MTPA based method for parameterless and position-sensorless control of a permanent magnet synchronous motor

Abstract

A parameterless and position-sensorless MTPA control of a permanent magnet synchronous motor including: using three rotating reference frames having different observation angles to parse the current vector; using a target current value and a preset current-rotor angle y that is between the current vector and the qr-axis of the (dr, qr) rotor reference frame to obtain the angles between the current vector, the voltage vector, and the rotor position; obtaining the target voltage value and the target voltage angle by using the obtained angles to obtain the target phase voltage values for regulation. The method is simple in controlling the motor, improves the control efficiency and reliability, and improves the control accuracy.

Description

BACKGROUND

The invention relates to a maximum torque per ampere (MTPA) based method for a position-sensorless control (which is carried out without a position sensor) of a permanent magnet synchronous motor.

Methods for a position-sensorless control of a permanent magnet synchronous motor typically include a constant-torque control mode, a constant-speed control mode, and a constant-air-volume control mode.

For example, U.S. Pat. No. 7,525,269 discloses an apparatus for a position-sensorless control of a permanent magnet synchronous motor drive system, in which a current-torque control is provided for performing a constant torque control.

Chinese patent publication No. 103929109A discloses a constant speed control method for a position-sensorless control of a permanent magnet synchronous motor.

FIG. 1 shows a block diagram of a conventional constant torque control. Referring to FIG. 1, the torque T is related only to the q-axis current, and thus a preset torque T can be converted to a preset q-axis current iq0 according to the formula T=K×iq0. Then a PI controller with the requisite corrective behavior can be used to perform a closed-loop control of the q-axis current, thus achieving the constant torque control.

The conventional field-oriented control (FOC) methods for a position-sensorless control of a permanent magnet synchronous motor are typically based on the rotor frame. The algorithm for deriving the rotor position in the FOC method highly depends on the accuracy of the motor parameters (e.g., the resistance Rs, the q-axis inductance Lq, the d-axis inductance Ld, and the magnetic flux λm), thus results in a large error due to for example the change in the rotor temperature. Also, the vector control algorithm in the FOC method is complicated, which requires time-consuming computation and large sources in the microcontroller unit (MCU). Therefore, a high-demanded microcontroller unit (MCU) is required to operates the FOC method, making the motor control costly. Moreover, because the FOC method is highly dependent on the accuracy of the motor parameters such as the resistance Rs, the q-axis inductance Lq, the d-axis inductance Ld, and the magnetic flux λm, the motor control based on the FOC method has a relative narrow range of applications.

SUMMARY

In view of the above-described problems, the disclosure provides a MTPA based control method for a parameterless and position-sensorless control of a permanent magnet synchronous motor. The method is operated without a position sensor (i.e., position-sensorless control) and does not depend on the motor parameters such as the resistance Rs, the q-axis inductance Lq, the d-axis inductance Ld, and the magnetic flux λm (i.e., parameterless control); its control algorithm is more uncomplicated and time-saving, and is more cost-saving because it does not require a demanded microcontroller unit (MCU).

To achieve the above objectives, in accordance with one embodiment of the invention, there is provided a MTPA based control method for a parameterless and position-sensorless control of a permanent magnet synchronous motor, the method comprising:

1). calculating a target d_r-axis current value I_dr and a target q_r-axis current value I_qr of a target current vector {right arrow over (Idq)} in a (d_r, q_r) rotor reference frame by using a target current value I_dq of the current vector {right arrow over (Idq)} and a preset current-rotor angle γ; in which the preset current-rotor angle y is measured between the current vector {right arrow over (Idq)} and the q_r-axis of the (d_r, q_r) rotor reference frame; the (d_r, q_r) rotor reference frame is a first (d, q) rotating reference frame that is rotating in synchronism with the rotating magnetic field of the rotor, and the rotor position (i.e., the north polar of the magnetic field of the rotor) is aligned with the d_r-axis of the (d_r, q_r) rotor reference frame; the target current value I_dq is input by a user or obtained by using the real-time phase current values I_a, I_b, and I_c of the real-time current vector {right arrow over (Iabc)}; and the preset current-rotor angle γ is input by the user;

2). looking up a target voltage-rotor angle α or a target voltage-current angle β through a MTPA look-up table by referring to the target d_r-axis current value I_dr and the target q_r-axis current value I_qr; in which the target voltage-rotor angle α is an angle between the d_v-axis of a (d_v, q_v) voltage reference frame and the d_r-axis of the (d_r, q_r) rotor reference frame; the (d_v, q_v) voltage reference frame is a second (d, q) rotating reference frame that is rotating in synchronism with the rotating magnetic field, and a target voltage vector {right arrow over (Vdq)} is aligned with the q_v-axis of the (d_v, q_v) voltage reference frame such that a target d_v-axis voltage value V_dv of the target voltage vector {right arrow over (Vdq)} equals to zero and a target q_v-axis voltage value V_qv of the target voltage vector {right arrow over (Vdq)} equals to a target voltage value V_dq of the voltage vector {right arrow over (Vdq)} (i.e., V_dv=0 and V_qv=V_dq); the target voltage-current angle β is an angle between the voltage vector {right arrow over (Vdq)} and the current vector {right arrow over (Idq)}; and the MTPA look-up table is obtained in the maximum torque per ampere (MTPA) mode and comprises correspondences between the target dr-axis current value I_dr, the target q_r-axis current value I_qr, the preset current-rotor angle γ, the target voltage-rotor angle α, and the target voltage-current angle β;

3) calculating a PI error Δ by using the target d_r-axis current value I_dr, the target qr-axis current value I_qr, the target voltage-rotor angle a, the target voltage-current angle β, and the real-time phase current values I_a, I_b, and I_c; and obtaining the target voltage value V_dq of the voltage vector {right arrow over (Vdq)} by regulating the PI error Δ to be zero through a PI controller via the formula V_dq=PI (Δ);

4) obtaining a target voltage angle θ_v of the target voltage vector {right arrow over (Vdq)}, in which the target voltage angle θ_v is an angle between the target voltage vector {right arrow over (Vdq)} and the A-axis of the orthogonal (A, B) stationary reference frame; the A-axis of the orthogonal (A, B) stationary reference frame is aligned with the a-axis of the (a, b, c) stationary reference frame (i.e., the phase-a winding having the real-time phase current Ia); and the target voltage angle θ_v is obtained by using the target dr-axis current value I_dr, the target q_r-axis current value I_qr, the target voltage-rotor angle α, and the real-time phase current values I_a, I_b, and I_c, or obtained by using a given rotating speed value S_pd of the rotor;

5) calculating the target A-axis voltage value V_A and the target B-axis voltage value V_B of the target voltage vector {right arrow over (Vdq)} in the orthogonal (A, B) stationary reference frame by using the target voltage value V_dq and the target voltage angle θ_v through the inverse Park transmission; converting the target A-axis voltage value V_A and the target B-axis voltage value V_B into target phase voltage values V_a, V_b, and V_c, and modulating PWM signals of the inverter by using the phase the target voltage values V_a, V_b, and V_c for regulating the real-time phase current values I_a, I_b, and I_c.

