Control method and system for motor

[0001] 1. Field of the Invention

[0002] The present invention relates to a control method and system for a motor, more particularly, to a control method and system using a micro-programming & memory controller for a motor.

[0003] 2. Description of the Related Art

[0004] FIG. 1 shows a conventional PI-PWM (Proportion Integration-Pulse Width Modulation) current control system 11 for a motor 10. The conventional PI-PWM current control system 11 comprises: a first difference operator 12, a PI controller 14, a second difference operator 15, a PWM comparator 16 and an inverter 18. The control system 11 receives three phase reference currents 20 to the first difference operator 12, and detects three phase actual currents 22 from a motor 10. The first difference operator 12 calculates three phase difference currents between the three phase reference currents 20 and three phase actual currents 22.

[0005] The three phase difference currents are amplified by the PI controller 14 to obtain three phase control voltages. The second difference operator 15 calculates difference voltages between the three phase control voltages and three phase triangle-wave signals 24. The PWM comparator 16 receives the three phase difference voltages and outputs three phase switch signals to the inverter 18. According to the three phase switch signals, the inverter 18 outputs three phase input voltages to the motor 10 and controls the motor 10.

[0006] However, the conventional PI-PWM current control system 11 has the current phase-lag problem. Referring to FIGS. 2A, 2B, 3A, 3B and 3C, they show the comparison between the three phase actual currents and the three phase reference currents, according to the conventional PI-PWM current control system. FIG. 2A shows A phase reference current I_{a1, refer}. FIG. 2B shows A phase actual current I_a1,actual. FIG. 3A shows a comparison between A phase reference current I_a1,refer and A phase actual current I_a1,actual. FIG. 3B shows a comparison between B phase reference current I_b1,refer and B phase actual current I_b1,actual. FIG. 3C shows a comparison between C phase reference current I_c1,refer and C phase actual current I_c1,actual. As shown in FIGS. 2A, 2B, 3A, 3B and 3C, the current phase-lag problem exists between the three phase reference currents I_a1,refer, I_b1,refer, I_c1,refer and three phase actual currents I_a1,actual, I_b1,actual, I_c1,actual.

[0007] Besides, when a software is utilized to perform the operation of the PI controller 14, the conventional PI-PWM current control system 11 has the high noise problem. FIGS. 4A, 4B, 5A, 5B and 5C show the comparison between the three phase actual currents and the three phase reference currents, according to the conventional PI-PWM current control system 11 using a software to perform the PI controller 14. FIG. 4A shows A phase reference current I_{a2, refer}, FIG. 4B shows A phase actual current I_a2,actual, FIG. 5A shows a comparison between A phase reference current I_a2,refer and A phase actual current I_a2,actual, FIG. 5B shows a comparison between B phase reference current I_b2,refer and B phase actual current I_b2,actual, FIG. 5C shows a comparison between C phase reference current I_c2,refer and C phase actual current I_c2,actual. As shown in FIGS. 4A, 4B, 5A, 5B and 5C, the three phase actual currents I_a1,actual, I_b1,actual, I_c1,actual have the high noise problem.

[0008] Therefore, it is necessary to provide an innovative and progressive control method and system to solve the above problem.

[0009] One objective of the present invention is to provide a control method and system for a motor. The control method comprises the steps of: (a) determining a motor parameter; (b) providing a d-axis reference current and a q-axis reference current; (c) detecting actual currents of the motor and converting to a d-axis actual current and a q-axis actual current; (d) calculating a d-axis counter electromotive force and a q-axis counter electromotive force at a sample period according to the motor parameter, the d-axis actual current and the q-axis actual current at the sample period, the d-axis actual current and the q-axis actual current at a last sample period, and a d-axis voltage and a q-axis voltage at the last sample period; and (e) calculating a d-axis voltage and a q-axis voltage at the sample period according to the motor parameter, the d-axis actual current and the q-axis actual current at the sample period, the d-axis reference current and the q-axis reference current at a next sample period, the d-axis counter electromotive force and the q-axis counter electromotive force calculated at the sample period.

