Method and device for injecting a high-frequency current vector

Abstract

The present invention relates to a method and device for injecting a high frequency current vector at an angle to a measured current vector flowing through a machine, comprising: measuring a motor current vector, determining a projection of the motor current vector on an axis orthogonal to a direction at an angle to the measured motor current vector, demodulating high frequency variations in the projection using a first high frequency demodulation signal, determining at least a first angle from the demodulated high frequency variations, and determining an injection voltage vector from at least the first angle, a predetermined voltage, and a second high frequency modulated signal, the second high frequency modulated signal having the same frequency as the first high frequency demodulation signal and a phase difference of π / 2 with the first high frequency demodulation signal.

Description

Technical Field

[0001] The present invention generally relates to a method and device for injecting a high-frequency current vector having a certain angle with respect to a measured current vector flowing through a machine.

Background Art

[0002] Electromechanical machines are widely used in any industry of factory automation or transportation. Many control techniques of machines as permanent magnet synchronous machines (PMSMs), synchronous reluctance machines (SyncRMs), wound rotor synchronous machines (WRSMs) often obtain the speed and position of the machine as feedback using a rotary encoder.

[0003] Due to the demand for low-cost and robust motor drives, the development of sensorless control is increasing. Without those sensors, the machine drive becomes less expensive and more robust against harsh environments with a lot of dust.

[0004] Many controller techniques rely on the injection of high-frequency signals. The injected high-frequency voltage signal is generally superimposed on the control voltage waveform of the machine, causing high-frequency fluctuations of the magnetic flux in the machine, and ultimately causing high-frequency fluctuations of the current flowing through the machine conductors.

[0005] The relationship between the high-frequency fluctuations of the magnetic flux and the high-frequency fluctuations of the current contains valuable information about the machine inductance. For example, even in the absence of a position sensor, using this information, the position of the rotor of the machine can be determined. In another example, the relationship between the code inductance and the incremental inductance can prompt the selection of the current shoot angle so as to reach the MTPA (maximum torque / current (maximum torque per current)) operating point. In the MTPA state, since torque is generated with a minimum level of current, a minimum level of loss can be obtained.

[0006] The ideal MTPA operating point is typically identified using a look-up table (LUT). The MTPA LUT can be determined manually or created on other inductance maps or flux maps, but if determined, a specific self-commissioning session is also required. LUT-less MTPA identifies the MTPA operating point in the absence of a look-up table.

[0007] Various injection techniques such as sine waves and square waves are known, and some techniques can be combined across different frequencies and axis frameworks and may be able to address different purposes such as circular injection or elliptical injection.

[0008] Injection in the dq framework rotates with the machine's rotor and heads towards the most salient direction / least salient direction of the rotor, so it is widely used to quickly identify incremental inductance in extreme saliency regions. In contrast, injection in the αβ framework is fixed and associated with the stator and thus does not require knowledge of the rotor position.

[0009] High-frequency signal injection can be implemented in voltage form or current form. High-frequency voltage injection is simple, but current injection requires a current regulator to generate a voltage pattern that makes the measured current into the desired high-frequency current injection waveform.

[0010] The current regulator must be designed with a high bandwidth so that it can follow the distance to the desired high-frequency injection signal. This high bandwidth causes a high level of noise for the estimation of the instantaneous injection direction.

[0011] When the injected signal is subsequently used for estimating the rotor position or estimating the ideal shoot angle of the motor current to achieve the MTPA state, any estimation has noise, which may cause instability of the sensorless controller of the machine. Therefore, generally, it is necessary to limit the bandwidth of the speed controller to avoid divergence, and the application range of sensorless control is limited to low-bandwidth applications.

[0012] Since the regulator bandwidth is necessarily limited by the switching frequency of the inverter, it is also impossible to inject a very high-frequency current, for example, a current with a vibration frequency close to the switching frequency.

