Received December 17, 2018, accepted December 28, 2018, date of publication January 1, 2019, date of current version

January 23, 2019.

Digital Object Identifier 10.1109/ACCESS.2018.2890539

High-Precision Servo Control of Industrial

Robot Driven by PMSM-DTC Utilizing
Composite Active Vectors
TIANQING YUAN 1, DAZHI WANG1 , XINGHUA WANG2 , XINGYU WANG1 , AND ZHENAO SUN1
1 School of Information Science and Engineering, Northeastern University, Shenyang 110819, China
2 Physical Education Department, School of Humanities & Law, Northeastern University, Shenyang 110819, China

Corresponding author: Dazhi Wang (noblefuture@163.com)

This work was supported in part by the National Key R&D Program of China under Grant 2017YFB1300900, and in part by the National
Natural Science Foundation of China under Grant 61433004 and Grant 51467017.

ABSTRACT In this paper, a novel direct torque control (NDTC) scheme utilizing composite active
vectors based on discrete duty ratio modulation method is presented for permanent magnet synchronous
motor (PMSM) in industrial robot servo control system. The errors in the PMSM driven by conventional
DTC (CDTC) are compensated by single active vector during each control period, leading to unsatisfied
steady-state performance. To improve the operation performance of the industrial robot servo control system
effectively, the errors in the PMSM are compensated independently in NDTC. Therefore, two active vectors
should be applied to compensate flux error and torque error, respectively, namely flux active vector and
torque active vector. In order to maintain the merit of the fast dynamic response in CDTC, the introduced
active vectors should be selected from the fixed six active vectors supplied by a two-level voltage source
inverter. To describe the scheme of NDTC clearly, the angle between the applied active vector and the stator
flux linkage is used to represent the error compensational effect provided by the applied active vector, namely
active angle. Additionally, the compensational effect is defined as active factor. Subsequently, the duty ratio
value of the active vector applied time can be derived easily. To simplify the calculations of the duty ratio
values, novel discrete calculation method for the active factor by hysteresis controllers is presented in this
paper. The effectiveness of the proposed NDTC strategy is verified through the experimental results on a
100-W PMSM drive system.

INDEX TERMS Industrial robot, direct torque control (DTC), permanent magnet synchronous
motor (PMSM), composite active vectors, discrete duty ratio modulation.

I. INTRODUCTION DTC (CDTC), the PMSM suffers from some drawbacks,

In manufacturing systems, precision servo control has vital
importance, especially in industrial robot motion control sys-
tems. To improve the servo control precision at industrial
robot, permanent magnet synchronous motors (PMSM) have
been applied in the industrial robot widely [1]–[3]. Direct
torque control (DTC) and field-oriented control (FOC) are
two widely applied high-performance control strategies for
the PMSM. Different from decoupled analyzing method in
FOC, the torque and the flux linkage are controlled directly
in DTC, therefore, the quickest dynamic response can be
obtained in the PMSM driven by DTC [4]–[6]. However,
as only six active vectors can be selected to compen-
sate the errors of flux linkage and torque in conventional
such as large torque and flux linkage ripples, leading to
axial tracking errors in the coordination control of multiple
robots [7]. To improve the operation performance of the
PMSM in industrial robot, many researchers have attempted
to reduce these ripples by adding the amount of the active
vectors through different methods.
Employing more appropriate active vectors in each sam-
pling period, a novel DTC strategy based on three-level
inverter is proposed for permanent magnet synchronous
motor drive system [8], [9], therefore, the ripples of the sys-
tem can be suppressed effectively. In [10] and [11], a PMSM
DTC system based on matrix converter is proposed. Based on
the four enhanced switching table for the optimal selection of