[0010] According to the control method of the invention, the d-axis voltage and the q-axis voltage are calculated and converted to three phase control voltages. The three phase control voltages are processed to control the motor. The control method of the invention not only preserves the merit of constant switching frequency of the conventional PI-PWM control system, but also can eliminate the motor current phase-lag problem of the conventional PI-PWM control system. Also the control method of the invention will have the following characteristics and economic advantages such as high accuracy, fast response, low cost and robustness etc.

[0032] The control method of the invention can be utilized to control an AC motor, more particularly to control an induction motor so that three phase actual currents are equal to three phase reference currents. In the embodiment, the induction motor is taken as an example to describe the operation of the control method and system of the invention. However, the control method and system of the invention are not limited to control the induction motor, the other AC motor and DC motor can be controlled by the control method and system of the invention.

[0033] Firstly, the induction motor dynamic equations on the synchronous rotating d-q frame are shown as follows Equation (1). $\begin{array}{cc}\frac{d}{dt}[\begin{array}{c}{i}_{\mathrm{ds}}\\ i_{\mathrm{qs}}\\ {\lambda }_{\mathrm{dr}}\\ {\lambda }_{\mathrm{qr}}\end{array}]=\hspace{1em}\left[\begin{array}{cccc}-\left(\frac{r_s}{L_{\sigma }}+\frac{r_r^{\prime }·L_M^{\mathrm{\prime 2}}}{L_{\sigma }·L_{\mathrm{rr}}^{\mathrm{\prime 2}}}\right)& {\omega }_e& \frac{r_r^{\prime }·L_M}{L_{\sigma }·L_{\mathrm{rr}}^{\mathrm{\prime 2}}}& {\omega }_r\frac{L_M}{L_{\sigma }{L}_{\mathrm{rr}}^{\prime }}\\ -{\omega }_e& -\left(\frac{r_s}{L_{\sigma }}+\frac{r_r^{\prime }·L_M^2}{L_{\sigma }·L_{\mathrm{rr}}^{\mathrm{\prime 2}}}\right)& -{\omega }_r\frac{L_M}{L_{\sigma }·L_{\mathrm{rr}}^{\prime }}& \frac{r_r^{\prime }·L_M}{L_{\sigma }{L}_{\mathrm{rr}}^{\mathrm{\prime 2}}}\\ r_r^{\prime }\frac{L_M}{L_{\mathrm{rr}}^{\prime }}& 0& -\frac{r_r^{\prime }}{L_{\mathrm{rr}}^{\prime }}& \left({\omega }_e-{\omega }_r\right)\\ 0& r_r^{\prime }\frac{L_M}{L_{\mathrm{rr}}^{\prime }}& -\left({\omega }_e-{\omega }_r\right)& -\frac{r_r^{\prime }}{L_{\mathrm{rr}}^{\prime }}\end{array}\right]·\hspace{1em}\left[\begin{array}{c}{i}_{\mathrm{ds}}\\ i_{\mathrm{qs}}\\ {\lambda }_{\mathrm{dr}}\\ {\lambda }_{\mathrm{qr}}\end{array}\right]+\left[\begin{array}{cc}\frac{1}{L_{\sigma }}& 0\\ 0& \frac{1}{L_{\sigma }}\\ 0& 0\\ 0& 0\end{array}\right]·\left[\begin{array}{c}{V}_{\mathrm{ds}}\\ V_{\mathrm{qs}}\end{array}\right]& \left(1\right)\end{array}$

[0034] wherein L_σ is defined as a leakage induction. $L_{\sigma }=\left(1-\frac{L_M^2}{L_{\mathrm{ss}}{L}_{\mathrm{rr}}}\right)L_{\mathrm{ss}}=\frac{L_{\mathrm{ss}}{L}_{\mathrm{rr}}-L_M^2}{L_{\mathrm{rr}}}=\mathrm{leakage}\mathrm{inductance}$

