Detection of offset errors in a phase current measurement for a motor control system

Abstract

System for detecting an offset error in a power steering system, wherein the system comprises: a module (204) with a programmable high-pass filter configured to filter a final voltage command (126) and to produce a filtered final voltage command (214) containing a sinusoidal component; a gain and phase compensation module (206) configured to perform gain compensation and phase compensation on the filtered final voltage command (214) to produce a compensated final voltage command (216); and a fault detection module (210) designed to detect an offset error in a current measurement based on the compensated final voltage command (216) in a stator reference frame.

Description

BACKGROUND OF THE INVENTION

[0001] Electric power steering systems (EPS systems) require that the electric motor used to provide steering assistance be operated using a torque control method. When a permanent magnet synchronous motor (PMSM) is used, field-oriented control (FOC) can be employed. The FOC transforms the motor's AC phase voltage and AC phase current signals in the stationary reference frame into the synchronously rotating reference frame, commonly referred to as the d / q-axis reference frame, in which the motor voltages and currents become direct current (DC) quantities. FOC torque control is typically implemented using a closed-loop current control method that employs current controllers to minimize the error between commanded and measured currents, ensuring perfect current tracking.Therefore, current control requires that the motor currents be measured, which can be achieved by measuring the phase currents of the electric machine, which are then transformed into the synchronous frame using the Park transformation in order to perform the control in the synchronous reference frame.

[0002] If an offset error of a given magnitude occurs in a phase current measurement, the closed-loop current control, operating in the synchronous reference frame, adjusts the motor voltage so that the measured motor current matches the command. Because the measurement is incorrect, the actual motor currents are also incorrect. This error mode results in motor position-dependent errors in motor torque and motor current, which can be perceived as a large torque ripple at the motor shaft and are potentially larger than the motor currents classified (for hardware design purposes). If the torque ripple caused by the offset error in the phase current measurement exceeds a certain threshold, the offset error can generate motor torque in a direction opposite to the motor torque command.When used in EPS systems, errors that generate torque in a direction opposite to the desired engine torque command result in driver effort exceeding that of manual operation.

[0003] The closest related document to the application, US 2015 0 291 207 A1, discloses a system for detecting an offset error in a power steering system, comprising a fault detection module configured to determine an offset error in a current measurement. Document US 2015 0 291 207 A1 further discloses a method in which an offset error in a current measurement is determined based on a compensated voltage difference in a stator reference frame.

[0004] One objective of the invention is to improve the steering behavior of vehicles with a power steering system.

[0005] This problem is solved according to the invention by the features of the independent claims. Advantageous embodiments are defined in the dependent claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] The subject matter, which is considered to be the invention, is specifically set out. The foregoing and further advantages of the invention will become apparent from the following detailed description in conjunction with the accompanying drawings, in which: Fig. 1 is an exemplary schematic representation of an engine control system in accordance with an embodiment; Fig. 2 A schematic drawing of a diagnostic module for phase current measurements in accordance with some embodiments illustrates; Fig. 3 A schematic drawing of a diagnostic module for phase current measurements in accordance with another embodiment illustrates; Fig. 4. Records illustrating processing results in accordance with some embodiments; and Fig. Figure 5 illustrates a method for determining an offset error in accordance with an embodiment. DETAILED DESCRIPTION

[0007] With reference now to the figures, in which the invention is described with reference to specific embodiments without limiting them, it shows Fig. Figure 1 shows a block diagram of a torque control module 100 of a multiphase synchronous permanent magnet machine (PMSM) which uses phase current measurements in a current control feedback circuit in accordance with some embodiments of the invention. As shown, the torque control module 100 includes a generator 102 for a motor reference current, a current control module 104, and a diagnostic module 106 for current measurements. Fig. Figure 1 also shows a current measurement module 108 and a motor 110.

[0008] The term "module" or "submodule," as used here, refers to an application-specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or grouped) with memory that executes one or more software or firmware programs, a combinational logic circuit, and / or other suitable components that provide the described functionality. In a software implementation, a module or submodule may be implemented in the memory of a non-temporary, machine-readable storage medium that can be read by a processing circuit and stores instructions for execution by the processing unit to carry out a procedure. Furthermore, the modules and submodules shown in the figures may be combined and / or further subdivided.

