the forward voltage and the quadrature voltage in the Park frame, and we seek the first rotational speed and the first angular position such that the forward voltage is zero. Indeed, we aim to regulate the zero forward voltage in such a way as to regulate the sine function of the difference between the estimated angular position and the zero actual angular position, which corresponds to an estimated angular position of the voltage vector approximately equal to the actual angular position of the voltage vector.

[0108] Preferably, in step b), a phase-locked loop (PLL) method can be applied. This loop allows the phase to be controlled by the three-phase voltages to precisely determine the first angular position of the voltage vector. Indeed, thanks to the use of the phase-locked loop, the first angular position can gradually converge to its actual value.

[0109] Advantageously, for the phase-locked loop method, at least the following substeps can be performed: 1) A Park transformation can be applied to the measured voltages to determine the forward voltage and the quadrature voltage, the voltage vector corresponding to the vector defined by the forward voltage and the quadrature voltage; 2) A proportional-integral controller can be used to determine a comparative rotational speed from the forward voltage: this type of controller is simple to implement. 3) Then we can estimate a comparative position of the tension vector by integrating said comparative rotational speed over time, the angular position being directly related to the rotational speed, 4) We can repeat substeps 1) to 3), taking into account, in substep 1), the last estimated comparative position of the voltage vector, until the forward voltage is less than a predetermined threshold, so as to make the estimated angular position of the voltage vector converge towards the actual angular position of the voltage vector (and therefore the sine function of the difference between these two positions tends towards zero), which implies that the forward voltage tends towards zero, 5) we can determine said first rotational speed and said first angular position as corresponding respectively to the last comparative rotational speed and the last comparative position of the tension vector, that is to say when the loop has converged.

[0110] These steps allow the first rotation speed and the first angular position of the tension vector to be defined precisely and quickly.

[0111] Fig. 2 illustrates, schematically and not in a limiting manner, an example of a phase-locked loop for determining the first rotational speed and the first angular position of the velocity vector according to the invention.

[0112] In a first step, we measure Mes of the instantaneous voltages U and currents I of each phase of the electrical machine.

[0113] At least the voltage measurements U are used to perform Park transforms and thus determine a voltage vector composed of the forward voltage vd and the quadrature voltage vq. To initialize the control loop, an arbitrary value of the rotor's angular position can be used (for example, equal to zero or to the last known position of the rotor).

[0114] We seek to regulate the zero forward voltage so as to regulate the sine function of the difference between the estimated angular position and the actual angular position so that it is zero, which corresponds to an estimated angular position of the voltage vector substantially equal to the actual angular position of the voltage vector. Thus, we can find the angular position of the voltage vector.

[0115] Thus, the forward voltage vd is compared to zero, and a proportional-integral PI controller is then applied to determine the rotational speed coest. By then integrating this rotational speed coest into Int, the angular position 0est of the voltage vector can be estimated. This angular position 0est of the voltage vector is then reused at the loop input for the Park transformation. When the forward voltage is zero or nearly zero, the first angular position of the voltage vector and the first rotational speed can then be determined, as corresponding to the last estimated values ​​of the estimated angular position 0est of the voltage vector and the rotational speed coest, respectively.

[0116] For example, the following equation can give the voltage vector, in the estimated Park frame dq^, as a function of the estimated angular position Qest of the voltage vector, sin(ev-0estï) ' -cos(dv-9est}).

[0118] As we seek to bring vd_PLL back to 0, the estimated position 9 becomes equal to when the phase-locked loop converges. Step c)

[0119] Step c) refines the accuracy of the rotor's angular position, starting from the estimation of the first angular position obtained for the voltage vector. This step may, for example, include a loop to progressively refine the rotor's angular position. To initialize this loop, one may, for example, consider, 'vd_PLL Vq_PLL '^est ) max Vdq_PLL ~ Initially, the angular position of the rotor corresponds either to the first angular position of the voltage vector, or to a first estimated value.

[0120] Fig. 3 illustrates, schematically and not in a limiting way, the angular phase shift between currents and voltages.

[0121] This figure illustrates a first frame (of Park) defined by the axes dred and qrei and a second frame (of Park) defined by the axes destimated and qestimated. The angular phase shift between these two frames corresponds to the angle of calibration 0caliage.

[0122] The first frame (dréei, qréei) corresponds to the Park frame of the currents in the rotor, and the second frame (destimated, qestimated) corresponds to the Park frame of the voltage vector. Thus, in the first frame (dréei, qréei), the current vector is in the direction qréei (indeed, we assume a smooth-pole synchronous machine operating at maximum torque). Therefore, the reference current Iref is equal to the quadrature current Iq in the direction qréei, and the forward current is zero in the direction dréei.