In a class of this embodiment, the method is operated under a current-control mode, in which:

in 1), the target current value I_dq is input by a user; and the target dr-axis current value I_dr and the target q_r-axis current value I_qr are calculated via the formulas:I_dr=−I_dq×sin(γ), andI_qr=I_dq×cos(γ);in 3), the PI error Δ equals to the target d_v-axis current value I_dv of the target current vector {right arrow over (Idq)} minus the d_v-axis current value i_{dv_real} of the real-time current vector {right arrow over (Iabc)} in the (d_v, q_v) voltage reference frame (i.e., Δ=I_dv−I_{dr_real}); in which the target d_v-axis current value I_dv and the target q_v-axis current value I_qv of the target current vector {right arrow over (Idq)} are calculated via the formulas:I_dv=I_dr*cos(α)+I_qr*sin(α), andI_qv=−I_dr*sin(α)+I_qr*cos(α);the d_v-axis current value I_{dv_real} and the q_v-axis current value I_{dv_real} of the real-time current vector {right arrow over (Iabc)} are obtained by projecting the real-time current vector {right arrow over (Iabc)} from the orthogonal (A, B) stationary reference frame onto the (d_v, q_v) voltage reference frame via the Park transmission, by using the formulas:I_{dv_real}=I_A*cos(θ_dv−qv)+I_B*sin(θ_dv−qv), and

I_{qv_real}=I_B*cos(θ_dv−qv)−I_A*sin(θ_dv−qv); θ_dv−qv is the observation angle (i.e., the azimuth angle) for the (d_v, q_v) voltage reference frame, θ_dv−qv is an angle between the d_v-axis of the (d_v, q_v) voltage reference frame and the A-axis of the orthogonal (A, B) stationary reference frame, and θ_v=θ_dv−qv+90°;

I_A and I_B are respectively the A-axis current value and the B-axis current value of the real-time current vector {right arrow over (Iabc)} in the orthogonal (A, B) stationary reference frame, and are converted from the real-time phase current values I_a, I_b, and I_c of the real-time current vector {right arrow over (Iabc)} in the (a, b, c) stationary reference frame via the Clarke transmission, by using the formulas:

$I_A=\frac{2}{3}{I}_a-\frac{1}{3}\left(I_b+I_c\right),\mathrm{and}$ $I_B=\frac{2}{\sqrt{3}}\left(I_b-I_c\right);$ in 4), the target voltage angle θ_v is obtained by inputting the target q_v-axis current value I_qv of the target current vector {right arrow over (Idq)} and the q_v-axis current value I_{qv_real} of the real-time current vector {right arrow over (Iabc)} in the (d_v, q_v) voltage reference frame into a phase lock loop (PLL) for processing; in which the phase lock loop is carried out via the formulas to decode θ_v:I_qv=I_{qv_real},I_{qv_real}=I_B*cos(θ_dv−qv)−I_A*sin(θ_dv−qv), andθ_v=θ_dv−qv+90°.

In a class of this embodiment, the method is operated under a speed-control mode, in which:

in 1), the target current value I_dq is obtained via:

1a) constructing a (d_i, q_i) current reference frame, in which the (d_i, q_i) current reference frame is a third (d, q) rotating reference frame that is rotating in synchronism with the rotating magnetic field, and the real-time current vector {right arrow over (Iabc)} is aligned with the d_i-axis of a (d_i, q_i) current reference frame; whereby a current angle θ_{i_real} of the real-time current vector {right arrow over (Iabc)} in the orthogonal (A, B) stationary reference frame is the same as the observation angle θ_di−qi for the (d_i, q_i) current reference frame (i.e., θ_{i_real}=θ_di−qi), and a q_i-axis current value I_{qi_real} of the real-time current vector {right arrow over (Iabc)} is zero; the current angle θ_{i_real} is an angle between the real-time current vector {right arrow over (Iabc)} and the A-axis of the orthogonal (A, B) stationary reference frame, and the observation angle θ_di−qi is an angle between the d_i-axis of the (d_i, q_i) current reference frame and the A-axis of the orthogonal (A, B) stationary reference frame;

1b) obtaining the d_i-axis current value I_{di_real} and the q_i-axis current value I_{qi_real} of the real-time current vector {right arrow over (Iabc)} in the (d_i, q_i) current reference frame by projecting the real-time current vector {right arrow over (Iabc)} in the orthogonal (A, B) stationary reference frame onto the (d_i, q_i) current reference frame via the Park transmission, by using the formulas:I_{di_real}=I_A*cos(θ_{i_real})+I_B*sin(θ_{i_real}), andI_{qi_real}=I_B*cos(θ_{i_real})−I_A*sin(θ_{i_real});

obtaining the current angle θ_{i_real} for the real-time current vector {right arrow over (Iabc)} by inputting the q_i-axis current value I_{qi_real} and a zero into a phase lock loop (PLL); wherein the phase lock loop decodes θ_{i_real} via the formulaI_{qi_real}=I_B*cos(θ_{i_real})−I_A*sin(θ_{i_real})=0; and

1c) calculating the d_i-axis current value I_{di_real} by using the current angle θ_{i_real}, and using the d_i-axis current value I_{di_real} as the target current value I_dq of a target current vector {right arrow over (Idq)} through I_dq=I_{di_real}.

• in 3), the PI error Δ equals to a real-time voltage-current angle θ_iv minus the target voltage-current angle β (i.e., Δ=θ_iv−β); wherein the real-time voltage-current angle θ_iv equals to the target voltage angle θ_v of the target voltage vector {right arrow over (Vdq)} minus the current angle θ_{i_real} of the real-time current vector {right arrow over (Iabc)} (i.e., θ_iv=θ_v−θ_{i_real}); and
• in 4), the target voltage angle θ_v is calculated by using a rotating speed value Spd of the motor that is input by a user through the formula θ_v=∫s_pd×(pole_pair×360×Δt÷60)·dt; in which pole pair is the number of the magnetic rotor poles, and Δt is the time variable.

In a class of this embodiment, in 2) the MTPA look-up table is obtained through experiments, theoretical calculations, or computer-aided finite-element-analysis software.

In a class of this embodiment, when the target voltage value V_dq (i.e., qv-axis voltage value V_qv) of the target voltage vector {right arrow over (Vdq)} is larger than or equal to the preset threshold Vmax, the PI controller is operated at a saturated state and the output voltage of the PI controller is limited to be the preset threshold Vmax, the d_v-axis current value I_dv of the target current vector {right arrow over (Idq)} is not useful for control, and the method is turns into a field-weakening control.

The benefits of the invention include:

1) The control method derives the rotor position without using a magnetic flux observer, thus requiring less CPU time to do the calculation, and being more intuitive and simpler to the position-sensorless control of the motor. The current-control mode and the speed-control mode of the disclosure are operated by two decoupled PI-controllers, achieving better control stability and dynamic response than the multi-stage nested control circuits.

2) In the current-control mode and the speed-control mode of the disclosure, the method regulates the current along the MTPA trajectory that can be calibrated to optimize the motor, and the motor is fully functional and allows starts with a full load. And the method of the invention is carried out over the full operating range from BEMF-free to field-weakening control.

3) The model used in the method of the invention is named a PLSL-MTPA mathematical model. The model does not rely on a single rotor reference frame, it converts the current vector of the motor to a current reference frame and a voltage reference frame, and parses the angles between the voltage vector, the current vector, and the rotor position, for performing the position-sensorless control. The mathematical model and the relative algorithms and calculations are simple in the method, which take less space on the chips, have low requirement for MCU, and make the motor more costly to control.