2169-3536 2019 IEEE. Translations and content mining are permitted for academic research only.
VOLUME 7, 2019 Personal use is also permitted, but republication/redistribution requires IEEE permission. 7577
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
 T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors

the switching states, the ripples of torque and flux linkage can analyzed in Section II. The compensational effects of torque
be reduced effectively. Despite multiple active vectors can be error and flux linkage error in SVM-DTC system are also
provided by the three-level inverter or the matrix converter, illustrated in Section II. The scheme of the proposed NDTC
the cost of the system is inevitably increased. utilizing flux active vector and torque active vector during
As the errors of torque and flux linkage are minor actually each control period is described in Section III. The discrete
in most cases, the errors will be over-compensated if the calculation method for the active factors is also illustrated
active vector is applied during the whole control period. in Section III. The description of experimental setup and
To solve these problems, duty ratio modulation strategy is discussions on experimental results are given in Section IV.
introduced into the DTC-fed PMSM. Different duty ratio The conclusion is analyzed in section V.
modulation methods are studied in [12]–[16], and the rip-
ples can be reduced effectively without decreasing the fast II. PRINCIPLE OF THE PMSM DTC STRATEGIES
dynamic response in CDTC. USED IN INDUSTRIAL ROBOT
The stator flux linkage currents are decoupled in d-q axes A. PRINCIPLE OF THE CONVENTIONAL DTC
and controlled independently in FOC, hence, the outstanding In the PMSM DTC system driven by a two-level voltage
operating performance of the PMSM can be obtained eas- source inverter, eight voltage vectors can be applied to com-
ily. The decoupled-analyses method is adopted in the novel pensate the errors of torque and flux linkage, including six
DTC based on space vector modulation (SVM) strategy with active vectors Vn (n = 1, 2, 3, 4, 5, 6) and two null vectors
simple proportional-integral (PI) regulator or sliding mode (V0 and V7 ). The spatial placements of the six active vectors in
observer [17]–[25]. With the independent control of torque αβ-reference frames are shown in Fig. 1. The whole rotation
and flux linkage in SVM-DTC, the amplitude and the phase space of stator flux linkage ϕs can be divided into six sectors
of the required active vector can be determined accurately, through section boundary lines li (i = 1, 2, 3, 4, 5, 6),
and the errors of torque and flux linkage can be compensated as shown in Fig. 1. The six sectors are represented with
precisely. However, the introduced PI regulator or sliding number ‘‘N ’’, and the sector vector Vs represents the active
mode observer will degrade the dynamic response of the vector in every sector which the stator flux linkage ϕs is
system. located in.
To obtain the best operation performance of PMSM,
a novel DTC strategy (NDTC) utilizing composite active vec-
tors during each control period is presented in this paper. The
introduced flux linkage and torque active vectors are used to
compensate flux linkage error and torque error, respectively.
It should be noted that the active vectors will be selected from
the fixed six basic active vectors supplied by the conventional
two-level voltage-source inverter. Without using space vector
modulation strategy or multiple-level inverters, the merit of
fast dynamic response in CDTC can be maintained. It is worth
mentioning that the switching table for the selections of the
active vectors is redesigned without considering the divided
small sectors in [6].
In order to bring the novel DTC smoothly, active angle is
introduced to represent the angle between the active vector
and the stator flux linkage, and active vector is introduced
to represent the error compensational effect provided by the
active vector. The relationships between the active factors
and the active angles are analyzed in this paper. Duty ratio
modulation strategy is used to determine the applied time of
the active vector based on the introduced active factor. To sim-
plify the calculations, hysteresis controllers are adopted for FIGURE 1. Active vectors in DTC system.
the determination of the discrete duty ratio values in NDTC.
The effectiveness of the NDTC scheme is validated through
the experimental results. It should be noted that the steady- The torque error eT is obtained by the comparison between
state and the dynamic response performances of the PMSM the reference value Tref and the real-time value T . The hys-
driven by CDTC, DDTC and SVM-DTC are also studied in teresis comparator is used to determine the property εT of
this paper. torque error eT . The property εT value is 1 or −1, which indi-
The organization of this paper comprises the following sec- cates torque T needs to be increased if the value of property
tions. The principle of the conventional DTC and the single εT is 1, while the torque needs to be decreased if the value
active vector DTC based on duty ratio modulation strategy are of property εT is −1. The determination process of another