[0123] In the second frame, the voltage vector V is in the direction qestimated. Given the non-zero angle of adjustment θ between the two frames (drei, qrei) and (destimated, qestimated), the voltage vector V decomposes into a direct voltage Vd = -ω - Lq - Iq in the direction dred of the first frame (drei, qrei) and a quadrature voltage Vq - rS'Iq"F (x? * Oans the direction q,ée of the first frame (drei, qrei)*

[0124] Using a method for finding (in particular, search-based optimization) the maximum power of the electrical machine is advantageous because the input power, especially at the inverter (the electrical machine advantageously including this inverter), is a readily determinable quantity. For example, one can measure the DC voltage (using a voltage measuring instrument such as a voltmeter) and the DC current (using an electric current measuring instrument such as an ammeter) at the inverter input, and deduce the DC power at the inverter input by multiplying the DC voltage and the DC current. Generally, the electrical machine is used to operate at its maximum torque, which corresponds to zero forward current and maximum quadrature current (after Park transformation, in the Park frame).For a smooth-pole machine, maximizing the torque of the electric machine is equivalent to maximizing the power. Thus, the search method (search-based optimization) for the maximum power of the electric machine allows us to find the maximum torque corresponding to the operation of the electric machine.

[0125] For a salient-pole machine, the flux deflection angle is taken into account to integrate the phase shift induced by this type of machine on the maximum torque. Indeed, for the salient-pole machine, the maximum current intensity is not in the quadrature direction q but shifted from this direction by a flux deflection angle. (which therefore corresponds to the angular shift between the quadrature direction q and the direction of maximum current intensity, after Park transformation). This flux-fluidization angle is generally known from other sources and depends on the electrical machine. It is therefore a predetermined parameter of the salient-pole electrical machine. Thus, the flux-fluidization angle can be taken into account to define the Park reference frame linked to the rotor for this type of electrical machine and thus control the electrical machine in the correct direction of the Park reference frame, for example.

[0126] The determination of the angular position of the rotor can in particular be used for the control of the electric machine.

[0127] According to one configuration of the invention, a first estimate of the angular offset between the angular position of the voltage vector and the angular position of the rotor can be made from parameters of the electric machine. Then, the angular offset of the rotor can be modified from the first estimate of the angular offset and the instantaneous electrical power of the electric machine to maximize the electrical power. Next, the angular position of the rotor can be determined from the modified angular offset. In other words, the angular offset is initialized with a first value. This first value can be that of the first angular position of the voltage vector or can be determined, for example, numerically from various parameters of the electric machine.This first value is imprecise, either because we know that the voltage vector is angularly offset from the rotor, or because the parameters of the electrical machine introduce errors.

[0128] For example, when determined numerically, the first value of the angular offset can be determined from the following equation:

[0129] , \ \ v calage “ ~ ^TCtail [ rsl /

[0130] This equation is particularly valid for salient pole machines.

[0131] When the electrical machine is a smooth-pole machine, one can simplify the previous equation and determine the first value of the angular offset from the following equation, assuming that the forward current is regulated to zero in order to maximize the torque of the electric machine (the stator is assumed to be unsaturated): FOI 371 f vd \ l \ ( -wLalq \ $ calage = U ) = *^11 ( ) « ^Ctail ( )

[0133] The preceding equations for determining the first value of the pitch angle depend heavily on parameters of the electrical machine.

[0134] Thus, a parametric error may induce an angle error which can lead to a loss of performance of the electrical machine (the torque at the point of (operation is not at its maximum), or even to the point of loss of control. This is why the angular offset is then modified to improve its accuracy.

[0135] For these reasons, we then seek to modify this first value to reduce the associated error. To improve the accuracy of the rotor's angular position, we then seek to modify the angular offset in order to maximize electrical power, which amounts to maximizing electrical torque. The (modified) angular offset gradually converges, and we can then deduce the rotor's angular position.

