Method and apparatus for estimating rotor position and for sensorless control of a switched reluctance motor

Abstract

A discrete rotor position estimation method for a synchronized reluctance motor is provided. A d.c.-link voltage Vdc and a phase current Iph are sensed. A flux-linkage λph of an active phase is calculated from the sensed d.c.-link voltage Vdc and the sensed phase current Iph. The calculated flux-linkage λph is compared with a reference flux-linkage λr. The reference flux-linkage λr corresponds to a reference angle θr which lies between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor. An estimated rotor position θcal is obtained only once when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

Description

TECHNICAL FIELD

The present invention relates to the development of closed loop control techniques for switched reluctance motors (SRMs) without a shaft position sensor.

BACKGROUND ART

A switched reluctance motor (SRM) is energized phase by phase in sequence to generate reluctance torque and enable smooth motor rotation. A schematic diagram of a three phase switched reluctance motor is shown in FIG. 1. The number of strokes (N) in SRM per one mechanical revolution is dependent on the number of phases (M) and the number of rotor poles (P) and is given by, N=M*P  (1).

Therefore, the stroke angle(S) in mechanical degrees is defined as, S=360°/N  (2).

When the number of poles is very large and the stroke angle is very small, the SRM is typically operated in open loop as a variable reluctance stepper motor and needs no knowledge of rotor position information during running condition. On the other hand, when the number of poles is small and the stroke angle is very large, the SRM is generally operated in closed loop during running condition and hence, the knowledge of accurate rotor position information is very important to rotate the motor.

Accurate rotor position information is typically obtained from a shaft position sensor. Shaft position sensors are expensive and have reliability problems and hence, the sensorless operation of SRM is desired.

In a typical position sensorless operation of SRM, rotor position estimation is carried out either discretely i.e. once per stroke angle or continuously. Discrete rotor position estimation is ideal for applications where slow speed response is required where as, continuous rotor position estimation is carried out in applications where fast speed response is desired. Rotor position estimation of the SRM can be carried out from the pre-determined knowledge of its non-linear per phase flux-linkage/current characteristics or the inductance/current characteristics as shown in FIG. 2 and FIG. 3 respectively. Several non-linear analytical models based on the flux-linkage and the inductance characteristics of the motor with respect to the rotor position have been proposed [S. Saha, K. Ochiai, T. Kosaka, N. Matsui and Y. Takeda “Developing a sensorless approach for switched reluctance motors from a new analytical model”, 1999 IEEE-IAS Annual meeting, Vol. 1, pp 525-532, and G. Suresh, B. Fahimi, K. M. Rahaman and M. Eshani “Inductance based position encoding for sensorless SRM drives”, 1999 IEEE-PESC Annual Meeting, pp 832-837.] to calculate the exact rotor position after sensing the d.c. link voltage V_dc and the phase current I_ph of the inverter drive circuit as shown in FIG. 4. A voltage sensor VS is used to sense the d.c. link voltage and three current sensors CS are used to sense the current flowing in each phase winding 1u, 1v or 1w of the SRM 1. The traditional lock and forced drive method with open loop operation can be applied for the starting of these motors at no load. Rotor position estimation scheme is enabled at a minimum speed (ω_min) which is normally 10% of the rated speed or higher. The motor moves from open loop to closed loop operation after the rotor position estimation scheme is enabled and then subsequently load is applied to the motor. The entire closed loop control scheme with the rotor position estimation is implemented with the help of a micro-controller or a digital-signal processor.

DISCLOSURE OF INVENTION

In a first aspect of the invention, provided is a discrete rotor position estimation method for a synchronized reluctance motor. According to this estimation method, d.c.-link voltage V_dc and a phase current I_ph are sensed. A flux-linkage λ_ph of an active phase is calculated from the sensed d.c.-link voltage V_dc and the sensed phase current I_ph. The calculated flux-linkage λ_ph is compared with a reference flux-linkage λ_r. The reference flux-linkage λ_r corresponds to a reference angle θ_r which lies between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor. An estimated rotor position θ_cal is obtained only once based on the comparison result when the calculated flux-linkage λ_ph is greater than the reference flux-linkage λ_r. From the estimated rotor position θ_cal, incremental rotor angle Δθ for every PWM interrupt is calculated. The knowledge of Δθ is finally used to control the motor.

In a second aspect of the invention, provided is a discrete rotor position estimation method for a synchronized reluctance motor. According to the method in the second aspect, a d.c.-link voltage V_dc and a phase current I_ph are sensed. A flux-linkage λ_ph of an active phase is calculated from the sensed d.c.-link voltage V_dc and the sensed phase current I_ph. The calculated flux-linkage λ_ph is compared with either two or three reference flux-linkages such as λ_r1, . . . . The reference flux-linkages λ_r1, . . . correspond to a reference rotor angles θ_r1, . . . all of them lying between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor. Rotor positions θ_cal1, . . . are obtained based on the comparison results, twice or thrice when the calculated flux-linkage λ_ph is greater than the reference flux-linkage λ_r1, . . . . From the estimated rotor positions θ_cal1, . . . incremental rotor angles for every PWM interrupt are also calculated twice or thrice such as Δθ₁, . . . . The values of the incremental rotor angles are averaged to obtain the final incremental rotor angle Δθ. The knowledge of the final incremental rotor angle Δθ is used to control the motor.

In the first aspect, the value of the estimated rotor position θ_cal obtained from comparison result typically equals the reference rotor angle θ_r. However, the estimated rotor position θ_cal may be further modified with an incremental angle φ corresponding to the reference rotor angle θ_r to obtain a more accurate estimated rotor position. This idea can also be extended to the second aspect.