[0035] Equation (1) is re-arranged, and a d-axis voltage V_ds and a q-axis voltage V_qs are obtained as follows. $\begin{array}{cc}\{\begin{array}{c}{V}_{\mathrm{ds}}=L_{\sigma }\frac{d{i}_{\mathrm{ds}}}{dt}-L_{\sigma }{\omega }_e{i}_{\mathrm{qs}}-{\omega }_r\frac{L_M}{L_{\mathrm{rr}}^{\prime }}{\lambda }_{\mathrm{qr}}+\left(r_s+r_r^{\prime }\frac{L_M^2}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}\right)·i_{\mathrm{ds}}-r_r^{\prime }\frac{L_M}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}{\lambda }_{\mathrm{dr}}\\ V_{\mathrm{qs}}=L_{\sigma }\frac{d{i}_{\mathrm{qs}}}{dt}-L_{\sigma }{\omega }_e{i}_{\mathrm{ds}}+{\omega }_r\frac{L_M}{L_{\mathrm{rr}}^{\prime }}{\lambda }_{\mathrm{dr}}+\left(r_s+r_r^{\prime }\frac{L_M^2}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}\right)·i_{\mathrm{qs}}-r_r^{\prime }\frac{L_M}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}{\lambda }_{\mathrm{qr}}\end{array}& \begin{array}{c}\left(2\right)\\ \left(3\right)\end{array}\end{array}$

[0036] A d-axis counter electromotive force E_ds and a q-axis counter electromotive force E_qs are defined as follows. $\begin{array}{cc}\{\begin{array}{c}{E}_{\mathrm{ds}}=-{\omega }_r\frac{L_M}{L_{\mathrm{rr}}^{\prime }}{\lambda }_{\mathrm{qr}}+\left(r_s+r_r^{\prime }\frac{L_M^2}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}\right)·i_{\mathrm{ds}}-r_r\frac{L_M}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}{\lambda }_{\mathrm{dr}}\\ E_{\mathrm{qs}}=+{\omega }_r\frac{L_M}{L_{\mathrm{rr}}^{\prime }}{\lambda }_{\mathrm{dr}}+\left(r_s+r_r^{\prime }\frac{L_M^2}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}\right)·i_{\mathrm{qs}}-r_r\frac{L_M}{L_{\mathrm{rr}}^{\mathrm{\prime 2}}}{\lambda }_{\mathrm{qr}}\end{array}& \begin{array}{c}\left(4\right)\\ \left(5\right)\end{array}\end{array}$

[0037] Therefore, Equation (2) and (3) can be simplified as follows. $\begin{array}{cc}{V}_{\mathrm{ds}}=L_{\sigma }\frac{d{i}_{\mathrm{ds}}}{dt}-L_{\sigma }{\omega }_e{i}_{\mathrm{qs}}+E_{\mathrm{ds}}& \left(6\right)\\ V_{\mathrm{qs}}=L_{\sigma }\frac{d{i}_{\mathrm{qs}}}{dt}+L_{\sigma }{\omega }_e{i}_{\mathrm{ds}}+E_{\mathrm{qs}}& \left(7\right)\end{array}$

[0038] Wherein di_ds/dt and di_qs/dt can be calculated by a d-axis reference current i_ds,ref, a q-axis reference current i_qs,ref, a d-axis actual current i_ds,actual and a q-axis actual current i_qs,actual as follows. $\begin{array}{cc}\frac{d{i}_{\mathrm{ds}}}{dt}=\frac{i_{\mathrm{ds},\mathrm{ref}}\left(t+T_s\right)-i_{\mathrm{ds},\mathrm{actual}}\left(t\right)}{T_s}=\frac{i_{\mathrm{ds2},\mathrm{error}}}{T_s}& \left(8\right)\\ \frac{d{i}_{\mathrm{qs}}}{dt}=\frac{i_{\mathrm{qs},\mathrm{ref}}\left(t+T_s\right)-i_{\mathrm{qs},\mathrm{actual}}\left(t\right)}{T_s}=\frac{i_{\mathrm{qs2},\mathrm{error}}}{T_s}& \left(9\right)\end{array}$