[0136] Advantageously, to modify the angular offset of the rotor, at least the following sub-steps can be carried out: i) the instantaneous power of the electrical machine can be calculated from the measured currents and voltages (which can be the currents and voltages measured in step a) or the DC currents and voltages measured at the inverter) and the first estimate of the angular offset; ii) we can determine a time variation of instantaneous power; indeed, since currents and voltages are measured instantaneously, we can, for two successive measurement instants, determine the power (by the product of voltage and current for example) of each instant and consequently the variation of power between these two successive instants; iii) Based on the time variation of instantaneous power, a comparative calibration angle can be determined; for example, if the power variation is positive, this means that the maximum power has not yet been reached, since the power tends to increase. Therefore, the comparative calibration angle can be increased. Conversely, if the power variation is negative, this means that the maximum power has been exceeded and the power then tends to decrease. Therefore, the comparative calibration angle can be reduced to get closer to the maximum power value. iv) we can recalculate the instantaneous power from the currents and voltages taking into account the comparative calibration angle; this allows us to re-evaluate the power with the comparative calibration angle to converge towards the maximum power; (v) Preferably, substeps i) to iv) can be repeated at least once, replacing in step i) the first estimate of the angular offset with the determined comparative pitch angle, and the pitch angle can be determined as the last of the determined comparative pitch angles. This iteration allows for gradual convergence towards the pitch angle that defines the angular position of the rotor. This also ensures good accuracy.

[0137] According to one configuration, substep v) can be implemented continuously: in other words, the loop of step c) runs continuously.

[0138] For example, one can seek to maximize the continuous input power PDC of the inverter of the electric machine. The continuous input power can be determined, for a smooth-pole machine, by the following equation: [01391 Pnc*l[P-^ + P'(^-^'h^^^

[0140] Preferably, the angular offset can be adjusted to a reference current I f— + / ' constant ct a constant rotational speed. Therefore, the Joule losses corresponding to the term 2^ + j2 and the mechanical losses are constant during the adjustment of the angular offset. Therefore, maximizing power is equivalent to maximizing electromagnetic torque.

[0141] When the electric machine has salient poles, the principle explained above applies but it is necessary to determine the deflux angle which corresponds to the angle of maximization of the torque, in order to generate the references of the currents Id and Iq in the control frame (corresponding to the Park frame defined by the determined angular position of the rotor).

[0142] In addition, if the delays caused by acquisition filters and calculation delays are known, they can be compensated separately and added as compensation to the calibration angle (for example, to allow its initialization).

[0143] The optimization method based on finding the maximum power of the electric machine consists of finding the compensation angle to be added to the first angular position of the voltage vector to compensate for the phase angle between voltages and current (and possibly filter delays). This method can use, for example, measurements of the DC voltage and current (e.g., that of the power bus) to maximize the DC input power (PDC).

[0144] Fig. 4 illustrates, schematically and not in a limiting way, an example of an optimization method based on finding the maximum power of the electric machine according to the invention.

[0145] From the instantaneous voltage U and current I measurements of each phase of the electric machine and a first value of the rotor's angular position 0est_rotor, the instantaneous electrical power P of the electric machine can be determined. The time variation of the instantaneous power Diff_P can then be determined.

[0146] We can then compare this time variation of the instantaneous power Diff_P with respect to zero and define an angular offset based on this comparison. For example, here, when the time variation of the instantaneous power Diff_P is less than zero Neg, we determine <e>cai2 the angular shift; when the time variation of the instantaneous power Diff_P is greater than zero Pos, we determine <e>The angular offset is calculated. Then, the impact on the voltage V+AV and on the current I+AI is calculated as a function of the previously calculated angular offset. These impacts on the input voltage and current can then be taken into account in the Power determination step of the instantaneous electrical power P of the electric machine.

[0147] Advantageously, in step iii), when the time variation is negative, the comparative calibration angle can be determined by the following formula: [° 14 81 Æ • Dtff_P

[0149] and when the time variation is positive, the comparative calibration angle can be determined by the following formula:

[0150] . Diff _P

[0151] with K1, K2, K3 and K4 being predefined parameters, s the Laplace operator and Diff_P the time variation of power. Therefore, the method's response can be adapted to the dynamics of the slope of the time variation of power Diff_P.

[0152] The comparative calibration angle is a parameterizable (first-order) filtering term used to manage the detection dynamics when the derivative Diff_P is positive or negative. This provides degrees of freedom for the dynamics when the derivative Diff_P is positive or negative.

[0153] K1, K2, K3 and K4 can be derived from pre-determined electrical machine maps, or from recalibration during the electrical machine's lifetime. These parameters can also depend on the value of the time-dependent power variation Diff_P to achieve faster convergence of the calibration angle.

[0154] The invention also relates to a method of controlling a synchronous electric machine in which the angular position of the rotor of said electric machine is determined according to one of the variants or combinations of variants described above and the electric machine is controlled from the determined angular position of the rotor.

[0155] Preferably, the synchronous electric machine is a variable reluctance machine, a smooth-pole electric machine or a salient-pole electric machine, the electric machine comprising or not permanent magnets.

[0156] According to one embodiment of the invention, when the electric machine has salient poles, a deflux angle can be determined to control the electric machine.

