This article has been accepted for publication in IEEE Transactions on Power Electronics.

This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

1
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

Active Flux Based Advanced Encoderless ac Drives:

a Tutorial Review
I. Boldea, Life Fellow, IEEE, A. A. Popa, Member, IEEE and F. Blaabjerg, Fellow, IEEE

Abstract—"Active flux” is an alternative unitary concept for

the control of practically all a.c. drives. As aligned to rotor d axis I. INTRODUCTION
(for synchronous and flux-modulation machines) and to rotor flux

C
ONVENTIONALLY signal injection or (and)
axis for induction machines (IMs), “active flux” is a
physical/technical concept that turns all travelling field physically fundamental model based encoderless control systems
salient electric a.c. machines into a functionally nonsalient model. for a.c. drives-field oriented control (FOC) [1-2], direct
As its position is independent of load torque (for known machine torque and flux control (DTFC) [3], feedback linearization
parameters) it leads to notable simplifications in encoderless control (FL) [4] and model (or model-free) predictive control
advanced control of a.c. drives. This tutorial review summarizes (MPC) [5] – process current vector or inductance or emf in
the fundamentals of the “active flux” concept as applied to observers to detect simultaneously rotor (or rotor flux for IM)
practically all a.c. machines and the progress from 2008 to today position and rotor speed, from zero to maximum speed, but
with sample literature results. It also introduces in draft the theory
strong filtering introduces delays. These delays have been
of applying “active flux” to MTPA, MaxCosFi, MTPF
alleviated in many ways, but the control became inevitably
integrated into smooth control of a.c. drives, for constant
more complicated, especially by introducing operation modes
and variable magnetic saturation, with promising recent
such as maximum torque per Ampere (MTPA),
results for a reluctance synchronous motor drive.
MaximumCosFi and maximum torque per flux (MTPF), to
Weighting “active flux” merits and demerits and offering
increase performance and torque–speed range within limited
new workfronts in its application the paper is hoped to
d.c. input voltage of the inverter that feeds the a.c.
assist in producing even better a.c. motor/generator drives
motor/generators.
in the future.
The alternative would be to “observe” the “flux” and to
process it directly to yield/estimate rotor position and speed.
Index Terms—"Active flux”, encoderless control systems,
maximum torque per Ampere, MaximumCosFi, maximum torque But this approach was considered problematic until the offset
per flux, a.c. drives. related to integrals etc. was alleviated by the introduction of
second order generalized integrals (SOGI or reduced order
Nomenclature generalized differentiators (ROGD)) etc. [6].
E – extended emf As the stator flux with respect to rotor (or rotor flux) axis
Ld, Lq – dq inductances of synchronous motors
vector position varies with load (even with constant machine
ΨPM – PM flux linkage
TePMSM, TeEESM, TeRMS, TeIM – electromagnetic torque of Permanent parameters), and thus its speed s is equal to rotor speed (for
Magnet Synchronous Motor (PMSM), Electrically Excited SMs) only during steady state, other candidates were looked
Synchronous Motor (EESM), Reluctance Synchronous Motor (RSM), for, in order to simplify the combined rotor position and speed
Induction Motor (IM). observer in advanced encoderless a.c. drives.
𝑑𝑃𝑀𝑆𝑀
𝑎
, 𝑑𝐸𝐸𝑆𝑀
𝑎
, 𝑑𝑅𝑆𝑀
𝑎
, 𝑑𝐼𝑀
𝑎
– active flux linkage vectors One of the first attempts was made with “extended emf” [1]
of respective motors defined above. of interior permanent magnet synchronous machine (IPMSM):
Ls, Lsc – no load and shortcircuit inductance of IM
rIM – electric angular rotor speed of IM 𝑑𝑖
𝐸̅ = ((𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 + 𝑃𝑀 ) 1 + 𝑗1 𝐿𝑞 𝑞 (1)
𝑑𝑡
𝑆1 – slip angular frequency of IM
𝑅𝑠 , 𝑉̅𝑠 , I𝑠̅ – stator phase resistance, stator voltage and current which managed to somewhat reduce the complexity of rotor and
space vectors speed observation. Next step was to use the first term in (1) but
s – Laplace operator only as a virtual flux of IPMSM [7]:

̅ 𝑠 – stator flux linkage space vector
PLL – phase locked-loop 𝑣𝑖𝑟𝑡𝑢𝑎𝑙 = (𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 + 𝑃𝑀 (2)
𝜃𝑒𝑟 – rotor d axis position with respect to stator phase “a” - where Ld, Lq are the dq axes inductances
𝑖𝑑 , 𝑖𝑞 – dq stator currents The “virtual flux” concept in [7] for IPMSM was generalized
ALA rotor – axially laminated anisotropic rotor (Ld>>Lq) ̅ 𝑑𝑎 (it is called
for all a.c. machine drives as “active flux” 
𝐹0 – electric excitation flux linkage amplitude
“active flux” as it contributes to the total torque of all a.c.
(traveling field) motor/generator drives), starting from torque
expression scrutiny [8]:

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

2
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

3 for IPMSM This vector equation (8) reflects the presumption that, in
𝑇𝑒𝑃𝑀𝑆𝑀 = 𝑝1 ((𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 + 𝑃𝑀 ) 𝑖𝑞
2 in rotor “active flux” terms, the a.c. machines manifest themselves as
3 coordinates nonsalient machines in the dq-model, with a single inductance,
= 𝑝1 𝑑𝑃𝑀𝑆𝑀 𝑖𝑞 ;
𝑎
2 𝐿𝑞 , while “active flux” position is rotor position (axis d here)
3 for EESM in for synchronous and flux-modulation machines and it is rotor
𝑇𝑒𝐸𝐸𝑆𝑀 = 𝑝1 ((𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 + 𝐿𝑑𝑚 𝑖𝐹 ) 𝑖𝑞
2 rotor flux position in IMs, indifferent of torque (load) value, in any
3 coordinates
= 𝑝1 𝑑𝐸𝐸𝑆𝑀
𝑎
𝑖𝑞 ; coordinate system and at any speed; for injection signal or for
2 (3) fundamental model.
3 for RSM in
𝑇𝑒𝑅𝑆𝑀 = 𝑝1 (𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 × 𝑖𝑞
2 rotor Note. On Fig. 1c related to PM assisted reluctance
3 coordinates synchronous machines (RSM) – the PM axis is still kept in axis
= 𝑝1 𝑑𝑅𝑆𝑀 𝑖𝑞 ;
𝑎
2 d (Ld<Lq), but we could simply apply the “active flux” concept
3 3 for IM in swapping the d-q axis as for the RSM (no PMs) in Fig. 3d.
𝑇𝑒𝐼𝑀 = 𝑝1 (𝐿𝑠 − 𝐿𝑠𝑐 )𝑖𝑑 𝑖𝑞 = 𝑝1 𝑑𝐼𝑀 𝑎
𝑖𝑞 ;
2 2 rotor flux
coordinates Based on equation (8) we may simply state that the “active
- where 𝐿𝑠 , 𝐿𝑠𝑐 – no load and shortcircuit induction machine flux model” with axis d along rotor axis (in synchronous
(IM) inductances; 𝑖𝐹 – field current machines and in rotor flux axis in induction machines)
As noticed in (3) the “active flux” is aligned to axis d (of dq represents phenomenologically the respective machines.
model) if the machine parameters are known, indifferent of The “active flux” is thus a technical not only a mathematical
torque (load) – and thus the speed of the “active flux” is rotor concept.
speed in synchronous (and flux-modulation) machine drives
and it is the rotor-flux speed for the IM, r . Note. It goes without saying that inverter nonlinearities,
magnetic saturation and parameters detuning causes a departure
Thus, for IMs the rotor speed rIM is:
of “active flux” axis (position) estimation from reality, as it did
rIM = 𝑑𝐼𝑀
𝑎 − 𝑆1 , where 𝑆1 − slip frequency (4) in the stator (rotor) flux observers and, consequently, should be
corrected.
For synchronous and flux – modulation – machine drives:
r = 𝑑𝑎 (5)