7578 VOLUME 7, 2019

T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors
TABLE 1. Conventional switching table.
FIGURE 2. Small sectors division of the rotation space.
parameter flux linkage is in the same way. The active vector
selection rules in SV-CDTC system are described in table 1.
B. ANALYSIS OF ERROR COMPENSATIONAL EFFECTS IN
S0 SECTOR IN DDTC SYSTEM
In [6], the compensational effects of torque error eT and flux
linkage error eF provided by different active vectors with
shifty active angle θ are described. The reclassified small
sector of the stator flux linkage rotation space that considered
the variation of the active angle θ , is shown in Fig. 2. In this
DTC strategy, the appropriate single active vector will be
selected from table 2 according to the small sector location
of the stator flux linkage and the property values of torque
error and flux linkage error. And then the duty ratio will be
determined through the active angle θ and used to control
the applied time of the active vector in each control period.
Therefore, this DTC strategy can also be called single active
vector duty ratio modulation DTC system (DDTC).
VOLUME 7, 2019
TABLE 2. Redesigned switching table for small sectors.
FIGURE 3. Errors of torque and flux linkage in α-β reference frame.
It should be noted that, the max compensation of torque
error eT 1 provided by the single active vector Vn is 1Tmax
during the entire control period Ts time, and the max com-
pensation of flux linkage error eF1 provided by the single
active vector Vn is 1ϕmax . As shown in Fig. 3, the stator flux
linkage ϕs is located in S0 small sector. If the torque error is
eT 1 , the flux linkage error is eF1 , and both properties of eT 1
and eF1 are 1, the active vector V1 will be selected and applied
to the control system in the next control period. If eF1 is lower
than 1ϕmax , while eT 1 is greater than 1Tmax , the duty ratio
d can be defined as
eF1
d = dF = <1 (1)
CF
where CF is the max compensation of flux linkage error
provided by the active vector Vn during the entire control
period Ts time, which can be given by
CF = 1ϕmax = Vn · Ts (2)
7579
 T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors
FIGURE 4. Errors of torque and flux linkage in d-q reference frame.
The real compensations of torque error eT 1 and flux link-
age error eF1 are
1T = CT · d · sin θ ≈ 0
1ϕ = CF · d · cos θ = eF1
where CT is the max compensation of torque error provided
by the active vector Vn during the entire control period Ts
time, which can be expressed as
3p
CT = 1Tmax = · ϕf · Vn · Ts (5)
2Ls
where p is the number of pole pairs, Ls is the stator induc-
tance, and ϕf is permanent magnet flux linkage.
It is obvious that V1 can supply the full compensation to
flux linkage error eF1 , while the compensation of torque error
eT 1 is nearly zero. From the variables in Fig. 3, it can be
observed that the torque error eT 1 can be compensated by
the active vectors V2 or V3 . The active vector V1 used to
compensate the flux linkage error eF is defined as flux linkage
active vector VF , and the active vector applied to compensate
the torque error eT is defined as torque active vector VT .
From the aforementioned analyses, it can be observed that
the key of the unsatisfied compensational effects of torque
error and flux linkage error in DDTC system is that single
active vector can be applied in each control period.
C. COMPENSATIONS OF TORQUE ERROR AND FLUX
LINKAGE ERROR IN SVM-DTC SYSTEM
The number and the direction of the active vectors are fixed
in CDTC and DDTC, therefore, the errors of torque and flux
linkage in the PMSM are difficult to be compensated effec-
tively in all conditions. To solve these problems, the decou-
pling control strategy used in FOC is introduced into DTC.
The PI controllers are used to obtain the amplitude of the
torque vector and the flux linkage vector based on torque error
and flux linkage error. And then, the space vector modula-
tion (SVM) is used to determine the precise vector.
As shown in Fig. 4, the stator flux linkage ϕs changes from
ϕs1 to ϕs2 during one control period. The variation of the
stator flux linkage ϕs is 1ϕs , which can be decoupled into
7580
1ϕsd and 1ϕsq in d-q axis. And then, a precise synthesized
vector which can compensate the error 1ϕs can be obtained.
From the above analyses, it can be found that the switch-
ing table in CDTC system is replaced by PI controllers in
SVM-DTC system. The active vectors used to compensate
the errors of torque and flux linkage are not limited to the
six basic active vectors in SVM-DTC system. And also,
the required steady-state performance of the PMSM can be
obtained easily.
III. NOVEL DTC UTILIZING TORQUE AND
FLUX LINKAGE VECTORS
In CDTC and DDTC, the torque error and the flux linkage
error are compensated by single active vector in every control
period, thus, the error compensational effect will be affected
by the active angle. Despite the precise of the wanted syn-
thetic vector can be ensured with the using of SVM strategy
in SVM-DTC, the dynamic response of the PMSM is affected
inevitably. Therefore, the errors of torque and flux linkage
(3)
should be compensated by different vectors independently,
(4) and the applied different active vectors should be selected
from the six basic active vectors without using SVM strategy.
FIGURE 5. Analysis of error compensational effects.
A. NOVEL DTC STRATEGY
In order to improve the performance of PMSM effectively,
the stator flux linkage error 1ϕs should be decoupled in d-q
axes firstly. As shown in Fig. 5, the error 1ϕs is decoupled
into eT and eF . It should be noted that the direction of eF is
the same as the direction of the stator flux linkage ϕs , and the
direction of eT is perpendicular to the direction of the stator
flux linkage ϕs .
It is obvious that eF can be compensated by the active
vector V1 , and eT can be compensated by the active vector V2 .
The active vector V1 used to compensate the flux linkage error
eF is defined as flux linkage vector VF , and the active vector
V2 used to compensate the torque error eT is defined as torque
VOLUME 7, 2019
T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors

FIGURE 6. Schematic diagram of the NDTC strategy for PMSM.

vector VT . The angle between the stator flux linkage ϕs and TABLE 3. Switching table for active vectors.
the flux linkage vector VF is defined as flux linkage angle θ,
and the angle between the stator flux linkage ϕs and the torque
vector VT is defined as torque angle δ.
The scheme diagram of the proposed NDTC utilizing both
flux linkage vector VF and torque vector VT in one control
period is shown in Fig. 6. It is worth mentioning that the
compensational effects of torque error eT and flux linkage
error eF provided by VT or VF can be determined based on
the active factors. It is also shown in Fig. 6 that torque vector
VT and flux linkage vector VF will be selected according to
the property values of torque error and flux linkage error, C. ANALYSIS OF ERROR COMPENSATIONS SUPPLIED BY
and the location Si of the stator flux linkage ϕs in the small DIFFERENT COMPOUND MODES OF ACTIVE VECTORS
sector. 1) MODE 1
In mode 1, the property values of both eT and eF are 1,
B. ACTIVE VECTOR SELECTION RULES therefore, the active vectors V1 and V2 should be selected
In Fig. 5, the stator flux linkage ϕs is located in N1 sector. as the flux linkage vector VF and the torque vector VT ,
It is obvious that the sector vector V1 is the nearest active respectively. The flux linkage angle and the torque angle are
vector to the flux linkage ϕs in the six basic active vectors. θ1 and δ1 , respectively, as shown in Fig. 7.
Therefore, V1 will be selected as flux linkage active vector The flux linkage error compensation 1ϕF1 and the torque
VF if the property value εF of flux linkage error is 1, while error compensation 1TF1 provided by VF can be calcu-
its opposite active vector V4 should be selected if the property lated as
value εF of flux linkage error is −1.
1ϕF1 = CF · cos θ1 (6)
In order to achieve the best error compensational effect,
the selection of the torque vector VT should be judged accord- 1TF1 = CT · sin θ1 (7)
ing to the compound modes of the flux linkage error property The flux linkage error compensation 1ϕT 1 and the torque
εF and the torque error property εT . error compensation 1TT 1 provided by VT can be described as
The selection rules of the torque active vector VT and the
flux linkage active vector VF for the proposed NDTC scheme 1ϕT 1 = CF · cos δ1 (8)
are shown in table 3. 1TT 1 = CT · sin δ1 (9)