4) The PLSL-MTPA mathematical model used in the disclosure is an optimized position-sensorless control technology that is not dependent on the motor parameters. The technology solves the conventional bottleneck problem that the position-sensorless control of the motor is highly dependent on the motor parameters such as the resistance Rs, the q-axis inductance Lq, the d-axis inductance Ld, and the magnetic flux λm, such that the mathematical model has a relative wider range of applications.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a conventional position-sensorless FOC method for controlling a permanent magnet synchronous motor;

FIG. 2 is a perspective view of a permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 3 is a perspective view of a motor controller of a permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 4 is a sectional view of a permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 5 is a diagram of a control logic in a motor controller of a permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 6 is a circuit diagram corresponding to FIG. 5;

FIG. 7 is a schematic diagram of a (a, b, c) stationary reference frame for a three-phase permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 8 is a schematic diagram of an orthogonal (A, B) stationary reference frame for a three-phase permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 9 shows a relationship between a (a, b, c) stationary reference frame, an orthogonal (A, B) stationary reference frame, and a (d, q) rotating reference frame for a vector control of a three-phase permanent magnet synchronous motor, in accordance with one embodiment of the invention;

FIG. 10 shows a relationship between an orthogonal (A, B) stationary reference frame and a (d, q) rotating reference frame, in accordance with one embodiment of the invention;

FIG. 11 shows the changes of three angles θ_v, θi and θr having the same frequency along with the phase-a voltage component Va and the phase-a current la in a time domain, in accordance with one embodiment of the invention;

FIG. 12 shows a voltage vector {right arrow over (Vabc)}, a current vector {right arrow over (Iabc)}, and a rotor position in a (a, b, c) stationary reference frame, in accordance with one embodiment of the invention;

FIG. 13 is a schematic diagram of the control of the voltage vector {right arrow over (Vdq)} and the current vector {right arrow over (Idq)} in the method, in accordance with one embodiment of the invention;

FIG. 14 shows the d_r-axis current value I_dr and the q_r-axis current value I_qr in a (d_r, q_r) rotor reference frame and the d_v-axis current value I_dv and the q_v-axis current value I_qv in a (d_v, q_v) voltage reference frame for the current vector {right arrow over (Idq)} in accordance with one embodiment of the invention;

FIG. 15 is a block diagram of the method under the current-control mode, in accordance with one embodiment of the invention;

FIG. 16 is a block diagram of the method under the speed-control mode, in accordance with one embodiment of the invention;

FIG. 17 is a correlation between the voltage-rotor angle α, the d_r-axis current value I_dr, and the q_r-axis current value I_qr, in accordance with one embodiment of the invention;

FIG. 18 is a correlation between the voltage-current angle β, the d_r-axis current value I_dr, and the q_r-axis current value I_qr, in accordance with one embodiment of the invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Detailed description of the invention will be given below in conjunction with the drawings.

Referring to FIGS. 2-4, the invention is used for controlling a three-phase permanent magnet synchronous motor that comprises a motor unit 1 and a motor controller 2. The motor unit 1 comprises a stator assembly 12, a rotor assembly 13 and a housing assembly 11. The stator assembly 12 is mounted on the housing assembly 11; the rotor assembly 13 is mounted within or outside the stator assembly 12; the motor controller 2 comprises a controller housing 22, and a control circuit board 21 mounted in the controller housing 22; the control circuit board 21 typically comprises a power supply circuit, a microprocessor, a bus voltage detection circuit, and an inverter; the power supply circuit supplies power to each part of the circuit. The DC bus voltage V_DC-bus is applied from the bus voltage detection circuit to the microprocessor; the microprocessor controls the inverter to turn on/off the currents to the phase windings of the stator assembly 12.

Referring to FIGS. 5 and 6, a three-phase permanent magnet brushless DC motor inputs the phase currents Ia, Ib, and Ic through the phase current detection circuit to the microprocessor. An alternating-current input (AC input) flows through a full-wave rectification circuit comprising four diodes D7, D8, D9, and D10, and a DC bus voltage V_{dc_bus} is output at one end of a capacitor C1, wherein the DC bus voltage V_{dc_bus} is related to the AC-input voltage. The microprocessor inputs pulse-width modulation (PWM) signals to the inverter comprising six electronic switching tubes Q1, Q2, Q3, Q4, Q5, Q6, and the six control ends of the electronic switching tubes are separately controlled by six PWM signals (P1, P2, P3, P4, P5, P6) transmitted from the microprocessor.

Referring to FIG. 7, the phase currents of phase-a winding, phase-b winding, and the phase-c winding in the motor are labeled as Ia, lb and Ic, respectively, with a phase difference of 120 degrees between the two adjacent windings in the time domain, and a reference frame formed by respectively aligning an axis a, an axis b, and an axis c along the phase-a current Ia, the phase-b current Ib, the phase-c current Ic, is called a (a, b, c) stationary reference frame. Referring to FIG. 8, the three time-domain current quantities Ia, Ib and Ic, can be simplified into two currents I_A and I_B in an orthogonal (A, B) stationary reference frame. FIG. 9 is a relationship between a (a, b, c) stationary reference frame, an orthogonal (A, B) stationary reference frame, and a (d, q) rotating reference frame for a vector control of the invention. The mathematical relationships between the currents Ia, Ib, Ic in the (a, b, c) stationary reference frame and the currents I_A and I_B in the orthogonal (A, B) stationary reference frame of the same current vector, are as follows:

$\begin{array}{cc}{I}_A=\frac{2}{3}{I}_a-\frac{1}{3}\left(I_b+I_c\right),\mathrm{and}{I}_B=\frac{2}{\sqrt{3}}\left(I_b-I_c\right);& \left(1\right)\end{array}$

The above mathematical transformation in the formulas (1) is the so-called Clarke Transformation, where IA and IB are two sinusoidal currents that change with time. When IA and IB are observed on a (d, q) rotating reference frame having the same angular frequency as the sinusoidal currents I_A and I_B, the sinusoidal characteristics are eliminated from the current values and only the phase characteristics are retained in the current values. The sinusoidal current values I_A and I_B of a current vector in the orthogonal (A, B) stationary reference frame are converted to the direct-current (DC) d-axis and q-axis current values I_d and I_q in a (d, q) rotating reference frame by performing the Park Transformation through the formulas:I_d=I_A*cos(θ_d−q)+I_B*sin(θ_d−q), andI_q=I_B*cos(θ_d−q)−I_A*sin(θ_d−q).   (2)

Referring to FIG. 10, the variable θ_d−q in the PARK transformation of the formulas (2) is the angle between the d-axis of the (d, q) rotating reference frame and the A-axis of the orthogonal (A, B) stationary reference frame, and is called the azimuth angle or observation angle for the (d, q) rotating reference frame. When observing the current values I_A and I_B in different (d, q) rotating reference frames having different azimuth angles θ_d−q, the current values I_A and I_B are projected onto the different (d, q) rotating reference frames to obtain different current values I_d and I_q. When a (d, q) rotating reference frame having an observation angle θ_d−q is selected for performing the Park transformation, the sinusoidal current values I_A and I_B are converted to the DC current values Id and Iq. Other sinusoidal A-axis and B-axis variable quantities (e.g., the voltage values of a voltage vector) in the orthogonal (A, B) stationary reference frame can also be converted to direct-current d-axis and q-axis variable quantities in a (d, q) rotating reference frame by using Park transformation through the formulas in (2), is the observation angle θ_d−q of the (d, q) rotating reference frame is given. In the position-sensorless control of a permanent magnet synchronous motor, the engineers are faced with an important problem of how to select a suitable observation angle θ_d−q for converting the sinusoidal A-axis and B-axis variable quantities (e.g., the current values of a current vector, and the voltage values of a voltage vector) in the orthogonal (A, B) stationary reference frame to the direct-current d-axis and q-axis variable quantities in a (d, q) rotating reference frame, for the purpose of convenient control of the current and voltage vectors of the motor. In a conventional motor control method, the rotor position that refers to the north polar of the magnetic field of the rotor is originated at 0 degree, and the rotor goes through a full circle is defined as a 360° of rotation. The relationship between the rotor position and the number of the magnetic rotor poles is θr=Pole_pair×θ₀, in which θr is the rotor angle representing the angle of the rotor position in the orthogonal (A, B) stationary reference frame with respect to the A-axis of the orthogonal (A, B) stationary reference frame, and is measured between the rotor position and the A-axis of the orthogonal (A, B) stationary reference frame; θ₀ is the mechanical angle of the rotor; and pole_pair is the number of the magnetic rotor poles.

