Method for calibrating model for estimating torque provided by electric machine

Abstract

The present application relates to a calibration method for an estimation model for estimating a torque provided by an electric machine (10), the calibration method comprising the steps of: (E1) measuring a current in the electric machine (10) at an operating point; (E2) estimating the torque provided by the motor (10) under the working condition according to the measured current and an estimation model, and associating the output torque with the input current by the estimation model; (E3) measuring a reference torque provided by the motor (10) at the operating point; (E4) comparing the estimated torque with a reference torque; and (E5) adjusting the estimation model to reduce or even eliminate the difference between the estimated torque and the reference torque.

Description

Technical Field

[0001] The invention relates to a method for calibrating a model for estimating the torque provided by an electric motor. The invention also relates to a calibration device for estimating the torque provided by an electric motor. Background Art

[0002] Heat or electrical machines can be electric motors or generators and are commonly used in industry to provide power to systems. In particular, electric machines are currently widely used to provide propulsion power to propulsion units of aircraft, such as helicopter rotors, propellers or turbojet engine fans. The machine provides mechanical power in the form of torque and speed on a rotating shaft. The machine is controlled using an electronic control system that is designed to track set points for speed or torque that meet the system requirements.

[0003] Controlling the torque provided by a working machine is essential for the proper functioning of the system. In particular, when electric motors are used for aviation propulsion, torque control is a matter of flight safety. Specifically, applying the wrong torque can have serious consequences for flight safety. Applying too much torque may result in exceeding the maximum torque designed for the mechanical transmission system, causing damage that may lead to failure of critical components such as the helicopter main gearbox or propellers. Applying too little torque, i.e. less than the maximum torque that the propulsion system should be able to provide, may result in insufficient propulsion power, which may be critical in certain flight phases.

[0004] The uncertainty of the torque information provided by the machine has been taken into account when designing the machine and its control system in order to ensure flight safety. However, significant uncertainty either leads to oversizing of the mechanical transmission, thus increasing the empty weight of the aircraft, or to a reduction in the performance of the aircraft (payload capacity, range, etc.), thus reducing its economic viability.

[0005] The torque information provided by a thermal or electric machine can be obtained by direct measurement using a torque sensor, also called a torque meter, which is placed, for example, on the engine shaft of a thermal machine or at another location in the mechanical transmission system. Such sensors usually measure the deformation of a mechanical component under the action of the torque provided by the machine. The measurements obtained are accurate. However, this requires dedicated electronics and there are related fault management and reliability issues. In order to obtain the required accuracy, it is usually necessary to perform a calibration procedure on the sensor. Finally, the torque sensor is a bulky and expensive device, which has a negative impact on the system design.

[0006] In the case of several electric machines, the electronic control systems of these electric machines are based on the measurement of the current flowing through the electric machine, realized by means of current sensors. The torque provided by the electric machine can then be estimated by means of the following relationship: T = Kt*I, where T is the torque, I is the current flowing through the electric machine and Kt is the characteristic torque constant of the electric machine under consideration. Since this estimation method by indirect torque measurement is based on the current sensors already provided in the design, the estimation method is simple and cheap in terms of integration (cost, quality, reliability). However, this torque estimation lacks accuracy. For the electric machines provided in aviation propulsion systems, the lack of accuracy of the estimated torque value will cause problems for flight safety.

[0007] The main factors of inaccuracy in indirectly measuring the torque provided by the motor by measuring the current are: - a measurement of the current itself, with an accuracy typically between about ±+ / -5% and 10% of the full measurement range, - the torque constant Kt of the motor, which is a characteristic related to the specific machine topology, but which may vary according to manufacturing and assembly tolerances, with the dispersion of Kt values ​​also being between about + / -5% and 10% of the "average" nominal value, and - Various other sources of inaccuracies, in particular those related to the control electronics of the motor, the operating conditions used (especially temperature), the supply voltage or aging. Summary of the invention

[0008] An object of the present invention is to provide a method and a corresponding calibration device for calibrating a model for estimating the torque provided by an electric motor, so that the accuracy of the torque estimation based on the measurement of the current flowing through the electric motor can be improved, and it is also possible to avoid the need for incorporating a specific torque measurement sensor that has integration (cost, quality, weight) and reliability problems, and the method is also easy to implement.

[0009] According to a first aspect, the present invention provides a calibration method for estimating a model for estimating a torque provided by an electric machine, in particular for estimating a torque provided by a permanent magnet electric machine, the method comprising the following steps: E1, measures the current flowing through the motor at the motor's operating point; E2, estimates the torque provided by the motor at that operating point based on the measured current and an estimation model that relates output torque to input current; E3, measuring the reference torque provided by the motor at the operating point of the motor; E4, comparing the estimated torque with a reference torque; and E5, based on the comparison, adjusting the estimation model to reduce or even eliminate the difference between the estimated torque and the reference torque.

[0010] Certain preferred but non-limiting features of the calibration method according to the first aspect are as follows, which features may be used alone or in combination.

[0011] The method further comprises a step E6 of storing the adjusted estimation model in a memory of the control system;

[0012] Steps E1 to E4 may be repeatedly performed at multiple different operating points of the motor;

[0013] The estimation model includes an affine function, wherein adjusting the estimation model includes: adjusting the gradient and / or y-intercept of the affine function;

[0014] the estimation model comprises a lookup table associating each of a plurality of input currents with a corresponding output torque, wherein estimating the torque in step E2 comprises: recalibrating at least one output torque value in the lookup table during step E5 based on the comparison of said estimated torque with a reference torque performed in step E4, so as to reduce or even eliminate the difference between the estimated torque and the reference torque;

[0015] The estimation model comprises a set of lookup tables, each lookup table being associated with a temperature in a set of operating temperatures of the motor, wherein estimating the torque in step E2 comprises the following sub-steps: selecting a lookup table associated with the temperature of the motor at the operating point;

[0016] The estimation model is adjusted to reduce or even eliminate the difference between the estimated torque and the reference torque within a given current range [Ik, In] and / or within a specific operating temperature range; and / or

[0017] The method includes the following steps: calibrating the motor, calibrating the control system, and storing the motor model and the shaping model of the control system in the control unit. Calibrating the motor includes the following sub-steps: measuring the no-load electromotive force of the motor; measuring the speed of the motor at the operating point of the motor; based on the measured electromotive force and speed, determining the motor model, the motor model relates the electromotive force of the motor to the speed of the motor. Calibrating the control system includes the following sub-steps: measuring the current flowing through the control system on a static load test bench equipped with a main current sensor to obtain a measurement value of the reference current; calculating the measurement error of the secondary current of the control system relative to the measurement value of the reference current; and creating a shaping model to correct the measurement value of the current controlled by the control system based on the measurement value of the reference current.

