Stator flux based MRAS speed and stator resistance

estimator for sensorless PMSM drive

Karthikeyan A1,*, Prabhakaran K K1, Nagamani C2
1
Department of Electrical and Electronics Engineering, National Institute of Technology Karnataka, Surathkal
2
Department of Electrical and Electronics Engineering, National Institute of Technology, Tiruchirappalli
Email ID: *[email protected].

Abstract — In this paper, a simple and robust speed and stator thermal variations in the motor. So, stator resistance
resistance estimator is proposed for sensorless permanent magnet estimation method is proposed to improve robustness of
synchronous motor drive to improve its performance at standstill PMSM drive. There are several rotor speed/ position schemes
and low speed regions. The speed and stator resistance estimator are proposed in the literature [2-8]. The MRAS based speed
is formed using stator flux based model reference adaptive
and position estimation methods are simple, robust and
system, which yields fast transient response, reduces steady state
error for low rotor speed variation and load disturbances. The requires less computational time [9], [10]. In this paper, stator
proposed method improves the performance of the drive at low flux based MRAS speed and resistance (stator) estimator is
speed regions. The simulations are carried out through presented to improve the robustness of speed and position
MATLAB/Simulink environment and results show the estimation at low speed and standstill operation. The proposed
effectiveness of the sensorless PMSM drive. The maximum method has significant advantage compared to previous speed
estimated rotor speed and position error are found to be are ±3 estimation methods [2-8] and the proposed method is suitable
rad/s, ±1°respectively, during low speed operation at transient for low cost and industrial applications.
and steady state conditions. In this paper, stator flux based MRAS speed and stator
Keywords: permanent magnet synchronous motor (PMSM), model
resistance estimator is presented to improve the robustness of
reference adaptive system (MRAS), sensorless speed and position
estimation, field oriented control (FOC). speed and position estimation at low speed and standstill
operation. This paper is organized as follows, the proposed
I. INTRODUCTION stator flux based MRAS speed and stator resistance estimator
In recent years, PMSM drives are the first choice in industrial is elaborated in section 2 and 3. Simulation results are
applications and home appliances such as fans, air demonstrated in section 4.
conditioners, washing machines and dishwashers. The field II. STATOR FLUX BASED MRAS SPEED ESTIMATOR
oriented control of PMSM drive [1] has drawn attention from
both industry and academia due to its high dynamic Fig. 1. shows the structure of stator flux based MRAS speed
performance. Sensorless rotor speed and position estimation estimator. In this work, the rotor speed is estimated by
schemes are broadly grouped as: fundamental excitation differences in stator flux. Real system is considered as
methods, saliency and signal injection methods. The reference model and it is independent with respect to
fundamental excitation methods can be grouped into open estimated quantity which provides ȥds and ȥqs. The
loop and closed loop estimators. The closed loop estimation mathematical flux model is considered as the adjustable model
schemes are divided into two categories: state observer based and it includes estimated quantity, which provides ȥըds and ȥըqs.
methods and MRAS. State observer based methods are further The error signal is obtained from comparing reference and
classified into: stator flux based observer methods [2-3], estimated stator flux and it is processed through the PI
sliding mode observer methods [4-5], Extended Kalman filter controller. The stator flux error signal is used to estimate the
method [6]. Observer based methods, require knowledge of rotor speed. The Structure of sensorless FOC of PMSM drive
machine parameters. At standstill or low speed regions, the with proposed MRAS speed and stator resistance estimator is
speed and position estimation performance are affected by depicted in Fig. 2.
noise, error and delay. Though Extended Kalman filter and vs is
sliding mode observer perform well at standstill and low speed ψs
operation, but they suffer from high computational complexity εψ ωˆ r θˆr
and chattering phenomenon. MRAS based methods discussed
in [6-8] for rotor speed and position estimation have drawn ψˆ s
attention due to their simplicity. MRAS speed and position
estimators based on stator current, stator flux, real power and
reactive power have been proposed for vector control of Fig. 1. Structure of the stator flux based MRAS speed estimator.
PMSM drive. The main drawback of these techniques is
machine parameter sensitivity which constraints its
performance at standstill and low speed regions. The stator
resistance value varies from nominal value due to ageing and