In most position-sensorless control of the permanent magnet synchronous motors, the mechanical angle θ₀ of the rotor is unknown, the task of the control algorithm is to estimate the angle θr that is the key element to execute the control algorithm of the motor. But the conventional FOC method to estimate θr used in the control is highly dependent on the motor parameters, thus its control algorithm and mathematical model are complicated, which requires time-consuming computation.

The disclosure provides a MTPA based method for parameterless and position-sensorless control of a permanent magnet synchronous motor, which is named a parameterless sensorless MTPA (PLSL-MTPA) method. It is important to note that the sinusoidal quantities are parsed by establishing different (d, q) rotating reference frames having different observation angles. In particular, the current vector and the voltage vectors are projected onto a (d_r, q_r) rotor reference frame, a (d_v, q_v) voltage reference frame, and a (d_i, q_i) current reference frame which all are (d, q) rotating reference frames having different observation angles θ_d−q. In particular, the rotor position is aligned with the d_r-axis of the (d_r, q_r) rotor reference frame such that the rotor angle θ_r equals to the observation angles θ_dr−qr of the (d_r, q_r) rotor reference frame. The voltage vector {right arrow over (Vdq)} is aligned with the q_v-axis of the (d_v, q_v) voltage reference frame, such that the voltage angle θ_v equals to the observation angles θ_dv−qv of the (d_v, q_v) voltage reference frame plus 90° (i.e., θ_v=θ_dv−qv+90°); the d_v-axis value V_dv of the voltage vector {right arrow over (Vdq)} is zero (i.e., V_dv=0); and q_v-axis value V_qv of the voltage vector {right arrow over (Vdq)} equals to the voltage value V_dq of the voltage vector {right arrow over (Vdq)} (i.e., V_qv=V_dq). The current vector {right arrow over (Idq)} is aligned with the d_i-axis of the (d_i, q_i) current reference frame, such that the current angle θ_i is the same as the observation angles θ_di−qi of the (d_i, q_i) current reference frame (i.e., θ_i=θ_di−qi); the q_i-axis value I_qi of the current vector {right arrow over (Idq)} is zero (i.e., I_qi=0); and d_i-axis value I_di of the current vector {right arrow over (Idq)} equals to the current value I_dq of the current vector {right arrow over (Idq)} (i.e., I_di=I_dq).

With reference to FIG. 12, the voltage angle θ_v represents the angle of the voltage vector {right arrow over (Vdq)} in the orthogonal (A, B) stationary reference frame with respect to the A-axis of the orthogonal (A, B) stationary reference frame, and is measured between the voltage vector {right arrow over (Vdq)} and the A-axis of the orthogonal (A, B) stationary reference frame. The current angle θi represents the angle of the current vector {right arrow over (Idq)} in the orthogonal (A, B) stationary reference frame with respect to the A-axis of the orthogonal (A, B) stationary reference frame, and is measured between the current vector {right arrow over (Idq)} and the A-axis of the orthogonal (A, B) stationary reference frame.

In this invention, the current and voltage vectors are parsed in a (d_v, q_v) voltage reference frame and a (d_i, q_i) current reference frame as claimed through the Park transformation, and a phase lock loop (PLL) is introduced to decode the observation angle θ_dv−qv of the (d_v, q_v) voltage reference frame or the observation angle θ_di−qi of the (d_i, q_i) current reference frame so as to arrive at the voltage angle θ_v or the current angle θ_i. The working principle of a phase lock loop is opposite to that of the Park Transformation. Particularly, the Park Transformation converts the sinusoidal variable quantities into the DC (direct-current) variable quantities by using the observation angle θ_d−q of a (d, q) rotating reference frame, while the phase lock loop decodes the observation angle θ_d−q of a (d, q) rotating reference frame by locking one DC variable quantity on an axis of the (d, q) rotating reference frame.

In the speed control mode of the method, the voltage angle θ_v is obtained by integrating the open-loop speed S_pd for calculating the A-axis and B-axis voltage value V_A and V_B in the orthogonal (A, B) stationary reference frame and the phase voltage values V_a, V_b and V_c in the (a, b, c) stationary reference frame. The phase lock loop outputs the current angle θ_{i_real} of the real-time current vector {right arrow over (Iabc)} that is generated by parsing the A-axis and B-axis current value I_A and I_B of the real-time current vector {right arrow over (Iabc)} in the orthogonal (A, B) stationary reference frame in a (d_i, q_i) current reference frame by letting the q_i-axis current value I_{qi_real} to be 0. In particular, the current angle θ_{i_real} is decoded by using the above formulas (2) to fulfill the relationship (I_{di_real,} 0)=Park transformation (I_A, I_B) by θ_{i_real}. In other words, θ_{i_real} is decoded in the PLL by fulfilling the relationship: 0=I_B×cos(θ_{i_real})−I_A×sin(θ_{i_real}). The d_i-axis current value I_{di_real} is the calculated by using the formula I_{di_real}=I_A×cos(θ_{i_real})+I_B×sin(θ_{i_real}) and is used as the target current value I_dq of the target current vector {right arrow over (Idq)}.

In the synchronous motors, the angles θ_v, θ_i and θ_r are three time-domain variables with the same frequency. Taking the phase a for example, supposing the phase-a voltage component Va(t) has a phase lead β with respect to the phase-a current component Ia(t), the phase-a voltage component Va(t) and the phase-a current component Ia(t) are expressed as: Va(t)=V_abc×cos(θ_v)=V_abc×cos(ωt+β) and Ia(t)=I_abc×cos(θ_i)=I_abc×cos(ωt); in which V_abc is the voltage value of a voltage vector {right arrow over (Vabc)} composing the phase voltage components V_a(t), V_b(t), V_c(t) in the (a, b, c) stationary reference frame; I_abc is the current value of a current vector {right arrow over (Iabc)} composing the phase current components I_a(t), I_b(t), I_c(t) in the (a, b, c) stationary reference frame; ω is the angular velocity of the rotor; t is the time; and β is the angle between the voltage component Va (t) and the current component Ia (t) of the phase a.

The relationship between the angles θ_v, θi and θr is as follows:β=θ_v−θ_i   (3), andθr=θ_v−α−90°  (4).

