Patent Application
20240313680
The present invention relates generally to a method and a device for injecting a high frequency current vector which has a fixed angle with the measured current vector flowing in the machine.
Electrical machines are widely used on the industry either for factory automation or transportation. Many control techniques for machines as Permanent Magnet Synchronous Machines (PMSM), Synchronous Reluctance Machines (SyncRM), Wounded Rotor Synchronous Machines (WRSM) often use a rotary encoder for obtaining the speed and the position of the machine as feedback.
The demands for low-cost and robust motor drives have increased the development of sensorless control. Without those sensors the machine drives become less expensive and more robust to dusty and harsh environments.
Many controller technologies rely on high frequency signal injection. The injected HF voltage signal is generally superposed to the control voltage waveforms of the machine, causing HF variations of flux in the machine, finally causing HF variations of currents flowing through the machine conductors.
The relation between HF variations of the flux and of the current contains precious information on machine inductances. As example, in absence of position sensor, these information can be used to determine the position of the rotor of the machine. As another example, relationship between chord and incremental inductances can drive the choice of current shoot angle, so as to reach MTPA (Maximum Torque Per ampere) operation point. MTPA conditions produce torque with minimal level of current, thus of losses.
Ideal MTPA operation point is typically identified using lookup tables (LUT). MTPA LUT may be determined manually, or may be built on other inductance or flux maps, which determination also requires specific self commissioning sessions. LUT-less MTPA consists in identifying MTPA operation point in absence of lookup table.
Different injection techniques are known such as sinusoidal, square, and several techniques can be combined across different frequencies and axis frameworks, possibly addressing different purposes, e.g. leading to circular or elliptic injection.
Injection in dq framework, which rotates together with the rotor of the machine and heads to most/least salient directions of the rotor, is popular as it can get used to quickly identify incremental inductances in extreme saliency regions. In contrast, injection in alpha-beta framework, that is fixed and related to stator, requires no knowledge on rotor position.
HF signal injection can be implemented as voltage or current form. While HF voltage injection is straightforward, current injection requires a current regulator, that produces a voltage pattern to drive the measured current towards the desired HF current injection waveform.
The current regulator must be designed with high bandwidth, to enable the tracking of the distance to the desired HF injection signal. This high bandwidth causes a high level of noise on the estimation of instantaneous direction of injection.
If the injected signal is then used for the estimation of rotor position, or for the estimation of ideal shoot angle of motor current to realise MTPA conditions, both noisy estimations can cause instability of the sensorless controller of the machine. This generally requires to limit the bandwith of speed controller to avoid divergence and limits the applicability of sensorless control to low bandwidth applications.
As the regulator bandwidth is necessarily limited by the switching frequency of the inverter, it is also impossible to inject a very high frequency current, e.g. oscillating close to the switching frequency.
In contrast, some LUT-less MTPA technologies rely on HF current injection strictly perpendicular to the measured current. However, aforementioned problems limit the use of HF current injection for LUT-less MTPA in practical applications in both literature and industry.
To improve the stability and bandwith of the sensorless control of the machine in absence of LUT, the invention proposes a method for injecting a high frequency current vector which has a fixed angle with the measured current vector flowing in the machine.
The present invention aims to provide a method for injecting a high frequency current vector which has a fixed angle with the measured current vector flowing in the machine.
To that end, the present invention concerns a method for injecting a high frequency current vector which has a fixed angle with the measured current vector flowing in the machine, characterized in that the method comprises the steps of:
The present invention concerns also a device for injecting a high frequency current vector which has a fixed angle with the measured current vector flowing in the machine, characterized in that the device comprises:
Thus, the first angle is driven to minimise the projection of HF current orthogonal to the desired direction. As the machine is mainly inductive, the resulting HF current necessarily flows in the desired direction, forming the fixed angle with the measured current vector flowing in the machine. The perturbations on controller stability, resulting from non constant angle, are much reduced, globally increasing the stability of the controller.