In a third aspect of the invention, provided is a discrete rotor position estimation method for a synchronized reluctance motor. According to the method, a d.c.-link voltage V_dc and a phase current I_ph are sensed. A flux-linkage λ_ph of an active phase is calculated from the sensed d.c.-link voltage V_dc and the sensed phase current I_ph. The calculated flux-linkage λ_ph is compared with a reference flux-linkage λ_r. The reference flux-linkage λ_r corresponds to a reference angle θ_r which lies between angles corresponding to aligned rotor position and non-aligned rotor position of the synchronized reluctance motor. An estimated rotor position θ_cal is calculated from the calculated flux-linkage λ_ph using either one of the inductance model or the flux linkage model of the active phase, only once when the calculated flux-linkage λ_ph is greater than the reference flux-linkage λ_r. In this case, the ideal instant to estimate the rotor position may be at one PWM interrupt before the next phase is turned ON. From the estimated rotor position θ_cal, incremental rotor angle Δθ for every PWM interrupt is calculated. The knowledge of Δθ is finally used to control the motor.

The above idea can be also extended to estimate the rotor position either twice or thrice such as θ_cal1, . . . from the calculated flux-linkage λ_ph using either one of the inductance model or the flux linkage model of the active phase, at every consecutive PWM interrupt when the calculated flux-linkage λ_ph is greater than the reference flux-linkage λ_r. From the estimated rotor positions θ_cal1, . . . incremental rotor angles for every PWM interrupt are also calculated twice or thrice such as Δθ₁, . . . . The values of the incremental rotor angles are averaged to obtain the final incremental rotor angle Δθ. The knowledge of the final incremental rotor angle Δθ is used to control the motor.

In a fourth aspect of the invention, provided is a discrete rotor position estimation method for a synchronized reluctance motor. According to the method, a phase inductance of the synchronized reluctance motor is detected. A minimum region of the phase inductance during turn-on of an active phase is identified. An approximate rotor position θ_app is determined from the identified minimum inductance region. From the approximate rotor position θ_app, incremental rotor angle Δθ for every PWM interrupt is calculated. The knowledge of Δθ is finally used to control the motor.

In a fifth aspect of the invention, provided is a control method of a synchronized reluctance motor and a technique to obtain the incremental rotor angle Δθ for every PWM interrupt. According to the control method, the estimated rotor position θ_cal is obtained by the previously described estimation methods. An absolute rotor position θ_abs is calculated from the estimated rotor position θ_cal by adding a stroke angle of the motor. The incremental rotor angle Δθ is determined by processing an error between the absolute rotor position θ_abs and a finally estimated rotor position θ_est through either one of a proportional-integral (PI) control and a proportional control. The finally estimated rotor position θ_est in every predetermined period is generated by adding the incremental rotor angle Δθ to the finally estimated rotor position θ_est in the previous cycle. Turn-on and turn-off angles of each phase is controlled based on the finally estimated rotor position θ_est.

In a sixth aspect of the invention, provided is a control method of a synchronized reluctance motor and a technique to obtain the incremental rotor angle Δθ for every PWM interrupt. According to the control method, an incremental rotor angle Δθ is calculated by counting the number of PWM interrupts between two consecutive instants when the estimated rotor position θ_cal is obtained by the estimation method according to the invention. Delays to turn-off an active phase and turn-on the next phase is generated, in which the delays are normally defined with respect to the reference rotor position θ_r. The delays are adjusted with the estimated rotor position θ_cal to turn-off the active phase and turn-on the next phase. A turn-on angle θ_on and a turn-off angle θ_off of each phase of the motor are controlled based on the adjusted delays decided by the incremental rotor angle Δθ.

In the above control methods, a speed ω of the motor may be calculated from the incremental rotor angle Δθ in a relatively slower timer interrupt compared to a PWM interrupt, and a turn-on angle θ_on and a turn-off angle θ_off of each phase of the motor may be varied continuously based on the speed ω and the torque demand of the motor.

In a seventh aspect of the invention, provided is a control method of a synchronized reluctance motor. According to the control method, a peak of a phase current and a negative change rate of phase current in each phase are continuously monitored. Turn-off angle is kept fixed and turn-on angle is advanced so that a pre-determined peak phase current and a negative rate of change of phase current corresponding to the maximum torque are achieved.

In the control method, instead of monitoring the negative rate of change of phase current, a lead angle φ between the peak current and the peak flux in each phase may be detected to judge the maximum torque at the rated speed condition.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 shows the cross-sectional view of a switched reluctance motor.

FIG. 2 shows the per phase flux-linkage/current characteristics with respect to rotor position of a switched reluctance motor.

FIG. 3 shows the per phase inductance/current characteristics with respect to rotor position of a switched reluctance motor.

FIG. 4 shows a typical inverter circuit for the switched reluctance motor.

FIG. 5 shows the region where the reference rotor position θ_r corresponding to the reference flux-linkage λ_r in rotor position estimation technique I can be defined

FIG. 6 shows the instant at which estimation of the exact rotor position θ_cal is carried out for the rotor position estimation technique II.

FIG. 7 shows the instant at which the measured actual phase current I_ph and the estimated phase current Is are compared in the discrete rotor position estimation technique III.

FIG. 8 shows the incremental change of per phase flux-linkage λ_ph in every PWM interrupt of the processor in accordance with the first exemplary embodiment.

FIG. 9 shows the block diagram for closed loop control using control strategy I in accordance with the first exemplary embodiment.

FIG. 10 shows the turn-on and the turn-off delay with respect to the reference rotor position θ_r in accordance with the second exemplary embodiment.

FIG. 11 shows the block diagram for closed loop control using control strategy II in accordance with the second exemplary embodiment.

FIG. 12A shows two reference flux linkages corresponding to two reference rotor positions in accordance with the third exemplary embodiment.

FIG. 12B shows three reference flux linkages corresponding to three reference rotor positions in accordance with the third exemplary embodiment.

FIG. 13 shows the block diagram for closed loop control using control strategy I in accordance with the third exemplary embodiment.