[0039] Wherein T_s is a sample time, and T_s=0.2 ms. The i_ds,actual (t) means a i_ds,actual value at a sample period. The i_qs,actual (t) means a i_qs,actual value at a sample period. The i_{ds, ref} (t+T_s) means a i_ds,ref value at a next sample period. The i_qs,ref (t+T_s) means a i_qs,ref value at a next sample period. Therefore, di_ds/dt and di_qs/dt can be calculated easily by Equations (8) and (9).

[0040] The d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs can be derived from Equations (6) and (7) as follows. $\begin{array}{cc}\{\begin{array}{c}{E}_{\mathrm{ds}}=V_{\mathrm{ds}}-L_{\sigma }\frac{d{i}_{\mathrm{ds}}}{dt}+L_{\sigma }{\omega }_e{i}_{\mathrm{qs}}\\ E_{\mathrm{qs}}=V_{\mathrm{qs}}-L_{\sigma }\frac{d{i}_{\mathrm{qs}}}{dt}-L_{\sigma }{\omega }_e{i}_{\mathrm{ds}}\end{array}& \begin{array}{c}\left(10\right)\\ \left(11\right)\end{array}\end{array}$

[0041] Wherein di_ds/dt and di_qs/dt can be calculated by the d-axis actual current i_{ds, actual} and the q-axis actual current i_{qs, actual} as follows. $\begin{array}{cc}\{\begin{array}{c}\frac{d{i}_{\mathrm{ds}}}{dt}=\frac{i_{\mathrm{ds},\mathrm{actual}}\left(t\right)-i_{\mathrm{ds},\mathrm{actual}}\left(t-T_s\right)}{T_s}=\frac{i_{\mathrm{ds}1,\mathrm{error}}}{T_s}\\ \frac{d{i}_{\mathrm{qs}}}{dt}=\frac{i_{\mathrm{qs},\mathrm{actual}}\left(t\right)-i_{\mathrm{qs},\mathrm{actual}}\left(t-T_s\right)}{T_s}=\frac{i_{\mathrm{qs}1,\mathrm{error}}}{T_s}\end{array}& \begin{array}{c}\left(12\right)\\ \left(13\right)\end{array}\end{array}$

[0042] Wherein T_s is a sample time, and T_s=0.2 ms. The i_ds,actual (t) means a i_ds,actual value at a sample period. The i_qs,actual (t) means a i_qs,actual value at a sample period. The i_{ds, actual} (t−T_s) means a i_ds,actual value at a last sample period. The i_qs,actual (t−T_s) means a i_qs,actual value at a last sample period. Therefore, di_ds/dt and di_qs/dt in Equations (10) and (11) can be calculated easily by Equations (12) and (13).

[0043] Therefore, the d-axis voltage V_ds and the q-axis voltage V_qs can be calculated by Equations (6) to (13) according to the leakage inductance, the d-axis reference current i_ds,ref, the q-axis reference current i_qs,ref, the d-axis actual current i_ds,actual and the q-axis actual current i_qs,actual.

[0044] Referring to FIG. 6A, FIG. 6B and FIG. 7, the control system 60 comprises: a micro-programming & memory controller 61, a PWM comparator 62 and a switch-mode inverter 63. The micro-programming & memory controller 61 comprises: an input means 611, a detecting means 612, a first calculating means 613, a second calculating means 614, a converting means 615 and a third calculating means 616. The input means 611 is used for receiving a motor parameter, a d-axis reference current i_ds,ref and a q-axis reference current i_qs,ref. The motor parameter can be a leakage inductance L_σ, and the expression is $L_{\sigma }=\left(1-\frac{L_M^2}{L_{\mathrm{ss}}{L}_{\mathrm{rr}}}\right)L_{\mathrm{ss}}=\frac{L_{\mathrm{ss}}{L}_{\mathrm{rr}}-L_M^2}{L_{\mathrm{rr}}}.$
The leakage inductance L_σ can be a constant. As shown in step 701 of FIG. 7, the motor parameter is determined.