[0009] The generator 102 for a motor reference current generates a bias command 120 and a current command 122 based on input signals that include a motor torque command 114, a motor speed 116, a source voltage signal 118, and motor parameters 112. The motor torque command 114 represents a commanded torque value and can be derived from another (not shown) torque control module or it can correspond to a torque value generated by an operator. The motor speed 116 is an angular velocity of the motor 110 measured by a (not shown) speed sensor. The speed sensor may, for example, include an encoder and a speed calculation circuit for calculating the angular velocity of a rotor of the motor 110 based on a signal it receives from the encoder.The source voltage signal 118 represents a bridge voltage from a (not shown) DC current source.

[0010] The motor parameters 112 are estimated values ​​for the motor 110, which include, for example, a motor constant K. e (Volt / rad / s), a motor circuit resistance R (Ohm), a direct shaft inductance L d (Henry) and a quadrature axis inductance L q (Henry) included.

[0011] In some embodiments, the current command 122, generated by the generator 102 for a motor reference current, includes a d-axis reference current command and a q-axis reference current command based on the motor torque command 114, the source voltage signal 118, and the angular velocity. The generator 102 for a motor reference current also calculates the bias command 120, which may include a d-axis bias command and a q-axis bias command. The bias command 120 and the current command 122 fulfill the motor torque command 114.

[0012] The current control module 104 determines a final voltage command 126 based on the bias command 120, the current command 122, and the measured motor currents 124. The measured motor currents 124 include a measured d-axis current and a measured q-axis current, which are transformed from the current measurement signals into the stator reference frame.

[0013] The bias commands are typically different for different types of current controller designs. It is understood that the bias terms used to implement the diagnosis of offset in current readings may be similar to or identical to the bias terms for a motor control system using purely static feedforward control. Furthermore, the final voltage commands may be issued by the current controller. However, the actual voltages applied to the motor by the inverter, or any other suitable voltages, may also be used instead of a final voltage command if a measurement of these voltages is available.

[0014] The current control module 104 sends the final voltage command 126 to the motor 110 for motor control. Specifically, in some embodiments, a polarity converter (not shown) receives the d-axis voltage command and the q-axis voltage command as inputs. Based on these inputs, the polarity converter determines a voltage command and a phase advance angle. A PWM inverter controller (not shown) then receives the voltage command and the phase advance angle from the polarity converter controller as input signals. The PWM inverter controller also receives a rotor angle value of the motor 110, measured by a motor position sensor (not shown). In some embodiments, the PWM inverter controller may include an overmodulation space vector PWM unit, which generates three respective duty cycle values.The duty cycle values ​​are used to drive gate driver circuits of a (not shown) inverter, which excite phases of motor 110.

[0015] For control purposes, the current sensing module 108 generates measured motor currents 124 based on the actual currents 128 from the motor 110. In some embodiments, the current sensing module 108 includes current sensors that generate the measured motor currents 124 from the actual currents 128 transmitted to the motor 110. The measured motor currents 124 therefore represent the values ​​of phase currents (e.g., two-phase, three-phase) measured by the current sensing module 108. In some embodiments, the current sensing module 108 converts measured values ​​of the AC phase currents into equivalent measured DC current components, which are a measured d-axis current and a measured q-axis current in the d / q reference frame (rotor reference frame).

[0016] During operation, the actual currents can exhibit 128 offset errors in the steady-state reference frame. The measured actual currents include the offset errors in the αβ frame, which is also the steady-state frame and represents the phase currents. The steady-state reference frame can be expressed in accordance with the following equations: lαm=lα−Δlα lβm=lβ−Δlβ where I αm and I βm the respective measured currents in the stationary reference frame are, I α and I β the actual currents in the stationary reference frame are, ΔI α and ΔI β The offset errors in the current measurements are expressed in the steady-state reference frame. The steady-state reference frame can be used to simplify the mathematical derivations shown here.

[0017] When transformed into the synchronous reference frame, the measured motor currents are available in the rotating reference frame and can be expressed as: ldm=ld−Δld=ld−l0cos(θ−ϕ)lqm=lq−Δlq=lq−l0sin(θ−ϕ) where lv=Δlα2−Δlβ2andϕ=tan−1(ΔlβΔlα).