[0018] According to a second aspect, the present invention provides a calibration device for estimating the torque provided by an electric machine, in particular for estimating the torque provided by a permanent magnet electric machine, the calibration device comprising:

[0019] A current measuring device adapted to measure the current flowing through the motor at the operating point of the motor;

[0020] an estimation model that relates the output torque to the input current, such that the estimation model is suitable for estimating the torque provided by the motor at said operating point of the motor based on the measured current;

[0021] A torque meter, adapted to measure a reference torque provided by the motor at said operating point of the motor;

[0022] A control unit is adapted to compare the estimated torque with a reference torque, the control unit being further adapted to adapt the estimation model based on said comparison in order to reduce or even eliminate the difference between the estimated torque and the reference torque.

[0023] Optionally, the calibration device according to the second aspect further includes:

[0024] a first chamber adapted to replicate a first external condition, the first chamber adapted to accommodate the motor during a calibration process;

[0025] A second chamber is adapted to replicate a second external condition, the second chamber being adapted to house a control system of the motor during a calibration process. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Other features, objects and advantages of the invention will become more apparent from a reading of the following detailed description given by way of non-limiting example and illustrated by the following drawings:

[0027] Figure 1 is a schematic diagram showing a calibration device for estimating the torque provided by an electric motor according to an embodiment.

[0028] Figure 2a and Figure 2b is a diagram showing the calibration of a model for estimating the torque provided by a motor obtained in a calibration method according to an embodiment.

[0029] Figure 3 is a schematic diagram illustrating a calibration method for estimating the torque provided by a motor according to an embodiment.

[0030] Figure 4 is a schematic diagram showing steps for estimating the torque provided by a motor in a calibration method according to an embodiment.

[0031] Figure 5 is a schematic diagram showing a step of calibrating a motor, a step of calibrating a control system, and a step of storing an estimation model in a control unit in a calibration method according to an embodiment. DETAILED DESCRIPTION

[0032] A non-limiting example of a calibration method (also called a shaping method) for estimating a model is Figure 3As shown, the model is used to estimate the torque provided by the motor 10 (especially a permanent magnet motor). The calibration method includes the following steps: E1: measuring the current flowing through the motor 10 at the operating point of the motor 10; E2: estimating the torque provided by the motor 10 at said operating point of the motor based on the measured current and an estimation model, the estimation model relating the output torque to the input current; E3: measuring the reference torque provided by the motor 10 at the operating point of the motor 10; E4: comparing the estimated torque with a reference torque; and E5: Based on the comparison, adjust the estimation model to reduce or even eliminate the difference between the estimated torque and the reference torque.

[0033] For example, the estimation model may be used during operation of the electric machine 10 in an aircraft, the electric machine being intended to provide torque to a propulsion system of the aircraft.

[0034] The torque is thus estimated by an indirect measurement based on measuring the current flowing through the motor 10. The accurate torque estimation thus makes it possible to avoid using a specific torque measuring sensor and instead use the sensor 30 that is necessarily present to measure the current flowing through the motor.

[0035] Therefore, the calibration of the torque estimation model is performed on the entire system, which includes, in particular: the current measurement device 30, the control system 20 of the motor 10, and the motor 10 itself, rather than on the part of the system that only includes the motor 10. In addition, the calibration is performed at a specific operating point, which represents the operating conditions of the motor 10 during its use. Therefore, once calibrated, the estimation model can accurately estimate the torque provided by the motor 10 during subsequent operation of the motor 10 based on the measurement of the input current of the motor.

[0036] This calibration method for the torque estimation model makes it possible to reduce or even eliminate the inaccuracies associated with the indirect measurement of torque by measuring the current, in particular the uncertainties associated with the current measurement itself and the manufacturing tolerances of the motor 10. Therefore, the torque provided by the motor 10 can be estimated with lower uncertainty, which is particularly compatible with the use of the motor 10 for aviation propulsion. In particular, due to the improved accuracy of the torque measurement, the calibration of the torque estimation model makes it possible to meet flight safety regulations and also allows the input current of the machine to accurately correspond to the output torque of the machine. Therefore, the torque provided by the motor 10 can be better regulated. This regulation makes it possible to control the motor 10 more accurately, so as to maximize the torque provided by the motor 10 in different flight phases, while ensuring that the maximum torque designed for the mechanical transmission system is not exceeded, thereby avoiding the risk of damage.

[0037] Calibration device

[0038] A calibration device for estimating the torque provided by an electric machine 10, in particular a permanent magnet electric machine, is shown by way of non - limiting example in Figure 1 . The calibration device comprises: a current measurement device 30, an estimation model, a calibration unit 50 and a test bench 60.

[0039] The current measurement device 30 is adapted to measure the current flowing through the electric machine at the operating point of the electric machine. Advantageously, the current measurement device 30 already present in the electric machine 10 can be used (and thus it can be used for other processes associated with the electric machine 10), which makes it possible to avoid increasing the volume or mass of the electric machine 10.

[0040] The estimation model correlates the output torque with the input current such that the estimation model is adapted to estimate the torque provided by the electric machine 10 at this operating point of the electric machine based on the current measured by the measurement device 30.

[0041] The calibration unit 50 comprises a torque meter 40, which is adapted to measure the reference torque provided by the electric machine 10 at this operating point of the electric machine 10 and is adapted to compare the estimated torque with the reference torque; and the control unit 50 is also adapted to adjust the estimation model based on this comparison to reduce or even eliminate the difference between the estimated torque and the reference torque.

[0042] The test bench 60 is adapted to accommodate the electric machine 10 for performing the calibration process of the torque estimation model.

[0043] In particular, the torque meter 40 can be integrated into the test bench 60. In any case, there is no need to integrate the torque meter into the electric machine 10 itself, and the torque meter 40 integrated on the test bench 60 will be sufficient to calibrate the estimation model for estimating the torque provided by the electric machine 10.