[0045] The input means 611 can receive three phase reference currents 601, then the three phase reference current are converted to the d-axis reference current i_ds,ref and the q-axis reference current i_qs,ref, as shown in step 702 of FIG. 7. Then, it is easy to obtained the d-axis reference current i_ds,ref (t+T_s) and the q-axis reference current i_qs,ref (t+T_s) at a next period in Equations (8) and (9).

[0046] The detecting means 612 is used for detecting three phase actual currents 602 from a motor 64 and converting the three phase actual currents to a d-axis actual current i_ds,actual and a q-axis actual current i_qs,actual, as shown in step 703 of FIG. 7. At a sample time, the d-axis actual current i_ds,actual (t) and the q-axis actual current i_qs,actual (t) can be obtained in real-time, and at a last sample time (t−T_s) the values of the d-axis actual current i_ds,actual (t−T_s) and the q-axis actual current i_qs,actual (t−T_s) are stored in a memory.

[0047] Therefore, the di_ds/dt and di_qs/dt in Equations (12) and (13) can be calculated. The i_ds,actual (t) is the i_ds,actual value at a sample period, the i_qs,actual(t) is the i_qs,actual value at a sample period, and the i_ds,actual (t−T_s) is the stored i_ds,actual value at a last sample period. The i_qs,actual (t−T_s) is the stored i_qs,actual value at a last sample period.

[0048] The first calculating means 613 is utilized to calculate a d-axis counter electromotive force E_ds and a q-axis counter electromotive force E_qs at a sample period according to the motor parameter L_σ, the d-axis actual current i_ds,actual (t) and the q-axis actual current i_qs,actual (t) at the sample period, the d-axis actual current i_ds,actual (t−T_s) and the q-axis actual current i_{qs, actual} (t−T_s) at a last sample period, and a d-axis voltage V_ds and a q-axis voltage V_qs at the last sample period. Therefore, the d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs are calculated by Equations (10) and (11). The method for calculating the d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs is shown in step 704 of FIG. 7.

[0049] The second calculating means 614 is used for calculating a d-axis voltage V_ds and a q-axis voltage V_qs at the sample period according to the motor parameter L_σ, the d-axis actual current i_ds,actual (t) and the q-axis actual current i_qs,actual (t) at the sample period, the d-axis reference current i_ds,ref (t+T_s) and the q-axis reference current i_qs,ref (t+T_s) at a next sample period, the d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs at the sample period calculated by the first calculating means 612. Therefore, the d-axis voltage V_ds and the q-axis voltage V_qs are calculated by Equations (6) and (7). The method for calculating the d-axis voltage V_ds and the q-axis voltage V_qs is shown in step 705 of FIG. 7. The d-axis voltage V_ds and the q-axis voltage V_qs can be used to control the motor 64.

[0050] At the first sample period, the d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs are equal to zero. Then, the d-axis voltage V_ds and the q-axis voltage V_qs can be calculated by Equations (6) and (7) at the first sample period using E_ds=0 and E_qs=0. At the second sample period, the d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs can be calculated by Equations (10) and (11) using the d-axis voltage V_ds and the q-axis voltage V_qs at the first sample period. Then, the d-axis voltage V_ds and the q-axis voltage V_qs can be calculated by Equations (6) and (7) at the second sample period using the d-axis counter electromotive force E_ds and the q-axis counter electromotive force E_qs calculated at the second sample period. The above steps are repeated to calculate the d-axis voltage V_ds and the q-axis voltage V_qs every sample period so as to control the motor 64 of FIG. 6A.