According to a particular feature, the injection voltage vector is determined from the first angle Ψ and the angle of measured current vector γs by executing an eJψ+βtransform where J is the matrix
that is multiplied by the high frequency signal νh (ωht) where νh is the amplitude of the voltage injection in order to obtain the injection voltage vector in the αβ framework.
Thus, the voltage vector forms a stable angle Ψ with the measured current vector. This angle is controlled to drive the HF response of the current in the desired direction having a fixed angle with the measured current vector. The angle Ψ does not need to follow the rotation of the measured current vector with the fundamental frequency, relaxing the constraints on the design of a PI filter. The direction of high frequency current injection is well stabilised with respect to measured current vector. The injection frequency can get increased, up to the switching frequency of the inverter. The ability of the controller to react to fast injection is increased, e.g. it can estimate more quickly the MTPA and the voltage levels needed to reach a desired torque. The controller performance is globally improved despite of absence of lookup tables of machine parameters.
According to a particular feature, the method further comprises the steps of:
Thus, both the quadrature and in phase components of the HF current response are minimised orthogonally to the desired injection direction. The HF current response can hold the two components due to the complex response of the machine, that contains both in phase (resistive) and quadrature impedances (inductive). The method needs no assumption on the impedance of the machine and equally operates at any injection frequency. The signal to noise ratio on the desired angle is improved.
According to a particular feature, the projection value of the motor current on the axis that is orthogonal to the direction having the fixed angle with the measured current vector is determined by transforming the measured current motor vector iαβ in the αβ framework into a measured current motor vector ixy in the xy framework using a measured angle γs between the measured current vector in the αβ framework and the α axis of αβ framework where
the xy framework being rotated from the αβ framework by an angle equal to sum of the measured angle γx and the fixed angle Δ minus Π/2, and by keeping only the measured current in the x axis.
Thus, the angle of the current vector in the machine is precisely calculated.
According to a particular feature, the determining of the first angle from the demodulated high frequency variation is performed using a Proportional Integral regulator with an integral gain of k1 and a proportional gain equal to null value.
Thus, the present invention enables to obtain an angle that ensures that the high frequency inductive component of the motor current vector has a fixed angle with respect to the measured current vector flowing in a machine.
According to a particular feature, the determining of the second angle is performed using a Proportional Integral regulator with the integral gain of k1 and a proportional gain equal to null value.
Thus, the present invention enables to obtain an angle that ensures that the high frequency inductive component of the motor current vector has a fixed angle with respect to the measured current vector flowing in a machine.
The characteristics of the invention will emerge more clearly from a reading of the following description of example embodiments, the said description being produced with reference to the accompanying drawings.
It has to be noted here that the present invention is disclosed in an example wherein an MTPA module is used. The present invention is also applicable for example for determining the position of the machine or estimating parameters of the machine under control.
In the system shown in
The DFVC module 110, from the reference current i*τ in the τ axis, an estimated flux {circumflex over (λ)} norm, a measured current ifτ in a fτ framework and the reference flux λ*MTPA from the MTPA tracking module 155, determine a reference voltage ν*fτ in the fτ framework.
The reference voltage ν*fτ in in the fτ framework is provided to a framework transformation module 115 that transforms the reference voltage v*fτ in the fτ framework into a reference voltage ν*αβ in the αβ framework using an estimated load angle {circumflex over (δ)}s.
The reference voltage ν*αβ in the αβ framework is provided to a summation module 120 that sums the reference voltage ν*αβ with high frequency injection voltage ναβ
The modified reference voltage ν**αβ in the αβ framework is provided to a voltage source inverter VSI 130 that is connected to the motor 135. The motor current vector iabc measured in the three phases abc is provided to a framework transformation module 140.
The framework transformation module 140 transforms the motor current iabc measured in the three phases abc in a measured motor current vector iαβ in the αβ framework.
The measured current motor vector iαβ in the αβ framework is provided to a flux estimation module 150, a j-axis injection module 125 and to a framework transformation module 145.