Note. Flux-modulation a.c. machines produce torque through

the fundamental and a few harmonics of stator mmf due to their
special topology [9] essentially could be considered as
synchronous machine.
And it is, in general, obtained through a phase-locked loop
(PLL) on the estimated rotor position θ̂𝑑𝑎 , which will yield a
better rotor position via θ̂ 𝑎 and its speed 
𝑑
̂  𝑎 , though still 𝑑
speed estimation “inherits” the position estimation error
inevitably.
Note. For surface permanent magnet synchronous machines
(SPMSM) (𝐿𝑑 = 𝐿𝑞 = 𝐿𝑆 ) the “active flux” coincides with
rotor (PM) flux 𝑃𝑀 . jq

It should be noted that apparently the “active flux” turns

machines with physical “magnetic saliency” into functionally d
nonsalient machine(s) model, which should be characterized by
a single inductance 𝐿𝑞 . If so, then, we consider again the dq
model of a.c. machines:
𝑑
̅𝑠
I𝑠̅ 𝑅𝑠 − 𝑉̅𝑠 = − − 𝑗𝑟 (𝑜𝑟 𝑟𝐼𝑀 )
̅ 𝑠 with: (6)
𝑑𝑡

𝑎

̅𝑑 = 
̅ 𝑆 − I𝑠̅ 𝐿𝑞 (7)
we obtain:
𝑑 𝑑𝑎 𝑎 𝑑𝐼𝑠̅
I𝑠̅ 𝑅𝑠 − 𝑉̅𝑠 = − − 𝑗𝑟 (𝑜𝑟 𝑟𝐼𝑀 )𝑑 − 𝐿𝑞 − 𝑗𝑟 𝐿𝑞 I𝑠̅
𝑑𝑡 𝑑𝑡
= −(𝑗𝑟 + 𝑠)𝐿𝑞 I𝑠̅ − (𝑗𝑟 + 𝑠)̅ 𝑑𝑎 (8)

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

3
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

induction and synchronous motor/generators control,

replacing torque for thrust and rotary speed for linear speed
and “adding” the longitudinal end effects by disturbance
observers.
B. “Active flux” challenges/demerits
 As already mentioned, “magnetic saturation” also causes
an error in “active flux” position observation if we use
constant machine parameters (as it did for other flux vector
choices: stator flux, rotor flux, airgap flux). So, corrections
for parameter detuning are required for precise fast torque
response of electric drives;
 The machine parameters may be online estimated and
corrected or treated together with inverter nonlinearities,
inertia, load torque variation as disturbances, to be online
estimated and added to the “active flux” observer [10] to
reduce position and speed coupling [57] estimation errors.
Fig. 1. “Active flux” illustration in the dq models of: (a) SPMSM, (b) and (c)  However, it should be noted that so far only the “active
IPMSMs, (d) Relsyn, (e) dc-excited SM, and (f) IM, after [8]. flux” position and speed were observed while its
Let us notice that Lq in wound rotor synchronous machines amplitude, especially with 𝐿𝑞 ≈ 𝑐𝑜𝑛𝑠𝑡 (as in WRSM,
(WRSM) and ALA-rotor reluctance synchronous motor (RSM) SPMSM, ALA-rotor RSMs, IMs), is still available from
(or Lsc for IMs) in (3) is weakly dependent on saturation but not the “active flux” observer for direct online correction of
so in IPMSM (with Lq >Ld , Fig. 3c, 3d and dependent on one machine parameter (Ld or Ldm).
magnetic saturation).  Reference [41] developed on “active flux” produced
position and speed observation to preserve stability at
An illustration of the “active flux” concept in permanent overload for a RSM drive.
magnet synchronous machines (PMSMs), WRSMs, RSMs, and  MaxCosFi was added as a distinct operation mode as its
IMs dq models in dq coordinates as discussed above, is shown stator flux for given torque is between MTPA and MTPF
in Fig.1. (as shown later in this paper) and all three operation modes
A. “Active flux” merits are independent of speed for given machine parameters and
 Turning magnetically (or virtually as in IM in rotor flux losses are neglected. The losses lead to a decrease of
coordinates) salient rotor synchronous, flux modulation efficiency but to an increase in actual power factor.
and induction machines into nonsalient dq models, “active MaxCosFi operation mode is mainly useful to base speed
flux” concept simplifies rotor position and speed observers for more torque (power) for wide CPSR in RSM and IMs
by the fact that “active flux vector” has rotor d axis (or rotor which have high physical saliency (Ld>>Lq (RSM),
flux in IMs) orientation indifferent of the load (torque) Ls>>Lsc (IM)).
value, in contrast to stator flux. It applies to all travelling So, this paper continues as:
field electric motor/generator drives.
- State of the art on “active flux” in a.c. drives control in
 “Active flux” concept allows rotor and position
section 2 like: a) IPMSMs, b) RSMs, c) IMs, d) d.c. excited
observation also for signal injection state observer even at
machines, e) flux – modulation (F-M) machines.
zero speed if physical magnetic saliency exists; including
- Extension of “active flux” concept to MTPA, MaxCosFi,
initial position with PM (or d.c. excitation) flux polarity.
MTPF wide torque-speed control without and with
 The use of SOGI-like schemes allows efficient “active
magnetic saturation from start, with sample results on an
flux” direct processing to observe rotor position and speed
ALA RSM drive, with experiments in Section 3.
[6].
- Discussion and conclusions in Section 4.
 The above simplifications were already acknowledged in
[8] and have been enriched subsequently since 2008 in
II. STATE OF THE ART ON “ACTIVE FLUX” IN A.C. MACHINE
numerous publications [9-63].
DRIVES
 In this paper, besides summarizing recent progress on
“active flux” implementation, we add the derivation of A. Fundamental “active flux”
MTPA, MaxcosFi, MTPF conditions using “active flux” The key to the encoderless “active flux” fundamental model
for all a.c. machine drives (with a case study for axially is the “active flux” observer, which should be rather similar for
laminated anisotropic reluctance synchronous motors all a.c. drives, if a notable simplification of control technology
(ALA-RSM)), without and with magnetic saturation is targeted.
considered in the derivations, from the beginning. A main competitor in this sense proved to be the combined
 The “active flux” concept may be extended to linear closed-loop voltage plus current model for stator flux  ̂ 𝑠 with

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

4
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

subtracted 𝐿𝑞 𝐼𝑠 followed by a PLL to improve its precision on The “active flux” observer may simply be included in FOC,
both estimated angle 𝜃̂𝑑𝑎 and estimated speed ̂ 𝑟 (Fig. 2). The DTFC, feedback linearization, MPC encoderless controls [11]
PI closed loop controllers in the observers may be with added disturbance observers for more robust control
complemented for robustness, by sliding mode components. responses.
The PLL may include the motion equation, to secure good Subsequently, numerous papers have improved/developed/
precision for speed estimation under load torque fast implemented [17-21] concepts of similar or equivalent “active
perturbations during speed transients (Fig. 2b, c [12]). flux” under the “name” of “virtual flux”, “extended flux”,
“equivalent flux” [13], “rotor flux model” [14], “fictitious flux”
Note. The amplitude of ̂ 𝑑𝑎 estimation, which is available as
[15], “auxiliary flux” [16], many inspired from the extended
a bonus, may be “exploited” to correct online one machine emf concept [1].
parameter (Ld or Ldm or ΨPM) which varies most due to magnetic
saturation or/and temperature.

a)

b)

c)
Fig. 2. Control systems: a) Generic encoderless DTFC control system for PMSM, b) and c) typical voltage plus current model close loop active flux observer
with PLL for PMSMs, IMs and RSMs [12].