FIG. 14 shows the block diagram using the discrete rotor position estimation technique III and the control strategy I.

FIG. 15 shows typical per phase current waveforms and per phase flux waveform at the rated speed and the maximum torque condition in accordance with the fifth exemplary embodiment.

FIG. 16 shows the peak current detector block introduced in the closed loop block diagram using discrete rotor position technique I and control strategy I.

FIGS. 17A through 17G show examples of applications using the proposed rotor position estimation schemes and control strategies for the switched reluctance motor.

BEST MODE FOR CARRYING OUT THE INVENTION

The present invention relates to the closed loop control techniques of switched reluctance motors (SRM) without a shaft position sensor where the discrete rotor position estimation techniques are followed. Three typical discrete rotor estimation techniques I, II and III are described briefly below.

(Discrete Rotor Position Estimation Technique I)

In the discrete rotor position estimation technique I, the per phase flux-linkage at rotor position θ_r as shown in FIG. 5 is chosen as a reference flux-linkage λ_r and is expressed as a function of phase current I_ph by a simple polynomial expression such as, λ_r=ΣA_nI_phⁿ  (3).

θ_r is defined as any rotor position near the mid-position between the aligned and the non-aligned rotor position which is shown by the shaded area in FIG. 5. Mathematically θ_r can be expressed as, θ_r=θ_m±α  (4) where, θ_m is the exact middle position between the aligned and the non-aligned rotor position and α is the deviation angle whose maximum value α_max is equal to 30° electrical from the middle position θ_m and α_max are expressed in mechanical degrees as follows, θ_m=(θ_a+θ_n)/2  (5) and α_max=30°/P  (6).

θ_a is defined as the aligned position i.e. when a rotor pole aligns with a stator pole. θ_n is defined as the non-aligned position i.e. when a rotor pole is in between two stator poles as shown in FIG. 1. The value of n in the polynomial equation (3) is dependent on the magnetic design of a particular motor. The reason why the near middle position θ_m is selected as a reference position is that the flux linkage varies with the position θ in a high changing rate in vicinity of the middle position θ_m and thus the position θ with respect to the flux linkage can be easily determined.

Use of fast processors will enable to calculate polynomial expressions easily. The reference flux-linkage value λ_r with respect to phase current I_ph is pre-determined off-line by locking the rotor at position θ_r and carrying out standard experiments. During motor rotation, the flux-linkage per phase λ_ph is calculated on-line at every PWM interrupt or at every half-cycle PWM interrupt of the processor by sensing the d.c.-link voltage V_dc and the phase current I_ph of the inverter circuit as shown in FIG. 4. It is given by, λ_ph=(V_dc−I_ph*R−V_p)*dt  (7) where R is the phase resistance of the SRM in the high frequency mode and V_p is the voltage drop across the power devices. An average value of mutual flux (maximum 10% of the maximum per phase flux) should be considered for more accurate calculation of the per phase flux-linkage λ_ph. The on-line calculated per phase flux-linkage λ_ph is continuously compared with the reference flux-linkage λ_r in the processor. At the instant when λ_ph is equal to λ_r, the absolute discrete rotor position θ_abs in mechanical degrees is determined or estimated as given below θ_abs=K*360°/N+θ_cal  (8) where, θ_cal=θ_r and K=(N−1).

From the information of θ_abs, closed loop control techniques are generated to rotate the motor.

(Discrete Rotor Position Estimation Technique II)

In the discrete rotor position estimation technique II, similarly the d.c.-link voltage V_dc and the phase current I_ph are sensed and the flux-linkage per phase λ_ph is calculated on-line from equation (7) at every PWM interrupt of the processor. From the knowledge of λ_ph the exact rotor position θ_cal is calculated only once either from the flux-linkage model or the inductance model of the active phase when the calculated flux-linkage λ_ph is greater than the reference flux-linkage λ_r. The ideal instant to estimate the rotor position may be at one PWM interrupt before the next phase is turned ON. This is shown in FIG. 6. This concept is ideal when the turn-on angle is advanced at the rated speed for high torque applications. The absolute discrete rotor position θ_abs in one mechanical cycle is always calculated from the information of the exact rotor position θ_cal and is expressed by equation (8). From the knowledge of θ_abs control techniques are again generated to rotate the motor in closed loop.

(Discrete Rotor Position Estimation Technique III)

The discrete rotor position estimation technique III is very simple and does not involve any exact rotor position estimation θ_cal as described in techniques I and II. This scheme includes identifying the minimum inductance region during the turn-on of a new phase so that the turn-on angle always lies in the minimum inductance region. This has been made possible by comparing the measured actual phase current I_ph with an estimated phase current I_s after a finite time interval to from the turn-on of a new phase as shown in FIG. 7. The estimated current I_s is given by, I_s=(V_dc−I_phR−V_p)*t₀/L_min  (9).

The value of minimum inductance L_min is constant for a particular motor and does not vary with the phase current I_ph as shown in FIG. 3. The minimum inductance region normally extends to few degrees and is typically dependent on the magnetic design of the motor. Hence, the approximate rotor position θ_app is known when the minimum inductance region is identified. Therefore, instead of exact rotor position θ_cal, the knowledge of approximate rotor position θ_app is used to calculate the absolute discrete rotor position θ_abs and rotate the motor in closed loop.

In the discrete rotor position estimation techniques of I, II and III, a rotor position estimation error less than one mechanical degree is difficult to achieve. A small error in rotor position estimation at a particular speed can decrease the motoring torque i.e. the motor performance to a large extent. Hence, optimum control techniques are invented to minimize rotor position estimation error so that high performance is always achieved for the sensorless drive of SRM's. In addition, control techniques are invented to obtain the high torque even if there is a rotor position estimation error.