[0044] The current measurement device 30 corresponds to the current sensor provided for the present invention. In fact, the current measurement device 30 can be a Hall - effect sensor, a shunt, a magnetoresistive sensor or a fluxgate sensor. The type of the current measurement device 30 is selected based on different criteria such as: the accuracy required for the measurement, the available integration space and the cost of the current measurement device 30. For example, it has been observed that the Hall - effect sensor is smaller and less costly than the fluxgate sensor, but has more limited accuracy. An object of the present invention is to enable the use of a less accurate and less costly current measurement sensor, such as a Hall - effect sensor, by means of a method and a device for calibrating a torque estimation model based on such current measurement.

[0045] In practice, the current measuring device 30 may be located at the motor 10, or at the control system 20. The measured current corresponds to the current flowing through one phase of the motor 10 at the operating point.

[0046] The calibration device is suitable for implementing the above method and thus has the same calibration accuracy advantages as the above method without increasing the cost, mass or volume of the system.

[0047] The control system 20 of the motor 10 is adapted to control the rotation speed and the supply voltage of the motor 10. In practice, the control system 20 of the motor 10 comprises control electronics and power electronics of the motor 10.

[0048] The power electronics comprises control means for controlling the current flowing through the electric machine 10, suitable for this type of electric machine 1, for example a power converter (eg an inverter or a rectifier). The control electronics enable the power electronics to be controlled.

[0049] The above method enables calibration to be performed on the entire system, including the complete electronic / power transmission system (and power harness if applicable) from the control system 20 to the motor 10. In this way, the uncertainty of the entire system is reduced, not just the uncertainty of the motor 10 or the control system 20 of the motor 10.

[0050] The method may further include step E6 of recording the adjusted estimation model into a memory of the control system 20 .

[0051] Steps E1 to E4 may be repeated at a plurality of different operating points of the motor 10. Each operating point may be defined by one or more parameters of the motor, including in particular the temperature, speed and supply voltage of the motor. In other words, each operating point is defined by a specific value or a combination of multiple values ​​(n-tuple) of an operating parameter of the motor.

[0052] In more detail, at each operating point of the motor 10, the calibration method includes the following steps:

[0053] In step E1, the input current of the motor 10 on the test bench 60 is measured, in step E2, the torque provided by the motor 10 is estimated, in step E3, the reference torque is measured by the torque meter of the test bench 60, and in step E4, the estimated torque is compared with the reference torque. In particular, different operating points of the motor 10 may correspond to different currents in the motor 10. Therefore, repeating steps E1 to E4 at different operating points makes it possible to estimate different torques, each torque corresponding to a different measured current. Therefore, the calibration will be more accurate and take into account the different operating conditions of the motor 10. The more operating points there are, the more accurate the calibration of the torque estimation model is, and the more the robustness of the torque estimation model is improved.

[0054] The above-described method for calibrating the model for estimating the torque provided by the electric machine can be performed after receipt of the electric machine 10 and before integration of the electric machine 10 into a propulsion system (e.g. an aircraft) to which it must be supplied with energy. In other words, the calibration of the estimation model is carried out on the ground (on the test bench 60) using the calibration unit 50. Such calibration of the torque estimation model makes it possible to obtain an accurate estimation of the torque provided by the electric machine 10 during flight, based on the input parameters of the estimator (in particular the current or the supply voltage and / or the temperature), which ensures, as will be further explained below, that the optimal torque is provided in each flight phase with respect to propulsion efficiency and flight safety.

[0055] If in the application in question (aircraft) a power harness 70 (i.e. a set of cables capable of transmitting power signals between the control system 20 (which is remote) and the motor 10 itself) is provided, then such a power harness 70 will be provided for the calibration device and the calibration method in order to take into account the entire system consisting of the complete electronic / power transmission system (from the control system 20 to the motor 10).

[0056] Estimating torque (step E2)

[0057] An estimation of the torque provided by the electric machine is carried out in a step E2 based on the current measured in step E1 and an estimation model relating the output torque to the input current provided by the calibration unit 50 .

[0058] The overall current range [I1, In] comprises a set of currents that can be applied to the input of the electric machine 10 during operation of the electric machine 10 (ie, different flight phases of the aircraft); it is therefore a set of currents measured in step E1.

[0059] Preferably, the calibration of the estimation model is performed over the entire global current range [I1, In] to obtain the best accuracy over the entire operating range of the motor (all flight phases). In step E5, the adjustment of the estimation model performed over the global current range [I1, In] makes it possible to minimize the difference between a set of estimated torques and a set of reference torques.

[0060] In a variant, the calibration of the estimation model can be performed within a given current range [Ik, In], which is a smaller range contained within the global current range [I1, In]. This allows a particularly accurate calibration within this smaller range, since the calibration is specifically optimized within this range. Compared to the global current range [I1, ..., In], the given current range [Ik, In] is smaller and generally includes a selection of the current measured in step E1. The adjustment performed on the estimation model in step E5 will then apply to the given current range [Ik, In], i.e., in step E5, only the torque estimated based on the measured current within this current range is taken into account, without taking into account the torque estimated based on the measured current falling outside the given current range [Ik, In]. Therefore, a particularly accurate calibration can be obtained within the given current range [Ik, In]; but the accuracy of the calibration outside this range [Ik, In] will be lower.

[0061] A given current range [Ik, In] may correspond to a preferred operating range of the motor 10, or to a critical operating current range of the motor 10, within which the calibration must be particularly accurate. For example, in the field of aerospace propulsion, the need for accuracy is particularly evident when approaching the maximum torque (and therefore approaching the maximum current that may be applied to the motor 10 when operating). In particular, the dimensions of power transmission and the maximum power generally correspond to a torque range close to the maximum torque. Therefore, a given current range [Ik, In] may correspond to a current range between 90% and 100% of the maximum current value of the motor 10 under consideration, and thus to a torque range between 80% and 100% of the maximum torque value defined by the propulsion system under consideration.

[0062] The described calibration method and calibration device thus enable calibration of an estimation model defined for an electric motor and a given propulsion system.

[0063] Calibrate the estimation model

[0064] There are several different available models that can be used.