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

1
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

Fig. 4. DTFC: steady state encoderless operation at lowest speed of 2 rpm (0.1
Fig. 3. Vector control: steady state encoderless operation at lowest speed of 2
Hz) and 50% rated torque: a) Actual speed, b) Estimated speed, c) Measured
rpm (0.1 Hz) and 50% rated torque: a) Actual speed, b) Estimated speed, c)
currents, d) Estimated torque, e) Active flux, f) Actual and estimated rotor
Measured currents, d) Estimated torque, e) Active flux, f) Actual and estimated
position, g) Rotor position error [11].
rotor position, g) Rotor position error [11].

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

2
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

Also, recent implementations to IPMSM drives using sliding

mode “active flux” (some with a disturbance) observer [22-32]
have been introduced to enhance the robustness to machine
parameters and load torque perturbations and to reduce the
lowest speed operation, under full step torque perturbation
down to 2 rpm proven in experiments [12], see Figs. 3, 4.
The so far mentioned literature is referring mainly to PMSM
drives (IPMSM, mostly), but very recently it was also applied a)
to RSM encoderless control (with PM=0 and 𝐿𝑑 > 𝐿𝑞 in
contrast to IPMSM (PM0 and 𝐿𝑑 𝐿𝑞 )) which, in terms of
fundamental “active flux” concept, is discussed in [34-43]. An
example of I-f for starting, combined with FOC active-flux-
based encoderless control of RSM [34] is presented in Fig. 5.
The difference between measured and estimated rotor position
during I-f to FOC transition (Fig. 5b) is to be expected.
However, Fig. 5c shows stable speed response over a large
speed range. b)

c)

a)

d)
Fig. 6. Encoderless V/f control with stabilizing loops versus FOC at ±200 rpm
reference speed under load torque transients: a) reference and estimated speed,
b) estimated d current, c) estimated q current, d) estimated torque [46].

Also, advanced scalar encoderless control (V/f and I-f control

methods) -with stabilizing loops- for IPMSMs, IMs and RSMs
based on “active flux” have been recently proposed, especially
for maximum power factor, to yield a better wide speed range
b)
energy conversion [46-49], as, again, notable simplifications in
the implementation have been found and providing rather close
to FOC and DTFC performance in control (Fig. 6).
B. The case of fundamental “active flux” concept in
encoderless electrically excited SM drives
As alluded to in Introduction the “active flux” in d.c. excited
(and given field current) SMs is similar to the one in IPMSM
but with 𝐿𝑑 > 𝐿𝑞 (normal saliency) [50, 51]. In drives where
c) below 10-15 rpm heavy load operation is not required, it
Fig. 5. I-f plus active flux encoderless RSM control: a) control scheme, b)
sample transition transients, c) wide speed range control, [34].
suffices to estimate the initial rotor position by injecting a d.c.
voltage pulse in the field winding with the inverter short-
Only a few attempts to use “active flux” for MTPA in circuited (zero voltage vector) and record 𝑖𝛼 , 𝑖𝛽 and notice that
IPMSM and RSM encoderless drives [44, 45], all without the d axis is opposite to the stator currents mmf position [50],
considering magnetic saturation, have been made so far, despite (Fig. 7).
possible simplifications.

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

3
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

Also, in electrically excited SMs there is one more variable

iF (field current), that is open to change for energy conversion
optimization, say, for 𝑖𝑑 =0 control during acceleration, below
base speed, and unity power factor above base speed, to max
speed, in wide constant power speed range (CPSR)
applications. Such an encorderless drive has been started on full
friction (load) torque -80 Nm- (from zero speed) (Fig. 7, [50]).

c)
Fig. 8. Complete “active flux” plus cross-coupling saturation flux model of
d.c. excited SM, for encoderless control: a) complete active flux observer, b)
cross-coupling saturation flux model marked in a), c) experimental on-line
equivalent unique magnetization curves 𝑑∗ (𝐼𝑚), 𝑞∗ (𝐼𝑚), 𝐼𝑚 =
√(𝑖𝑑 + 𝑖𝐹 ) + 𝑖𝑞 , [50].
2 2

The “active flux” observer (considering here cross-coupling

saturation estimation online) for this drive is shown in Fig. 8a,
b, c.
The generic control scheme is shown in Fig. 9.

Fig. 7. Electrically excited SM, initial rotor position searching and start to
80 rpm, 100% load: (1) estimated and (2) measured speed (10 rad/s/div), (3)
stator phase current (100 A/div), (4) rotor (d.c. excitation) current (10A/div);
time scale 500 ms/div [50].

8a

Fig. 9. Encoderless d.c. excited SM FOC with active flux observer, [53].

No other work (besides [51-52]) on “active flux” control of

8b) electrically excited synchronous motors was found so far in the
mainstream literature.
Signal injection encoderless control with “active flux” will
now be presented.
C. “Signal injection” with “active flux”
As pointed out in the Introduction, basically “active flux” is
aligned to d axis in a machine with physical d-q saliency even
at standstill and thus it is able to estimate rotor position by
signal injection, too. An early implementation of “active flux”
for an ALA – RSM encoderless control from zero speed [53]
a) that already integrates signal injection (but still with injected
current processing for rotor position) with “active flux” FOC,
offering zero to 3000 rpm operation, is presented in [53].
In a more recent implementation for RSM [54] it was proved
that by using a fundamental “active flux observer” for rotor
position but processing q axis flux for signal injection, zero
speed operation is improved (smaller position error) in
comparison with the injected current processing.
Then [55] is dedicated to signal injection by “active flux” in
an RSM as it is applied only at standstill and small speeds, but
it considers the magnetic saturation via transient inductances
𝐿𝑑𝑑 , 𝐿𝑞𝑞 while neglecting cross-coupling saturation.
It needed a LUT table of magnetization curves, but the results
b)
at standstill are outstanding (Fig. 10).