978-1-5386-4318-1/17/$31.00 ©2017 IEEE

Authorized licensed use limited to: Carleton University. Downloaded on January 12,2021 at 18:19:20 UTC from IEEE Xplore. Restrictions apply.
 d ε ds − Rs (11)
ω*
r iqs* vqs* vα* = ε + ωrψ qs − ωˆ rψ qs
dt Ld ds

3-ࢥ Inverter
iqs ωˆ r (L di d +ψ m ) vβ*
ωˆ r
d ε qs − Rs (12)
= ε qs − ωrψ ds + ωˆ rψ ds
*
ids* = 0 vds
θˆr iα s
ωˆ r L q i q
ids ids dt Lq
iqs iβ s
Adding and subtracting of the terms ωˆ rψ qs in (11) and
θˆr
ωˆ rψ ds in (12). The state error can be written as follow
Rˆ s
d ε ds − Rˆ s (13)
= ε ds + ωˆ r ε qs +ψ qs (ωr − ωˆ r )
Fig. 2. Structure of sensorless FOC of PMSM drive with proposed MRAS dt Ld
speed and stator resistance estimator.
d ε qs − Rˆ s (14)
The stator d-q voltages in terms of machine variables are = ε qs − ωˆ r ε ds − ψ ds (ωr − ωˆ r )
dt Lq
expressed as [1].
Eq. (13) and (14) can be written in matrix form as below
vqs = Rˆs iqs + ωψ
r ds + pψ qs (1)
ª − Rˆ s º
« ωˆ r » (15)
vds = Rˆsids − ωψ
r qs + pψ ds (2) d ªε ds º « Ld » ªε ds º ª ψ qs º
« »= »« »+« » (ωr − ωˆ r )
Stator flux linkages can be expressed as dt ¬ε qs ¼ « − Rˆ s » ¬ε qs ¼ ¬ −ψ ds ¼
« −ωˆ
ψds = Ld ids +ψm (3) « r Lq »¼
¬
ψqs = Lqiqs (4) State error model is simplified as follows
• (16)
by substituting (3) and (4) in (1) and (2), the stator flux can be εψ = A2εψ + W1
written in matrix form as given
T ª ψ qs º
ª Rˆ s º where εψ = ª¬ε ds ε qs º¼ , W1 = « » (ωr − ωˆ r )
«− ωr » ª ψ m Rˆ s º ¬−ψ ds ¼
d ªψ ds º « Ld » ªψ ds º ª 1 0 º «Vds + »
«ψ » = « » « » +« » « Ld » (5) To ensure the system is stable, which requires the state error
dt ¬ qs ¼ ˆ ψ 0 1¼
« −ωr − Rs » ¬ qs ¼ ¬ « Vqs » (İ) to be near to zero. The stability of the system is analyzed
«¬ Lq »¼ ¬ ¼ using Popov super hyperstability theorem. The state error
depends on two criterions, where the first is linear time variant
The above (5) is considered as reference model, which is
further simplified and given as below matrix, H (s) = (sI− A 2 ) −1 should be a strictly positive matrix
ψ = A1ψ + Bu (6) and the second is nonlinear
t0
γ 0 can
T
ψ = Cψ (7) term, η (0, t 0 ) = ³ ª¬εψ º¼ [W1 ] dt ≥ − γ 02 , ∀t0 ≥ 0 where
0
Where be finite positive number.
ª Rˆ s º The above expression can be modified using W1
«− ωr » ª ψ m Rˆ s º
« Ld » T ª1 0 º «Vds + »
t0 (17)
A1 = « ,ψ = ªψ ψ º , B = C = , u = Ld »
³ ª¬ε dsψ qs − ε qsψ ds º¼ (ωr − ωˆ r ) dt ≥ −γ 0
2
¬ ds qs ¼ « » «
« −ωr Rˆ s »» ¬0 1¼ « »
− ¬ Vqs ¼ 0
«¬ Lq »¼
Finally, the estimated rotor speed is obtained using stator flux
In (6) the parameters stator flux and rotor speed are replaced error which is given below
with estimated quantities which forms adjustable model and t
given as ωˆ = ³ k1 (ε dsψ qs − ε qsψ ds ) dt + k2 ( ε dsψ qs − ε qsψ ds ) (18)
ψˆ = A2ψˆ + Bu (8) 0