[0051] The converting means 615 is used for converting the d-axis voltage V_ds and the q-axis voltage V_qs to three phase voltage V_as,V_bs, V_cs, as shown in step 706 of FIG. 7. The three phase voltage are converted by the third calculating means 616 to obtain three phase control voltage V_a,control, V_b,control and V_c,control, as shown in step 707 of FIG. 7, according to the three phase voltages V_as,V_bs, V_cs, a peak voltage of a comparing voltage V_Tri—p and a DC voltage of the switch-mode inverter V_d as follows. $\begin{array}{cc}\{\begin{array}{c}{V}_{a,\mathrm{control}}=\left(V_{\mathrm{as}}*V_{\mathrm{Tri\_P}}\right)/\left(1/2*V_d\right)\\ V_{b,\mathrm{control}}=\left(V_{\mathrm{bs}}*V_{\mathrm{Tri\_P}}\right)/\left(1/2*V_d\right)\\ V_{c,\mathrm{control}}=\left(V_{\mathrm{cs}}*V_{\mathrm{Tri\_P}}\right)/\left(1/2*V_d\right)\end{array}& \begin{array}{c}\begin{array}{c}\left(14\right)\\ \left(15\right)\end{array}\\ \left(16\right)\end{array}\end{array}$

[0052] The three phase control voltages are compared with a comparing voltage 603 by the PWM comparator 62, as shown in step 708 of FIG. 7. The comparing voltage 603 is a triangle wave voltage signal. After the PWM comparator 62, three phase switching signals are obtained to input and control the switch-mode inverter 63, as shown in step 709 of FIG. 7. The switch-mode inverter 63 outputs three phase input voltages to the motor 64, as shown in step 710 of FIG. 7.

[0053] Therefore, the three phase input voltages of motor can be calculated by the control method and the control system, according to the invention. The motor 64 can be controlled by the three phase input voltages so as to perform the objective of the three phase actual currents being equal to the three phase reference current.

[0054] FIGS. 8A, 8B, 9A, 9B and 9C show the comparison between the three phase actual currents and the three phase reference currents, according to the control system 60 and the control method of the invention. FIG. 8A shows A phase reference current I_{a3, refer}, FIG. 8B shows A phase actual current I_a3,actual, FIG. 9A shows a comparison between A phase reference current I_a3,refer and A phase actual current I_a3,actual, FIG. 9B shows a comparison between B phase reference current I_b3,refer and B phase actual current I_b3,actual, and FIG. 9C shows a comparison between C phase reference current I_c3,refer and C phase actual current I_c3,actual. FIGS. 8A, 8B, 9A, 9B and 9C are obtained by utilizing the Matlab software to perform the operation of the control system and using a correct leakage inductance. As shown in FIGS. 9A, 9B and 9C, the three phase actual currents I_a3,actual, I_b3,actual, I_c3,actual are almost equal to the three phase reference currents I_a3,refer, I_b3,refer, I_c3,refer, respectively. Therefore, the control system can solve the current phase-lag problem in the conventional PI-PWM current control system.

[0055] To prove the robustness of the control system according to the invention, the Matlab software is also utilized to perform the operation of the control system using an excessive value of leakage induction and an inadequate value of leakage induction.

[0056] FIGS. 10A, 10B, 11A, 11B and 11C show the comparison between the three phase actual currents and the three phase reference currents, according to the control system 60 and the control method of the invention and using an excessive value of leakage induction 1.5 L_σ. FIG. 10A shows A phase reference current I_a4,refer, FIG. 10B shows A phase actual current I_a4,actual, FIG. 11A shows a comparison between A phase reference current I_a4,refer and A phase actual current I_a4,actual, FIG. 11B shows a comparison between B phase reference current I_b4,refer and B phase actual current I_b4,actual, and FIG. 11C shows a comparison between C phase reference current I_c4,refer and C phase actual current I_c4,actual. As shown in FIGS. 11A, 11B and 11C, even using the excessive leakage induction 1.5 L_σ, three phase actual currents I_a4,actual, I_b4,actual, I_c4,actual are still almost equal to the three phase reference currents I_a4,refer, I_b4,refer, I_c4,refer, respectively.