[0013] In contrast, some LUT-less MTPA techniques rely on injecting a high-frequency current that is exactly perpendicular to the measured current. However, due to the aforementioned problems, the use of high-frequency current injection for LUT-less MTPA is limited in actual applications in both the literature and the industry. Summary of the Invention Problems to be Solved by the Invention

[0014] In order to improve the stability and bandwidth of the sensorless control of the machine in the case without a LUT, the present invention proposes a method of injecting a high-frequency current vector having a certain angle with respect to the measured current vector flowing through the machine.

[0015] An object of the present invention is to provide a method of injecting a high-frequency current vector having a certain angle with respect to the measured current vector flowing through the machine. Means for Solving the Problems

[0016] To that end, the present invention is a method of injecting a high-frequency current vector having a certain angle with respect to the measured current vector flowing through the machine, comprising: measuring a motor current vector; Determining a projection value of a motor current vector on an axis orthogonal to a direction having a constant angle with respect to the measured current vector; Demodulating high-frequency fluctuations of the projection value using a first high-frequency demodulation signal; Determining at least a first angle from the demodulated high-frequency fluctuations; Determining an injection voltage vector from at least the first angle, a predetermined voltage, and a second high-frequency modulation signal, wherein the second high-frequency modulation signal has the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; A method, characterized by comprising the above.

[0017] The present invention also relates to a device for injecting a high-frequency current vector having a constant angle with respect to a measured current vector flowing through a machine, means for measuring a motor current vector; means for determining a projection value of the motor current vector on an axis orthogonal to a direction having a constant angle with respect to the measured current vector; means for demodulating high-frequency fluctuations of the projection value using a first high-frequency demodulation signal; means for determining at least a first angle from the demodulated high-frequency fluctuations; means for determining an injection voltage vector from at least the first angle, a predetermined voltage, and a second high-frequency modulation signal, wherein the second high-frequency modulation signal has the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; A device, characterized by comprising the above.

[0018] Accordingly, the first angle is driven to minimize the projection of the high-frequency current orthogonal to the desired direction. Since the machine is mainly inductive, the resulting high-frequency current necessarily flows in a direction forming a constant angle with respect to the measured current vector flowing through the machine. The disturbance to the controller stability caused by the non-constant angle is significantly reduced, and the overall stability of the controller is improved.

[0019] According to a specific feature, the injection voltage vector is the first angle Ψ and the angle γ of the measured current vector s from which

Number

Number

[0020] Therefore, the voltage vector forms a stable angle Ψ with the measured current vector. By controlling this angle, the high-frequency response of the current is driven in a desired direction having a constant angle with respect to the measured current vector. Since the angle Ψ does not need to follow the rotation with respect to the fundamental frequency of the measured current vector, the constraints on the design of the PI filter are relaxed. The direction of the high-frequency current injection is sufficiently stabilized with respect to the measured current vector. The injection frequency can be increased up to the switching frequency of the inverter. The ability of the controller to respond to high-speed injection is enhanced. For example, the controller can estimate the MTPA and the voltage level required to reach the desired torque more quickly. The performance of the controller is improved overall despite the absence of a look-up table of machine parameters.

[0021] According to a specific feature, the method is demodulating the high-frequency variation of the projection value using a second high-frequency demodulation signal having the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; obtaining a second angle from the high-frequency variation of the projection value demodulated using the second high-frequency demodulation signal; Weighting the first angle and the second angle by respective weights obtained from the demodulated high-frequency variations of the value of the motor current on an axis orthogonal to a direction having a constant angle with respect to the measured current vector, using the first high-frequency demodulated signal and the second high-frequency demodulated signal; Adding the weighted first angle and the weighted second angle; further comprising The required injection voltage vector is obtained from the sum of the weighted first angle and the weighted second angle.

[0022] Accordingly, both the orthogonal component and the in-phase component of the high-frequency current response are minimized orthogonal to the desired injection direction. The high-frequency current response can hold two components including both the in-phase impedance (resistive) and the orthogonal impedance (inductive) due to the complex response of the machine. This method does not require estimating the impedance of the machine and operates similarly at any injection frequency. The signal-to-noise ratio at the desired angle is improved.