ψˆ = Cψˆ (9) where k 1 , k 2 ≥ 0

The estimated rotor position is achieved by integrating the
ª − Rˆ s Ld ωˆ r º T estimated speed.
where A2 = « » ,ψˆ = ª¬ψˆ ds ψˆ qs º¼
«¬ −ωˆ r − Rˆ s Lq »¼ θˆ = ³ ωˆ dt (19)
The stator flux error is obtained between reference model and Fig. 3 shows the pole-zero plot for the transfer matrix H(s) for
adjustable model outputs (i.e. reference stator flux and the speed range ( ωˆ r = 0 → ±180 rad/s). It is observed that, all
estimated stator flux).
ε ds = ψ ds −ψˆ ds , ε qs = ψ qs −ψˆ qs (10)
subtracting (8) from (6) and written as

Authorized licensed use limited to: Carleton University. Downloaded on January 12,2021 at 18:19:20 UTC from IEEE Xplore. Restrictions apply.
 εψ
= G (s) =
(
−ωrψ ds 0 + (s + Rˆ s Lq )ψ qs 0 *ψ qs
+
) (28)

Δωr (s + Rˆ s Lq ) *(s + Rˆ s Ld ) + ωr2

( (s+ Rˆ s )
Ld )ψ ds 0 + ωˆ rψ qs 0 *ψ ds
(s + Rˆ s Lq ) *(s + Rˆ s Ld ) + ωr2

finally, the closed loop transfer function of proposed MRAS

speed estimator is given as
K (29)
G (s)(K pMRAS + iMRAS )
ωr s
=
∧
K
ωr 1 + G (s)(K pMRAS + iMRAS )
s
∧
ωr Δωr εψ ωr
Fig. 3. Pole placement for the transfer matrix H(s), ωˆ = 0 → ±180 rad/s.
r

the poles are located in the left half of the s plane. So, it
confirms that (H(s)) is completely positive matrix.
Fig. 4. Closed representation of MRAS speed and position estimator
2.1 Stability analysis of the proposed MRAS speed estimator
To validate the dynamic response of the stator flux based
MRAS speed estimator stability analysis is performed by
linearizing the stator flux equation (6) and (7) around a stable
point ȥ0., which is as follows:
• (20)
Δψ = A1Δψ + ΔA1ψ 0
Δψ = C Δψ (21)
Δψ = C (sI− A1 )−1 ΔA1ψ 0 (22)

where ψ 0 = [ψ sd 0 ψ sq 0 ]T
now consider small variation in speed, then Δ A1 can be
expressed as
ª 0 1º (23) Fig. 5. Loci for the closed loop transfer function of proposed MRAS speed
ΔA1 = « » Δω r estimator, Ȧࡂr = 0ĺ±180rad/s
¬− 1 0 ¼
Fig. 4 shows the closed loop representation of proposed
by substituting (23) into (22), the Δψ can be expressed as
MRAS rotor speed estimator. The stability of the proposed
ª Δψ sd º ª1 0 º ª 0 1º ªψ sd 0 º (24) system is analyzed through pole placement for the closed loop
« Δψ » = « » (sI − A1 ) −1 « » Δω r « » transfer function of proposed MRAS speed estimator. Fig 5
¬ sq ¼ ¬0 1 ¼ ¬ −1 0 ¼ ¬ψ sq 0 ¼ shows loci of the closed loop poles and zeros for the rated
§ Δψ sd · § Δψ sq · speed range (Ȧࡂr = 0ĺ±180 rad/s). It is observed from fig. 5
From (24), the transfer function of ¨ ¸ and ¨ ¸ can
© Δω r ¹ © Δω r ¹ that all the poles are located in the left half of the s plane and it
shows the proposed MRAS speed estimator is stable for
be expressed as
motoring and speed reversal operation for entire operating
Δψ ds −ωrψ ds 0 + (s + Rˆ s Lq )ψ qs 0 region.
= (25)
Δω r (s + Rˆ L ) * (s + Rˆ L ) + ω 2
s q s d r
III. STATOR FLUX BASED MRAS STATOR
Δψ qs −(s + Rˆ s Ld )ψ ds 0 − ωˆ rψ qs 0 RESISTANCE ESTIMATOR
=r (26)
Δω r (s + Rˆ s Lq ) * (s + Rˆ s Ld ) + ωr2 Fig. 6. shows the structure of stator flux based MRAS stator
resistance estimator. In this work, the stator resistance is
By linearizing the stator flux error (17) and dividing by
estimated by differences in stator flux. Real system is
Δωr yields the transfer function between Δεψ and Δωr considered as reference model and it is independent with
εψ Δψ ds *ψ qs Δψ qs *ψ ds (27) respect to stator resistance which provides ȥds1 and ȥqs1. The
= − mathematical flux model is considered as the adjustable model
Δω r Δω r Δω r
and it includes stator resistance, which provides ȥըds1 and ȥըqs1.
Substituting (25)-(26) in (27), the stator flux error transfer
The error signal is obtained from comparing reference and
function can be expressed as