[0018] During operation, the measured motor currents 124 can become equal to the current command 122, leading to a distortion of the actual currents 128. This distortion propagates through the torque control module 100, causing a d / q offset error between the bias command 120 and the final voltage command 126 in the synchronously rotating reference frame. Assuming accurate parameter estimation, the d / q offset error (offset error in the rotating reference frame) can be expressed as: ΔVd=−ΔlD(Rcos(θ−ϕ)+ωe(Lq−Ld)sin(θ−ϕ)) ΔVq=−ΔlD(Rsin(θ−ϕ)+ωe(Lq−Ld)cos(θ−ϕ)) where ω e , R, Ld , L q Each of these values ​​is a synchronous frequency (electric motor speed), the motor resistance, the d-axis inductance, and the q-axis inductance. When these voltages are transformed back into the steady-state frame, the offset in the voltage signals of the steady-state reference frame is preserved.

[0019] In the stationary reference frame, the voltage offset signals can be represented as ΔV α , ΔV β The stress offsets ΔV can be expressed. α , ΔV β The stationary reference frame contains respective constant components and respective sinusoidal components. The sinusoidal components can be equal to twice the synchronous frequency of the motor and generally have a small effect on the offsets ΔV. α , ΔV β within the stationary reference framework.

[0020] Due to parameter errors, which are transformed into sinusoidal components when transformed into the stationary frame, the d / q voltage offsets contain a constant term in addition to the sinusoidal term. Furthermore, these terms can be functions of the synchronous frequency depending on the specific parameter and increase with increasing velocity.

[0021] Module 106 for diagnosing current measurements removes the constant component of the d / q voltage offsets to increase the accuracy of fault condition detection after transformation to the steady-state reference frame.

[0022] Specifically, module 106 for diagnosing current measurements determines whether the measured motor currents 124 generated by the current measurement module 108 represent correct measured values. That is, module 106 for diagnosing current measurements determines whether one or more of the phase current measurement sensors of the current measurement module 108 exhibit an acceptable level of offset error in current measurements by measuring the voltage offsets ΔV α , ΔV β analyzed within a stationary reference frame.

[0023] The current measurement diagnostic module 106 compares the final voltage command 126 with the bias commands 120 to determine whether the phase current measurement has a steady-state offset error large enough to indicate sensor malfunction or failure. In some embodiments, the current measurement diagnostic module 106 determines that the received voltage commands do not indicate an offset error in the current measurement. For example, the current measurement diagnostic module 106 generates a diagnostic flag 130 if the steady-state offset errors in the current measurement exceed a threshold. Diagnostic errors can introduce an offset error into phase current measurements. The systems and procedures described below increase the robustness of the diagnostics.

[0024] Fig. Figure 2 illustrates the current measurement diagnostic module 106 of some embodiments in greater detail. As shown, the current measurement diagnostic module 106 includes a module 204 with a programmable high-pass filter, a gain and phase compensation module 206, a transformation module 208, and a fault detection module 210.

[0025] Module 204 with a programmable high-pass filter receives the final voltage command 126. Module 204 with a programmable high-pass filter is configured to filter the final voltage command 126, resulting in the removal of constant components from the final voltage command 126. Accordingly, Module 204 with a programmable high-pass filter is configured to generate a filtered final voltage command 214, which contains the sinusoidal component of the final voltage command 126 without the constant components. Consequently, Module 204 with a programmable high-pass filter can have a cutoff frequency proportional to a synchronous frequency ω. e of the engine 110 ( Fig. 1) The module 204 with a programmable high-pass filter can be programmed based on a calibratable constant value. The calibratable constant value can be fine-tuned to reduce gain and phase error in the sinusoidal voltage component. A transfer function of the module 204 with a programmable high-pass filter can be written as follows: H(s)=sε+k˙ωe where k is the calibratable constant that can be fine-tuned to reduce the gain and phase error in the sinusoidal component. Module 204, with a programmable high-pass filter, can adjust the synchronous frequency ω. eThe gain and phase error introduced by the filter is static. Furthermore, in the module 204 with a programmable high-pass filter, the calibratable constant k can be independently fine-tuned to ensure minimal distortion of the input signal. The module 204 with a programmable high-pass filter is not limited to the above embodiment. For example, a higher-order filter can be used as long as the programmable filter removes the constant component of the final voltage command 126.