First Exemplary Embodiment

In the discrete rotor position estimation technique I, the per phase flux-linkage λ_ph is calculated on-line in every PWM interrupt or every PWM half-cycle interrupt by sensing the d.c.-link voltage V_dc and the phase current I_ph and is defined by the equation (7) above. The incremental change of per phase flux-linkage λ_ph at the valley of every PWM interrupt of the processor is shown in FIG. 8 where dt is the time interval between two PWM interrupts and θ_on is the turn-on angle and θ_off is the turn-off angle. Any digital technique cannot exactly define the instant when the per phase flux-linkage λ_ph is equal to the reference flux-linkage λ_r. In this embodiment, the processor within a PWM interrupt identifies an instant when the per phase flux-linkage λ_ph is greater than λ_r as shown in FIG. 8. Therefore, for more accurate rotor position estimation, θ_abs of equation (8) in the discrete rotor position technique I has to be modified. In the simplest technique, the modification of θ_abs in equation (8) can be done by adding an angle t such that, θ_abs=K*360°/N+(θ_r+φ)  (10).

The combined angle (↓_r+φ) now defines the exact rotor position θ_cal. The angle φ can be obtained from a pre-determined two dimensional look-up table in which Δλ and phase current I_ph are the variable parameters. Δλ is defined as the difference between λ_ph and λ_r.

Alternatively, by the discrete rotor position estimation technique II, exact rotor position θ_cal should be calculated at the instant when the per phase flux-linkage λ_ph is greater than the reference flux-linkage λ_r. Exact rotor position estimation can be carried out by analytical methods based on the flux-linkage model or the inductance model of the SRM. The absolute discrete rotor position θ_abs in one mechanical cycle is always calculated from the information of the exact rotor position θ_cal and is expressed by equation (8). The value of θ_abs which is defined in mechanical degrees in equation (8) can also be expressed in electrical degrees by considering the number of rotor poles (P).

FIG. 9A shows an entire block diagram of a driving apparatus of SRM for achieving the closed loop control with control strategy I according to the invention.

The SRM 1 is driven by the inverter 2 which includes switching devices T1 to T6 as shown in FIG. 4. The gate drive circuit 22 generates pulse width modulated base drive signals to turn-on and turn-off the switching devices T1 to T6 and generate a phase current. PWM control is done by the block 21 with PWM interrupt timer. Turn-on angle θ_on and turn-off angle θ_off are calculated by the block 19. The turn-on and turn-off controller 18 determines the turn-on and the turn-off timings of each phase based on the calculated turn-on and turn-off angles θ_on and θ_off and the estimated rotor position θ_est.

In the control strategy I involving the discrete rotor position estimation technique I, phase current I_ph and d.c.-link voltage V_dc are sensed to calculate the per phase flux-linkage λ_ph (by block 11). The flux-linkage λ_ph is compared with the reference flux-linkage λ_r (by block 11a). Referring to the comparison result, when the phase flux-linkage λ_ph is greater than the reference flux-linkage row the rotor position θ_cal is once calculated (by block 12). The absolute discrete rotor position θ_abs is calculated (by block 13) and continuous rotor position θ_est is estimated from the absolute discrete rotor position θ_abs in the following way (by block 14). At the instant when the exact rotor position θ_cal is calculated (by block 12), the error between the absolute rotor position θ_abs and the estimated rotor position θ_est is calculated and processed in a proportional-integral (PI) control method (by block 15) to give the incremental rotor angle Δθ in every PWM interrupt. Δθ is expressed as Δθ=Kp*θ_err+ΣKI*θ_err  (11) where, θ_err=(θ_abs−θ_est). At the start of rotor position estimation, θ_est is initialized as θ_abs and then calculated as follows θ_est(n)=θ_est(n−1)+Δθ  (12).

Instead of a proportional-integral control method, a proportional control method can also be used.

The turn-on angle θ_on and turn-off angle θ_off of each phase is controlled by θ_est which can be expressed either in mechanical or electrical degrees (by block 18). For achieving the turn-on and the turn-off of each phase at any point in between two PWM interrupts, a very fast timer interrupt compared to the PWM interrupt has been defined.

The speed (ω) of the motor is calculated (by block 16) from the incremental rotor angle Δθ in a relatively slow timer interrupt compared to the PWM interrupt. It is given by, ω=Δθ*fs  (13) where, fs is the PWM carrier frequency. The speed ripple is filtered by processing it through a low-pass filter (LPF) 17. If the per phase flux-linkage λ_ph is calculated in every half-cycle PWM interrupt, the incremental rotor displacement Δθ has to be defined for every half-cycle PWM interrupt and the speed (ω) of the motor is expressed as, ω=2*Δθ*fs  (14).

The turn-on and the turn-off angle together with the phase voltage reference (V_ph*) or phase current reference (I_ph*) are continuously calculated within the processor and are varied depending on the speed (ω) and the torque demand of the motor (by blocks 18-20). In closed loop control, the difference between the commanded speed (ω*) and the calculated speed (ω) is processed through a proportional-integral (PI) control block 20 to generate either the phase voltage reference (V_ph*) or the phase current reference (I_ph*). Either the phase voltage reference (V_ph*) is compared with the PWM carrier waveform or the phase current reference (I_ph*) is compared with the original phase current I_ph to finally generate the pulse-width modulated (PWM) base drive waveforms (by block 21).

A gate drive circuit 22 controls an inverter 2 using the pulse-width modulated (PWM) waveforms and turn-on and turn-off angles from the turn-on and turn-off controller 18 to drive the SRM 1.

Second Exemplary Embodiment

FIG. 11 shows an entire block diagram of a driving apparatus of SRM for achieving the closed loop control with a control strategy II according to the invention.