FIG. 11 shows the changes of three angles θ_v, θi and θr along with the phase-a voltage Va and the phase-a current Ia in a time domain. In one example, the PLSL-MTPA method of the invention control the result of θ_v−θ_i to result in a given voltage-current angle β according to the above formula (3), thus forming a speed-control mode. In another example, the PLSL-MTPA method of the invention regulate the projections of the sinusoidal currents I_A and I_B on a (d_v, q_v) voltage reference frame to result in a given voltage-rotor angle α according to the above formula (4), thus forming a current-control mode. In the PLSL-MTPA method of the invention, the d_v-axis voltage value and the q_v-axis voltage value of the voltage vector in the (d_v, q_v) voltage reference frame are controlled as: V_dv=0 and V_qv=V_dq, and the relationships between the voltage vector, the current vector, and the rotor position are controlled as that in FIG. 13. As shown in FIG. 13, the relation α=β+γ is obtained in the method of the invention.

Referring to FIGS. 14 and 15, in one example of the invention, the method is performed under a current-control mode, a (d_v, q_v) voltage reference frame in which the voltage vector is aligned with the q_v-axis of the frame (i.e., V_dv=0 and V_qv=V_dq), is used to parse the orthogonal sinusoidal currents I_A and I_B of the real-time current vector {right arrow over (Iabc)}. The q_v-axis projection I_qv_real of the real-time current vector {right arrow over (Iabc)} and the q_v-axis projection I_qv of the target current vector {right arrow over (Idq)} in the (d_v, q_v) voltage reference frame are processed with the phase lock loop (PLL) to obtain the voltage angle θv of the target voltage vector {right arrow over (Vdq)}. And the d_v-axis projection I_{dv_real} of the real-time current vector {right arrow over (Iabc)} and the d_v-axis projection I_dv of the target current vector {right arrow over (Idq)} in the (d_v, q_v) voltage reference frame are processed with the PI controller to obtain the voltage value V_dq of the target voltage vector {right arrow over (Vdq)}. Then, the DC quantities V_dv and V_qv of the target voltage vector {right arrow over (Vdq)} in the (d_v, q_v) voltage reference frame are converted into the sinusoidal quantities V_A and V_B in the orthogonal (A, B) stationary reference frame by performing a dq to AB transformation (i.e., a transformation from a (d, q) rotating reference frame to an orthogonal (A, B) stationary reference frame) through the inverse Park transmission by using the voltage value V_dq and the voltage angle θ. Finally, the sinusoidal voltages V_A and V_B are converted to the phase voltage values V_a, V_b, and V_c for modulating PWM signals of the inverter so as to regulate the real-time phase currents I_a, I_b, and I_c.

Referring to FIGS. 14 and 16, in another example of the invention, the method is performed under a speed-control mode, which makes the voltage vector rotate at a given speed S_pd to form an open-loop control. The voltage angle θv of the voltage vector {right arrow over (Vdq)} is obtained by: θv=∫s_pd×(pole_pair×360×Δt÷60)·dt, in which Spd is the speed value, pole_pair is the number of the magnetic rotor poles, and Δt is the time variable. The phase lock loop (PLL) coincides the (real-time or target) current vector {right arrow over (Idq)} with the d_i-axis of a (d_i, q_i) current reference frame (i.e., I_qi=0 and I_di=I_dq) to generate the current angle θi of current vector {right arrow over (Idq)}; then the angle difference θiv is obtained according to θiv=θv−θi and the voltage value V_dq of the target current vector {right arrow over (Vdq)} is obtained according to V_dq=PI (θiv−β); and the DC quantities (V_dv=0, and V_qv=V_dq) in the (d_v, q_v) voltage reference frame are converted to the sinusoidal quantities voltage values V_A and V_B in the orthogonal (A, B) stationary reference frame by using the voltage angle θ_v and the voltage value V_dq.

The PI controller of the two control modes in the method of this invention regulates an angle error or a current error to produce the voltage value V_dq through V_dq=PI (Δ); in the speed-control mode, Δ=θ_iv−β; and in the current control mode, Δ=I_dv−I_{dv_real}. The phase current values Va, Vb, and Vc are derived from the voltage angle θ_v and the voltage value V_dq, and the voltage angle θ_v and the voltage value V_dq are calculated by different means in these two control modes.

A static full-load start-up process of the motor with the PLSL-MTPA control of the invention includes: the motor is started with the maximum current value I_{dq_max} and is in a resistive state when the speed is low, the current vector and the voltage vector are in the same phase, and the PI controller tends to be operated at the maximum current value I_{dq_Max} to drive the motor. Then the back electromotive force (BEMF) increases with the increasing rotational speed, causing the phase difference β between the voltage vector {right arrow over (Vdq)} and the current vector {right arrow over (Idq)} to gradually become non-zero, such that the PI controller is operated in a normal condition. The phase difference β between the voltage vector {right arrow over (Vdq)} and the current vector {right arrow over (Idq)} varies with the actual load, and the motor currents change accordingly. The start-up ability of the motor depends on the target speed, the speed increasing slope, current PI gain, and the maximum current limit.

The conventional FOC method is typically carried out through a rotor reference frame, and a position-sensorless algorithm is used to derive the rotor position for performing the FOC method. Compared with the conventional FOC method, the PLSL-MTPA control method of the invention is not dependent on the motor parameters, it employs an angle conversion scheme and regulates phase difference between the current vector and the voltage vector, for realizing the synchronous control of the motor. The PLSL-MTPA scheme of the invention greatly simplifies the processes of a sensorless control of a motor.

The PLSL-MTPA control method of the disclosure employs the following four control schemes:

• 1) Under synchronous operation conditions, an adjustable angle β between the current vector and the voltage vector is unchanged, ensuring the comprehensive control of the synchronous motor, in which the angle β is also commonly referred to as the power factor angle;
• 2) Under synchronous operation conditions, the rotor position of a synchronous motor is derivable by using the voltage vector and the adjustable angle α, in which the angle α may lead or lag the voltage vector. For a given angle α, a control of a synchronous motor is fulfilled by applying a voltage vector to the motor;
• 3) a control under the maximum torque per current (MTPA) mode is carried out by converting the target current into the angles α and β, and the two angles α and β are determined according to the MTPA principle to be used the schemes in 1) and 2);
• 4) only when the target current is controlled along the MTPA trajectory and is converted to the angles α and β under the MTPA criterion, the rotor position derived through the scheme in 2) is the actual rotor position of the motor, and the torque generated by the motor is called MTPA torque; and
• 5) When the motor runs at a small current, the ratio of voltage to speed can be approximated as back electromotive force (BEMF). The operating conditions of a synchronous motor can be determined by comparing the ratio with a threshold.

The PLSL-MTPA method under the speed-control mode of the invention is an open-loop speed control that can be carried out under a load and through the field-weakening area. The speed and the position of the rotor only change with control commands, so that the control method is simple and optimizable and has a wide range of applications on electric machinery and electric drives.

The MTPA look-up table, also known as MTPA_Angle_Lookup, is mainly used to look up the angles α and β angle in FIG. 13. The relationship between the angles α and β and the MTPA look-up table is as follows:

$\left(\begin{array}{c}\alpha \\ \beta \end{array}\right)=\mathrm{MTPA\_Angle}\mathrm{\_Lookup}\left(I_{\mathrm{dr}},I_{\mathrm{qr}}\right).$

The angles α and β angle are determined by looking up the MTPA look-up table by referring the current I_dr and the current I_qr.