[0023] According to a particular feature, the projected value of the motor current on an axis orthogonal to a direction having a constant angle with respect to the measured current vector is the measured angle γ between the measured current vector in the αβ framework and the α axis of the αβ framework s using to convert the measured motor current vector i αβ in the αβ framework to the measured motor current vector i xy in the xy framework. Here, γ s = arctan(i β / i α ), and the xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of the measured angle γ s and a constant angle Δ, and only the measured current on the x-axis is maintained.

[0024] Accordingly, the angle of the current vector in the machine is accurately calculated.

[0025] According to a specific feature, obtaining the first angle from the demodulated high-frequency fluctuations is performed using a proportional-integral regulator having an integral gain of k1 and a proportional gain of a null value.

[0026] Therefore, according to the present invention, it becomes possible to obtain an angle that guarantees that the high-frequency induction component of the motor current vector has a constant angle with respect to the measured current vector flowing through the machine.

[0027] According to a specific feature, obtaining the second angle is performed using a proportional-integral regulator having an integral gain of k1 and a proportional gain of a null value.

[0028] Therefore, according to the present invention, it becomes possible to obtain an angle that guarantees that the high-frequency induction component of the motor current vector has a constant angle with respect to the measured current vector flowing through the machine.

[0029] The features of the present invention will become more apparent by reading the following description of the exemplary embodiments. This description is made with reference to the accompanying drawings.

Brief Description of the Drawings

[0030]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5a

Figure 5b

Figure 6

Figure 7

Figure 8a

Figure 8b

[0031] FIG. 1 shows a first example of direct magnetic flux vector control of a motor using maximum torque / current at which the present invention is implemented.

[0032] Here, it should be noted that the present invention is disclosed in an example where an MTPA module is used. The present invention is also applicable to, for example, obtaining the position of a machine or estimating the parameters of a machine under control.

[0033] In the system shown in FIG. 1, a reference torque T * is supplied to a divider 100 together with a reference flux λ * MTPA provided by an MTPA tracking module 155 and three-halves times the number of pole pairs of a motor 135, and a reference current i * τ on the τ-axis provided to a DFVC module 110 is obtained. NUMERICAL EXPRESSION

[0034] The DFVC module 110 includes a reference current i * τ on the τ-axis and an estimated flux NUMERICAL EXPRESSION The norm, the measured current i in the fτ framework fτ and the reference flux linkage λ from the MTPA tracking module 155 * MTPA are used to obtain the reference voltage v in the fτ framework * fτ .

[0035] The reference voltage v in the fτ framework * fτ is provided to the framework conversion module 115. The framework conversion module 115 uses the estimated load angle

Number

[0036] The reference voltage v in the αβ framework * αβ is provided to the addition module 120. The addition module 120 adds the reference voltage v * αβ to the high-frequency injection voltage

Number

[0037] The modified reference voltage v in the αβ framework ** αβ is provided to the voltage source inverter VSI130 connected to the motor 135. The motor current vector i abc measured in the three-phase abc is provided to the framework conversion module 140

[0038] The framework conversion module 140 converts the motor current i measured in the three-phase abc abc into the motor current vector i αβ measured in the αβ framework.

[0039] The motor current vector i αβ measured in the αβ framework is provided to the flux estimation module 150, the j-axis injection module 125, and the framework conversion module 145.

[0040] The framework conversion module 145 uses the estimated load angle

Number

[0041] The current vector i fτ measured in the fτ framework is provided to the DFVC module 110.

[0042] The injection module 125 determines the injection voltage vector αβ in the αβ framework from the motor current vector i h measured in the αβ framework and the high-frequency signals sin(ω h t) and cos(ω

Number

[0043] The high-frequency injection voltage

Number

[0044] High-frequency injection voltage

Number

[0045] The flux estimation module 150 estimates the flux αβ and the reference voltage v * αβ in the αβ framework from the measured motor current vector i

Number

Number

[0046] The speed estimation module 152 obtains the speed of the motor 135 from the estimated load angle.