Authorized licensed use limited to: Carleton University. Downloaded on January 12,2021 at 18:19:20 UTC from IEEE Xplore. Restrictions apply.
 vs is Eq. (38) and (39) can be written in matrix form as given
ψ s1
ª − Rˆ s º ª −ψ ds1 + ψ m º
εψ Rˆ s « ωˆ r » « » (40)
d ε ª ds1 º « d
L ε » ª ds1 º «
L d » ( R − Rˆ )
« »=« » « »+
¬ε qs1 ¼ « −ωˆ − Rˆ s » ¬ε qs1 ¼ «
ψˆ s1
dt −ψ qs1 » s s

« »
«¬ r Lq »¼ ¬« Lq ¼»
Fig. 6. Structure of the stator flux based MRAS stator resistance estimator. The above (40) is stator flux error, which is further simplified
and given as below
estimated stator flux and it is processed through the PI • (41)
controller. The stator flux error signal is used to estimate the εψ 1 = A2εψ 1 + W2
stator resistance. The stator flux can be written in matrix form ª −ψ ds1 +ψ m º
given as « »
Ld
ª Rs º T
W = « » ( R − Rˆ )
« − ωˆ r » ª ψ R º where εψ 1 = ª¬ε ds1 ε qs1 º¼ , 2 « −ψ » s s
d ªψ ds1 º « Ld ψ V + m s»
» ª ds1 º + ª1 0 º « ds Ld «
qs1
»
« »= Rs » «¬ψ qs1 »¼ «¬ 0 1 »¼ « » (30)
dt ¬ψ qs1 ¼ « ¬« Lq ¼»
« −ωˆ r − » «¬ Vqs »¼
¬ Lq ¼
To ensure the system stability, it requires that state error ( εψ 1 )
The above (30) is considered as reference model, which is
further simplified and given as below should be near to zero. The stability of the system is analyzed
using Popov super hyperstability theorem, state error depends
ψ1 = A3ψ 1 + Bu1 (31)
on two criterions. The first is linear time variant matrix,
ψ 1 = Cψ 1 (32)
H1 (s) = (sI− A 2 ) −1 should be a strictly positive matrix and the
ª Rs º t0 T
«− ωˆ r » ª ψ R º
Vds + m s » second is nonlinear term, η (0, t 0 ) = ³ ª¬εψ 1 º¼ [W2 ] dt ≥ − γ 02 ,
where A = « Ld » ,ψ = ªψ ψ qs1 º¼ , u1 = ««
T
Ld
0
3
« Rs » 1 ¬ ds1 »
∀t0 ≥ 0 where γ 0 can be finite positive number.
« −ωˆ r − » «¬ Vqs »¼
¬ Lq ¼
The above expression can be modified by using W2
In (31) the parameters stator flux and resistance are replaced t0 ª
§ ψ qs1 · º (42)
§ −ψ ds1 + ψ m · ˆ ) dt ≥ −γ 2
with estimated quantities which forms adjustable model and
given as
³0 «© Ld ¹
« ¨ ¸ ε ds 1 − ¨
¨ Lq ¸
¸ ε qs 1 »
»¼
( Rs − Rs 0
¬ © ¹
ˆ
ψ1 = A2ψ1 + Bu
ˆ (33) Fig. 7 shows the pole-zero plot for the transfer matrix H1(s)
for the stator resistance variation from -50% to +50% with
ψˆ 1 = Cψˆ 1 (34)
respect to nominal value for the speed range (Ȧࡂr = 0ĺ±180
T rad/s). It is observed that all the poles are located in the left-
where ψˆ1 = ª¬ψˆ ds1 ψˆ qs1 º¼
half of the s plane. So, it confirms that (H1(s)) is completely
The stator flux error is obtained between reference model and positive matrix.
adjustable model outputs (i.e. reference stator flux and
estimated stator flux).
ε ds1 = ψ ds1 −ψˆ ds1 , ε qs1 = ψ qs1 −ψˆ qs1 (35) +50% Rs
-50% Rs
Subtracting (33) from (31) and written as
d ε ds1 − Rsψ ds1 Rψ Rˆ ψˆ Rˆ ψ (36)
= + ωˆ r ε qs1 + s m + s ds1 − s m
dt Ld Ld Ld Ld
d ε qs1 Rψ Rˆ ψˆ (37)
= −ωˆ r ε ds − s qs + s qs
dt Lq Lq
Rˆ sψ ds1
Adding and subtracting of the terms in (36) and
Ld
Rˆ sψ qs1
in (37). The state error can be written as follow
Lq
Fig. 7. Pole placement for the closed loop transfer function of proposed
d ε ds1 − Rˆ s ª −ψ + ψ º (38)
MRAS stator resistance estimation, ωˆ = 0 → ±180 rad/s, rs= -50% to
= ε ds1 + ωˆ r ε qs1 + « ds1 m » ( Rs − Rˆ s ) r
dt Ld ¬ Ld ¼ +50% of their nominal value.
d ε qs1 − Rˆ s ψ (39)
= ε qs1 − ωˆ r ε ds1 − qs1 (R s − Rˆ s )
dt Lq Lq