In the control strategy II involving the discrete rotor position estimation technique I, instead of calculating θ_est continuously to control the turn-on and turn-off angle the incremental rotor displacement Δθ in every PWM interrupt can be calculated in an alternative way which can generate appropriate delays to turn-off the active phase and turn-on the next phase. The speed (ω) of the motor is also calculated from the incremental rotor angle Δθ (by block 16), and the closed loop control can be executed in the similar manner described in the first exemplary embodiment. In the control strategy II, Δθ can be calculated by counting the number of PWM interrupts between two consecutive instants when the per phase flux-linkage λ_ph is greater than the reference flux-linkage λ_r (by blocks 13a and 15a), that is, between the instants of θ_cal(n) and θ_cal(n−1). It is given by, Δθ=stroke angle(S)/number of PWM interrupts  (15).

In equation (15), Δθ is expressed in mechanical degrees but also can be expressed in electrical degrees by considering the number of rotor poles. The exact rotor position θ_cal at the instant when the per phase flux-linkage λ_ph is greater than the reference flux-linkage λ_r is calculated in a similar way as described in the first exemplary embodiment. Calculation of θ_cal helps to adjust the turn-on and the turn-off delay and this is explained as follows. The turn-off delay (x°) of the active phase and the turn-on delay (y°) of the next phase are always defined with respect to the reference rotor position θ_r as shown in FIG. 10 for a specific speed (ω) of the motor (by block 18b). The difference φ between the exact rotor position θ_cal and the reference rotor position θ_r is always estimated so that the turn-off delay (x°) of the active phase and the turn-on delay (y°) of the next phase are modified as X₁° and Y₁° respectively (by block 18a). The modified turn-off delay (X₁°) and the turn-on delay (Y₁°) are now redefined with respect to the exact rotor position θ_cal, and can be expressed as, X₁°=x°−φ  (16) and Y₁°=y°−φ  (17).

For achieving the turn-on and the turn-off of each phase at any point in between two PWM interrupts, a very fast timer interrupt compared to the PWM interrupt has been defined as explained in the first exemplary embodiment. In addition, if the per phase flux-linkage λ_ph is calculated in every half-cycle PWM interrupt, the incremental rotor displacement Δθ has to be defined for every half-cycle PWM interrupt.

The above control technique will help to achieve both high performance as well as maximum torque of the motor at the rated speed.

Third Exemplary Embodiment

The discrete rotor position estimation technique I described in the first and the second exemplary embodiment includes comparing the per phase flux-linkage λ_ph with only one reference flux-linkage λ_r defined for a rotor position θ_r. Hence, in both cases, exact rotor position θ_cal and incremental rotor angle Δθ are calculated only once at the instant when the per phase flux-linkage λ_ph is greater than the reference flux-linkage λ_r during the conduction of each active phase. The discrete rotor position estimation technique I can be extended by defining two reference flux-linkages λ_r1 and λ_r2 at rotor positions θ_r1 and θ_r2 respectively as shown in FIG. 12A or three reference flux-linkages λ_r1, λ_r2 and λ_r3 at rotor positions θ_r1, θ_r2 and θ_r3 respectively as shown in FIG. 12B by polynomial expression in phase current I_ph. All these rotor positions lie near the mid-position θm with a deviation angle α_max of 30°. Hence, the on-line estimated per phase flux-linkage λ_ph is compared with two or three reference flux-linkages and accordingly two exact rotor positions (θ_cal1 and θ_cal2) or three exact rotor positions (θ_cal1, θ_cal2 and θ_cal3) are calculated during the active conduction of each phase based on the comparison result.

Alternatively, the discrete rotor position estimation technique II can be extended by calculating the exact rotor positions either twice (θ_cal1 and θ_cal2) or thrice (θ_cal1, θ_cal2 and θ_cal3) from the calculated flux-linkage θ_ph by using either one of the inductance model or the flux linkage model of the active phase, at every consecutive PWM interrupt when the calculated flux-linkage λ_ph is greater than the reference flux-linkage λ_r corresponding to reference rotor angle θ_r.

Now from θ_cal1, . . . by following either the control strategy I or the control strategy II, incremental rotor angle for every PWM interrupt is also calculated twice (Δθ₁ and Δθ₂) or thrice (Δθ₁, Δθ₂ and Δθ₃) during the active conduction of each phase. The final incremental rotor angle Δθ for every PWM interrupt is the average of all the calculated incremental rotor angle. The final incremental rotor angle Δθ is used for the estimation of rotor position θ_est in control strategy I or providing the necessary delay for the turn-on or the turn-off a phase as described in control strategy II. The speed (ω) of the motor is also calculated from the final incremental rotor angle Δθ. A control block diagram using the control strategy I and calculating the incremental rotor angle for every PWM interrupt twice is shown in FIG. 13. In this figure, absolute discrete rotor positions θ_abs1 and θ_abs2 are calculated, respectively (by blocks 13a and 13b). Then the incremental rotor angles (Δθ₁ and Δθ₂) are calculated respectively and averaged to obtain Δθ (by blocks 15a to 15c). A control block diagram for calculating the incremental rotor angle thrice can be similarly obtained.

Fourth Exemplary Embodiment

In the discrete rotor position estimation technique II, the d.c-link voltage V_dc and the phase current I_ph are sensed and the flux-linkage per phase λ_ph is calculated on-line from equation (7) at every PWM interrupt of the processor. From the knowledge of λ_ph, the exact rotor position θ_cal is calculated either from the flux-linkage model or the inductance model of the active phase at one PWM interrupt before the next phase is turned ON. This is shown in FIG. 6. This concept is ideal when the turn-on angle is advanced at the rated speed for high torque applications. The absolute discrete rotor position θ_abs in one mechanical cycle is always calculated from the information of the exact rotor position θ_cal and is expressed by equation (7).

From the knowledge of θ_abs, control strategies I and II are again applied in the discrete rotor position estimation technique II to rotate the motor in closed loop.

Fifth Exemplary Embodiment

There exist many industrial applications which require motor operation only at the rated speed and the maximum torque condition instead of the continuous variable speed-torque operation. For such applications, the discrete rotor position estimation technique III is proposed.