[0047] For example, the speed

Number

[0048] The estimated flux in the αβ framework

Number

[0049] The MTPA tracking module 155 estimates the flux in the αβ framework

Number

[0050] FIG. 2 shows an example of a block diagram of a direct flux vector control module.

[0051] The direct flux vector control module 110 includes a subtraction module 220 that subtracts the measured current i * τ on the τ-axis from the reference current i τ on the τ-axis.

[0052] The output of the subtraction module 220 is provided to a PI regulator 225. The output of the PI regulator 225 is provided to a multiplication module 230. The multiplication module 230 multiplies the output of the PI regulator 225 by a decoupling constant value 1 / b.

[0053] The direct flux vector control module 110 includes a subtraction module 200 that subtracts the estimated flux norm

Number

[0054] The outputs of the multiplication modules 230 and 210 are added by an addition module 235.

[0055] The output of the addition module 235 is provided to an addition module 240. The addition module 240 adds the result of the addition performed by the addition module 235

Number

Number

[0056] The stator resistance R s is obtained, for example, from a self-commissioning procedure. The motor speed is the angle of the estimated flux vector

Number

[0057] The output of the PI regulator 205 is provided to the addition module 215. The addition module 215 adds the result of the output of the PI regulator 205 to R s i f and provides the reference voltage v on the f axis * f .

[0058] Figure 3 shows an example of a block diagram of the flux estimation module.

[0059] The flux estimation module 150 includes a subtraction module 300 that subtracts the product of the current i on the α axis and the resistance R α from the reference voltage v on the α axis s multiplied by. * α

[0060] The result of the subtraction module 300 is provided to the subtraction module 305.

[0061] The subtraction module 305 subtracts the result provided by the multiplication module 315 from the result of the subtraction module 300.

[0062] ​The output of the subtraction module 305 is provided to the integrator 310, providing the estimated flux on the α-axis. [Number] is provided.

[0063] The estimated flux on the α-axis [Number] is provided to the divider 320 and the multiplication module 315. The multiplication module 315 multiplies the estimated flux on the α-axis [Number] by the coefficient k which is the observation gain. obs is multiplied.

[0064] The flux estimation module 150 includes a subtraction module 330 that subtracts the product of the current i on the β-axis β and the resistance R of the motor 135 s from the reference voltage v on the β-axis * β .

[0065] The result of the subtraction module 330 is provided to the subtraction module 335.

[0066] The subtraction module 335 subtracts the result provided by the multiplication module 345 from the result of the subtraction module 330.

[0067] The output of the subtraction module 335 is provided to the integrator 340, providing the estimated flux on the β-axis [Number] is provided.

[0068] The estimated flux on the β-axis [Number] is provided to the divider 320 and the multiplication module 345. The multiplication module 345 multiplies the estimated magnetic flux in the β-axis

Number

[0069] The division module 320 divides the estimated magnetic flux in the β-axis

Number

Number

[0070] The result of the division module 320 is converted by the arctangent by the module 350, and the estimated load angle

Number

[0071] Figure 4 shows an example of a block diagram of the maximum torque / current module.

[0072] The maximum torque / current module 155 uses the measured angle γ s between the measured current and the α-axis of the αβ framework to convert the estimated magnetic flux in the αβ framework

Number

Number

[0073] The estimated magnetic flux on the j-axis

Number

Number

[0074] The output of the multiplication module 415 is processed by a low-pass filter 420.

[0075] Modules 415 and 420 form a heterodyne demodulation 410.

[0076] The output of the heterodyne modulation 410 is provided to a PI regulator 425 having an integral gain of k λ and a proportional gain of 0, and a reference magnetic flux λ * MTPA is provided.

[0077] The reference magnetic flux λ * MTPA is thus controlled to ensure that the magnetic flux response to the high-frequency injection voltage is null on the j-axis and thus exists only on the i-axis. The high-frequency magnetic flux response to the injected voltage is aligned with the measured current vector.