[0065] Estimation model based on affine function

[0066] In a first example, the estimation model to be calibrated can be an affine function, ie a law that is a linear function of the measured current and relates the output torque to the input current. Such an affine function is defined by two parameters: the y-intercept and the gradient.

[0067] The calibration method and calibration device based on such an affine function enable the parameters of the affine function to be adjusted.

[0068] To this end, the initial values ​​of these parameters (y-intercept and gradient) define an initial affine function, which is stored in the memory of the calibration unit 50 in step E6, and the calibration method makes it possible to obtain the final affine function based on the measured values ​​of the reference torque for different measured values ​​of input current, and the measured values ​​of different input currents correspond to different operating points of the motor.

[0069] The initial affine function of the estimation model to be calibrated is denoted as Tini. The initial affine function Tini is defined by the initial y-intercept Bini and the initial gradient Ktini, and provides a torque value Tini(x) based on the measured input current by applying the equation EQ1: Tini(x)=Ktini*Iini+Bi.

[0070] The adjustment performed on the estimation model in step E5 includes adjusting the initial gradient Ktini and / or the initial y-intercept Bini of the initial affine function Tini. More specifically, by comparing the estimated torque and the measured reference torque at different operating points of the motor, the parameters and / or the intercept values ​​of the initial affine function Tini can be recalibrated to minimize or even eliminate the total distance between the estimated torque and its corresponding measured reference torque T1, ..., Tn. The final affine function Tfin obtained by calibrating the estimation model is defined by new parameters Ktfin and / or Bfin so that a more accurate estimated torque value can be obtained for the measured value of the current Ii (i is an integer from 1 (or k in a smaller range) to n, n is the number of operating points used), written as T*, and the T* value is calculated by the following equation: T*=Ktfin*Ii+Bfin.

[0071] In practice, the final affine function Tfin can be obtained by performing a linear regression on a set of reference torques measured in step E3 or a selection in the set of reference torques (corresponding to the range of interest above).

[0072] Figure 2a By way of non-limiting example, an estimation model is shown which is adjusted in step E5 to minimize the distance between the torque estimated in step E2 and the reference torques T1, ..., Tn measured in step E3 over the global current range [I1, In]. Starting from a set of different torques estimated by an initial affine function Tini, the final affine function Tfin is obtained by linear regression of a set of corresponding measured reference torques T1, ..., Tn. Thus, the y-intercept and / or the gradient of the torque estimation affine function are adjusted so that the final affine function Tfin corresponds to a linear regression of a set of measured reference torques T1, ..., Tn within the global current range [I1, In] considered.

[0073] Figure 2bBy way of non-limiting example, an estimation model is shown which is adjusted in step E5 so as to minimize the distance between the choice of the torque estimated in step E2 and the corresponding choice of the reference torque Tk, ..., Tn measured in step E3 within a given current range [Ik, In] corresponding to the range of interest. The final affine function Tfin is obtained by linear regression of the choice of the reference torque Tk, ..., Tn measured within the given current range [Ik, In]. Thus, the initial y-intercept Bi and / or the initial gradient Ktini of the torque estimation affine function are adjusted so that the final affine function Tfin corresponds to the linear regression of the choice of the reference torque Tk, ..., Tn measured for the currents considered within the given current range [Ik, In]. Thus, within the given current range [Ik, In] of interest, the calibration of the torque estimation model is particularly accurate.

[0074] Estimation model based on polynomial function

[0075] In another example, the estimation model is based on a polynomial function of degree strictly greater than 1, such as a quadratic function or a cubic function. The estimation model is suitable for estimating the output torque according to the input current and according to the coefficients associated with each degree of the polynomial function. Therefore, the estimation model contains parameters corresponding to the coefficients associated with each degree of the polynomial function. Then, in step E6, these parameters are stored in the memory of the control system 20 of the electric machine 10.

[0076] Such a polynomial function of degree strictly greater than 1 enables to take into account the nonlinearity of the torque estimate based on the current (for example when the torque constant of the motor 10 depends on the current). In particular, when the motor reaches the magnetic saturation "bend" region (close to the maximum current), a decrease in the torque constant is usually observed. In step E5, the adjustment of the parameters of such a polynomial function of degree strictly greater than 1 can be performed in a global current range [I1, In] or in a more limited given current range [Ik, In], which is more specifically similar to the case concerned with the affine function described above.

[0077] Estimation model based on any other function relating measured current to estimated torque

[0078] More generally, the estimation model may include any function that enables the torque provided by the motor 10 to be estimated based on the current measurements flowing through the motor 10. The estimation model contains one or more parameters that define a function that relates the measured current to the estimated torque. In step E6, these parameters may be stored in a memory of the control system 20 and may be adjusted in a manner similar to that described above.

[0079] Estimation Model Based on Lookup Table

[0080] In another example, the estimation model comprises a lookup table associating each of a plurality of input currents with a corresponding estimated output torque. Thus, the lookup table of the estimation model contains a plurality of input currents and a plurality of estimated output torques corresponding thereto. Thus, the adjustment performed on the estimation model in step E5 comprises: adjusting one or more input currents and / or output torques associated with the lookup table. In other words, during step E5, one or more output torque values ​​in the lookup table are recalibrated according to the comparison of the estimated torque with the reference torque achieved in step E4, so as to reduce or even eliminate the difference between the estimated torque and the reference torque. Thus, when the current measurement value recorded during the calibration test matches the initial input current value exactly, the torque value is only updated; on the other hand, in the opposite case, the input current and the associated torque value are updated so that the lookup table matches the value recorded during the calibration test. Thus, the use of a lookup table means that there is no need to compare the torque estimated by the initial lookup table with the actual measured torque. In fact, inputting the measured torque value into the lookup table means that the error between the estimated torque obtained using the final lookup table and the actual (measured) torque will be minimized. Therefore, for operating points that exactly match the input value (current) of the lookup table, the error must be zero; between these points, the error depends on the interpolation function used, the distance between the points, and the nonlinearity of the actual torque in the interval between the points of the lookup table.

[0081] The use of a lookup table enables accurate calibration of the estimate of the torque provided by the motor 10, particularly when the relationship between the estimated torque and the measured current cannot be easily approximated by a mathematical function. Therefore, the accuracy of the calibration also depends on the number of points in the lookup table (corresponding to the measured current values ​​and torque values). For a given current range contained in the lookup table, the greater the number of measured input current values ​​associated with the output estimated torque, the more accurate the calibration.