The MTPA look-up table is a data set obtained from experiments. For example, a motor of 1/3 HP is used for experiments and its power is measured a dynamometer. A test for determining the included angle γ includes the steps of: the rotational speed of the motor is set at 1450 rpm, and the torque interval of the dynamometer is set at 10, 15, 20, 26, 31 oz-in, respectively. A maximum system efficiency or a maximum MTPA Index is searched under each torque range of the dynamometer, the corresponding current values I_dr, I_qr and the angles α, β, and γ are recorded. Next, the test results are subjected to verification, five sets of data are obtained under the rotational speed of 1450 rpm, the set of data that allows for the maximum torque is selected and written to the MTPA look-up table. As shown in Table 1, the set of data including the angle γ, the current I_dr, the current I_qr, the angle α, and the angle β that allows for the maximum torque 31.64 are selected and written to the MTPA look-up table as the data correspondence at rotational speed of 1450 rpm. Similarly, the data set of the angle γ, the current I_dr, the current I_qr, the angle α, and the angle β that allows for the maximum torque is determined for a different rotational speed of 1400 rpm, 1350 rpm, 1300 rpm, etc., and written to the MTPA look-up table. The data in Table 1 are obtained in the speed control mode.

TABLE 1
Sdq	Torque	γ	Idr	Iqr	α	β
1450 rpm	10.46	1728	500	1652	3391	13356
1450 rpm	15.98	2640	1240	2344	6592	11308
1450 rpm	20.94	3500	1900	2964	8640	10456
1450 rpm	26.46	4483	2608	3672	10177	9988
1450 rpm	31.64	5498	3356	4388	11328	9604

For the current control mode, the MTPA look-up table is obtained in a similar way from experiments. For example, a motor of 1/3 HP is used for experiments and its power is measured a dynamometer. A test for determining the included angle γ includes the steps of: the current value I_dq flowing through the motor is set to 3.2A, and the torque interval of the dynamometer is set at 12, 18, 24, 28, 33 oz-in, respectively. A maximum system efficiency or a maximum MTPA_Index is searched under each torque range of the dynamometer, and the corresponding current values I_dr, I_qr and the angles α, β, and γ are recorded. The test results are subjected to verification for the current value I_dq that defaults to 3.2 A, five sets of data are obtained under the current value I_dq of 3.2 A, the set of data as shown in FIG. 2 that allows for the maximum torque is selected and written to the MTPA look-up table.

TABLE 2
Idq	Torque	γ	Idr	Iqr	α	β
3.2 A	12	1455	650	2253	2351	13450
3.2 A	18	2228	1440	2698	4323	12011
3.2 A	24	3400	1800	3298	6543	10244
3.2 A	28	4578	2708	3789	8678	951
3.2 A	33	5677	3676	4565	10233	8867

The MTPA look-up table can also be obtained from computer-aided finite element analysis software. As shown in FIGS. 17 and 18, the relevant data is obtained by computer-aided finite element analysis, and, the angles α or β are obtained under various current I_dr and current I_qr. The MTPA look-up table can also be obtained from theoretical calculations.

The control method derives the rotor position without using a magnetic flux observer, thus requiring less CPU time to do the calculation, and being more intuitive and simpler to the position-sensorless control of the motor. The current control mode and the speed control mode of the motor are operated by two decoupled PI-controllers, achieving better control stability and dynamic response than the multi-stage nested control circuits. In the current control mode and the speed control mode of the disclosure, the control method regulates the current to flow along the MTPA trajectory that can be calibrated. The motor is fully functional and allows for starting up with a full load, over the full operating range from BEMF-free to field-weakening control. The PLSL-MTPA mathematical model of the disclosure does not reply on a single rotor reference frame, it converts the current vector of the motor to a current reference frame and a voltage reference frame, thus parsing the included angle of the vectors to perform the position-sensorless control. The mathematical models, the relative algorithms, and the calculations in the method are simple, and thus the method requires less space on the chips and low requirement for MCU and makes the motor more costly to control. The PLSL-MTPA mathematical model used in the disclosure is an optimized position-sensorless control technology that is not dependent on the motor parameters. The technology solves the bottleneck problem that the position-sensorless control of the motor is highly dependent on the motor parameters such as the resistance Rs, the d-axis inductance Ld, the q-axis inductance Lq, and the magnetic flux λm, such that the mathematical model has a relative wider range of applications.

Comparison of the field-weakening control method of the disclosure and that of the conventional FOC control theory:

The principle of the field-weakening control method of the conventional FOC control theory is that the control process of a permanent magnet synchronous motors is typically performed in two areas including MTPA area and a field weakening area. As the speed of the motor increases, there may be a maximum torque area or a maximum current area outside the field weakening area, but the two maximum areas are rarely used in practice. The FOC control method of the motor is a method to regulate the current values Id and Iq, that is, a control method with two degrees of freedom. The MTPA control is a theory that controls the two current values Id and Iq to keep a motor in optimized operational condition. When the current enters the field-weakening area, the freedom in the Id direction is locked, and only the current value Iq is proportional to the output torque, so that the motor may no longer be optimized. The relationship between the current value Id and field weakening is: since a permanent magnet is embedded in the rotor, the rotor magnetic field induces a back electromotive force (BEMF) that is offset by the stator voltage when the motor is rotating, the back electromotive force (BEMF) is proportional to the speed. Once the rotational speed exceeds a threshold, the BEMF will also exceed the stator voltage, causing the motor to fail to operate satisfactorily in the electric state. The so-called field-weakening control is a method to increase the current value Id in the negative direction when the back electromotive force (BEMF) exceeds a threshold, thereby generating a magnetic field to specifically weaken the rotor magnetic field. The current value Id depends on the rotational speed and the motor load, but the ultimate goal is to make the back electromotive force (BEMF) less than the maximum stator voltage.

In the classic field-weakening control theory, the following motor electromagnetic equations are used, in which V_d is a d-axis voltage of the voltage vector; V_q is the q-axis voltage; I_d is the d-axis current of the current vector; I_q is the q-axis current; r is the motor resistance; O)r is the angular speed of the rotor; λ_d is the d-axis flux linkage; λ_q is the q-axis flux linkage; λ_pm is the permanent magnetic flux linkage; L_d is the d-axis inductance; and L_q is the q-axis inductance; the quantities V_d, V_q, I_d, I_q, λ_d, λ_q, L_d, and L_q are components in a (d, q) rotating reference frame.V_d=rI_d−ω_rλ_q=rI_d−ω_rL_qI_q;V_q=rI_q+ω_rλ_d=rI_q+ω_r(L_dI_d+λ_pn).

Because the motor resistance r is very small, the components rI_d and rI_q are negligible in the above electromagnetic equations. And due to the voltage limitation, the current values I_d and I_q satisfy the following elliptic equation:

${\left(\frac{V_0}{{\omega }_r{L}_q}\right)}^2=I_q^2+{\left(\frac{L_d}{L_q}\right)}^2{\left(I_d+\frac{{\lambda }_{\mathrm{pm}}}{L_d}\right)}^2.$

Under the field-weakening condition with

$V_0=\frac{2V_{\mathrm{DC}}}{\pi },$ and the single solution of the current value I_q is obtained by the following formula, and the current value I_q can be calculated according to the known current I_d

$I_q=\sqrt{{\left(\frac{V_0}{{\omega }_r{L}_q}\right)}^2-{\left(\frac{L_d}{L_q}\right)}^2{\left(I_d+\frac{{\lambda }_{\mathrm{pm}}}{L_d}\right)}^2}.$

It can be seen from the actual test simulated result that, the currents I_d and I_q can flow along the MTPA trajectory at low speed, and satisfy the voltage-ellipse limit at high speed. The solution of the current value I_d is given by the control logic (such as a PI controller, a look-up table and so on) instead of a given formula. In summary, the key to the field-weakening control is to find the current value I_d that ensures the voltage does not exceed the preset threshold, and then the current value I_q is calculated by using the I_d, thus achieving the current control of a motor under the FOC theory.