[0078] FIG. 5a shows a first example of a block diagram of an injection module according to the present invention.

[0079] The injection module 125 includes a framework conversion module 500 that uses the measured angle γ s between the measured current vectors in the αβ framework to convert the measured motor current vector i αβ in the αβ framework to the measured motor current vector i xy in the xy framework. Here, γ s = arctan(i β / i α) It is so. The xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of the measured angle and a constant angle Δ. Δ is included between 0 and π / 2. Only the measured current on the x-axis is provided to the multiplier 505. In other words, the measured motor current vector is projected onto the x-axis. The measured current on the x-axis corresponds to the projection of the measured current vector in the direction orthogonal to the constant angle Δ. The multiplier 505 multiplies the high-frequency sine wave sin(ω h t) by the measured current on the x-axis to demodulate the high-frequency fluctuations of the norm of the measured motor current vector.

[0080] The output of the multiplier 505 is provided to the low-pass filter 510.

[0081] The output of the low-pass filter 510 is provided to the proportional-integral regulator 515 having an integral gain of k1 and a proportional gain of a null value. The value of k1 is between 0.01 and 0.5 depending on the machine rating, the injection frequency and injection scale, and the sampling frequency. The output of the proportional-integral regulator 515 is the angle Ψ provided to the conversion module 520.

[0082] The conversion module 520

Number

Number

[0083] The output of the conversion module 520 is provided to the multiplier 525. The multiplier 525 multiplies the output of the conversion module 520 by the high-frequency signal v h cos(ω h t) to obtain the voltage injection signal

Number

[0084] Figure 5b shows a second example of a block diagram of an injection module according to the present invention.

[0085] The injection module 125 includes a framework conversion module 550 that uses the measured angle γ between the measured current vector in the αβ framework and the α-axis of the αβ framework to convert the measured motor current vector i in the αβ framework into the measured motor current vector i in the xy framework. Here, γ s = arctan(i αβ / i xy ). The xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of the measured angle and a fixed angle Δ. Δ is included between 0 and π / 2. s = arctan(i β / i α ) and the xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of the measured angle and a fixed angle Δ. Δ is included between 0 and π / 2.

[0086] The measured current on the x-axis is provided to a multiplier 552. The measured current on the x-axis corresponds to the projection of the measured current vector in a direction orthogonal to the fixed angle Δ. The multiplier 552 multiplies the measured current on the x-axis by an orthogonal high-frequency sine wave sin(ω h t) to demodulate the high-frequency fluctuations of the measured current on the x-axis. The output of the multiplier 552 is provided to a low-pass filter 554.

[0087] The low-pass filter 554 has a cut-off frequency lower than the frequency ω h of the injected signal.

[0088] The output of the low-pass filter 554 is provided to a proportional-integral regulator 556 having an integral gain of k1 and a proportional gain of a null value. The value of k1 is between 0.01 and 0.5 depending on the machine rating, the injection frequency and scale, and the sampling frequency. The output of the proportional-integral regulator 556 is the first angle provided to the multiplier 558.

[0089] The measured current on the x-axis, i.e., the current projected onto the x-axis, is provided to multiplier 562. Multiplier 562 multiplies the measured current on the x-axis by a high-frequency cosine waveform cos(ω h t) in phase to demodulate the high-frequency fluctuations of the measured current on the x-axis. The output of multiplier 562 is provided to low-pass filter 564.

[0090] Low-pass filter 564 has a cut-off frequency lower than the frequency ω h of the injected signal.

[0091] The output of low-pass filter 564 is provided to proportional-integral regulator 568 having an integral gain of k1 and a proportional gain of a null value. The value of k1 is between 0.01 and 0.5. The output of proportional-integral regulator 568 is provided to multiplier 570, which is the second angle.

[0092] The demodulated and filtered measured current β Q and β I on the x-axis are provided to weight calculation module 560. Weight calculation module 560 determines a weight α Q to be provided to multiplier 558 and a weight α I to be provided to multiplier 570.