The framework transformation module 145 transforms the measured current motor vector iαβ in the αβ framework into a measured current motor vector ifτ the fτ framework using an estimated load angle {circumflex over (δ)}s.
The measured current vector ifτ in the fτ axis is provided to the DFVC module 110.
The injection module 125 determines the injection voltage vector ναβ
The high frequency injection voltage (ναβ
The high frequency injection voltage (ναβ
The flux estimation module 150 determines, from the measured current motor vector iαβ in the αβ framework and the voltage references ν*αβ, the estimated flux {circumflex over (λ)}αβ in the αβ framework and the estimated load angle {circumflex over (δ)}s.
The speed estimation module 152 determines the speed of the motor 135 from the estimated load angle.
For example, the speed {circumflex over (ω)} is estimated using a phase lock loop and a low pass filtering of the output of the phase lock loop.
The estimated flux {circumflex over (λ)}αβ in the αβ framework is provided to a MTPA tracking module 155 and to the DFVC module 110.
The MTPA tracking module 155 determines, from the estimated flux {circumflex over (λ)}αβ in the αβ framework and the high frequency sinewave signal sin(ωht), the reference flux λ*MTPA.
The direct flux vector control module 110 comprises a subtracting module 220 that subtracts from the reference current i*τ in the τ axis, the measured current iτ in the τ axis.
The output of the subtracting module 220 is provided to a PI regulator 225 the output of which is provided to a multiplication module 230 that multiplies the output of the PI regulator 225 by a decoupling constant value 1/b.
The direct flux vector control module 110 comprises a subtracting module 200 that subtracts from the reference flux λ*MTPA, the estimated flux norm
The output of the subtracting module 200 is provided to a PI regulator 205 the output of which is provided to a multiplication module 210 that multiplies the output of the PI regulator 255 by a decoupling constant value −a/b that is dependent of the motor. As example, values for a and b are a=4 and b=13.
The outputs of the multiplication modules 230 and 210 are summed by a summation module 235.
The output of the summation module 235 is provided to a summation module 240 that sums the result of the summation performed by the summation module 235 to Rxiτ+{circumflex over (λ)}{circumflex over (ω)} in order to provide the reference voltage ν*τ in the τ axis, where Rx is the stator resistance, and {circumflex over (ω)} is the estimated motor speed.
The Stator resistance Rs is for example obtained from a self-commissioning procedure. The motor speed is estimated from the variations of the angle {circumflex over (δ)}s of the estimated flux vector.
The output of the PI regulator 205 is provided to a summation module 215 that sums the result of the output of the PI regulator 205 to Rsif in order to provide the reference voltage ν*f in the f axis.
The flux estimation module 150 comprises a subtracting module 300 that subtracts from the reference voltage ν*α in the α axis the current iα in the α axis multiplied by the resistance Rs.
The result of the subtracting module 300 is provided to a subtracting module 305.
The subtracting module 305 subtracts from the result of the subtracting module 300 a result provided by a multiplication module 315.
The output of the subtracting module 305 is provided to an integrator 310 in order to provide the estimated flux {circumflex over (λ)}α in the α axis.
The estimated flux {circumflex over (λ)}α in the α axis is provided to a divider 320 and to the multiplication module 315 that multiplies the estimated flux {circumflex over (λ)}α in the α axis by a coefficient kobs that is an observation gain.
The flux estimation module 150 comprises a subtracting module 330 that subtracts from the reference voltage ν*β in the β axis the current iβ in the β axis multiplied by the resistance Rs of the motor 135.
The result of the subtracting module 330 is provided to a subtracting module 335.
The subtracting module 335 subtracts from the result of the subtracting module 330 a result provided by a multiplication module 345.
The output of the subtracting module 335 is provided to an integrator 340 in order to provide the estimated flux {circumflex over (λ)}β in the β axis.