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

4
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

TABLE I “ACTIVE FLUX” CONCEPT: FUNDAMENTAL MODELS (FM)

Machine Electrically F-M
IPMSM RSM IM
type Excited SM Machines
Extended emf [1], 2000
First
Fictitious PM flux [7] 2005 Active flux [8] 2007
proposals
Active Flux [8], 2008
With SM or disturbance [35-
Subsequent
observers, [2], [11] [12-33] 43] NA [50-51] NA
proposals
[ 58-63] [59]
[22], [2]
Position Robustness of active-flux
[39], [10],
speed sensorless control NA NA
[41], [19],
observer [65], [66], [68]
[58] [24]
[45],
MTPA [44], [64] NA NA NA
[48]
Scalar
[46] [49] [47] NA NA
Phase current [A]

Control
MTPA
Maxcos𝜑 NA [45] NA NA NA
MTPF
TABLE II
“ACTIVE FLUX” CONCEPT: SIGNAL INJECTION (SI) PLUS INTEGRATED
SIGNAL INJECTION AND FUNDAMENTAL MODEL (SI+FM)
DC
Machine F-M
IPMSM RSM IM Excited
type Machines
SM
First [ 53-55],
NA NA NA NA
proposals [67]

 The “active flux” concept was applied in a unified manner

for all synchronous and induction motor drives with notable
a)
performance in encoderless control with rather precise and
robust (after correction for magnetic saturation) response
(mainly using closed loop voltage plus current model and
sliding mode “active flux” observers);
 So far mainly the fundamental “active flux” model was
particularly investigated for IPMSMs and RSMs and less
for IMs and d.c. excited SMs and not yet for flux-
Phase current [A]

modulation (F-M) machine drives [9].

 Among the “active flux” observers the voltage – current
closed loop (PI+SM) observer with PLL or with active
disturbance compensation has been proven adequate for the
scope: though other observers with disturbance estimation
components may be suitable for same scope.
 Despite of efforts, so far, to use signal injection, for heavy
starts and prolonged 1-5 rpm operation, fusing “active flux”
both for signal injection and fundamental model,
considering cross-coupling saturation, is yet due;
 Only a few very recent attempts have been made to use
“active flux” also to simplify MTPA, MaxCosFi, MTPF, all
fused smoothly for optimal operation for the entire torque-
b)
speed envelope of a.c. drives.
Fig. 10. Active flux signal injection encoderless RSM speed control test: a) at
standstill with step full loading, b) and no-load acceleration/deceleration The next paragraph discusses such an endeavor.
to/from 50 rpm [55].
III. “ACTIVE FLUX” FOR MTPA, MAXCOSFI, MTPF, IN A.C.
D. Discussion on state of the art
DRIVES
 Table I and Table II summarize the mainstream literature on
“active flux” so far. A. Magnetic saturation is considered constant
 The active flux, proved to be a phenomenological (physical) MTPA in synchronous (IPMSMs plus RSMs) has been
concept based on the simple formula:  ̅ 𝑑𝑎 = ̅ 𝑆 − I𝑠̅ 𝐿𝑞 , investigated thoroughly by a multitude of methodologies [56],
valid in any system of coordinates and aligned to rotor d axis in order to reduce losses at given speed-torque operation modes
(for SMs, and, in principle to flux-modulation machines) when the required stator voltage is smaller than the maximum
and to rotor flux axis in IMs.

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

5
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

value available from the inverter, with, in general, a smooth b. Proposed MTPF
passing to MTPF (is required).
As long as overmodulation is available and then for six pulse This time simply we express 𝑠2 as a function of 𝑑𝑎 :
operation, rather reconfigurable control is required especially 2
𝐿𝑑 (𝑑𝑎 −𝐹 ) 2𝑇𝑒∗ 𝐿𝑞 2
for wide CPSR as in electric power trains. 𝑠2 (𝑑𝑎 ) = (𝐹 + 𝐿𝑑 −𝐿𝑞
) +(
3𝑝1 𝑑𝑎
2
) < 𝑉𝑠𝑚𝑎𝑥 /2𝑟 (12)
While widely accepted in theory, a full torque – speed
envelope control for low losses in the electric machine drives Making the derivative of (12) with respect to 𝑑𝑎 zero we
observing both stator current and voltage limitations is still due. obtain:
Here, aiming at such a goal, we first describe as an alternative (𝐿𝑑 − 𝐿𝑞 )
the use of “active flux” for integrated MTPA, MaxCosFi, (𝑑𝑎 )4 − (𝑑𝑎 )3 (1 − ) 𝐹 −
𝐿𝑑
MTPF control in a.c. machine drives, first considering a 2𝑇𝑒∗ 𝐿 2
constant magnetic saturation in the machine. −( (𝐿𝑑 − 𝐿𝑞 ) 𝑞 ) = 0 (13)
3𝑝1 𝐿𝑑
Note. In general, MTPA and MTPF are used in the literature
as fused operation modes to optimize the entire torque-speed An iterative solution of (13) may be obtained similarly as
range in a.c. drives. Here we add in between (especially for done in (11).
RSM, IM and wide CPSR IPMSM drives) the max ideal power c. Proposed MaxCosFi
factor mode because it leads to higher torque (for same stator
flux) and given speed and voltage and because it is also weakly Here we use 𝑡𝑎𝑛𝜑 (𝜑 − power factor angle):
dependent on speed. In contrast to efficiency, though losses
𝑑 𝑖𝑞 − 𝑞 𝑖𝑑
increase actual power factor in motoring and reduce it in 𝑡𝑎𝑛𝜑(𝑑𝑎 ) = =
generating mode. 𝑑 𝑖𝑑 + 𝑞 𝑖𝑞
It should be mentioned that MTPA for IPMSMs may be used 2
up to full (max.) torque and up to base speed, while for RSMs 𝑇𝑒∗
3𝑝1
(or IMs) going towards MaxCosFi will produce more torque for 2 (14)
a given voltage with high saliency: 𝐿𝑑 /𝐿𝑞 (RSM) or 𝐿𝑠 /𝐿𝑠𝑐 (𝑑𝑎 − 𝐹 ) ( 𝑎 − 𝐹 ) 2𝑇𝑒∗ 2
𝑑𝑎 + 𝐿𝑑 ( 𝑑 ) +( )
(IM). 𝐿𝑑 − 𝐿𝑞 𝐿𝑑 − 𝐿𝑞 3𝑝1 𝑑𝑎
In what follows we will derive the conditions for MTPA,
Zeroing the derivative of (14) with respect to 𝑑𝑎 we
MaxCosFi, MTPF directly for IPMSM; RSM and IM are
obtain:
particular cases with 𝐹0 = 0; magnetic saturation is still
considered constant. (𝐿𝑑 − 𝐿𝑞 )
(𝑑𝑎 )4 − (𝑑𝑎 )3 𝐹 (1 − )
a. Existing MTPA [44] 2𝐿𝑑
2
For a given torque, based on the active flux expression (3) 3𝑇𝑒∗ 𝐿𝑞
−( (𝐿𝑑 − 𝐿𝑞 )) =0 (15)
the stator current of IPMSM 𝑖𝑠 (peak phase value) is: 3𝑝1 𝐿𝑑