[0082] Alternatively, if Figure 4 As shown in the non-limiting example in FIG. 1 , estimating the torque in step E2 may include sub-steps E21 and E22. In sub-step E21, the current measured in step E1 is compared with one or more of the plurality of input currents, and in sub-step E22, an interpolation is performed to derive therefrom an estimated torque provided by the motor 10.

[0083] For example, the interpolation in sub-step E22 may be a linear interpolation. In particular, if the current measured in step E1 is between adjacent first and second input currents in the lookup table, the torque provided by the motor 10 may be estimated by performing a linear interpolation between adjacent first and second output torques, wherein the first and second output torques are associated with the first and second input currents, respectively. Thus, the relative distance between the measured current and the first input current is equal to the relative distance between the estimated torque and the first output torque, or the relative distance between the measured current and the second input current is equal to the relative distance between the estimated torque and the second output torque.

[0084] Consider other variables

[0085] The lookup table may be a one-dimensional lookup table whose dimension corresponds to the current measured in the motor 10. Such a one-dimensional lookup table enables compensation for the sensitivity of the magnetic flux to the current to improve the accuracy of the estimate of the torque provided by the motor 10. Specifically, the magnetic field saturation bending phenomenon causes the torque change under high current to be smaller than the torque change under low current.

[0086] In a variant, the lookup table may be a multi-dimensional lookup table. The lookup table may include a first dimension corresponding to the current measured in the motor 10, and one or more additional dimensions, the additional dimensions corresponding to: a temperature to which the motor 10 is exposed and / or a temperature to which the control system 20 is exposed; and / or The supply voltage and rotation speed of the motor 10.

[0087] Then, for each temperature and / or supply voltage and speed in the lookup table, a plurality of input currents associated with a plurality of output torques are determined. In other words, the estimation model comprises one or more sets of lookup tables, a first set of lookup tables associated with temperature, and / or a second set of lookup tables associated with supply voltage and speed. Subsequently, the torque estimation in step E2 comprises a sub-step E23. In sub-step E23, a lookup table associated with the temperature and / or supply voltage and speed of the motor 10 at the operating point of the motor (at which the current was measured in step E1) is selected.

[0088] Such a lookup table with a dimension greater than 1 makes it possible to compensate for the sensitivity of the determination process of the torque provided by the motor 10 to different environmental factors, such as the variability of the torque constant of the motor with temperature (the influence of temperature on the magnetic field generated by the magnets and the magnetic resistance of the metal stator sheets), or the variability of the measurement error of the current sensor with temperature, which makes it possible to further improve the accuracy of the calibration.

[0089] The temperature to which the motor 10 and / or control system 20 are exposed may be, for example, as close as possible to the measured temperature of electronic components of the motor 10 and / or control system, or the measured temperature of the external environment located near the motor 10 and / or control system.

[0090] Therefore, a lookup table including an additional dimension corresponding to the temperature to which the motor 10 and / or the control system 20 are exposed makes it possible to: compensate for the sensitivity of the measurement of the current flowing through the motor 10 to temperature, and / or, the effect of temperature on the torque provided by the motor 10 at a specific current, thereby further improving the accuracy of the calibration of the torque estimation model.

[0091] For example, the lookup table may be a two-dimensional lookup table, the second dimension corresponding to the temperature to which the electric machine 10 and / or the control system 20 are exposed. In other words, the estimation model comprises a set of lookup tables, each associated with a temperature from a set of operating temperatures of the electric machine 10. Estimating the torque in step E2 comprises a sub-step E23 in which a lookup table associated with the temperature of the electric machine 10 at the operating point is selected.

[0092] The selection step E23 may include applying a time domain filter that represents a dynamic model of temperature variations. The time domain filter is configured to model the temperature deviation between a temperature measurement point (usually the stator winding head of the motor 10) and a point of interest for the temperature parameter (usually located at a magnet on the rotor of the motor 10). Therefore, the selected lookup table is the one that best matches the temperature of the point of interest, because it affects the torque constant Kt of the motor, which makes it possible to further improve the accuracy of the calibration of the torque estimation model. In this way, the time domain filter models the inertia of the temperature variation between the measurement point and the point of interest. The time domain filter can be calibrated by measurements on the test bench 60 (using specific instruments) so that the temperature at the temperature measurement point and the point of interest can be accurately measured, and the temperature measurement point will be available (measurable) under working conditions (during the flight phase).

[0093] The lookup table comprises an additional dimension corresponding to the supply voltage and the rotational speed of the motor 10. This lookup table comprising the additional dimension is particularly useful in case the motor 10 is of the permanent magnet motor type, which may experience strong variations of the supply voltage and / or the rotational speed during operation. In particular, in this case, a so-called defluxing command is usually used, so that the operation of the motor 10 can be adapted to different supply voltages and rotational speeds. In this way, the torque provided by the motor 10 depends on the combination of the supply voltage and the rotational speed of the motor 10. The lookup table may, for example, be a single lookup table in which the additional dimension corresponds to the ratio of the supply voltage to the rotational speed; or in a variant, the lookup table may comprise a first additional dimension corresponding to the supply voltage and a second additional dimension corresponding to the rotational speed.

[0094] In step E6 , one or more look-up tables of the estimation model may be stored in a memory of the control system 20 of the electric machine.

[0095] In step E5, one or more lookup tables of the estimation model are adjusted to minimize the distance between the torque estimated in step E2 and the reference torque T1, ..., Tn measured in step E3 within the global current range [I1, In] or within the given current range [Ik, ..., In]. In other words, the correction of the output torque in step E5, and the correction of the associated input current (where applicable), can be performed within the global current range [I1, In] or within the given current range [Ik, In] in a manner similar to that described above with respect to the estimation model containing an affine function.

[0096] It will be understood that, in addition to providing the estimated output torque value of the motor by means of the measured current measurement, the operating parameters of the motor are also taken into account, in particular the temperature to which the motor 10 and / or the control system 20 are exposed, and / or the supply voltage and the speed of rotation of the motor 10, this consideration also applies in a similar way to the case where the relationship between current and torque can be modeled by a mathematical function (in particular an affine function or a polynomial function), as well as to other cases as long as the relevant measurements are available. Therefore, everything that has been stated about measuring and taking into account the temperature or about measuring and taking into account the supply voltage and the speed (involving demagnetization) can also be applied in a similar way to these estimation functions.