FIG. 14 shows an entire block diagram of driving apparatus of SRM using the discrete rotor position estimation technique III and the control strategy I. It should be noted that identical block diagram can be drawn for the discrete rotor position estimation technique III and the control strategy II.

The discrete rotor position technique III is very simple and does not involve any exact rotor position estimation θ_cal from the flux-linkage model or the inductance model as described in techniques I and II. The discrete rotor position technique III includes identifying the minimum inductance region during the turn-on of a new phase so that the turn-on angle always lies in the minimum inductance region. This has been made possible by comparing the measured actual phase current I_ph with the estimated phase current I_s after a finite time interval t₀ from the turn-on of a new phase as shown in FIG. 7. The estimated current I_s is calculated from equation (9) in the block 11a.

The locking of a particular phase and open loop starting technique with a forced drive always ensures that the turn-on angle is initially synchronized with the minimum inductance region. Coinciding the turn-on angle always with the minimum inductance region also guarantees perfect synchronous operation of the motor. Initially the turn-on angle can be anywhere between θ₁ and θ₂ as shown in FIG. 7. However, the turn-on angle is initialized with θ₀ which is the middle position between θ₁ and θ₂. Therefore, the approximate rotor position θ_app at t₀ can be easily calculated. In this embodiment, instead of exact rotor position θ_cal, approximate rotor position θ_app is obtained (in block 12a), and the knowledge of approximate rotor position θ_app is used to calculate the absolute discrete rotor position θ_abs (in block 13b) and rotate the motor in closed loop by applying either control strategy I or II.

Sixth Exemplary Embodiment

In the discrete rotor position estimation techniques of I, II and III, a rotor position estimation error less than one mechanical degree is difficult to achieve. At the rated speed and the maximum torque condition, the SRM operates in the single pulse mode. Typical per phase current waveforms with a variable turn-on angle and a fixed turn-off angle at the rated speed are shown in FIG. 15. The per phase flux waveform corresponding to the rated speed and the maximum torque condition is also shown in FIG. 15. From FIG. 15 it can be easily understood that the peak phase current (I_ph(peak)) which is immediately followed by the negative rate of change of phase current has to be continuously monitored if a specific application requires maximum torque operation at the rated speed. Therefore, keeping the turn-off angle fixed, the turn-on angle is advanced until the pre-determined reference peak phase current (I_p*) and a reference negative rate of change of phase current (dI_ph*/dt) corresponding to the maximum torque are achieved. Instead of monitoring the negative rate of change of phase current, the lead angle (φ) between the peak current (I_ph(peak)) and the peak flux (λ_ph(peak)) can be also monitored to judge the maximum torque at the rated speed condition. As shown in FIG. 16, the peak current detector block 30 is introduced in the closed loop block diagram using discrete rotor position technique I and control strategy I.

INDUSTRIAL APPLICABILITY

The proposed discrete position sensorless estimation schemes for the switched reluctance motors are ideal for several automotive applications such as the compressor drive in car air-conditioners for a conventional gasoline vehicle, an electric vehicle and a hybrid electric vehicle.

Examples of applications using the proposed rotor position estimation schemes and control strategies for the switched reluctance motor are shown in FIGS. 17A through 17G.

FIG. 17A is an explanatory diagram of an application of the invention to an electric vehicle. The drive unit 100 includes a switched reluctance motor 1, an inverter 2, a current sensor 5, and SRM control system 50. Typically it means that the SRM control system 50 has the same structure as that shown in FIG. 9, 11, 13, 14 or 16 except for the SRM 1. The drive unit 100 is coupled with a main ECU (Electrical Control Unit) 61 and a gear 63.

FIG. 17B also shows an application of the invention to a hybrid electric vehicle. The engine 65 is further provided to the structure of the FIG. 17B. FIG. 17C shows an application of the invention to a water pump.

FIG. 17D shows an application of the invention to a compressor drive of an air conditioner or a refrigerator. The drive unit 100 drives a compressor 70 which compresses refrigerant circulating in a refrigeration cycle 71.

FIG. 17E shows an application of the invention to a machine tool. FIG. 17F shows an application of the invention to a cleaner. FIG. 17G shows an application of the invention to a fan device.

However, the proposed discrete position sensorless estimation techniques can be extended to any industrial applications involving these types of motors where the speed response required is low.

Although the present invention has been described in connection with specified embodiments thereof, many other modifications, corrections and applications are apparent to those skilled in the art. Therefore, the present invention is not limited by the disclosure provided herein but limited only to the scope of the appended claims.

Claims

1. A discrete rotor position estimation method for a synchronized reluctance motor comprising: sensing a d.c.-link voltage Vdc and a phase current Iph; calculating a flux-linkage λph of an active phase from the sensed d.c.-link voltage λdc and the sensed phase current Iph; comparing the calculated flux-linkage λph with a reference flux-linkage λr, the reference flux-linkage λr corresponding to a reference angle θr which lies between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor; and typically obtaining an estimated rotor position θcal equal to θr only once during the active conduction of a phase based on the comparison result when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

2. A discrete rotor position estimation method for a synchronized reluctance motor comprising: sensing a d.c.-link voltage Vdc and a phase current Iph; calculating a flux-linkage λph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; comparing the calculated flux-linkage λph with either two or three reference flux-linkages such as λr1, . . . , corresponding to reference rotor angles θr1, . . . all of them lying between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor; typically obtaining rotor positions θcal1, . . . equal to θr1, . . . based on the comparison results, twice or thrice during the active conduction of a phase when the calculated flux-linkage λph is greater than the reference flux-linkages λr1, . . . .

3. The discrete rotor position estimation method according to claim 1 or 2, further comprising modifying the estimated rotor position θcal with an incremental angle φ corresponding to the reference angle θr to obtain a more accurate estimated rotor position.