VOLUME 7, 2019 7581

 T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors

4) MODE 4
In mode 4, the property values of both eT and eF are −1,
the active vectors V4 and V5 should be selected as the flux
linkage vector VF and the torque vector VT , respectively.
The flux linkage angle and the torque angle are θ4 and δ4 ,
respectively.
The compensations of flux linkage error eF and torque
error eT provided by flux linkage vector VF can be written as
1ϕF4 = CF · cos θ4 (18)
1TF4 = CT · sin θ4 (19)
The compensations of flux linkage error eF and torque
error eT provided by torque vector VT are
1ϕT 4 = CF · cos δ4 (20)
1TT 4 = CT · sin δ4 (21)
From above analyses, it can be found that the compensa-
FIGURE 7. Active angles (θ and δ) in different modes. tional effects of torque error eT and flux linkage error eF
provided by flux linkage vector VF or torque vector VT are
2) MODE 2 merely related to the active angle (θ or δ). Therefore, the com-
In mode 2, the property value εF of eF is 1 and the property pensational effects of flux linkage error eF and torque error eT
value εT of eT is −1, therefore, the active vectors V1 and V6 provided by flux linkage vector VF can be represented by flux
should be selected as the flux linkage vector VF and the torque linkage active factor λF and torque active factor µF , which
vector VT , respectively. The flux linkage angle and the torque can be described as
angle are θ2 and δ2 , respectively, as shown in Fig. 7. 
λF = cos θ
The compensations of eF and eT provided by flux linkage (22)
µF = sin θ
vector VF can be obtained as
And defining the compensational effects of flux linkage
1ϕF2 = CF · cos θ2 (10) error eF and torque error eT provided by torque vector VT
1TF2 = CT · sin θ2 (11) are flux linkage active factor λT and torque active factor µT ,
which can be expressed as
Although the torque angel δ2 is more than 180◦ , the com-
λT = cos δ

pensations of eF and eT provided by torque active vector VT (23)
can also be calculated as µT = sin δ

1ϕT 2 = CF · cos δ2 (12) Table 4 shows the variation range of the compensational
coefficients for different compound modes of active vectors.
1TT 2 = CT · sin δ2 (13) The compensations of flux linkage error eF and torque error
eT provided by flux linkage vector VF are
3) MODE 3
In mode 3, the property value εF of eF is −1 and the property 1ϕF = CF · λF (24)
value εT of eT is 1, therefore, the active vectors V4 and V3 1TF = CT · µT (25)
should be selected as the flux linkage vector VF and the torque
vector VT , respectively. The flux linkage angle and the torque The compensations of flux linkage error eF and torque
angle are θ3 and δ3 , respectively, as shown in Fig. 7. error eT provided by torque vector VT are
Although both the flux linkage angle θ3 and the torque 1ϕT = CF · λT (26)
angel δ3 are more than 90◦ , the compensations of eF and eT
provided by flux linkage vector VF can be evaluated by 1TT = CT · µT (27)