[0097] More specifically, the estimation model may include a function, such as an affine function, a polynomial function (of degree strictly greater than 1), or any other type of function that can be envisioned. The function contains one or more independent variables for associating the corresponding estimated value of the output torque with the measured input current at different values ​​or different value ranges of these additional parameters. For example, the first variable may be associated with the temperature and / or with the supply voltage and the speed. Subsequently, in step E2, estimating the torque includes a sub-step E23, in which the following variables are determined: an independent variable associated with the temperature; and / or, an independent variable associated with the supply voltage and the speed of the motor 10 at the operating point of the motor 10, wherein the current is measured at this operating point of the motor in step E1; and / or, a variable associated with the supply voltage and the speed ratio. Thus, the estimation model makes it possible to compensate for the sensitivity of the measurement of the current flowing through the motor 10 to the temperature and / or to the supply voltage and the speed, and / or, to compensate for the influence of the temperature and / or the supply voltage and the speed on the torque provided by the motor 10, thereby further improving the accuracy of the calibration of the torque estimation model.

[0098] Calibration method sequence

[0099] First embodiment - simultaneous calibration of the electronic control system 20 and the electric machine 10

[0100] The method includes a preliminary step of installing the motor 10 on a test bench 60 .

[0101] The method comprises a step E0 of controlling the calibration device, in particular controlling the motor 10 via the control system 20 to reach an operating point.

[0102] In practice, each operating point of the motor 10 may be defined by the speed of the motor 10, and / or the power level provided by the motor 10, and / or the temperature to which the motor 10 is exposed, and / or the temperature to which the control system 20 of the motor 10 is exposed, and / or the current flowing through the motor 10. In particular, both the torque estimation based on current measurement and the method used to measure the current may be sensitive to the conditions of use, in particular the temperature. Therefore, an operating point that depends in particular on the temperature allows for a more accurate calibration of the torque estimation.

[0103] In order to establish a calibration of the estimation function or a lookup table and take into account the influence of temperature, the calibration device includes at least a first thermal chamber, in which at least the motor 10 is placed, so that the reference torque at the output of the motor can be measured at different operating points and different temperatures.

[0104] Preferably, the calibration apparatus comprises at least two thermal chambers, and even three thermal chambers, so as to reproduce as closely as possible the operating temperature conditions of the motor 10 and its control electronics for different installation types / design types, the motor 10 and its control electronics possibly being close to each other and at substantially the same temperature, or, on the contrary, the motor 10 and its control electronics possibly being remote and connected via a power harness.

[0105] Then, under temperature conditions representative of the actual working environment, each of the functional units in the motor 10, the control system 20 and the power harness 70 can be placed:

[0106] A first chamber 110, adapted to reproduce a first external condition, the first chamber 110 being adapted to accommodate the motor 10 during a calibration process;

[0107] A second chamber 120, adapted to reproduce a second external condition, the second chamber 120 being adapted to accommodate a control system 20 of the motor 10 during a calibration process;

[0108] Optionally, the third chamber 130 is suitable for reproducing a third external condition, and the third chamber 130 is suitable for: accommodating the power harness 70 when the control system 20 is away from the motor 10, and the power harness 70 connects the motor 10 to the control system 20. This is because the cable temperature of the power harness 70 will affect the resistance of the cable, thereby affecting the measured current. By providing a dedicated thermal chamber 130 for the power harness 70, the effect of temperature on the cable resistance of the power harness 70 can be better considered, thereby better considering the effect on the current.

[0109] The first external condition may include a first temperature T°1. The second external condition may include a second temperature T°2, which may be equal to or different from the first temperature T°1. Therefore, the conditions reproduced by the first chamber 110 may be the same as or different from the conditions reproduced by the second chamber 120. In this way, the two chambers 110 and 120 make it possible to simulate the situation where the motor 10 and the control system 20 are exposed to the same external conditions, or to simulate the situation where the motor 10 and the control system 20 are exposed to different external conditions.

[0110] The first external condition reproduced by the first chamber 110 may substantially correspond to the conditions that the motor 10 may experience during its operation. The second external condition reproduced by the second chamber 120 may substantially correspond to the conditions that the control system 20 of the motor 10 may experience during its operation. Therefore, the motor 10 and the control system 20 of the motor 10 may be exposed to the same or different external conditions, respectively, which are representative of the respective operating conditions of the motor 10 and the control system 20. The third external condition may include a third temperature T°3, which may be equal to or different from the first temperature T°1 and / or the second temperature T°2.

[0111] When the motor 10 and the control system 20 are likely to be exposed to substantially the same external conditions during their operation, the motor 10 and the control system 20 may both be placed in the first chamber 110, or respectively in the first chamber 110 and the second chamber 120, in which case the first external condition and the second external condition are the same. Alternatively, when the power harness 70 is likely to experience substantially the same external conditions as those of the motor 10 and the control system 20, the power harness 70 may be placed in the first chamber 110, or in the second chamber 120, or in the third chamber 130, in which case the first external condition, the second external condition, and the third external condition are the same.

[0112] When control system 20 is embedded in electric machine 10 , these two functional blocks are in close proximity to each other and can be considered to be exposed to consistent temperatures.

[0113] Thus, the two functional blocks may be placed in the same chamber 110, or in two chambers 110 and 120, respectively, which, although separate, are configured to reproduce substantially the same external conditions (within the range of variation of these conditions), in particular the same temperature conditions. The temperatures to which the motor 10 and the control system 20 may be exposed (according to the following items) may vary widely:

[0114] The ambient atmospheric temperature, which in aviation can vary between -50°C and +70°C;

[0115] the temperature of the immediate surroundings of the electric machine 10 , which temperature may be affected by the operation of other items of equipment, for example the temperature in a nacelle (such as an aircraft nacelle or a helicopter nacelle) depends greatly on the operation of the gas turbine;

[0116] The temperature of the electrical components and electronic assemblies of the motor 10 and the control system 20, which are caused by the self-heating effect caused by the flow of current, will vary significantly as the operating power of the motor 10 itself changes.