Authorized licensed use limited to: Carleton University. Downloaded on January 12,2021 at 18:19:20 UTC from IEEE Xplore. Restrictions apply.
The estimated stator resistance can be obtained as follows Substitute (49)-(50) into (51), the stator flux error transfer
t ª§ −ψ +ψ · §ψ qs1 · º function can be expressed as
Rˆs = ³ k1 «¨ ds1 m ¸ε ds1 − ¨ ε » dt + (43) εψ εψ 1 (52)
«¬© Ld ¨ L ¸¸ qs1 » = G1 (s) = =
0 ¹ © q ¹ ¼ ΔRs ΔRs
ª§ −ψ +ψ · §ψ qs1 · º § −ψ ds1 + ψ m ( ) (
· § − s + Rs Lq (ψ ds 01 Ld ) − ωˆ r *ψ qs 01 Lq ) ¸·
k2 «¨ ds1 m ¸εds1 − ¨ ε » ¸¨
¨ L ¸¸ qs1 »
¨
«¬© Ld ¹ © Ld ¹ ¨© (s + Rs Lq ) * (s + Rs Ld ) + ωˆ r2 ¸
¹
© q ¹ ¼
where k 1 , k 2 ≥ 0
−¨ ¸¨
(
§ ψ qs1 · § (ωˆ r Ld ) (ψ ds 01 ) − ( s + Rs Ld ) * ψ qs 01 Lq ) ·¸
In the above equation ȥds1 and ȥqs1 can be calculated through ¨ Lq ¸ ¨ (s + Rs Lq ) * (s + Rs Ld ) + ωˆ r2 ¸
© ¹© ¹
the reference model, ȥըds1 and ȥըqs1 can be obtained through the
The closed loop representation of proposed MRAS stator
adjustable model.
resistance estimator
3.1 Stability analysis of the proposed MRAS stator resistance K (53)
G1 (s)(K p,res + i,res )
estimator Rs s
=
To validate the dynamic response of the stator flux based
∧
K
Rs 1 + G1 (s)(K p,res + i,res )
MRAS resistance estimator stability analysis is performed by s
linearizing the stator flux equation (31) and (32) around a ∧
stable point ȥ01., which is as follows: Rs ΔRˆ s εψ 1 Rs
Δψ1 = A3 Δψ1 + ΔA3ψ 01 (44)