[0057] FIGS. 12A, 12B, 13A, 13B and 13C show the comparison between the three phase actual currents and the three phase reference currents, according to the control system 60 and the control method of the invention and using an inadequate value of leakage induction 0.5 L_σ. FIG. 12A shows A phase reference current I_a5,refer, FIG. 12B shows A phase actual current I_a5,actual, FIG. 13A shows a comparison between A phase reference current I_a5,refer and A phase actual current I_a5,actual, FIG. 13B shows a comparison between B phase reference current I_b5,refer and B phase actual current I_b5,actual, and FIG. 13C shows a comparison between C phase reference current I_c5,refer and C phase actual current I_c5,actual. As shown in FIGS. 13A, 13B and 13C, even using the inadequate leakage induction 0.5 L_σ, three phase actual currents I_a5,actual, I_b5,actual, I_c5,actual are still almost equal to the three phase reference currents I_a5,refer, I_b5,refer, I_c5,refer, respectively.

[0058] Therefore, as shown in FIG. 8A to FIG. 13C, within ±50% error of the leakage inductance L_σ, the three phase actual currents can be almost equal to the three phase reference currents. The control method of the invention not only preserves the merit of constant switching frequency of the conventional PI-PWM control system, but also can eliminate the motor current phase-lag problem of the conventional PI-PWM control system. Also, the control method of the invention will have the following characteristics and economic advantages such as high accuracy, fast response, low cost and robustness etc.

[0059] In the above embodiments, the leakage inductance L_σ is the only variable. The leakage inductance L_σ is determined usually by the parameter test of the motor. If the parameter test is not correct, the value of the leakage inductance L_σ may be above the correct value in the error ±50%. If the incorrect leakage inductance L_σ with above ±50% error is utilized to calculate the d-axis voltage V_ds and the q-axis voltage V_qs, the three phase actual currents may be not equal to the three phase reference currents. Therefore, the leakage inductance L_σ must be real-time auto-tuning as follows.

[0060] At a sample period, Equations (6) and (7) are modified as follows. $\begin{array}{cc}{V}_{\mathrm{ds1}}=L_{\sigma }\frac{d{i}_{\mathrm{ds1}}}{dt}-L_{\sigma }{\omega }_e{i}_{\mathrm{qs1}}+E_{\mathrm{ds1}}=m_{\mathrm{d1}}{L}_{\sigma }+E_{\mathrm{ds1}}& \left(17\right)\\ V_{\mathrm{qs1}}=L_{\sigma }\frac{d{i}_{\mathrm{qs1}}}{dt}+L_{\sigma }{\omega }_e{i}_{\mathrm{ds1}}+E_{\mathrm{qs1}}=m_{\mathrm{q1}}{L}_{\sigma }+E_{\mathrm{qs1}}& \left(18\right)\end{array}$

[0061] Wherein in Equations (17) and (18) the variables are the leakage inductance L_σ, E_ds1, and E_qs1, the m_d1 and m_q1 are defined and obtained as follows $m_{\mathrm{d1}}=\left(\frac{d{i}_{\mathrm{ds1}}}{dt}-{\omega }_e{i}_{\mathrm{qs1}}\right)·m_{\mathrm{q1}}=\left(\frac{d{i}_{\mathrm{qs1}}}{dt}+{\omega }_e{i}_{\mathrm{ds1}}\right).$