[0117] When the control system 20 of the motor 10 is remote from the motor 10, the two functional blocks are connected by the power harness 70. At this time, the motor 10, the control system 20 and the power harness 70 are generally exposed to substantially different temperature conditions.

[0118] In particular, the motor 10 and its electromagnetic components may be exposed to extreme temperatures, for example, greater than 100° C., in particular when the motor 10 is located in an engine compartment, such as an aircraft cabin or a helicopter engine compartment. On the other hand, the control system 20 can be exposed to lower temperatures when the control system 20 is located away from the motor 10 to be placed in a more protected environment, the temperature of which is kept in a lower or even very low range, for example, below 20° C. or below -40° C. For example, for a specific rotation speed of the motor 10 and / or a specific supply voltage, the operating point of the motor 10 can be reached, wherein the motor 10 is placed in the first chamber 110 (which reproduces a temperature greater than 100° C.), the control system 20 is placed in the second chamber 120 (which reproduces a temperature less than -40° C.), and, where applicable, the power harness 70 is placed in one of the first chamber 110 and the second chamber 120 or in a dedicated third chamber 130 under the conditions that the power harness 70 may experience.

[0119] This calibration method using one or more chambers 110, 120, 130 (containing the motor 10 and / or the control system 20 and / or the power harness 70) enables global calibration of the components including the motor 10, the control system 20 of the motor 10, and the power harness 70 where applicable, and eliminates uncertainties related to the environmental parameters and topological parameters of the system as much as possible, thereby achieving a more reliable, more accurate and easy-to-implement calibration. By characterizing the control system 20 and the motor 10 in a global manner, the calibration method is not only simple and easy to implement, but also ensures that an accurate and robust torque estimation model for the use conditions is obtained.

[0120] Second embodiment - calibrating the electronic control system 20 and the motor 10 separately

[0121] In a second exemplary embodiment, as Figure 5 In the non-limiting example shown, the electronic control system 20 and the motor 10 are calibrated separately. In particular, this makes it possible to simplify the implementation of the invention in the case of maintenance, especially when the control system 20 is remote and is in a fault and needs to be changed. In particular, independent calibration of the control system 20 and the motor 10 is provided here. More precisely, in this embodiment, the current measurement of the control system 20 is characterized using the main current measurement system to obtain a shaping table of the current measurement according to the current actually controlled by the power electronics of the control system 20 and the operating conditions (such as ambient temperature, supply voltage, etc.) that may affect the accuracy of the current measurement of the control system where applicable (step E200). Therefore, the shaping table is specific to each control system 20. In this way, the motor 10 can be simply associated with the control system 20 by loading the shaping table obtained for any control system 20. Similarly, the torque constant of the motor 10 is characterized according to the applied current, the torque constant of the motor 10 is measured using the main current measurement system, and it is compared with the torque value measured using the test tool (torque meter of the test bench) (step E100). Thus, a torque shaping table can be obtained according to the current flowing through the motor and, where applicable, the operating conditions. The torque shaping table is therefore specific to each motor 10. In this way, the control system can be associated with any motor 10 simply by loading the torque shaping table obtained for that motor 10.

[0122] Thus, by using a reference system to determine the estimation model, a model can be obtained that is valid for the specific electric motor 10, but which can be adapted to any electronic control system by simple mapping, without requiring any recalibration of the control system when it has to be changed.

[0123] To this end, the calibration method comprises steps E100 , E200 and E300 .

[0124] The motor 10 is calibrated in step E100, which includes sub-steps E101, E102 and E103. In sub-step E101, the no-load electromotive force of the motor 10 is measured; in sub-step E102, the speed of the motor 10 at the operating point of the motor 10 is measured; in sub-step E103, based on the measured electromotive force of the motor 10 and the measured speed of the motor 10, a motor model associated with the electromotive force and the speed is determined.

[0125] The control system 20 is calibrated in a step E200, which includes sub-steps E201, E202 and E203. At sub-step E201, the current flowing through the power electronics of the control system 20 is measured on a static load test bench equipped with a main current sensor used as a reference (and, where applicable, with the wiring harness 70); at sub-step E202, the measurement error of the secondary current of the control system 20 relative to the main current sensor used as a reference is calculated. These measurements can be performed at different current setting values ​​to cover the current range of interest, and optionally, these measurements are performed under different use conditions that may affect the accuracy of the current measurement of the control system (such as ambient temperature, supply voltage, etc.). Advantageously, the reference current is measured on a static load test bench, so that robust and very accurate measurement results can be obtained. In addition, at sub-step E203, a shaping model associated with the reference current measurement is created based on the current actually controlled by the power electronic device and, optionally, the operating conditions (temperature, rotational speed, etc.) to determine the shaping parameters of the current controlled by the power electronic device of the control system 20 based on the current measured by the main current sensor used as a reference.

[0126] In step E300 , the motor model and the shaping model in the memory of the command system 20 are stored in the control unit 50 .

[0127] Therefore, when the control system 20 or the motor 10 changes, for example, when the control system 20 or the motor 10 fails, the calibration of the motor 10 in step E100 or the calibration of the control system 20 in step E200 is still valid. In particular, if the control system 20 is far away from the motor 10, the control system 20 often needs to be changed during system maintenance, and the change is independent of the motor 10. In this way, the command system can be associated with any motor under the condition that the shaping table of the torque estimator associated with the driven motor is loaded.

[0128] The estimation model is defined as considering the motor model and the control system model independently. The estimation model may correspond to the sum of the motor model and the control system model. Therefore, under the condition that the motor model or the control system model is changed when any motor 10 or any control system 20 is replaced, the control unit 50 may be associated with the motor 10 and / or the control system 20.

[0129] By driving the motor 10 to no-load (i.e. without any load), the no-load electromotive force is measured in sub-step E101 to collect data of the no-load electromotive force, and if applicable, the data will be collected according to the temperature of the motor 10. The electromotive force of the motor 10 can be measured multiple times at different operating temperatures to obtain a table of changes in the electromotive force according to temperature.

[0130] The electromotive force (FEM) is expressed as a voltage generated by the motor 10 when the motor 10 is driven to rotate at a certain speed. The relationship between the rotation speed of the motor 10 and the electromotive force is the same as the relationship between the current and the torque.