2 2
(𝑑𝑎 −𝐹 ) 2𝑇𝑒∗ Note. It could be simply proven that for given torque the
𝑖𝑠2 (𝑑𝑎 ) = ( ) +( 2
) < 𝑖𝑠𝑚𝑎𝑥 (9)
𝐿𝑑 −𝐿𝑞 3𝑝1 𝑑𝑎 minimum stator flux 𝑠 (12) (minimum stator voltage for given
speed) is obtained for MTPF, then for MaxCosFi and, then max
By making the derivative of (9) with respect to 𝑑𝑎 (active
𝑠 for MTPA:
flux) zero, for given reference torque 𝑇𝑒∗ , we obtain simply [44]:
2 (𝑠 )𝑀𝑇𝑃𝐴
𝑇𝑒∗ > (𝑠 )MaxCosFi
𝑇𝑒∗ > (𝑠 )𝑀𝑇𝑃𝐹
𝑇𝑒∗ (16)
2𝑇𝑒∗
(𝑑𝑎 )4 − (𝑑𝑎 )3 𝐹 −( (𝐿𝑑 − 𝐿𝑞 )) = 0 (10)
3𝑝1 Consequently, it is sufficient to post calculate the
voltage 𝑉𝑠∗ ≈ 𝜔𝑟 𝑠∗ (with 𝑠∗ from (12) for given 𝑇𝑒∗ and speed
with 𝐹 = 𝐿𝑑𝑚 𝑖𝐹∗ for d.c. excited SM; 𝐹 = 𝑃𝑀𝑑 for IPMSMs
and 𝐹 = 0 for RSM (for IM also, 𝐿𝑑 → 𝐿𝑠 , 𝐿𝑞 → 𝐿𝑠𝑐 ). and 𝑑𝑎 from (10), (13), (15)) and to use the control strategy
that leads to 𝑉𝑠 ≤ 𝑉𝑠∗ . If two of them satisfy the condition the
Equation (10) was solved in [44] in 4-5 iterations (by Newton
– Raphson method) online for IPMSMs: one with minimum total electric losses will be applied:
3 3
𝑑𝑎 (𝑘) = 𝑑𝑎 (𝑘 − 1) − ∑ 𝑝 = 𝑅𝑠 𝑖𝑠2 + 𝑅𝑟 𝑖𝑟2 + 𝑘𝑖𝑟𝑜𝑛 𝑉𝑠2 (17)
2 2
2
4 2𝑇𝑒∗
(𝑑𝑎 (𝑘 − 1)) − (𝑑𝑎 (𝑘 − 1))3 𝑃𝑀𝑑 − ( (𝐿𝑑 − 𝐿𝑞 )) where the core losses are considered simply
3𝑝1
(11) proportional to 𝑉𝑠2 ; which covers approximately for the
4(𝑑𝑎 (𝑘 − 1))3 − 3(𝑑𝑎 (𝑘 − 1))2 𝑃𝑀𝑑 harmonic core losses as it considers all core losses to be eddy
current losses. For a single given 𝑠 and 𝜔1 the core losses have
More than 10% stator current reduction, with respect to to be known, in order to determine approximately the
conventional methodologies has been reported experimentally coefficient 𝑘𝑖𝑟𝑜𝑛 .
[44]. For the RSM equations (10), (13), (15) get simplified (as
𝐹0 = 0):

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

6
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

2|𝑇𝑒∗ |
(𝑑𝑎 )𝑀𝑇𝑃𝐴 = √ (𝐿𝑑 − 𝐿𝑞 ) ;
3𝑝1
4 𝐿𝑞
(𝑑𝑎 )𝑀𝑎𝑥𝑐𝑜𝑠𝜑 = (𝑑𝑎 )𝑀𝑇𝑃𝐴 √ ; (18)
𝐿𝑑

𝐿𝑞
(𝑑𝑎 )𝑀𝑇𝑃𝐹 = (𝑑𝑎 )𝑀𝑇𝑃𝐴 √
𝐿𝑑

Note. For IMs basically, equation (18) is valid with 𝐿𝑑 →

𝐿𝑞 𝐿𝑠𝑐
𝐿𝑠 , 𝐿𝑞 → 𝐿𝑠𝑐 ; < 1 and ≪ 1.
𝐿𝑑 𝐿𝑠
For d.c. excited SMs it is possible to repeat the calculation
above for a few values of field current and choose the desired
control strategy. However, for more precision, a direct
derivation of MTPA, MaxCosFi=1 and MTPF conditions is
recommended for such a case.
B. An RSM case study
In [45] the integrated MTPA – MaxCosFi – MTPF control
strategies have been applied on an ALA RSM (Table III) with
rather smooth transitions and promising results for the entire
torque and speed range, but still with constant machine
parameters.
TABLE III ALA ROTOR SYNREL PARAMETERS

Sample results from [45] are given in Figs. 11 and 12.

Torque [Nm]
Current [A]

Fig. 12. ALA rotor RSM – MTPA + Maxcos𝜑 + MTPF “active flux” control:
experimental test results at 1000 rpm, [45].

The simulation results in Fig. 11 prove that the successful

transition between the three strategies that cover all torque–
Voltage [V]

speed range is rather smooth (stable, too) and that the max.
available voltage is observed (Fig. 12). Experiments (Fig. 12)
confirm rather satisfactorily the simulation results.

Fig. 11 ALA rotor RSM – MTPA + Maxcos𝜑 + MTPF “active flux”:

simulation results at 3000 rpm

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

7
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

The apparently complex equations (19) – (21) may be solved

in a few iterations by the Newton-Raphson method (as equation
(11)) [44]. 𝐹0 and 𝐿𝑞0 mark the fact that these parameters are
considered constant.
Note. Identifying 𝐿𝑑 (𝑑𝑎 ) is “a fruit” of | ̂ 𝑑𝑎 | = 𝐹0 +
(𝐿𝑑 − 𝐿𝑞 )𝑖𝑑0 which is the amplitude of |̂ 𝑑𝑎 |obtained already
𝑎
in the 
̂ 𝑑 observer, but not used so far in the literature. With