[0026] The filtered final voltage command 214 is sent to a gain and phase compensation module 206. The module 204, with a programmable high-pass filter, can introduce magnitude and phase errors into the filtered final voltage command 214, requiring compensation by the gain and phase compensation module 206. In some embodiments, the gain and phase compensation module 206 includes a unity-gain, first-order programmable low-pass filter (PLPF) with the following transfer function: L(s)=k′ωes+k′ωe

[0027] The programmable first-order low-pass filter with unity gain can compensate for the filtered final voltage command 214 to restore the gain and phase of the final voltage command 126.

[0028] The compensation can be performed in accordance with constraint equations, which are written as: ∠H(jωe)−∠L(jωe)=0 |H(jωe)L(jωe)|=1

[0029] In other embodiments, the gain and phase compensation module 206 can use a second-order low-pass filter with a transfer function shown below. L(s)=k′2ωe2s2+2ζk′ωes+k′2ωe2

[0030] The second-order low-pass filter shown above can enable more specific transient responses by using two fine-tuning parameters, k' and ζ. Gain and phase compensation will each require different k' and ζ values ​​to achieve the steady-state specifications, which must be appropriately calculated using the method presented above.

[0031] The gain and phase compensation module 206 generates a compensated final voltage command 216, which is sent to a transformation module 208 for processing. Specifically, the transformation module 208 typically uses an inverse Park transformation or a similar mathematical transformation to convert the measured phase currents in the DC frame (rotor reference frame or d / q reference frame) to the AC frame (stator reference frame), thereby simplifying fault detection analysis. The output of the transformation module 208, as a transformed voltage offset represented in a stator frame, is sent to the fault detection module 210. The fault detection module 210 determines the magnitude of the transformed voltage offset, for example, by comparing it to a threshold value in a stator reference frame, and determines whether an offset error is present in the current measurement.

[0032] The module 106 for diagnosing current measurements provides robust diagnostics for high synchronous frequencies as well as for a full spectrum of synchronous frequencies.

[0033] Module 204, with a programmable high-pass filter, and module 206, with gain and phase compensation, can be digitally implemented to attenuate differences in the magnitude and phase of a continuous-time filter. The digital implementation of module 204, with its programmable high-pass filter, and module 206, with gain and phase compensation, can be based on the synchronous frequency but is implemented in a discrete domain, as shown in the following equation, where ε e a synchronous frequency of a motor of the power steering system and T S a sampling time is: S=ωetan(ωeTs2)1−z−11+2−1,

[0034] Although the module 204 is equipped with a programmable high-pass filter in Fig. 2 is a first-order filter, it is further considered and it is understood that the module 204 can be a programmable high-pass filter of a higher order, for example a second-order filter.

[0035] Fig. Figure 3 illustrates a module for diagnosing phase current measurements of another embodiment in greater detail. As shown, the module 106A for diagnosing phase current measurements includes a voltage difference calculation module 302, a module 304 with a programmable high-pass filter, a gain and phase compensation module 306, a transformation module 308, and a fault detection module 310.

[0036] The stress difference calculation module 302 calculates a stress difference 312 based on the pre-stress command 120 and the final stress command 126. In some embodiments, the stress difference 312 is represented in the d / q domain as a d / q offset by determining the pre-stress command 120 and the final stress command 126. The stress difference 312 can have a constant component and a sinusoidal component, both of which are represented in the d / q reference frame.

[0037] Module 304 with a programmable high-pass filter receives the voltage difference 312. Module 304 with a programmable high-pass filter is designed to filter the voltage difference 312, resulting in the removal of the constant component. Consequently, module 304 with a programmable high-pass filter is designed to generate a filtered voltage difference 314, which contains the sinusoidal component of the voltage difference 312 without the constant component. Therefore, module 304 with a programmable high-pass filter can have a cutoff frequency proportional to a synchronous frequency ω. e of the engine 110 ( Fig. 1) The Module 304 with a programmable high-pass filter can be programmed based on a calibratable constant value. The calibratable constant value can be fine-tuned to reduce gain and phase error in the sinusoidal voltage component. A transfer function of the Module 304 with a programmable high-pass filter can be written as follows: H(s)=ss+kωe where k is the calibratable constant that can be fine-tuned to reduce gain and phase error in the sinusoidal component. Module 304, with a programmable high-pass filter, can set the synchronous frequency ω. eThe gain and phase error introduced by the filter is static. Furthermore, in the module 304 with a programmable high-pass filter, the calibratable constant k can be independently fine-tuned to ensure minimal distortion of the input signal. The module 304 with a programmable high-pass filter is not limited to the above embodiment. For example, a higher-order filter can be used as long as the programmable filter removes the constant component of the voltage difference 312.