[0131] Thus, by identifying the relationship between the rotation speed and the electromotive force of the motor 10, the motor 10 is calibrated in step E100, which makes it possible to accurately deduce therefrom the relationship between the measured current and the torque provided by the motor 10, these relationships being identical. In other words, a torque estimation model (table or function) can be constructed based on the estimation of the relationship between the current and the torque obtained by the measurement of the electromotive force and the reference current measurement obtained by the main current sensor as a reference on the static load test bench. The motor model determined in sub-step E103 comprises one or more parameters relating the torque provided by the motor 10 to the measured current (and, where applicable, also to other operating parameters such as temperature, orthogonal current, etc.).

[0132] The current flowing through the motor 10 measured in sub-step E201 may include three current measurements Ia, Ib and Ic, which are the currents flowing through each phase of the motor 10. Based on these three current values, the reference current Iq can be calculated in a known manner: Iq = function (Ia, Ib, Ic), which is a 3×3 matrix (Park transformation) depending on the position measurement of the motor rotor. Thus, based on the phase current measurement of the motor 10, the current measurement is reconstructed by calculation.

[0133] The control system model relates the measured currents to the calculated quadrature currents and the currents controlled by the control system 20: Imes = function (Iq, Iref).

[0134] Furthermore, the control system model created in sub-step E203 can be associated with the measured currents and the temperature of the control system 20 : Imes=function (Iq, Iref, T° 2 ).

[0135] When replacing the control system 20 (or replacing the motor 10 where applicable), the old control system model is replaced with a new control system model (or motor model), which is obtained by calibrating the new control system 20 (or the new motor in step E100) in step E200.

[0136] The calibration parameters of the motor model may be stored in a memory of the control system 20 for the motor 10. Thus, programming the control system 20 that must be paired with the motor 10 will be simple using the calibration parameters of the current measurements.

[0137] As an option, for the calibration parameters of the motor 10 itself (in particular the torque coefficient Kt), it is advantageously provided that a memory card is physically attached to the motor 10 (with an ad hoc interface between the memory card and the control system 20) to store these parameters.

[0138] Other embodiments may be envisioned, and a person skilled in the art may easily modify the above-described embodiments or the exemplary embodiments or envision other embodiments, all of which are within the scope of the present invention.

Claims

1. A calibration method for an estimation model, wherein the estimation model is used to estimate the torque provided by an electric machine (10), in particular, to estimate the torque provided by a permanent magnet motor, wherein the method The following steps are involved: E1: measuring the current flowing through the motor (10) at the operating point of the motor (10); E2: estimating the torque provided by the motor (10) at the operating point based on the measured current and the estimation model, wherein the estimation model relates the output torque to the input current; E3: measuring a reference torque provided by the motor (10) at the operating point of the motor (10); E4: comparing the estimated torque with the reference torque; as well as E5: Based on the comparison, adjust the estimation model to reduce or even eliminate the difference between the estimated torque and the reference torque.

2. The calibration method according to claim 1, further comprising a step E6: storing the adjusted estimation model in a memory of the control system (20).

3. The calibration method according to any one of claims 1 and 2, in, Steps E1 to E4 are repeatedly performed at a plurality of different operating points of the motor (10).

4. The calibration method according to any one of claims 1 to 3, in, The estimation model includes an affine function, and adjusting the estimation model includes: adjusting a gradient and / or a y-intercept of the affine function.

5. The calibration method according to any one of claims 1 to 3, in, The estimation model comprises a lookup table associating each of a plurality of input currents with a corresponding output torque, and estimating the torque in step E2 comprises recalibrating at least one output torque value in the lookup table during step E5 based on the comparison of the estimated torque with the reference torque performed in step E4 to reduce or even eliminate the difference between the estimated torque and the reference torque.

6. The calibration method according to claim 5, in, The estimation model includes a set of lookup tables, each lookup table is associated with a temperature in a set of operating temperatures of the motor (10), and estimating the torque in step E2 includes a sub-step (E23): selecting a lookup table associated with the temperature of the motor (10) at the operating point.

7. The calibration method according to any one of claims 1 to 6, in, The estimation model is adjusted to reduce or even eliminate the difference between the estimated torque and the reference torque within a predetermined current range [Ik, In] and / or a predetermined operating temperature range.

8. The calibration method according to any one of claims 1 to 7, include: - Step (E100) of calibrating the motor (10), including sub-steps (E101), (E102) and (E103), wherein, at the sub-step (E101), the no-load electromotive force of the motor (10) is measured, at the sub-step (E102), the rotational speed of the motor (10) at the operating point of the motor (10) is measured, and at the sub-step E103, based on the measured electromotive force and rotational speed, a motor model is determined, the motor model associating the electromotive force with the rotational speed of the motor (10); - Step (E200) of calibrating the control system (20), including sub-steps (E201), (E202) and (E203), wherein, at the sub-step (E201), the current flowing through the control system (20) is measured on a static load test bench equipped with a main current sensor to obtain a measured value of the reference current, at the sub-step (E202), the measurement error of the secondary current of the control system (20) relative to the measured value of the reference current is calculated, and at the sub-step (E203), a shaping model is created to correct the measured value of the current controlled by the control system (20) based on the measured value of the reference current; - Step (E300) of storing the motor model and the shaping model of the control system in the control unit (50).

9. A calibration device for estimating the torque provided by a motor (10), in particular for estimating the torque provided by a permanent magnet motor, the calibration device comprising: - a current measuring device adapted to measure the current flowing through the motor (10) at the operating point of the motor (10); - an estimation model associating the output torque with the input current such that the estimation model is adapted to estimate the torque provided by the motor (10) at the operating point of the motor (10) based on the measured current; - a torque meter (40) adapted to measure the reference torque provided by the motor (10) at the operating point of the motor (10); - a control unit (50) adapted to compare the estimated torque with the reference torque, the control unit (50) further adapted to adjust the estimation model according to the comparison to reduce or even eliminate the difference between the estimated torque and the reference torque.

10. The calibration device according to claim 9, further comprising: - a first chamber (110) adapted to reproduce first external conditions, the first chamber (110) being adapted to accommodate the motor (10) during the calibration process; - a second chamber (120) adapted to reproduce second external conditions, the second chamber (120) being adapted to accommodate the control system (20) of the motor (10) during the calibration process.

Images