[0062] At a next sample period, Equations (6) and (7) are modified as follows. $\begin{array}{cc}{V}_{\mathrm{ds2}}=L_{\sigma }\frac{d{i}_{\mathrm{ds2}}}{dt}-L_{\sigma }{\omega }_e{i}_{\mathrm{qs2}}+E_{\mathrm{ds2}}=m_{\mathrm{d2}}{L}_{\sigma }+E_{\mathrm{ds2}}& \left(19\right)\\ V_{\mathrm{qs2}}=L_{\sigma }\frac{d{i}_{\mathrm{qs2}}}{dt}+L_{\sigma }{\omega }_e{i}_{\mathrm{ds2}}+E_{\mathrm{qs2}}=m_{\mathrm{q2}}{L}_{\sigma }+E_{\mathrm{qs2}}& \left(20\right)\end{array}$

[0063] Wherein in Equations (19) and (20) the variables are the leakage inductance L_σ, E_ds2 and E_qs2, the m_d2 and M_q2 are defined and obtained as follows $m_{\mathrm{d2}}=\left(\frac{d{i}_{\mathrm{ds2}}}{dt}-{\omega }_e{i}_{\mathrm{qs2}}\right)·m_{\mathrm{q2}}=\left(\frac{d{i}_{\mathrm{qs2}}}{dt}+{\omega }_e{i}_{\mathrm{ds2}}\right).$

[0064] Equation (19) minus Equation (17) is as follows.
(V_ds2−V_ds1)=(m_d2−m_d1)L_σ+(E_ds2−E_ds1)≈(m_d2−m_d1)L_σ  (21)
In Equation (21), because the sample time is very small, the difference between the E_ds2 and E_ds1 can be ignored so that E_ds2≈E_sd1. Therefore, the leakage inductance L_σ,d is estimated using d-axis Equations (17) and (19). The expression of the leakage inductance L_σ,d is as follows.
L_σ,d=(V_ds2−V_ds1)/(m_d2−m_d1)  (22)
Similarly, Equation (20) minus Equation (18), and the difference between the E_qs2 and E_qs1 can be ignored. The leakage inductance L_σ,q is estimated using q-axis Equations (18) and (20). The expression of the leakage inductance L_σ,q is as follows.
 L_σ,q=(V_qs2−V_qs1)/(m_q2−m_q1)  (23)

[0065] Therefore, the leakage inductance L_σ is defined as an average of the leakage inductance L_σ,d and the leakage inductance L_σ,q as follows.
L_σ=0.5*(L_σ,d+L_σ,q)  (24)

[0066] The leakage inductance L_σ is calculated by the above Equations (17) to (24). The calculated leakage inductance L_σ is inputted to Equations (6), (7), (10) and (11) to calculate V_ds, V_qs, E_ds and E_qs for two sample periods as V_ds1, V_qs1, V_ds2 and V_qs2. Therefore, the value of V_ds1, V_qs1, V_ds2 and V_qs2 in Equations (17), (18), (19) and (20) can be obtained to calculated a new leakage inductance L_σ. Every two sample periods, the leakage inductance L_σ is calculated and updated to input to Equations (6), (7), (10) and (11) to calculate V_ds, V_qs, E_ds and E_qs. Therefore, Equations (17) to (24) is calculated every two sample periods to obtained a new leakage inductance L, so that the leakage inductance L_σ can approach the correct value.

[0067] Referring to FIG. 14A to FIG. 14D, it shows a simulation using the Matlab software. As shown in FIG. 14D, the leakage inductance L_σ can approach to the correct value and be stable within 0.25 second.

[0068] Referring to FIG. 6B, the micro-programming & memory controller 61 further comprises an fourth calculating means 617 for estimating the leakage inductance L_σ according to the above method. Therefore, by using the estimation of the leakage inductance L_σ the control system 60 can improve the robustness.

[0069] While an embodiment of the present invention has been illustrated and described, various modifications and improvements can be made by those skilled in the art. The embodiment of the present invention is therefore described in an illustrative but not restrictive sense. It is intended that the present invention may not be limited to the particular forms as illustrated, and that all modifications which maintain the spirit and scope of the present invention are within the scope as defined in the appended claims.