1ϕF3 = CF · cos θ3 (14) D. CONTINUOUS DUTY RATIO MODULATION

1TF3 = CT · sin θ3 (15) In each control period, the duty ratio values of flux linkage
vector VF and torque vector VT are dF and dT , respectively.
The compensations of flux linkage error and torque error Therefore, the relationships between the errors and the duty
provided by torque vector VT can also be defined as ratio values can be written in the following
1ϕT 3 = CF · cos δ3 (16) eF = 1ϕF · dF + 1ϕT · dT
1TT 3 = CT · sin δ3 (17) = CF · cos θ · dF + CF · cos δ · dT (28)

7582 VOLUME 7, 2019

T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors

eT = 1TF · dF + 1TT · dT
= CT · sin θ · dF + CT · sin δ · dT (29)
The duty ratio values of the flux linkage vector applied time
and the torque vector applied time can be evaluated by
CT · eF · µT − CF · eT · λT

 dF = CT · CF · µT · λF − CT · CF · λT · µF
 CT · eF · µT − CF · eT · λT
=

CTF (µT · λF − λT · µF )


 (30)

d = CT · eF · µF − CF · eT · λF
 T
CT · CF · µF · λT − CT · CF · λF · µT
CT · eF · µF − CF · eT · λF


=
CTF (µF · λT − λF · µT )
E. DISCRETE CALCULATION METHOD FIGURE 9. Discrete active vectors in S2+ small sector.
FOR ACTIVE FACTORS
It can be seen in (30), the determination of the active factors
is the key to obtain the duty ratio. Although the precise values
of active factors can be obtained by (22) and (23), the calcula-
tions are increased inevitably. To simplify the determination
process of the active factors, a novel discrete calculation
method based on hysteresis controllers for NDTC is proposed
in this section.
Since the compensational effects of flux linkage error eF
and torque error eT are merely related to the flux linkage
angle θ and the torque angle δ, the flux linkage location in
the divided small sectors can be used to determine the active
factors. Therefore, the compensational effect of torque vector
and flux linkage vector on torque error and flux error in
different small sectors can be described in the following.

FIGURE 10. Discrete active vectors in S2− small sector.

The active factors of eF and eT provided by the torque

vector VT are calculated as

 1
λT = cos δ ≈
 √2 (32)
3

µT = sin δ ≈
2
FIGURE 8. Discrete active vectors in S0 small sector.

2) SECTOR S2+
1) SECTOR S0
In S2+ small sector, the flux linkage angle θ and the torque
The sector range of S0 will be expanded if the flux linkage
angle δ can be fixed at 30◦ , as shown in Fig. 9.
error eF is much greater than the torque error eT [6]. As the
The active factors of eF and eT provided by the flux linkage
flux linkage angle θ is tiny, the compensation of torque
vector VF can be obtained as
error eT provided by flux linkage vector VF can be ignored,
as shown in Fig. 8. Therefore, the active factors of eF and eT  √
provided by the flux linkage vector VF are 3
 λF = cos θ ≈ 2

λF = cos θ ≈ 1 (33)
(31)
 1
µF = sin θ ≈ 0 µF = sin θ ≈ −
2
VOLUME 7, 2019 7583
 T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors

FIGURE 11. Discrete calculation method for active factors.

FIGURE 12. Experimental setup of control system.

FIGURE 14. The speed trajectory from 200 rpm to 400 rpm using:
(a) CDTC, (b) DDTC, (c) SVM-DTC, and (d) NDTC.

3) SECTOR S2−
FIGURE 13. Comparison of transient responses of CDTC, DDTC, SVM-DTC, In S2− small sector, the flux linkage angle θ and the torque
and NDTC. angle δ can be fixed at 30◦ and 90◦ , respectively, as shown
The active factors of eF and eT provided by the torque in Fig. 10. The active factors of eF and eT provided by the
vector VT can be expressed as flux linkage vector VF can be given as
 √ √
3 
 λT = cos δ ≈ 2
3
(34)  λF = cos θ ≈ 2