Equation

[0093] The output α Q of weight calculation module 560 is multiplied by the first angle provided by proportional-integral regulator 556 by multiplier 558.

[0094] The output α I of weight calculation module 560 is multiplied by the second angle provided by proportional-integral regulator 568 by multiplier 570.

[0095] The outputs of multiplier 558 and multiplier 570 are added by adder module 572.

[0096] The output of the summing module 572 is the angle Ψ, which is provided to a transformation module 576. The transformation module 576

number

number

[0097] The output of the transform module 576 is provided to a multiplier 578. The multiplier 578 multiplies the output of the transform module 576 by a high frequency signal v h cos(ω h t) to obtain the injection signal voltage in the αβ framework.

number

[0098] FIG. 6 shows a third example of a block diagram of an injection module according to the invention.

[0099] 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 FIG. 8a or 8b.

[0100] The bus 601 links the processor 600 to a read-only memory ROM 602 , a random access memory RAM 603 and an input / output interface I / O IF 605 .

[0101] Input / output interface I / O IF 605 allows injection module 125 to sense signals representative of the current flowing through motor 135 and predetermined signals.

[0102] The memory RAM603 includes registers intended to receive variables and instructions of a program related to an algorithm as disclosed in FIG. 8a or FIG. 8b.

[0103] The read-only memory ROM602, or in some cases a flash memory, contains program instructions related to an algorithm as disclosed in FIG. 8a or FIG. 8b. This program is loaded into the random access memory RAM603 when the injection module 125 is powered on. Alternatively, this program can also be executed directly from the ROM602.

[0104] The calculations performed by the injection module 125 can also be realized in software by executing a set of instructions or programs by a programmable computing machine such as a PC (personal computer), DSP (digital signal processor), or microcontroller, or can be realized in hardware by a machine or dedicated components such as an FPGA (field programmable gate array) or ASIC (application specific integrated circuit).

[0105] In other words, the injection module 125 is provided with a circuit or a device having a circuit for causing the injection module 125 to execute a program related to an algorithm as disclosed in FIG. 8a or FIG. 8b.

[0106] FIG. 7 represents the motor framework used by the present invention.

[0107] FIG. 7 shows an αβ two-phase stator framework. The αβ framework is static with respect to the stator of the motor.

[0108] FIG. 7 shows a dq two-phase rotor framework. The dq framework is dynamic and follows the rotor position θ.

[0109] Fig. 7 shows an ij two-phase current framework. The i-axis follows the current vector and forms an angle γ with the α-axis, while the j-axis is perpendicular to the current vector. s while the j-axis is perpendicular to the current vector.

[0110] Fig. 7 shows an fτ two-phase flux framework. The f-axis follows the estimated flux vector and forms an angle δ with the α-axis, while the τ-axis is perpendicular to the estimated flux vector. s while the τ-axis is perpendicular to the estimated flux vector.

[0111] Fig. 7 shows an xy framework. The y-axis follows a fixed angle Δ with the vector current on the x-axis.

[0112] Fig. 7 shows the high-frequency current response i to high-frequency voltage injection. HF The high-frequency current response i HF is perpendicular to the x-axis and thus forms a fixed angle Δ with the measured current vector flowing through the machine.

[0113] There is a high-frequency response λ of the estimated flux to the high-frequency injection voltage. HF The high-frequency response λ HF is aligned with the i-axis and thus is aligned with the measured current vector.

[0114] Fig. 8a shows a first example of an algorithm for obtaining an injection voltage vector added to a signal for driving a motor according to the present invention.

[0115] This algorithm is disclosed in an example executed by the processor 600 of the injection module 125.

[0116] In step S800, the processor 600 acquires a measured value of the motor current vector.

[0117] In step S801, the processor 600 measures, for example, the measured angle γ between the measured currents in the αβ framework. sUsing αβ , the measured motor current vector i in the αβ framework αβ is converted to the measured motor current vector i in the xy framework xy and by maintaining only the measured current on the x-axis, the projection of the motor current vector in the direction orthogonal to a certain angle Δ is obtained. Here, γ s = arctan(i β / i α ), and the xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of the measured angle and a certain angle Δ.