a)
𝐹0 , 𝐿𝑞 given and 𝑖𝑑 known from the FOC diagram, we may
invert the above equation point by point to get the 𝐿𝑑 (𝑑𝑎 )
function.
A very recent paper includes magnetic saturation for direct
active flux and torque deadbeat control for flux weakening and
max. torque up to base speed with stator resistance voltage drop
consideration [64].
b)
Fig. 13. Magnetization curves of the motor in Table III: a) axis d, b) axis q. IV. FINAL DISCUSSIONS AND CONCLUSIONS
From Fig. 13 we retained average linear curves of d and q  “Active flux” concept 𝑑𝑎 = 𝐹𝑑 + (𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 has been
axes measured flux linkages. introduced for all a.c. machines as a generalization in 2008,
However, tempting as they are, the above simplifications by of “virtual PM flux” derived, only for IPMSM [7], in 2005.
using “active flux” for MTPA – MaxCosFi – MTPF for all a.c.  “Active flux” is a technical concept representing an
drives in absence of magnetic saturation variation, the equivalent ac machine without functional saliency while the
consideration of the latter has to be thoroughly investigated too. original a.c. machine may have saliency (physical or virtual
This is done in the following paragraph. (IM)).
C. Proposed Magnetic saturation as accounted for via “active  The “active flux” space phasor of a.c. machines (traveling
flux”for all a.c. drives field and flux-modulation types) is ideally (with constant
machine parameters) aligned to axis d of all synchronous and
The key to a simplified version of considering magnetic F-M machines and along the rotor flux in IMs and shows only
saturation is to assume 𝐿𝑞 = 𝑐𝑜𝑛𝑠𝑡 but 𝐿𝑑 as a given function the inductance 𝐿𝑞 .
of active flux 𝐿𝑑 (𝑑𝑎 ), thus neglecting the cross-coupling  The “active flux” may be applied also along axis q 𝑞𝑎 =
saturation. Consequently, we simply restart from equations (9), 𝐹𝑞 + (𝐿𝑞 − 𝐿𝑑 )𝑖𝑞 when the model will show, again, only
(12), (14) but in the zeroing of their derivatives with respect to
one inductance: 𝐿𝑑 [35].
𝑑𝑎 we will consider now that 𝐿𝑑 is 𝐿𝑑 (𝑑𝑎 ), to obtain finally:
 Being aligned to the rotor d axis for correct machine
For MTPA: parameters, in SMs and F-M machines with physical saliency
𝜕𝐿𝑑 and fractional in any system of coordinates (at any speed),
𝜕𝑑𝑎 (𝑑𝑎 − 𝐹0 ) “active flux” simplifies the rotor position estimation in
− 3
(𝑑𝑎 − 𝐹0 )2 + 2 encoderless drives. So “active flux” may be used not only as
(−𝐿𝑞 ) (𝐿𝑑 (𝑑𝑎 ) − 𝐿𝑞 )
“fundamental active flux” but also in “signal injection active
(19) flux” observers, avoiding the low-speed large errors by SOGI
2𝑇𝑒∗ 2 1 variants or by adaptive robust disturbance observers
−( ) =0
3𝑝1 (𝑑𝑎 )3 integrated with “active flux” close loop voltage plus current
For MaxCosFi: model PI or PI+SM observers.
 As of now, “active flux” has been implemented very recently
𝜕𝐿𝑑 (𝑑𝑎 − 𝐹0 )2 ∙ 𝐿𝑞 𝑑𝑎 ∙ 𝐿2𝑑 (𝑑𝑎 ) in better and better configurations for encoderless IPMSM
−𝐿𝑑 (𝑑𝑎 ) ∙ +
𝜕𝑑 (𝐿𝑑 (𝑑𝑎 ) − 𝐿𝑞0 )3 (𝐿𝑑 (𝑑𝑎 ) − 𝐿𝑞0 )2
𝑎
and RSM drives, including stator resistance adaptation [65,
(20) 66], a new combined active flux with signal injection control
from zero to maximum speed [67] and compensation of
2𝑇𝑒∗ 𝐿𝑞 2 1
−( ) =0 inverter nonlinearity voltage errors [68].
3𝑝1 (𝑑𝑎 )3  However, Tables I, II showed many “NA” positions, so there
For MTPF: is still a lot of work to do.
𝜕𝐿𝑑 (𝑑𝑎 − 𝐹0 )2 In order to improve the technology, the following may be
− (𝐿𝑑 (𝑑𝑎 ) + 𝐿𝑞 )
𝜕𝑑𝑎 (𝐿𝑑 (𝑑𝑎 ) − 𝐿𝑞0 )3 considered:
- Even better, more robust and more precise “active flux”
(21)
observers for the entire speed-torque range in various
2(𝑑𝑎 − 𝐹0 )𝐿𝑑 (𝑑𝑎 ) 2𝑇𝑒∗ 2 2𝐿𝑞0 applications
+ −( ) ∙ =0
(𝐿𝑑 (𝑑𝑎 ) − 𝐿𝑞0 )
2
3𝑝1 (𝑑𝑎 )3