[0038] The filtered voltage difference 314 is sent to a gain and phase compensation module 306. The module 304, with a programmable high-pass filter, can introduce magnitude and phase errors into the filtered voltage difference 314, which necessitates compensation by the gain and phase compensation module 306. In some embodiments, the gain and phase compensation module 306 includes a unity-gain, first-order programmable low-pass filter (PLPF) with the following transfer function: L(s)=k'ωes+k'ωe

[0039] The programmable first-order low-pass filter with unity gain can compensate for the filtered voltage difference 314 to restore the gain and phase of the voltage difference 312. The compensation can be performed according to limiting equations written as: ∠H(jωe)−∠L(jωe)=0 |H(jωe)L(jωe)|=1

[0040] In other embodiments, the gain and phase compensation module 306 can use a second-order low-pass filter with a transfer function shown below. L(s)=k'2ωe2s2+2ζk'ωes−k'2ωe2

[0041] The second-order low-pass filter shown above can enable more specific transient responses by using two fine-tuning parameters, k' and ζ. Gain and phase compensation will each require different ζ and k' values ​​to achieve the steady-state specifications, which must be appropriately calculated using the method presented above.

[0042] The gain and phase compensation module 306 generates a compensated voltage difference 316, which is sent to a transformation module 308 for processing. Specifically, the transformation module 308 typically uses an inverse Park transformation or a similar mathematical transformation to convert the measured phase currents in the DC frame (rotor reference frame or d / q reference frame) to the AC frame (stator reference frame), thereby simplifying fault detection analysis. The output of the transformation module 308, as a transformed voltage offset represented in a stator frame, is sent to the fault detection module 310. The fault detection module 310 determines a magnitude of the transformed voltage offset, for example, by comparison with a threshold value in a stator reference frame.The fault detection module 310 can determine, based on the transformed voltage offset, whether an offset error exists in the current measurement.

[0043] The 106A module for diagnosing phase current measurements provides robust diagnostics for high synchronous frequencies as well as for a full spectrum of synchronous frequencies.

[0044] Module 304, with a programmable high-pass filter, and module 306, with gain and phase compensation, can be digitally implemented to attenuate the magnitude and phase difference of a continuous-time filter. The digital implementation of module 304, with its programmable high-pass filter, and module 306, with its gain and phase compensation, can be based on the synchronous frequency but implemented in a discrete domain, as shown in the following equation: s=ωetan(ωeTs2)1−z−11+z−1.

[0045] Fig. Figure 4 shows two records, the first record 402 illustrates V d and V q in the presence of an offset error in current measurements. A result of the processing by module 106 for diagnosing current measurements (e.g. Fig. 3) is shown in a second recording 404. As can be seen, the phase current offset is expressed as voltage offsets in V. α and V β .

[0046] Fig. Figure 5 is a flowchart for a control procedure that is controlled by the torque control module 100 ( Fig. 1) can be carried out in accordance with some embodiments of the invention.

[0047] As can be seen in light of the disclosure, the order of operations within the procedure is not limited to the sequential execution described in Fig.5 is illustrated, but it can be carried out in one or more varying sequences, provided this is applicable and in accordance with the present disclosure.

[0048] In block 510, the torque control module 100 determines a voltage difference between a reference voltage command and a final voltage command in a rotor reference frame. The final voltage difference contains a constant component and a sinusoidal component. In block 520, the torque control module 100 filters the final voltage difference to produce a filtered final voltage difference that includes the sinusoidal component.

[0049] In block 530, the torque control module 100 performs gain compensation and phase compensation on the filtered final voltage difference to generate a compensated final voltage difference. The phase compensation can restore the gain and phase of the voltage difference, for example, using a first- or second-order filter.