[0157] The control method according to the invention can notably be implemented, at least partially, by computer means, such as a computer, a server, or a tablet, in particular by a computer means in direct communication (wired or wifi for example) with the electrical machine (in particular with the means of measuring instantaneous voltages and currents of the electrical machine) in order to work in real time and to be able to control the electrical machine in real time.

[0158] In addition, the invention relates to a synchronous electric machine in which the control method is implemented according to one of the variants or combinations of variants described above or in which the angular position of the rotor is determined from the method according to one of the variants or combinations of variants described above.

[0159] The invention also relates to a computer program product downloadable from a communication network and / or recorded on a medium readable by a computer means (such as a computer or a calculator) and / or executable by a processor, comprising program code instructions for implementing the method of determining the angular position of the rotor according to one of the variants or combinations of variants described above or the control method according to one of the variants or combinations of variants described above, when said program is executed on the computer means.

[0160] The invention also relates to a storage medium readable by a computer means (such as a computer or a calculator), and storing instructions, which, when executed by the computer means, involve the computer means implementing the method of determining the angular position of the rotor according to one of the variants or combinations of variants described above or the control method according to one of the variants or combinations of variants described above. Examples

[0161] The method for determining the angular position of the rotor according to the invention has been tested for a smooth-pole synchronous electric machine, and the control of the electric machine has been carried out based on the angular position of the rotor determined according to this method of the invention.

[0162] The angular position of the rotor is determined by the method according to the invention at the following operating point: • Rotation speed: 50,000 revolutions per minute • Input current: 50 A AC • Maximum power (when the angular position of the rotor determined by the method according to the invention corresponds to the actual position): 180W

[0163] The model of the electrical machine is as follows: • Number of pole pairs: 1 • Ld and Lq inductances: 12pH • Magnetic flux of the magnets through the stator windings: 450 pWb

[0164] For the tested operating point, the speed is imposed by an external system (here a motor coupled to the same shaft as the electric machine, but another system could have been used instead). The method for determining the angular position of the rotor according to the invention is implemented from time t = 0.01 s. Figures 6 and 7 respectively represent the evolution of the forward current ([Fig. 6]) and the quadrature current ([Fig. 7]), in amperes, over time (time(s)) in seconds. The forward current Idreai corresponds to the actual forward current of the electric machine, while the forward current Idcontroic corresponds to the forward current determined by the method according to the invention.Similarly, the quadrature current Iq reai corresponds to the actual quadrature current of the electric machine, while the quadrature current Iq controi corresponds to the quadrature current determined by the method according to the invention.

[0165] As shown in Figures 6 and 7, at the time when the method of the invention is implemented, i.e., t=0.01s, the actual frame of reference of the electric machine (direct and quadrature directions of the actual Park frame of the electric machine) is not in phase with the frame of reference of the determined angular position of the rotor. Indeed, it is observed that at t=0.At 01s and immediately after, the actual direct and quadrature currents Idreai and Iqreai are significantly offset from the direct and quadrature currents resulting from the determined angular position of the rotor according to the method of the invention, Idcontroi and Iqconnoi. This is due to the fact that the angular offset between acquired voltages and measured currents is not compensated at the beginning of the implementation of the method according to the invention. It can be seen that the method of the invention nevertheless allows the direct and quadrature currents estimated by the Idcontroi and Iqconnoi method to converge rapidly with the actual direct and quadrature currents Idreai and Iqreai, respectively, thanks to the phase-locked loop shown in [Fig. 2] on the one hand, and to the optimization method based on finding the maximum power (which here runs in a loop according to the example shown in [Fig. 4]).This demonstrates the quality of the method in determining the angular position of the rotor accurately and quickly, without the need for sensors to be integrated into the electric machine, and without the need for electric machine parameters that may change over time and are therefore unreliable.

[0166] Figure 5 illustrates the comparison over time between the theoretical rotor angular position angle Ang_th and the rotor angular position angle determined by the optimization method based on finding the maximum power according to the invention. It can be seen that, thanks to the method according to the invention, the determined angle converges rapidly to the theoretical angle, accurately, without the need for sensors to be integrated into the electric machine, and without the need for parameters of the Electric machines that are subject to change over time and are therefore unreliable. The optimization method based on finding the maximum power allows for rapid and reliable convergence to determine the rotor's angular position angle.< / e> < / e>

Claims

Demands

1. A method for determining an angular position of a rotor of a three-phase synchronous electric machine, in particular a smooth-pole or salient-pole electric machine, wherein at least the following steps are carried out: a) the instantaneous voltages (U) and currents (I) of each phase of the electric machine are measured; b) a first rotational speed (œest) and a first angular position (0est) of a voltage vector of the electric machine are estimated from at least the voltages measured in step a); c) said angular position of the rotor (0rotor) is determined from said first rotational speed (œest) and said first angular position (0est) of the voltage vector, the instantaneous currents (I) measured in step a) and from a method for finding the maximum power of the electric machine.