The control method of the disclosure, also known as the PLPS-MTPA control method, has different control means in the field-weakening control compared to the conventional FOC control. The field-weakening control method of the invention directly regulates the voltage instead of finding the current value I_d in the conventional method. The voltage value V_dq (V_dq=Vq, Vd=0) is limited below the DC bus voltage V_{dc_bus}, the resulting current value Id is completely snubbed by the inverter, and the phase angle of the current and voltage vectors are regulated to generate the required torque. The control strategy in the invention completely meets the field-weakening theory, its implementation process is greatly simplified, and it allows for a more reliable control of the motor.

EXAMPLE 1

As shown in FIGS. 14 and 15, the PLSL-MTPA control method of the disclosure is carried out under a current-control mode as follows:

a), a target current value I_dq and a preset current-rotor angle γ are input by a user; second, the target dr-axis current value I_dr and q_r-axis current value I_qr are calculated by projecting target current value I_dq onto the d_r-axis and the q_r-axis of the (d_r, q_r) rotor reference frame; the target current value I_dq and a preset current-rotor angle γ are data conforming to the MTPA mode.

b), a target voltage-rotor angle a is looked up by using a MTPA look-up table according to the target d_r-axis current value I_dr and q_r-axis current value I_qr, in which the angle α is an included angle between the (d_r, q_r) rotor reference frame and the (d_v, q_v) voltage reference frame, the MTPA look-up table refers to a set of data obtained in the maximum torque per ampere (MTPA) mode.

c), the target d_v-axis current value I_dv and the target qv-axis current value I_qv of the target current vector {right arrow over (Idq)} are calculated by using the angle α, the d_r-axis current value I_dr, and the q_r-axis current value I_qr, in which the d_v-axis current value i_dv and the q_v-axis current value I_qv are obtained by projecting the target current value I_dq onto the (d_v, q_v) voltage reference frame;

d), the target q_v-axis current value I_qv and the feedback q_v-axis current value I_{qv_real} reflecting the operating conditions of the rotor in real time are inputted to a phase lock loop (PLL) to obtain the target voltage angle θ_v; target d_v-axis current value I_dv and the feedback dv-axis current value I_{dv_real} are regulated with a PI controller to obtain the target voltage value V_qv. Because V_dv=0 and V_dq=V_qv, the voltage V_A and the voltage V_B in the orthogonal (A, B) stationary reference frame are calculated according to the angle θ_v and the target voltage value V_dq, and are used for controlling the phase currents of the motor.

The following formulas are used to calculate the d_v-axis current I_dv and the q_v-axis current I_qv:I_dv=I_dr*cos(α)+I_qr*sin(α), andI_qv=−I_dr*sin(α)+I_qr*cos(α).

The following formulas are used to calculate the d_r-axis current value I_dr and the q_r-axis current value I_qr:I_dr=−I_dq×sin(γ), andI_qr=I_dq×cos(γ).

In the current control mode, when the target current value Vdq (i.e., the q_v-axis voltage value V_qv) is larger than or equal to the preset threshold Vmax, the PI controller is operated at a saturated state, the output voltage of the controller is limited to the preset threshold Vmax, and the d_v-axis current value I_dv cannot be used for control. Such condition is termed as a “field-weakening control.”

The MTPA look-up table is a data set obtained from experiments, theoretical calculations, or computer-aided finite element analysis softwares.

EXAMPLE 2

As shown in FIGS. 14 and 16, the PLSL-MTPA control method of the disclosure is carried out under a current-control mode as follows:

a), a given speed Spd and a preset angle γ are inputted by a user, and the vector angle of the voltage vector {right arrow over (Vdq)} is rotated at the given speed Spd to determine the angle θ_v, where the angle θ_v is an included angle between the voltage vector Vdq and the A-axis the (A, B) stationary coordinate system;

b), a (d_i, q_i) current reference frame is used and the real-time current vector {right arrow over (Iabc)} is aligned with the d_i-axis of the (d_i, q_i) current reference frame, such that the q_i-axis current value θ_{i_real} in the frame is 0. A phase lock loop (PLL) is used to decode the current angle θ_{i_real} for the real-time current vector {right arrow over (Iabc)}. The d_i-axis current value I_{di_real} is calculated and used as the target current value I_dq.

c), the target d_r-axis current value I_dr and the target qi-axis current value I_qr are calculated by using the angle γ and the target current value I_dq, in which the target d_r-axis current value I_dr and the target q_r-axis current value I_qr are obtained by converting the target current value I_dq to the (d_r, q_r) rotor reference frame, and the current values I_dr and I_qr are data conforming to the MTPA mode.

d), an target voltage-current β is determined by using the MTPA look-up table with reference to the current values I_dr and I_qr, in which the angle β is an included angle between the voltage vector {right arrow over (Vdq)} and the current vector {right arrow over (Idq)} in the MTPA mode; and

e) an angle difference θiv between the target voltage angle θ_v and the current angle θ_{i_real} for the feedback real-time current vector Iabc is calculated according to θ_iv=θ_v−θ_{i_real}. The target angle β and the angle difference θiv are regulated with the PI controller to obtain a target voltage value V_dq. Because V_dv=0 and V_dq=V_qv, the voltage V_A and the voltage V_B in the orthogonal (A, B) stationary reference frame are calculated according to the angle θ_v and the target voltage value V_dq, and are used for controlling the phase currents of the motor.

The MTPA look-up table is a data set obtained from experiments, theoretical calculations, or computer-aided finite element analysis softwares.

In the feedback q_i-axis current I_{qi_real} reflecting the operating condition of the motor in real time is inputted to the phase lock loop (PLL), and another input current Iq* defaults to zero, in which the feedback q_i-axis current I_{qi_real} is a projection of the real-time current vector {right arrow over (Iabc)} onto the q_i-axis of the (d_i, q_i) current reference frame. Then the phase lock loop (PLL) outputs the angle θ_{i_real} that is generated by parsing the currents I_A and the I_B with the phase lock loop (PLL) by fulfilling I_{qi_real}=0.

The following formula is used to calculate the angle θ_v:θv=∫s_pd×(pole_pair×360×αt÷60)·dt, in which Spd is the speed value, pole_pair is the number of the magnetic rotor poles, and Δt is the time variable.

In the speed control mode, when the target current value Vdq (i.e., the q_v-axis voltage value V_qv) is larger than or equal to the preset threshold Vmax, the PI controller is operated at a saturated state and the output voltage of the controller is limited to the preset threshold Vmax, and thus the current control mode is automatically converted to the filed-weakening control mode.

While particular embodiments of the invention have been shown and described, it will be obvious to those skilled in the art that changes and modifications may be made without departing from the invention in its broader aspects, and therefore, the aim in the appended claims is to cover all such changes and modifications as fall within the true spirit and scope of the invention.