[0050] In block 540, the torque control module 100 determines an offset of the compensated final voltage difference in a stator reference frame. The torque control module 100 can also send the direct-axis voltage signal and the quadrature-axis voltage signal to the motor for control purposes. In some embodiments, the torque control module 100 causes a status to be displayed indicating the correctness or incorrectness of the current measurement. For example, the torque control module 100 outputs the status to an operator or technician to diagnose the function of the motor current sensor.

Claims

[1] A system for detecting an offset error in a power steering system, the system comprising: a module (204) having a programmable high-pass filter configured to filter a final voltage command (126) and generate a filtered final voltage command (214) including a sinusoidal component; a gain and phase compensation module (206) configured to perform gain compensation and phase compensation on the filtered final voltage command (214) to generate a compensated final voltage command (216); and an error detection module (210) configured to determine an offset error in a current measurement based on the compensated final voltage command (216) in a stator reference frame. [2] The system of claim 1, wherein the programmable high-pass filter module (204) includes a transfer function based on a synchronous motor frequency and a calibratable constant. [3] The system of claim 2, wherein the calibratable constant is fine-tuned to reduce magnitude and phase error of the sinusoidal component. [4] The system of claim 3, further comprising a transformation module (208) that performs an inverse Park transformation on the compensated final voltage command (216) to transform the compensated final voltage command (216) into the stator reference frame. [5] The system of claim 1, wherein the gain compensation and / or the phase compensation is performed with a first-order filtering technique using a transfer function, the transfer function being written as: L(s)=k'ωes+k'ωe, where k' is a constant value and ω e a synchronous frequency of a motor of the power steering system. [6] The system of claim 1, wherein the gain compensation and the phase compensation are performed in a discrete region. [7] The system of claim 6, wherein the gain compensation and the phase compensation are performed in a discrete range in accordance with: s=ωetan(ωwTe2)1−z11+z−1, are implemented, where ω e is a synchronous frequency of a motor of the power steering system and T S is a sampling time. [8] The system of claim 1, wherein the offset error in a current measurement is further determined based on the final voltage command (126) in a rotor reference frame. [9] A method for detecting an offset error in a power steering system, the method comprising: determining (510) a voltage difference between a reference voltage command and a final voltage command (126) in a rotor reference frame, the voltage difference including a constant component and a sinusoidal component; the voltage difference is filtered (520) to produce a filtered voltage difference containing the sinusoidal component; a gain compensation and a phase compensation are performed on the filtered voltage difference (530) to produce a compensated voltage difference; and an offset error in a current measurement is determined based on the compensated voltage difference in a stator reference frame (540). [10] The method of claim 9, wherein the filtering is based on a transfer function including a synchronous motor frequency and a calibratable constant. [11] The method of claim 10, wherein the calibratable constant is fine-tuned to reduce magnitude and phase error of the sinusoidal component. [12] The method of claim 11, further comprising performing an inverse Park transformation on the compensated voltage difference to transform the compensated voltage difference into the stator reference frame. [13] A method according to claim 9, wherein the gain compensation and / or the phase compensation are carried out with a first-order filtering technique using a transfer function, the transfer function being written as: L(s)=k'ωes+k'ωe, where k' is a constant value and ω e a synchronous frequency of a motor of the power steering system. [14] The method of claim 9, wherein a transformed voltage offset is based on the difference between a bias command and the final voltage command (126) in a rotor reference frame, the offset in a current measurement being determined at least in part based on the transformed voltage offset. [15] A system for detecting an offset error in a power steering system, the system comprising: a voltage difference calculation module (302) configured to determine a voltage difference (312) between a reference voltage command and a final voltage command (126) in a rotor reference frame, the voltage difference (312) including a constant component and a sinusoidal component; a module (304) having a programmable high-pass filter configured to filter the voltage difference (312) and generate a filtered voltage difference (314) containing the sinusoidal component; a gain and phase compensation module (306) configured to perform gain compensation and voltage compensation on the filtered voltage difference (314) to generate a compensated voltage difference (316); and a fault detection module (310) configured to determine an offset of the compensated voltage difference (316) in a stator reference frame.

Images