The estimated flux {circumflex over (λ)}β in the β axis is provided to the divider 320 and to the multiplication module 345 that multiplies the estimated flux {circumflex over (λ)}β in the β axis by the coefficient kobs.
The division module 320 divides the estimated flux {circumflex over (λ)}β in the β axis by the estimated flux {circumflex over (λ)}α in the α axis.
The result of the division module 320 is transformed by arctangent by the module 350 in order to provide the estimated load angle {circumflex over (δ)}s.
The maximum torque per ampere module 155 comprises a framework transformation module 400 that transforms the estimated flux {circumflex over (λ)}αβ in the αβ framework into an estimated flux {circumflex over (λ)}ij in the ij framework using a measured angle γs between the measured currents and the α axis of αβ framework.
The estimated flux {circumflex over (λ)}j in the j axis is provided to multiplier 415 that multiplies the estimated flux {circumflex over (λ)}j in the j axis by the high frequency sinewave signal sin(ωht).
The output of the multiplication module 415 is processed by a low pass filter 420.
Modules 415 and 420 form a heterodyne demodulation 410.
The output of the heterodyne modulation 410 is provided to a PI regulator 425 with an integral gain of kλ and proportional gain zero in order to provide the reference flux λ*MTPA.
The reference flux λ*MTPA is thus controlled to ensure that the flux response to the high frequency injection voltage is null on the j axis, and thus only located on the i axis. The high frequency flux response to the injected voltage is aligned with the measured current vector.
The injection module 125 comprises a framework transformation module 500 that transforms the measured current motor vector iαβin the αβ framework into a measured current motor vector ixy in the xy framework using a measured angle γs between the measured current vector in the αβ framework where
The xy framework is rotated from the αβ framework by an angle equal to sum of the measured angle and a fixed angle Δ minus Π/2. Δ is comprised between 0 and Π/2. Only the measured current in the x axis is provided to a multiplier 505. In other words, the measured current motor vector is projected in the x axis. The measured current in the x axis corresponds to the projection of the measured current vector in the direction orthogonal to the fixed angle Δ. The multiplier 505 multiplies the high frequency sinewave sin(ωht) to the measured current in the x axis in order to demodulate the high frequency variations of the norm of the measured current motor vector.
The output of the multiplier 505 is provided to a low pass filter 510.
The output of the low pass filter 510 is provided to a Proportional Integral regulator 515 with an integral gain of k1 which value is between 0.01 and 0.5 depending on the machine ratings, the injection frequency and magnitude and the sampling frequency, and a proportional gain equal to null value. The output of the proportional integral regulator 515 is an angle Ψ that is provided to a transformation module 520.
The transformation module 520 performs an eJψ+γs transform where J is the matrix
The output of the transformation module 520 is provided to a multiplier 525. The multiplier 525 multiplies the output of the transformation module 520 by the high frequency signal νh cos(ωht) where νh is the amplitude of the voltage injection in order to obtain the voltage injection signal ναβ
The injection module 125 comprises a framework transformation module 550 that transforms the measured current motor vector iαβ in the αβ framework into a measured current motor vector ixy in the xy framework using a measured angle γs between the measured current vector in the αβ framework and the a axis of αβ framework where
The xy framework is rotated from the αβ framework by an angle equal to sum of the measured angle and the fixed angle Δ minus Π/2. A is comprised between 0 and Π/2.
The measured current in the x axis is provided to a multiplier 552. The measured current in the x axis corresponds to the projection of measured current vector in the direction orthogonal to the fixed angle Δ. The multiplier 552 multiplies the quadrature high frequency sine waveform sin(ωht) to the measured current in the x axis in order to demodulate the high frequency variations of the measured current in the x axis. The output of the multiplier 552 is provided to a low pass filter 554.
The low pass filter 554 has a cut frequency which is smaller than the frequency ωh of the injected signal.