Δψ 1 = CΔψ 1 (45)

Δψ 1 = C (sI − A 3 ) −1 ΔA3ψ 01 (46) Fig. 8. Closed representation of MRAS stator resistance estimator

Where ψ 01 = [ψ sd 01 ψ sq 01 ]T
+50% Rs -50% Rs
now consider small variation in stator resistance, then ¨A3 can
be expressed as
ª −1 º (47)
«L 0 »
ΔA3 = « » ΔR
d

« −1 » s
« 0 »
¬ Lq ¼
By substituting (47) into (46), the Δy can be expressed as
ª 1 º (48)
« − 0 »
ª Δψ sd 1 º ª1 0 º Ld ψ
» ΔR «ª sd 01 »º
« Δψ » = « » (sI− A 3 )−1 «
¬ sq1 ¼ ¬0 1¼ « 1 » s
¬ψ sq 01 ¼ Fig. 9. Pole placement for the closed loop transfer function of proposed
« 0 − »
L MRAS stator resistance estimator for the stator resistance variation from -
¬ q ¼
50% to +50% with respect to nominal value for the speed range (Ȧࡂr =
§ Δψ sd 1 · § Δψ sq1 · 0ĺ±180 rad/s).
From (48), the transfer function of ¨ ¸ and ¨ ¸ can Fig. 8 shows the closed loop representation of proposed stator
© ΔRs ¹ © ΔRs ¹ flux based MRAS stator resistance estimator. The stability of
be expressed as the proposed system is analyzed through pole placement for
( )
Δψ ds1 − s + Rs Lq (ψ ds 01 Ld ) − ωˆ r *ψ qs 01 Lq
=
( (49) ) the closed loop transfer function of proposed MRAS based
stator resistance estimator. Fig 9 shows loci of the closed loop
ΔRs (s + Rs Lq ) * (s + Rs Ld ) + ωˆ r2 poles and zeros where stator resistance is varied from -50 % to
Δψ qs1
=
(ωˆ r (
Ld ) (ψ ds 01 ) − ( s + Rs Ld ) * ψ qs 01 Lq ) (50) +50% with respect to nominal value for the speed range (Ȧࡂr =
0ĺ±180 rad/s rad/s). It is observed that all the poles are
ΔRs (s + Rs Lq ) * (s + Rs Ld ) + ωˆ r2 located in the left-hand side of the s plane and it shows the
By linearizing the stator flux error (42) and dividing by proposed stator flux based MRAS stator resistance estimator is
ΔRs yields the transfer function between Δεψ and ΔRs stable for motoring and speed reversal operation for entire
(51) operating region.
εψ 1 ª§ −ψ + ψ · Δψ § ψ qs1 · Δψ qs1 º
= Ǭ ds1 m
¸
ds1
−¨ ¸¸ » IV. SIMULATION RESULTS AND DISCUSSIONS
ΔRs «¬© ¨
Ld ¹ ΔRs © Lq ¹ ΔRs »¼
The proposed stator flux based MRAS speed and stator
resistance estimator for PMSM drive is simulated in the