[0118] In step S802, the processor 600 demodulates the high-frequency variation of the measured current on the x-axis using the first high-frequency demodulation signal sin(ω h t).

[0119] In step S803, the processor 600 performs first low-pass filtering of the demodulated high-frequency variation of the measured current on the x-axis using the first high-frequency signal.

[0120] In step S804, the processor 600 obtains at least a first angle Ψ from the measured current on the x-axis that has been demodulated and first low-pass filtered by the first high-frequency signal using, for example, a proportional-integral regulator having an integral gain of k1 and a proportional gain of a null value.

[0121] In step S805, the processor 600 obtains an injection voltage vector from at least the first angle, a predetermined voltage, and the first high-frequency modulation signal. The first high-frequency modulation signal has the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal.

[0122] For example, the processor 600

Equation

Number

Number

[0123] Figure 8b shows a second example of an algorithm for obtaining the injection voltage vector applied to the signal for driving the motor according to the present invention.

[0124] This algorithm is disclosed in an example executed by the processor 600 of the injection module 125.

[0125] In step S850, the processor 600 acquires the measured value of the motor current vector.

[0126] In step S851, the processor 600, for example, uses the measured angle γ s between the measured currents in the αβ framework to obtain the projection of the motor current vector in the direction orthogonal to a certain angle Δ by converting the measured current motor vector i αβ in the αβ framework to the measured current motor vector i xy in the xy framework and maintaining only the measured current on the x-axis. Here, γ s =arctan(i β / i α ) and the xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of a certain angle Δ and the measured angle.

[0127] In step S852, the processor 600 applies the first high-frequency demodulation signal sin(ωh Using (t), the high-frequency fluctuations of the measured current on the x-axis are demodulated.

[0128] In step S853, the processor 600 performs first low-pass filtering of the high-frequency fluctuations of the measured current on the x-axis, which are demodulated using the first high-frequency demodulation signal.

[0129] In step S854, the processor 600 determines a first angle from the high-frequency fluctuations of the measured current on the x-axis, which are demodulated by the first high-frequency demodulation signal and first low-pass filtered, using, for example, a proportional-integral regulator having an integral gain of k1 and a proportional gain of a null value.

[0130] In step S855, the processor 600 demodulates the high-frequency fluctuations of the measured current on the x-axis using the second high-frequency demodulation signal cos(ω h t).

[0131] In step S856, the processor 600 performs second low-pass filtering of the high-frequency fluctuations of the measured current on the x-axis, which are demodulated using the second high-frequency demodulation signal.

[0132] In step S857, the processor 600 determines a second angle from the high-frequency fluctuations of the measured current on the x-axis, which are demodulated by the second high-frequency demodulation signal and second low-pass filtered, using, for example, a proportional-integral regulator having an integral gain of k1 and a proportional gain of a null value.

[0133] In step S858, the processor 600 determines weights α I and α Q using the demodulated and filtered measured current β I and β Q on the x-axis, respectively provided by low-pass filtering steps S853 and S856.

Equation

[0134] In step S859, the processor 600 multiplies the first angle by the weight α Q .

[0135] In step S860, the processor 600 multiplies the second angle by the weight α I .

[0136] In step S861, the processor 600 adds the output of step S859 and the output of step S860.

[0137] The output of the addition step is the injection angle Ψ. This injection angle is converted by the processor 600 in step S862 using

Number

Number

Number

[0138] Of course, many changes can be made to the embodiments of the present invention described above without departing from the scope of the present invention.