The output of the low pass filter 554 is provided to a Proportional Integral regulator 556 with an integral gain of k1, which value is between 0.01 and 0.5 depending on the machine ratings, the injection frequency and magnitude and the sampling frequency, and a proportional gain equal to null value. The output of the proportional integral regulator 556 is a first angle that is provided to a multiplier 558.
The measured current in the x axis, or in other words projected in the x axis, is provided to a multiplier 562. The multiplier 562 multiplies the in-phase high frequency cosine waveform cos(ωht) to the measured current in the x axis in order to demodulate the high frequency variations of the measured current in the x axis. The output of the multiplier 562 is provided to a low pass filter 564.
The low pass filter 564 has a cut frequency which is smaller than the frequency ωh of the injected signal ωh.
The output of the low pass filter 564 is provided to a Proportional Integral regulator 568 with an integral gain of k1, the value of which is between 0.01 to 0.5 and a proportional gain equal to null value. The output of the proportional integral regulator 568 is provided to a multiplier 570 and is a second angle.
The demodulated and filtered measured current in the x axis βQ and βI is provided to a weight calculation module 560 that determines a weight αQ that is provided to the multiplier 558 and a weight αQ that is provided to the multiplier 558.
The output αQ of the weight calculation module 560 is multiplied to the first angle provided by the proportional integral regulator 556 by a multiplier 558.
The output αI of the weight calculation module 560 is multiplied to the second angle provided by the proportional integral regulator 568 by a multiplier 570.
The outputs of the multipliers 558 and 570 are summed by a summation module 572.
The output of the summation module 572 is an angle Π that is provided to a transformation module 576. The transformation module 576 performs an eJψ+γs transform where J is the matrix
The output of the transformation module 576 is provided to a multiplier 578. The multiplier 578 multiplies the output of the transformation module 574 by the high frequency signal νh cos(ωht) where νh is the voltage injection amplitude in order to obtain the injection signal voltage ναβ
The injection module 125 has, for example, an architecture based on components connected by a bus 601 and a processor 600 controlled by a program as disclosed in
The bus 601 links the processor 600 to a read only memory ROM 602, a random access memory RAM 603, an input output I/O IF interface 605.
The input output I/O IF interface 605 enables the injection module 125 to sense signals representative of current flowing through the motor 135 and predetermined signals.
The memory 603 contains registers intended to receive variables and the instructions of the program related to the algorithm as disclosed in
The read-only memory, or possibly a Flash memory 602, contains instructions of the programs related to the algorithm as disclosed in
The calculation performed by the injection module 125 may be implemented in software by execution of a set of instructions or program by a programmable computing machine, such as a PC (Personal Computer), a DSP (Digital Signal Processor) or a microcontroller; or else implemented in hardware by a machine or a dedicated component, such as an FPGA (Field-Programmable Gate Array) or an ASIC (Application-Specific Integrated Circuit).
In other words, the injection module 125 includes circuitry, or a device including circuitry, causing the injection module 125 to perform the program related to the algorithm as disclosed in
In
The dq two phase rotor framework is represented. The dq framework is dynamic and follows the rotor position θ.
The ij two phase current framework is represented. The i axis follows the current vector, forming an angle γs with the α axis, while the j axis is perpendicular to the current vector.
The fτ two phase flux framework is represented. The f axis follows the estimated flux vector, forming an angle γS with the α axis, while the τ axis is perpendicular to the estimated flux vector.
The xy framework is represented. The y axis follows the fixed angle Δ with the vector current in the x axis.
The HF current response iHF to the high frequency voltage injection is represented. The HF current response iHF is perpendicular to the x axis, thus forming the fixed angle Δ with the measured current vector flowing in the machine.
The HF response λHF of the estimated flux to the high frequency injection voltage is represented. The HF response λHF is aligned with the i axis, thus aligned with the measured current vector.
The present algorithm is disclosed in an example wherein it is executed by the processor 600 of the injection module 125.
At step S800, the processor 600 obtains measurements of the motor current vector.