Claims

1. A maximum-torque-per-ampere (MTPA) based control method for a parameterless and position-sensorless control of a permanent magnet synchronous motor, the method comprising: 1) calculating a target dr-axis current value Idr and a target qr-axis current value Iqr of a target current vector {right arrow over (Idq)} in a (dr, qr) rotor reference frame by using a target current value Idq of the target current vector {right arrow over (Idq)} and a preset current-rotor angle γ; wherein the preset current-rotor angle γ is an angle between the target current vector {right arrow over (Idq)} and the qr-axis of the (dr, qr) rotor reference frame; the (dr, qr) rotor reference frame is a first (d, q) rotating reference frame that is rotating in synchronism with the rotating magnetic field, and the rotor position is aligned with the dr-axis of the (dr, qr) rotor reference frame; the target current value Idq is input by a user or obtained by using the real-time phase current values Ia, Ib, and Ic of a real-time current vector {right arrow over (Iabc)}; and the preset current-rotor angle γ is input by the user;2) looking up a target voltage-rotor angle α or a target voltage-current angle β through a MTPA look-up table by referring to the target dr-axis current value Idr and the target qr-axis current value Iqr; wherein the target voltage-rotor angle α is measured between the dv-axis of α (dv, qv) voltage reference frame and the dr-axis of the (dr, qr) rotor reference frame; the (dv, qv) voltage reference frame is a second (d, q) rotating reference frame that is rotating in synchronism with the rotating magnetic field, and a target voltage vector {right arrow over (Vdq)} is aligned with the qv-axis of the (dv, qv) voltage reference frame; the target voltage-current angle β is measured between the target voltage vector {right arrow over (Vdq)} and the target current vector {right arrow over (Idq)}; and the MTPA look-up table is obtained in the MTPA mode and comprises correspondences between the target dr-axis current value Idr, the target qr-axis current value Iqr, the preset current-rotor angle γ, the target voltage-rotor angle α, and the target voltage-current angle β;3) calculating a PI error Δ by using the target dr-axis current value Idr, the target qr-axis current value Iqr, the target voltage-rotor angle α, the target voltage-current angle β, and the real-time phase current values Ia, Ib, and Ic; and obtaining a target voltage value Vdq of the target voltage vector {right arrow over (Vdq)} by regulating the PI error Δ through a PI controller;4) obtaining a target voltage angle θv of the target voltage vector {right arrow over (Vdq)}, wherein the target voltage angle θv is measured between the target voltage vector {right arrow over (Vdq)} and the A-axis of the orthogonal (A, B) stationary reference frame; the target voltage angle θv is obtained by using the target dr-axis current value Idr, the target qr-axis current value Iqr, the target voltage-rotor angle α, and the real-time phase current values Ia, Ib, and Ic, or obtained by using a given rotating speed value Spd of the rotor; and5) calculating a target A-axis voltage value VA and a target B-axis voltage value VB in the orthogonal (A, B) stationary reference frame by using the target voltage value Vdq and the target voltage angle θv through the inverse Park transmission, converting the target A-axis voltage VA and the target B-axis voltage VB into target phase voltage values Va, Vb, and Vc, and modulating PWM signals of the inverter by using the target voltage values Va, Vb, and Vc, for regulating the real-time phase current values Ia, Ib, and Ic.

2. The method of claim 1, the method is operated in a current-control mode, wherein: in 1), the target current value Idq is input by the user; and the target dr-axis current value Idr and the target qr-axis current value Iqr are calculated via the formulas: Idr=−Idq×sin(γ), and Iqr=Idq×cos(γ).in 3), the PI error Δ equals to a target dv-axis current value Idv of the target current vector {right arrow over (Idq)} minus a dv-axis current value Idv_real of the real-time current vector {right arrow over (Iabc)} in the (dv, qv) voltage reference frame by using Δ=Idv−Idv_real; wherein the target dv-axis current value Idv is calculated via the formula Idv=Idr*cos(α)+Iqr*sin(α); the dv-axis current value Idv_real is obtained via the park transmission by using an A-axis current value IA and a B-axis current value Is of the real-time current vector {right arrow over (Iabc)} in the orthogonal (A, B) stationary reference frame; the A-axis current value IA and the B-axis current value IB are converted by the real-time phase current values Ia, Ib, and Ic; and the dv-axis current value Idv_real is calculated by using the formula Idv_real=IA*cos(θdv−qv)+IB*sin(θdv−qv), θdv−qv is the observation angle of the (dv, qv) voltage reference frame, θdv−qv is measured between the dv-axis of the (dv, qv) voltage reference frame and the A-axis of the orthogonal (A, B) stationary reference frame, and θv=θdv−qv+90°; andin 4), the target voltage angle θv is obtained by inputting a target qv-axis current value Iqv of the target current vector {right arrow over (Idq)} and a qv-axis current value Iqv_real of the real-time current vector {right arrow over (Iabc)} in the (dv, qv) voltage reference frame into a phase lock loop; wherein the target qv-axis current value Iqv is calculated via the formula Iqv=−Idr*sin(α)+Iqr*cos(α); the qv-axis current value Iqv_real is calculated via the formula Iqv_real=IBcos(θdv−qv)−IA*sin(θdv−qv); and the phase lock loop decodes θv by fulfilling the formulas: Iqv=Iqv_real and θv=θdv−qv+90°.

3. The method of claim 1, the method is operated in a speed-control mode, wherein: in 1), the target current value Idq is obtained via: constructing a (di, qi) current reference frame, wherein the (di, qi) current reference frame is a third (d, q) rotating reference frame that is rotating in synchronism with the rotating magnetic field, and the real-time current vector {right arrow over (Iabc)} is aligned with the di-axis of a (di, qi) current reference frame; whereby a current angle θi_real of the real-time current vector {right arrow over (Iabc)} is the same as the observation angle θdi−qi for the (di, qi) current reference frame, and a qi-axis current value Iqi_real of the real-time current vector {right arrow over (Iabc)} is zero; the current angle θi_real is an angle between the real-time current vector {right arrow over (Iabc)} and the A-axis of the orthogonal (A, B) stationary reference frame, and the observation angle θdi−qi is an angle between the di-axis of the (di, qi) current reference frame and the A-axis of the orthogonal (A, B) stationary reference frame;2b) obtaining a qi-axis current value Iqi_real of the real-time current vector {right arrow over (Iabc)} in the (di, qi) current reference frame via the Park transmission by using the formula Iqi_real=IB*cos(θi_real)−IA*sin(θi_real); and obtaining the current angle θi_real by inputting the qi-axis current value Iqi_real and a zero into a phase lock loop, wherein the phase lock loop decodes θi_real via the formula IB*cos(θi_real0−IA*sin(θi_real)=0;1c) obtaining a di-axis current value Idi_real of the real-time current vector {right arrow over (Iabc)} via the Park transmission by using the formula Idi_real=IA*cos(θi_real)+IB*sin(θi_real), and using the di-axis current value Idi_real as the target current value Idq;in 3), the PI error Δ equals to a real-time voltage-current angle θiv minus the target voltage-current angle β; wherein real-time voltage-current angle θiv equals to the target voltage angle θv minus the real-time current angle θi_real; andin 4), the target voltage angle θv is obtained by using the formula θv=∫spd×(pole_pair×360×Δt÷60)·dt; wherein Spd is a given rotating speed value that is input by the user, pole_pair is the number of the magnetic rotor poles, and Δt is the time variable.

4. The method of claim 1, the MTPA look-up table is obtained through experiments, theoretical calculations, or computer-aided finite-element-analysis software.

5. The method of claim 2, when the target voltage value Vdq is larger than or equal to the preset threshold Vmax, the dv-axis current value Idv of the target current vector {right arrow over (Idq)} is not useful for control, and the method is a field-weakening control.

6. The method of claim 3, when the target voltage value Vdq is larger than or equal to the preset threshold Vmax, the dv-axis current value Idv of the target current vector {right arrow over (Idq)} is not useful for control, and the method is a field-weakening control.