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

8
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

- MTPA+MaxCosFi+MTPF full torque-speed envelope [16] A. Varatharajan, G. Pellegrino, E. Armando, “Sensorless Synchronous
Reluctance Motor Drives: Auxiliary Flux-Based Position Observer,”
control, observing the limited current and voltage of the
IEEE-JESTPE, vol. 9, no. 4, pp. 4330 – 4339, 2021;
inverter, by applying “active flux”; a draft theory of such an [17] M. Mamdouh, S. M. Ismaeel, S. M. Allam, E. M. Rashad, “A Proposed
endavour is included in this paper without and with Active Flux Based Sensorless Control of Cage-Rotor Synchronous
considering magnetic saturation variation with torque and Reluctance Motor,” Record of IEEE-MEPCON (22)-2021, pp. 21541589,
(IEEEXplore).
speed.
[18] M. Costin, C. Lazar, “Active Flux Based Predictive Control of Interior
- Novel, “active flux” based, self-commissioning schemes for Permanent Magnet Synchronous Machine,” IEEE-ISFEE-2020, pp.
all a.c. encoderless drives. 21759618, (IEEEXplore).
- “active flux” based encoderless control of electric [19] Q. Lin, L. Liu, D. Liang, “Hybrid Active Flux Observer to Suppress
Position Estimation Error for Sensorless IPMSM Drives,” IEEE Trans.
generators in wind energy and marine slip power systems
Power Electron., vol. 38, no. 1, pp. 872 - 886, 2023.
- Extension of “active flux” control to dc excited SM drives [20] A. Ejlali, S. Taravat, J. Soleimani, “Comparative study on control strategy
for e-transport; and to F-M machine drives of DTFC based on active flux concept and DTC for IPMSM applied to
- Potential generalization of “active flux” usage in scalar traction motor,” SPEEDAM, 2012, pp. 12934694, (IEEEXplore).
[21] S. M. S. Islam Shakib, D. Xiao, R. Dutta, M. F. Rahman, “Online deadbeat
control and also model predictive control of ac drives.
predictive control for IPMSM drive,” Record of IEEE-IPEC-Himeji 2022-
We should end, in a note of humbleness since “active flux” ECCE Asia, pp. 21861050, (IEEEXplore).
concept is just one potentially useful/practical alternative for [22] Y.-Ch. Liu, S. Laghrouche, A. N’Diaye, M. Cirrincione, “Active-Flux-
better control of unified a.c. motor/generator drives and not “the Based Super-Twisting Sliding Mode Observer for Sensorless Vector
Control of Synchronous Reluctance Motor Drives,” Record of IEEE-
one that solves it all”.
ICRERA 2018, pp. 18308554, (IEEEXplore).
[23] I. Boldea, M. C. Paicu, G.-D. Andreescu, F. Blaabjerg, ““Active Flux”
REFERENCES DTFC-SVM Sensorless Control of IPMSM,” IEEE Trans. Energy
Convers., vol. 24, no. 2, pp. 314-322, Jun. 2009.
[1] Z. Chen, M. Tomita, S. Ichikawa, S. Doki, S. Okuma, “Sensorless control
[24] A. T. Woldegiorgis, X. Ge, M. Hassan, “Advanced disturbance observer-
of interior permanent magnet synchronous motor by estimation of an
based active flux estimation for sensorless control of IPMSM”, Record of
extended electromotive force,” Record of IEEE-IAS, Rome, Italy, 2000, pp.
IEEE-PEMD, 2020, pp. 21016690, (IEEEXplore).
1814–1819.
[25] M. Zhang, C. Wu, H. Xiao, “Second-sliding mode adaptive active flux
[2] J. Chen, Xin Yuan, Frede Blaabjerg, Christopher H. T. Lee, “Overview of
control for Interior Permanent Magnet Synchronous Motor,” Record of
Fundamental Frequency Sensorless Algorithms for AC Motors: A Unified
IEEE – ICHCI, 2021, pp. 21722954, (IEEEXplore).
Perspective,” IEEE-JESTPE, vol. 11, no. 1, pp. 915 – 931, 2023.
[26] Z. Guoqiang, W. Gaolin, Ni Ronggang, Xu Dianguo, “Active flux based
[3] C. Lascu, I. Boldea, F. Blaabjerg,“Direct torque control of sensorless
full-order discrete-time sliding mode observer for position sensorless
induction motor drives a sliding-mode approach,” IEEE Trans. Ind. Appl.,
IPMSM drives,” Record of IEEE-ICEMS (17)-2019, pp. 14865140,
Nov./Dec. 2012. , vol. 40, no. 2, pp. 582-590, Mar.-Apr. 2004.
(IEEEXplore).
[4] C. Lascu, I. Boldea, F. Blaabjerg, “Direct torque control via feedback
[27] K. Zhao, R. Zhou, J. She, P. Li, “Active – flux – based sensorless control
linearization for permanent magnet synchronous motor drives,” Record of
for permanent magnet synchronous motor by variable – gain sliding –
OPTIM 2012 (IEEEXplore), Brasov, Romania, 2012, pp. 338-343.
mode observer,” Record of IEEE-ICCA(16), 2020, pp. 20237224,
[5] N.-D. Nguyen, N. N. Nam, C. Yoon, Y. Il Lee, “Speed Sensorless Model
(IEEEXplore).
Predictive Torque Control of Induction Motors Using a Modified Adaptive
[28] M.-C. Ancuti, A.-S. Isfanuti, G.-D. Andreescu, L.-N. Tutelea, I. Boldea
Full-Order Observer”, IEEE Trans. Ind. Electron., vol. 69, no. 6, pp. 6162-
“Referencing position versus speed active flux based encoderless control
6172, Jun. 2022.
of PM-RSM drives at ultra-low speeds without signal injection,”, Record
[6] R. Sreejith, B. Singh, "Sensorless Predictive Current Control of PMSM EV
of IEEE-OPTIM-ACEMP, 2021, pp. 21275304, (IEEEXplore).
Drive Using DSOGI-FLL Based Sliding Mode Observer," IEEE Trans.
[29] A. Ejlali, D. A. Khaburi, H. Behnia, “DTFC based indirect matrix
Ind. Electron., vol. 68, no. 7, pp. 5537-5547, 2021.
converter using active flux concept for IPMSM drive,” Record of IEEE-
[7] S. Koonlaboon, S. Sangwongwanich, "Sensorless control of interior
UPEC (46) 2011, (IEEEXplore).
permanent-magnet synchronous motors based on a fictitious permanent-
[30] C. R. S. Bartsch, L. F. F. de Campos, A. G. Bartsch, J. de Oliveira, A.
magnet flux model," Record of IEEE IAS, pp. 8671645, 2005.
Nied, “Low computational cost technique for SPMSM sensorless drive
[8] I. Boldea, M. C. Paicu, G.-D. Andreescu, “Active Flux Concept for Motion-
using active flux concept,” Record of IEEE-COBEP/SPEC (15) 2019, pp.
Sensorless Unified AC Drives,” IEEE Trans. Power Electron., vol. 23, no.
19534468, (IEEEXplore).
5, pp. 2612-2618, Sep. 2008.
[31] W. Liu, S. G. Ding, Y. H. Dong, H. Jin, K. Huang, Z. W. Hu, “Sensorless
[9] I. Boldea, L. Tutelea, “Reluctance electric machine design and control”,
control strategy for IPMSM based on ADRC and differential of active
book, CRC Press, Taylor and Francis, 2021.
flux,” Record of IEEE-ICEIEC (9) 2019, pp. 18886165, (IEEEXplore).
[10] S. Zhou, Z. Zhang, H. Li, Z. Li, Q. Xing, X. Liu, F. Wang, J. Rodriguez,
[32] A. T. Woldegiorgis, X. Ge, H. Wang, M. Hassan, “A New Frequency
“A Robust Encoderless Control for PMSM Drives: a Revised Hybrid
Adaptive Second-Order Disturbance Observer for Sensorless Vector
Active Flux-Based Technique,” IEEE Trans. Power Electron., vol. 38, no.
Control of Interior Permanent Magnet Synchronous Motor,” IEEE Trans.
11, pp. 14438-14449, Nov. 2023.
Ind. Electron., vol. 68, no. 12, pp. 11847 - 11857, 2021.
[11] I. Boldea, S. C. Agarlita, “The active flux concept for motion-sensorless
[33] F. Toso, P. G. Carlet, M. Preindl, S. Bolognani, “Active-Flux-Based
unified AC drives: A review,” Record of ACEMP-OPTIM 2011, pp.
Motion-Sensorless Control of PMSM Using Moving Horizon Estimator,”
13429763, (IEEEXplore).
2018 IEEE SLED (9), pp. 18163761, (IEEEXplore).
[12] M. C. Paicu, I. Boldea, G. D. Andreescu, F. Blaabjerg “Very low speed
[34] S.-C. Agarliţă, M. Fătu, L. N. Tutelea, F. Blaabjerg, I. Boldea, “I-f starting
performance of active flux based sensorless control: interior permanent
and active flux based sensorless vector control of reluctance synchronous
magnet synchronous motor vector control versus direct torque and flux motors, with experiments,” Record of IEEE-OPTIM (12), 2010, pp.
control,” Proc of IET, vol. EPA-3, no. 6, pp. 551-561, 2009. 11431647, (IEEEXplore).
[13] M. Koteich, G. Duc, A. Maloum, G. Sandou, “A unified model for low- [35] Z. Zhang, J. Lamb, “Active Q Flux Concept for Sensorless Control of
cost high-performance AC drives: The Equivalent Flux concept,” Record Synchronous Reluctance Machines,” IEEE Trans. Ind. Electron., vol. 70,
of IEEE-PEMC-2021, pp. 16004762, (IEEEXplore). no. 5, pp. 4526 - 4536, 2023.
[14] D. O. Kisck, D. S. Anghel, M. Kisck, “Rotor-flux model for sensorless [36] H. Hadla, S. Cruz, “Active flux based finite control set model predictive
vector-control of super-high-speed IPMSM,” Record of IEEE-ATEE (10), control of synchronous reluctance motor drives,” Record of IEEE-ECCE-
2017, pp. 16836073, (IEEEXplore). EUROPE (18), 2016, pp. 16412826, (IEEEXplore).
[15] D. O. Kisck, M. Kisck, D. S. Anghel, “Sensorless control of high-speed [37] F. J. W. Barnard, W. Villet, M. Kamper, “Hybrid active-flux and arbitrary
IPMSM based on rotor-flux and fictitious-flux estimation,” Record of injection position sensorless control of reluctance synchronous
IEEE-IEMDC, 2017, pp. 17097051, (IEEEXplore). machines”, IEEE Trans. Ind. Appl., vol. 51, no. 5, pp. 3899 - 3906, 2015.

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
 This article has been accepted for publication in IEEE Transactions on Power Electronics. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2024.3405073