At step S801, the processor 600 determines the projection of the motor current vector in the direction orthogonal to the fixed angle Δ for example by transforming the measured current motor vector iαβ in the αβ framework into a measured current motor vector ixy in the xy framework using a measured angle γs between the measured currents in the αβ framework where
the xy framework being rotated from the αβ framework by an angle equal to sum of the measured angle and the fixed angle Δ minus Π/2, and keeping only the measured current in the x axis.
At step S802, the processor 600 demodulates the high frequency variations of the measured current in the x axis using a first high frequency demodulation signal sin(ωht).
At step S803, the processor 600 executes a first low pass filtering of the demodulated high frequency variations of the measured current in the x axis using the first high frequency signal.
At step S804, the processor 600 determines at least a first angle Π from the first low pass filtered demodulated measured current in the x axis by the first high frequency signal, for example using a Proportional Integral regulator with an integral gain of k1 and a proportional gain equal to null value.
At step S805, the processor 600 determines the injection voltage vector at least from the first angle, a predetermined voltage and a first high frequency modulating signal, the first high frequency modulating signal having the same frequency as the first high frequency demodulating signal and a phase difference with the first high frequency demodulating signal that is equal to Π/2.
For example, the processor 600 determines the injection voltage vector by executing an eJψ+γs transform where J is the matrix
that is multiplied by the high frequency signal νh cos(ωht) where νh is the amplitude of the voltage injection in order to obtain the injection voltage vector ναβ
The present algorithm is disclosed in an example wherein it is executed by the processor 600 of the injection module 125.
At step S850, the processor 600 obtains measurements of the motor current vector.
At step S851, the processor 600 determines the projection of the motor current vector in the direction orthogonal to the fixed angle Δ for example by transforming the measured current motor vector iαβ in the αβ framework into a measured current motor vector ixy in the xy framework using a measured angle γs between the measured currents in the αβ framework wherein
the xy framework being rotated from the αβ framework by an angle equal to sum of the measured angle and the fixed angle Δ minus Π/2 and keeping only the measured current in the x axis.
At step S852, the processor 600 demodulates the high frequency variations of the measured current in the x axis using a first high frequency demodulation signal sin(ωht).
At step S853, the processor 600 executes a first low pass filtering of the demodulated high frequency variations of the measured current in the x axis using the first high frequency demodulation signal.
At step S854, the processor 600 determines a first angle from the first low pass filtered demodulated high frequency variations of the measured current in the x axis by the first high frequency demodulation signal, for example using a Proportional Integral regulator with an integral gain of k1 and a proportional gain equal to null value.
At step S855, the processor 600 demodulates the high frequency variations of the measured current in the x axis using a second high frequency demodulation signal cos(ωht).
At step S856, the processor 600 executes a second low pass filtering of the demodulated high frequency variations of the measured current in the x axis using the second high frequency demodulation signal.
At step S857, the processor 600 determines a second angle from the second low pass filtered demodulated high frequency variations of the measured current in the x axis by the second high frequency demodulation signal, for example using a Proportional Integral regulator with an integral gain of k1 and a proportional gain equal to null value.
At step S858, the processor 600 using the demodulated and filtered measured currents in the x axis β1 and βIQ provided respectively by the low pass filtering steps S853 and S856 determines weights α1 and αQ.
At step S859, the processor 600 multiplies the first angle by the weight
At step S860, the processor 600 multiplies the second angle by the weight αQ.
At step S861, the processor 600 sums the outputs of the steps S859 and S860.
The output of the summation step is an injection angle Π that is transformed at step S862 by the processor 600 using an eJψ+γs transform where J is the matrix
and modulated by the second high frequency signal νh cos(ωht) where νh is the voltage injection amplitude in order to obtain the injection signal voltage ναβ
Naturally, many modifications can be made to the embodiments of the invention described above without departing from the scope of the present invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 21305975.1 | Jul 2021 | EP | regional |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2022/023029 | 6/1/2022 | WO |