Authorized licensed use limited to: Carleton University. Downloaded on January 12,2021 at 18:19:20 UTC from IEEE Xplore. Restrictions apply.
MATLAB / SIMULINK environment. The proposed system is V. CONCLUSIONS
investigated under the following condition: low speed A simple and robust rotor speed and stator resistance
variation with stator resistance variation. The parameters of estimation for sensorless PMSM drive is proposed and
the PMSM used to validate the proposed system are given in investigated in this paper. Stator flux based MRAS speed and
Table I. stator resistance estimator is designed and implemented. The
A. Low speed variation stability of the proposed MRAS based speed and stator
resistance estimator method is confirmed through small signal
Fig.10 shows the speed is gradually varied between 0 to 20 analysis and proposed method performs well at standstill and
rad/s, -20, 10 and 30rad/s at rated load. The estimated rotor low speed region. Simulation results demonstrate the
speed follows the measured speed during speed reversal and effectiveness of the sensorless speed and stator resistance
zero speed operation satisfactorily, as shown in Fig.10a. The estimation of FOC with PMSM drive.
measured and estimated rotor position are shown in Fig. 10b. APPENDIX
From this analysis, it is clear that estimated rotor position is TABLE I. PMSM RATING AND PARAMETERS
virtually similar to the measured rotor position. The estimated Stator: 3ph, 4 pole, 220v, 1.2hp, 1700 rpm
Parameter Measured value in SI units
rotor speed and position error are ±3 rad/s, ±1° respectively Rs 4.2 Ÿ
as shown in Fig.10c and d. The estimated stator resistance is Ld 30 mH
Lq 65 mH
shown in Fig. 10e. It is observed that, the proposed speed
Ȝm 0.272 Wb
and stator resistance estimator perfoms well at low speed J 0.00018 kg m2
regions with 50% of increase in stator resistance.
References
[1] Pwgasan Pillay, R. Knshnan, “Modeling of Permanent Magnet Motor
Drives”, IEEE Trans. Ind. Electron., vol. 35, no. 4, pp. 537-541, 1988
[2] Sakorn Po-Ngam, Somboon Sangwongwanich, “Stability and Dynamic
Performance Improvement of Adaptive Full-Order Observers for
Sensorless PMSM Drive”, IEEE Trans. Power Electron., vol. 27, no. 2,
pp. 588-600, 2012.
[3] Toni Tuovinen, Marko Hinkkanen, Lennart Harnefors, Jorma Luomi,
“Comparison of a Reduced-Order Observer and a Full-Order Observer
for Sensorless Synchronous Motor Drives”, IEEE Trans. Ind. Appl., vol.
a 48, no. 6, pp. 1959-1967, 2012.
[4] Song Chi, Zheng Zhang, Longya Xu, “Sliding-Mode Sensorless Control
of Direct-Drive PM Synchronous Motors for Washing Machine
Applications”, IEEE Trans. Ind. Appl. Vol. 45, no. 2, pp. 582-590, 2009.
[5] Hongryel Kim, Jubum Son, Jangmyung Lee, “A High-Speed Sliding-
Mode Observer for the Sensorless Speed Control of A PMSM”, IEEE
Trans. Ind. Electron., vol. 58, no. 9, pp. 4069 – 4077, 2011.
[6] Yuchao Shi, Kai Sun, Lipei Huang, Yongdong Li, “Online Identification
of Permanent Magnet Flux Based on Extended Kalman Filter for
IPMSM Drive with Position Sensorless Control”, IEEE Trans. Ind.
Electron. Vol.59, no. 11, pp 4169-4178, 2012.
b [7] Y.S. Kim, Y.K. Choi, J.H. Lee, “Speed-sensorless vector control for
permanent-magnet synchronous motors based on instantaneous reactive
power in the wide-speed region”, IEE Proc.-Electr. Power Appl. Vol.
152, no. 5, pp 1343-1349, 2005.
[8] ZhangWenjuan, HuangShoudao, GaoJian, “Direct Torque Control for
PMSM Based on Model Reference Adaptive System, in: proceedings of
International Conference on Electrical Machines and Systems, pp. 1113-
1117, 2013.
[9] S.J. Underwood, I. Husain, “Online parameter estimation and adaptive
control of permanent-magnet synchronous machines”, IEEE Trans. Ind.
Electron. Vol.57, no. 07, pp 2435-2443, 2010
c d [10] B. Nahid-Mobarakeh, F. Meibody, F.M. Sargos, “Mechanical sensorless
e
control of PMSM with online estimation of stator resistance”, IEEE
Fig. 10. Simulation results of ramp type rotor speed response Trans. Ind. Appl., vol. 40, no. 2, pp. 457-471, 2004.
a. measured and estimated rotor speed
b. measured and estimated rotor position
c. speed estimation error
d. position estimation error
e. stator resistance

Authorized licensed use limited to: Carleton University. Downloaded on January 12,2021 at 18:19:20 UTC from IEEE Xplore. Restrictions apply.