9
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <

[38] B. Zhenfeng, W. Sun, D. Jiang, “Speed Sensorless Control of motors, with experiments from zero speed,” IEEE Trans. Ind. Appl., vol.
Synchronous Reluctance Motor At Full Speed Range,” Record of IEEE- 48, no. 6, pp. 1931-1939, 2012.
CIEEC (4), 2021, pp. 21416152, (IEEEXplore). [54] A. Y.-Talouki, P. Pescetto, G. Pellegrino, I. Boldea, “Combined Active
[39] A. Varatharajan, G. Pellegrino, “Sensorless Synchronous Reluctance Flux and High-Frequency Injection Methods for Sensorless Direct-Flux
Motor Drives: A Projection Vector Approach for Stator Resistance Vector Control of Synchronous Reluctance Machines,” IEEE Trans.
Immunity and Parameter Adaptation,” IEEE Trans. Ind. Appl., vol. 56, Power Electron., vol. 33, no. 3, pp. 2447-2457, 2018.
no. 5, pp. 5003 - 5012, 2020. [55] P. Pescetto, A. Y.-Talouki, G. Pellegrino, “Novel Sensorless Control
[40] B. Nikmaram, H. Pairo, “Comparison Study of Active Flux based Sliding- Algorithm for SyR Machines Based on Low Speed Active Flux,” Record
Mode Observer and PLL based Sliding-Mode Observer Sensorless of IEEE – SLED (10), 2019, pp. 19137078, (IEEEXplore).
Control of SynRM,” Record of IEEE-PEDSTC (12), 2021, pp. 20552580, [56] A. Dianov, F. Tinazzi, S. Calligaro, Silverio Bolognani, “Review and
(IEEEXplore). Classification of MTPA Control Algorithms for Synchronous Motors,”
[41] A. Varatharajan, G. Pellegrino, “Sensorless Control of Synchronous IEEE Trans. Power Electron., vol. 37, no. 4, pp. 3990-4007, 2022.
Reluctance Motor Drives: Improved Modeling and Analysis Beyond [57] A. T. Woldegiorgis, X. Ge, H. Wang, Y. Zuo, “An Active Flux Estimation
Active Flux,” Record of IEEE-IEMDC, 2019, pp. 18884971, in the Estimated Reference Frame for Sensorless Control of IPMSM,”
(IEEEXplore). IEEE Trans. Power Electron., vol. 37, no. 8, pp. 9047-9060, 2022.
[42] J. Matoušek, J. Duník, M. Brandner, V. Elvira, “Comparison of Discrete [58] F. Gao, Z. Yin, C. Bai, D. Yuan, J. Liu “A Lag Compensation-Enhanced
and Continuous State Estimation with Focus on Active Flux Scheme,” Adaptive Quasi-Fading Kalman Filter for Sensorless Control of
Record of IEEE – FUSION, 2021, pp. 21481160, (IEEEXplore). Synchronous Reluctance Motor,” IEEE Trans. Power Electron., vol. 37,
[43] N. Bianchi, S. Bolognani, F. Tinazzi, M. Zigliotto, “The influence of rotor no. 12, pp. 15322-15337, 2022.
design on active flux-based sensorless synchronous reluctance motor [59] J. Chen, J. Mei, X. Yuan, Y. Zuo, J. Zhu, C. H. T. Lee, “Online Adaptation
drives,” Record of IEEE-SLED, 2017, pp. 17287222, (IEEEXplore). of Two-Parameter Inverter Model in Sensorless Motor Drives,” IEEE
[44] S. Liu, W. Huang, X. Lin, Y. Zhao, W. Jiang, J. Yang, “Maximum Torque Trans. Ind. Electron., vol. 69, no. 10, pp. 9860-9871, 2022.
per Ampere Control Based on Active Flux Concept for DTC of IPMSMs,” [60] Y. Li, Y. Hu, X. Ma, L. Liu, “Sensorless control of dual three-phase
Record of IEEE-ECCE, 2018, pp. 18293364, (IEEEXplore). IPMSM based on frequency adaptive linear extended state observer”,
[45] S. C. Agarlita, A. S. Isfanuti, I. Boldea, “Active Flux based control of IEEE Trans. Power Electron., vol. 38, no. 11, pp. 14492-14503, 2023.
ALA Rotor SynRel with MTPA, Maximum Power Factor and Maximum [61] F. Butaru, C. Carstea, A-I. Ratiu, M.-C. Ancuti, S. Musuroi, “Active Flux
Torque per Flux”, Record of ACEMP-OPTIM 2023 (IEEEXplore). Observer-Based High-Speed Control of a Permanent Magnet
[46] A.-S. Isfanuti, M.-C. Paicu, L.-N. Tutelea, T. Staudt, I. Boldea, “V/f with Synchronous Motor”, Record of 2023 13 th International Symposium of
stabilizing loops versus FOC of Spoke-PM rotor SM drive: control with Advanced Topics of Electrical Engineering (ATEE), (IEEEXplore).
experiments,” Record of IEEE-PEMC (18), pp. 18248279, 2018, [62] A. Wang, D. Chen, L. Tao, F. Han, Z. Hu, H. Wang, H. Qiao, J. Wang
(IEEEXplore). “Two stage active flux observer for position sensorless control of PMSM
[47] A. A. Popa, A. S. Isfanuti, L. N. Tutelea, I. Boldea, “Induction Machine drives”, 2023 26th International Conference on Electrical Machines and
V/f Control with Stator Flux and Slip Frequency Compensation for Systems (ICEMS) (IEEEXplore).
Maximum Power Factor, with Experiments,” Record of ACEMP-OPTIM [63] M. Fontana, N. Bianchi, “Design and analysis of normal saliency IPM
2023 (IEEEXplore). spoke motor,” IEEE Trans. Ind. Appl., vol. 56, no. 4, pp. 3625-3635, 2020.
[48] M. Vidlak, S. C. Agarlita, I. Boldea, “MTPA control strategy for ALA [64] S M Sh. I. Shakib, D. Xiao, R. Dutta, M. F. Rahman, “An optimal
rotor RSM based on reactive and apparent power calculation,” Record of maximum torque per active flux and field weakening operation for
IEEE-EDPE 2023, (IEEEXplore). deadbeat direct torque control based IPMSM drive”, IEEE Trans. Ind.
[49] A. A. Popa, A. S. Isfanuti, L. N. Tutelea, I. Boldea, “V/f with stabilizing Appl., vol. 60, no. 2, pp. 3210 - 3220, 2024.
loops fast response control of induction motor (IM) drives,” Record of [65] S. Zhou, Z. Zhang, H. Li, Z. Li, Q. Xing, X. Liu, F. Wang, J. Rodriguez,
ACEMP-OPTIM 2021, pp. 21415281, (IEEEXplore). “A Robust Encoderless Control for PMSM Drives: A Revised Hybrid
[50] I. Boldea, G. D. Andreescu, C. Rossi, A. Pilati, D. Casadei, “Active flux Active Flux-Based Technique“, IEEE Trans. Power Electron., vol. 38, no.
based motion-sensorless vector control of DC-excited synchronous 11, pp. 14438-14449, Nov. 2023.
machines,” Record of IEEE-ECCE, 2009, pp. 10965333, (IEEEXplore). [66] Q. Lin, L. Liu, D. Liang, “Joint Estimation of Resistance and Speed with
[51] V. Coroban-Schramel, I. Boldea, G.-D. Andreescu, F. Blaabjerg, “Active- Auxiliary State of Active Flux for Sensorless IPMSM Drives”, IEEE
flux based motion sensorless vector control of biaxial excitation Trans. Ind. Electron. (Early Access), pp. 1-10, December 2023.
generator/motor for automobiles (BEGA),” IEEE Trans. Ind. Appl., vol. [67] F. Lee, Z. Zhao, Y. Wang, C. Shen, “Combined Active Flux and High-
47, no. 2, pp. 812-819, 2011. Frequency Injection Methods for Sensorless Control of Synchronous
[52] I. Boldea, S. C. Agarlita, “The active flux concept for motion-sensorless Reluctance Machines”, 2023 IEEE 2nd International Power Electronics
unified AC drives: A review,” Record of ACEMP-OPTIM 2011 and Application Symposium (PEAS), pp. 2332-2336.
(IEEEXplore), pp. 13429763. [68] H. Li, S. Zhou, Z. Zhang, Z. Li, “A Hybrid Active Flux Observer Based
[53] S. C. Agarlita, I. Boldea, F. Blaabjerg, “High frequency injection assisted Encoderless Control Method for AC Motor Drives”, 2023 26th
“active flux” based sensorless vector control of reluctance synchronous International Conference on Electrical Machines and Systems (ICEMS).

Authorized licensed use limited to: China University of Petroleum. Downloaded on June 04,2024 at 15:48:13 UTC from IEEE Xplore. Restrictions apply.
© 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.