4. A discrete rotor position estimation method for a synchronized reluctance motor comprising: sensing a d.c.-link voltage Vdc and a phase current Iph; calculating a flux-linkage λph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; comparing the calculated flux-linkage λph with a reference flux-linkage λr, the reference flux-linkage λr corresponding to a reference angle θr which lies between angles corresponding to aligned rotor position and non-aligned rotor position of the synchronized reluctance motor; and calculating an estimated rotor position θcal from the calculated flux-linkage λph using either one of the inductance model or the flux linkage model of the active phase, only once during the active conduction of a phase when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

5. The discrete rotor position estimation method according to claim 4, wherein the estimated rotor position is calculated at one PWM interrupt before the next phase is turned ON.

6. A discrete rotor position estimation method for a synchronized reluctance motor comprising: sensing a d.c.-link voltage Vdc and a phase current Iph; calculating a flux-linkage λph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; comparing the calculated flux-linkage λph with a reference flux-linkage λr, the reference flux-linkage λr corresponding to a reference angle θr which lies between angles corresponding to aligned rotor position and non-aligned rotor position of the synchronized reluctance motor; and calculating estimated rotor positions either twice or thrice during the active conduction of a phase such as θcal, . . . from the calculated flux-linkage λph using either one of the inductance model or the flux linkage model of the active phase, at every consecutive PWM interrupt when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

7. The discrete rotor position estimation method according to any one of claims 1 to 6, wherein the θr reference flux-linkage λr at the reference rotor position θr is predetermined experimentally and is expressed as a polynomial expression of phase current Iph. The reference rotor position θr is typically defined at any region near the mid-position θm of the aligned and the non-aligned position with a maximum deviation angle αmax of 30° electrical. The reference flux-linkage λr involving the polynomial expression in phase current Iph is calculated within a processor.

8. The discrete rotor position estimation method according to claim 1 or 3, wherein the incremental rotor angle Δθ for every PWM interrupt is obtained only once from the knowledge of θcal during the active conduction of a phase when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

9. The discrete rotor position estimation method according to claim 2 or 5, wherein the incremental rotor angles such as Δθ1, . . . for every PWM interrupt are obtained either twice or thrice from the knowledge of θcal1, . . . during the active conduction of a phase when the calculated flux-linkage λph is greater than the reference flux-linkage λr, and the incremental rotor angles Δθ1, . . . are averaged to obtain the final incremental rotor angle Δθ.

10. A discrete rotor position estimation method for a synchronized reluctance motor comprising: detecting a phase inductance of the synchronized reluctance motor; identifying an minimum region of the phase inductance during turn-on of an active phase; determining a rotor position θapp from the identified minimum region, as an estimated rotor position θcal and using it to calculate incremental rotor angle Δθ for every PWM interrupt.

11. A control method of a synchronized reluctance motor comprising: obtaining the estimated rotor position θcal by the estimation method according to one of claims 1 to 6 and 9; calculating an absolute rotor position θabs from the estimated rotor position θcal by adding a stroke angle of the motor; determining the incremental rotor angle Δθ by processing an error between the absolute rotor position θabs and a finally estimated rotor position θest through either one of a proportional-integral (PI) control and a proportional control; generating the finally estimated rotor position θest in every predetermined period by adding the incremental rotor angle Δθ to the finally estimated rotor position θest in the previous cycle; and controlling turn-on and turn-off angles of each phase based on the finally estimated rotor position θest.

12. A control method of a synchronized reluctance motor comprising: calculating an incremental rotor angle Δθ by counting the number of PWM interrupts between two consecutive instants when the estimated rotor position θcal is obtained by the method according to one of claims 1 to 6 and 9; generating delays to turn-off an active phase and turn-on the next phase, the delays normally defined with respect to the reference rotor position θr; adjusting the delays with the estimated rotor position θcal to turn-off the active phase and turn-on the next phase; and controlling a turn-on angle θn and a turn-off angle θoff of each phase of the motor based on the incremental rotor angle Δθ and the adjusted delays.

13. The control method according to claim 11 or 12, further comprising calculating a speed ω of the motor from the incremental rotor angle Δθ in a relatively slower timer interrupt compared to a PWM interrupt, and varying continuously a turn-on angle θon and a turn-off angle θoff of each phase of the motor based on the speed ω and the torque demand of the motor.

14. The control method according to claim 11 or 12, further comprising defining a timer interrupt faster than the PWM interrupt for achieving the turn-on and the turn-off of each phase at any point in between two PWM interrupts.

15. A control method of a synchronized reluctance motor comprising: monitoring continuously a peak of a phase current and a negative change rate of phase current in each phase; and keeping the turn-off angle fixed, and advancing the turn-on angle so that a pre-determined peak phase current and a negative rate of change of phase current corresponding to the maximum torque are achieved.

16. The control method according to claim 15, wherein instead of monitoring the negative rate of change of phase current a lead angle φ between the peak current and the peak flux in each phase is monitored, to judge the maximum torque at the rated speed condition.

17. An apparatus for estimating discretely a rotor position for a synchronized reluctance motor comprising: sensor operable to sense a d.c.-link voltage Vdc and a phase current Iph; section operable to calculate a flux-linkage λph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; section operable to calculate the reference flux-linkage λr from the polynomial expression in phase current Iph; section operable to compare the calculated flux-linkage λph with the reference flux-linkage λr, the reference flux-linkage λr corresponding to a reference angle θr which lies between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor; and section operable to obtain an estimated rotor position θcal only once when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

18. An apparatus for estimating discretely a rotor position for a synchronized reluctance motor comprising: sensor operable to sense a d.c.-link voltage Vdc and a phase current Iph; section operable to calculate a flux-linkage λph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; section operable to calculate reference flux-linkages λr1, . . . from the polynomial expression in phase current Iph; section operable to compare the calculated flux-linkage λph with reference flux-linkages λr1, . . . the reference flux-linkages λr1, . . . corresponding respectively to reference angles θr1, . . . which lie between angles corresponding to aligned rotor position and non-aligned rotor position in the synchronized reluctance motor; section operable to obtain rotor positions θcal1, . . . based on the comparison result, twice or thrice when the calculated flux-linkage λph is greater than the reference flux-linkages λr1, . . . .