Claims

1. A method of injecting a high-frequency current vector having a constant angle with respect to a measured motor current vector flowing through a machine, comprising: measuring the motor current vector; obtaining a projection value of the motor current vector on an axis orthogonal to a direction having a constant angle with respect to the measured motor current vector; demodulating high-frequency fluctuations of the projection value using a first high-frequency demodulation signal; obtaining at least a first angle from the demodulated high-frequency fluctuations; obtaining an injection voltage vector from at least the first angle, a predetermined voltage, and a second high-frequency modulation signal, wherein the second high-frequency modulation signal has the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; and the injection voltage vector is obtained by performing a conversion from the first angle Ψ and the angle γs of the measured motor current vector 【Number 1】 and multiplying this by a high-frequency signal vhcos(ωht), where J is the matrix 【Number two】 and vh is the amplitude for obtaining the injection voltage vector in the αβ framework, characterized by the method.

2. The method further comprises: demodulating high-frequency fluctuations of the projection value using a second high-frequency demodulation signal having the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; obtaining a second angle from the high-frequency fluctuations of the projection value demodulated using the second high-frequency demodulation signal; weighting the first angle and the second angle by respective weights obtained from the demodulated high-frequency fluctuations of the value of the motor current vector on an axis orthogonal to the direction having the constant angle with respect to the measured current vector, using the first high-frequency demodulation signal and the second high-frequency demodulation signal; adding the weighted first angle and the weighted second angle; and the obtained injection voltage vector is obtained from the sum of the weighted first angle and the weighted second angle, characterized by the method according to claim 1.

3. The projection value of the motor current vector on the axis orthogonal to the direction having the fixed angle with respect to the measured current vector is the measured angle γ between the measured motor current vector in the αβ framework and the α axis of the αβ framework ｓ is used to convert the measured motor current vector i αβ in the αβ framework to the measured motor current vector i ｘｙ in the xy framework, and is obtained by maintaining only the measured current on the x axis, where γ ｓ = arctan(i β / i α ), and the xy framework is rotated from the αβ framework by an angle obtained by subtracting π / 2 from the sum of the measured angle γ ｓ and the fixed angle Δ. The method according to claim 1 or 2, characterized in that

4. Obtaining the first angle from the demodulated high-frequency variation is performed using a proportional-integral regulator having an integral gain of k １ and a proportional gain of a null value, the method according to claim 1 or 2.

5. Obtaining the second angle is performed using a proportional-integral regulator having an integral gain of k １ and a proportional gain of a null value, the method according to claim 2, characterized in that it is performed.

6. A device for injecting a high-frequency current vector having a constant angle with respect to a measured motor current vector flowing through a machine, comprising: means for measuring the motor current vector; Means for obtaining a projection value of the motor current on an axis orthogonal to a direction having a fixed angle with respect to the measured motor current vector; Means for demodulating high-frequency fluctuations of the projection value using a first high-frequency demodulation signal; Means for obtaining at least a first angle from the demodulated high-frequency fluctuations; Means for obtaining an injection voltage vector from at least the first angle, a predetermined voltage, and a second high-frequency modulation signal, wherein the second high-frequency modulation signal has the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; Comprising; The injection voltage vector is obtained from the first angle Ψ and the angle γs of the measured motor current vector, 【Number 3】 by performing the transformation of, and multiplying this by a high-frequency signal vhcos(ωht), where J is the matrix 【Number 4】 and vh is the amplitude for obtaining the injection voltage vector in the αβ framework, a device characterized by this.

7. The device is Means for demodulating high-frequency fluctuations of the projection value using a second high-frequency demodulation signal having the same frequency as the first high-frequency demodulation signal and a phase difference of π / 2 from the first high-frequency demodulation signal; Means for obtaining a second angle from the high-frequency fluctuations of the projection value demodulated using the second high-frequency demodulation signal; Means for weighting the first angle and the second angle by respective weights obtained from the demodulated high-frequency fluctuations of the value of the motor current vector on an axis orthogonal to a direction having a fixed angle with respect to the measured current vector, using the first high-frequency demodulation signal and the second high-frequency demodulation signal; Means for adding the weighted first angle and the weighted second angle; Further comprising; The obtained injection voltage vector is obtained from the sum of the weighted first angle and the weighted second angle, the device according to claim 6, characterized by this.

Images