2. Method of determining an angular position of a rotor of a synchronous electric machine according to claim 1, wherein an acquisition filter is used which generates a time delay between the physical voltage or current signal measured in step a and the corresponding digital signal) and the time delay generated by the acquisition filter is compensated.

3. Method for determining an angular position of a rotor of a synchronous electric machine according to any one of the preceding claims, wherein in step b), a phase-locked loop method is applied.

4. A method for determining the angular position of a rotor of a synchronous electric machine according to claim 3, wherein, in step b), at least the following substeps are carried out: 1) a Park transform (Transf) is applied to the measured voltages (U) to determine the forward voltage (vd) and the quadrature voltage (vq), said voltage vector corresponding to the vector defined by the forward voltage (vd) and the quadrature voltage (vq), 2) a proportional-integral (PI) controller is used to determine a comparative rotational speed from the forward voltage, 3) then we estimate a comparative position of the voltage vector by integrating (Int) said comparative rotational speed over time, 4) we repeat substeps 1) to 3) taking into account, in substep 1), the last estimated comparative position of the voltage vector, until the forward voltage (vd) is less than a predetermined threshold, 5) we determine said first rotational speed (coest) and said first angular position (0est) as corresponding respectively to the last comparative rotational speed and the last comparative position of the voltage vector.

5. Method for determining an angular position of a rotor of a synchronous electric machine according to any one of the preceding claims, wherein, in step c), a first estimate of the angular offset between the angular position of the voltage vector and the angular position of the rotor is made, from parameters of the electric machine, then the angular offset of the rotor is modified from the first estimate of the angular offset and the instantaneous electrical power of the electric machine to maximize the electrical power, and the angular position of the rotor is determined from the modified angular offset.

6. A method for determining the angular position of a rotor of a synchronous electric machine according to claim 5, wherein, in step c), to modify the angular offset of the rotor, at least the following substeps are carried out: i) the instantaneous power (P) of the electric machine is calculated (Power) from the measured currents (I) and voltages (U) and the first estimate of the angular offset (0est_rotor); ii) a time variation of instantaneous power (Diff_P) is determined (Var); iii) as a function of the time variation of instantaneous power (Diff_P), a comparative pitch angle is determined; iv) the instantaneous power is recalculated from the currents and voltages taking into account the comparative pitch angle;v) preferably, we repeat at least once the substeps i) to iv) by replacing in step i) the first estimate of the angular offset with the determined comparative calibration angle and we determine; the angle of adjustment as the last of the comparative angles determined.

7. A method for determining the angular position of a rotor of a synchronous electric machine according to claim 6, wherein, in substep iii), when the time variation (Diff_P) is negative (Neg), the comparative pitch angle ( <bcai_comp) par la formule suivante : ^cal comp (¾¾ ■ Diff Pet lorsque la variation temporelle est positive (pos), on détermine l’angle de calage comparatif (<bcai_comp) par la formule suivante : ^cal comp - 'Diff Pavec Kl, K2, K3 et K4 des paramètres prédéfinis, s l’opérateur de Laplace et Diff_P la variation temporelle de puissance.

8. A method of controlling a synchronous electric machine in which the angular position of the rotor of said electric machine is determined according to the method according to any one of claims 1 to 7 and the electric machine is controlled (Cont) from the determined angular position of the rotor.

9. A control method according to claim 8, wherein the synchronous electric machine is a variable reluctance machine, a smooth-pole electric machine or a salient-pole electric machine, the electric machine comprising or not permanent magnets.

10. A control method according to any one of claims 8 or 9, wherein, when the electric machine is salient pole, a deflux angle is determined to control the electric machine.

11. Synchronous electric machine in which the control method according to one of claims 8 to 10 is implemented or in which the angular position of the rotor (0rotor) is determined from the method according to one of claims 1 to 7.

12. Product computer program downloadable from a communication network and / or stored on a medium readable by a computer, such as a computer or a calculator, and / or executable by a processor, comprising program code instructions for implementing the method for determining the angular position of the rotor according to any one of claims 1 to 7 or of the control method according to any one of claims 8 to 10, when said program is executed on a computer.

13. A computer-readable storage medium, such as a computer or calculator, for storing instructions which, when executed by a computer, involve the computer implementing the method for determining the angular position of the rotor according to any one of claims 1 to 7 or the control method according to any one of claims 8 to 10.

Images