19. The apparatus according to claim 17 or 18, further comprising section operable to modify the estimated rotor position θcal with an incremental angle φ corresponding to the reference angle θr to obtain a more accurate estimated rotor position.

20. An apparatus for estimating discretely a rotor position for a synchronized reluctance motor comprising: sensor operable to sense a d.c.-link voltage Vdc and a phase current Iph; section operable to calculate a flux-linkage Xph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; section operable to calculate the reference flux-linkage λph from the polynomial expression in phase current section operable to compare the calculated flux-linkage λph with a reference flux-linkage λr, the reference flux-linkage θr corresponding to a reference angle θr which lies between angles corresponding to aligned rotor position and non-aligned rotor position of the synchronized reluctance motor; and section operable to calculate an estimated rotor position θcal from the calculated flux-linkage λph using either one of the inductance model or the flux linkage model of the active phase, only once when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

21. The apparatus according to claim 20, wherein the estimated rotor position is calculated at one PWM interrupt before the next phase is turned ON.

22. An apparatus for estimating discretely a rotor position for a synchronized reluctance motor comprising: sensor operable to sense a d.c.-link voltage Vdc and a phase current Iph; section operable to calculate a flux-linkage λph of an active phase from the sensed d.c.-link voltage Vdc and the sensed phase current Iph; section operable to calculate the reference flux-linkage λr from the polynomial expression in phase current Iph; section operable to compare the calculated flux-linkage λph with two or three reference flux-linkages λr1, . . . the reference flux-linkages λr1, . . . respectively corresponding to reference angles θr1, . . . which lie between angles corresponding to aligned rotor position and non-aligned rotor position of the synchronized reluctance motor; and section operable to calculate estimated rotor positions θcal1, . . . either twice or thricefrom the calculated flux-linkage λph using either one of the inductance model or the flux linkage model of the active phase, at every consecutive PWM interrupt when the calculated flux-linkage λph is greater than the reference flux-linkage λr1, . . . .

23. The apparatus according to claim 17 or 20, further comprising a section operable to estimate the incremental rotor angle Δθ for every PWM interrupt only once from the knowledge of θcal during the active conduction of a phase when the calculated flux-linkage λph is greater than the reference flux-linkage λr.

24. The apparatus according to claim 18 or 22, further comprising a section operable to estimate the incremental rotor angles Δθ1, . . . for every PWM interrupt either twice or thrice from the knowledge of θcall, . . . during the active conduction of a phase when the calculated flux-linkage λph is greater than the reference flux-linkage λr1, . . . and a section operable to average the incremental rotor angles Δθ1, . . . to obtain the final incremental rotor angle Δθ.

25. An apparatus for estimating discretely a rotor position for a synchronized reluctance motor comprising: section operable to detect a phase inductance of the synchronized reluctance motor; section operable to identify an minimum region of the phase inductance during turn-on of an active phase; section operable to determine a rotor position θapp from the identified minimum region, as an estimated rotor position θcal; and section operable to obtain the incremental rotor angle Δθ from θapp.

26. An apparatus for controlling a synchronized reluctance motor comprising: section operable to obtain the estimated rotor position θcal by the estimation method according to any one of claims 12 to 17; section operable to calculate an absolute rotor position θabs from the estimated rotor position θcal by adding a stroke angle of the motor; section operable to determine the incremental rotor angle Δθ by processing an error between the absolute rotor position θabs and a finally estimated rotor position θest through either one of a proportional-integral (PI) control and a proportional control; section operable to generate the finally estimated rotor position θest in every predetermined period by adding the incremental rotor angle Δθ to the finally estimated rotor position θest in the previous cycle; and section operable to control turn-on and turn-off angles of each phase based on the finally estimated rotor position θest.

27. An apparatus for controlling a synchronized reluctance motor comprising: section operable to calculate an incremental rotor angle Δθ by counting the number of PWM interrupts between two consecutive instants when the estimated rotor position θcal is obtained by the method according to any one of claims 12 to 17; section operable to generate delays to turn-off an active phase and turn-on the next phase, the delays normally defined with respect to the reference rotor position θr; section operable to adjust the delays with the estimated rotor position θcal to turn-off the active phase and turn-on the next phase; and section operable to control a turn-on angle o and a turn-off angle θoff of each phase of the motor based on the adjusted delays decided by the incremental rotor angle Δθ.

28. The apparatus according to claim 18 or 19, further comprising section operable to calculate a speed ω of the motor is calculated from the incremental rotor angle Δθ in a relatively slower timer interrupt compared to a PWM interrupt, and section operable to vary continuously a turn-on angle θon and a turn-off angle θoff of each phase of the motor based on the speed ω and the torque demand of the motor.

29. An apparatus for controlling a synchronized reluctance motor comprising: section operable to monitor continuously a peak of a phase current and a negative change rate of phase current in each phase; and section operable to keep the turn-off angle fixed, and advance the turn-on angle so that a pre-determined peak phase current and a negative rate of change of phase current corresponding to the maximum torque are achieved.

30. The apparatus according to claim 29, wherein instead of monitoring the negative rate of change of phase current, the section operable to monitor monitors a lead angle φ between the peak current and the peak flux in each phase, to judge the maximum torque at the rated speed condition.

31. A motor drive system comprising a synchronized switched reluctance motor to provide a driving power to a compressor drive and driving the synchronized switched reluctance motor by the control method according to any one of claims 11 to 16.

32. An air conditioner comprising the motor drive system according to claim 31.