1 (35)
µT = sin δ ≈ 1

2 µF = sin θ ≈
2
7584 VOLUME 7, 2019
T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors
FIGURE 15. Experimental torque and flux linkage of the PMSM when using: (a) CDTC, (b) DDTC, (c) SVM-DTC, and (d) NDTC.
The active factors of eF and eT provided by the torque
vector VT are
λT = cos δ ≈ 0

(36)
µT = sin δ ≈ 1
It can be found that the active factors will vary among
the setting values while the stator flux linkage is located in
S1+ or S1− . To simplify the calculations of the active factors
when the stator flux linkage ϕs is located in S1+ or S1− ,
a discrete active factor determination method is proposed.
The schematic diagram of the proposed discrete factor deter-
mination method is shown in Fig. 11. The active angle σ
between the selected active vector and the stator flux linkage
is separated by a two-level hysteresis controller.
It should be noted that, the small sector S1+ will be
divided into S1++ and S1+− , and the small sector S1− will be
divided into S1−+ and S1−− . In the four divided small sector,
the active factors of flux linkage vector VF or torque vector
VT are viewed as the same value. The factor switching table
in mode 1 is shown in table 5.
IV. EXPERIMENTAL RESULTS
A. EXPERIMENTAL SYSTEM SETUP
Experimental studies are carried out on a 100-W PMSM
drive system to validate the feasibility and effectiveness of the
proposed NDTC strategy. The experimental hardware setup is
illustrated in Fig. 12. The parameters of the PMSM are given
as follows: Rs = 0.76 ; Ls = 0.00182 H ; the number of pole
pairs p = 4. The DC voltage is 36 V . This study compares
the steady-state and the dynamic response performance of
CDTC, DDTC, SVM-DTC, and NDTC. The experiments are
implemented in a TMS320F28335 DSP control system with
a sampling period of 100 µs.
VOLUME 7, 2019
B. DYNAMIC PERFORMANCE
To validate the fast dynamic response of the proposed NDTC,
the torque command is changed from 0 to 0.5 N·m when the
rotor speed is 600 rpm. The dynamic torque responses of the
PMSM driven by CDTC, DDTC, and SVM-DTC are also
tested. The torque responses of the four control strategies are
compared in Fig. 13. It can be found that the torque increases
to the setting value within 0.023 s when CDTC is used, and
the settling time of the PMSM driven by DDTC, SVM-DTC
and NDTC are 0.022, 0.025, and 0.023 s, respectively.
The speed responses of the PMSM driven by the four
control strategies are also tested when the torque is set as
0.5 N·m. In these tests, a step change from 200 to 400 rpm is
applied on the speed reference, as shown in Fig. 14. It can be
seen that the ripples of the rotor speed are 35 rpm when using
CDTC, while the speed ripples of the PMSM can be reduced
to 30, 25, and 29 rpm with the using of DDTC, SVM-DTC,
and NDTC. Moreover, the settling time of the rotor speed
using the four different control strategies are 0.013, 0.012,
0.019, and 0.012 s.
Therefore, the main advantage of CDTC, i.e., the fast
dynamic response, is maintained in NDTC. To obtain the
satisfied dynamic response, the dividing criterion of duty
ratio discrete values can be adjusted appropriately.
C. STEADY-STATE PERFORMANCE
The steady-state performances of CDTC, DDTC, SVM-DTC,
and NDTC are compared under the same operating con-
ditions. The PMSM is operated at 500 rpm and the refer-
ence values of torque and flux linkage are 0.8 N·m and 0.3
Wb, respectively. The torque and flux linkage waveforms of
the PMSM driven by different control strategies are shown
in Fig. 15.
7585
 T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors
Experimental results show that the torque ripples of the
PMSM driven by CDTC, DDTC, SVM-DTC, and NDTC are
0.56, 0.4, 0.32, and 0.36 N·m, respectively, and the flux rip-
ples of the PMSM are 0.08, 0.06, 0.04, and 0.035 Wb, respec-
tively. Consequently, compared with CDTC, DDTC and
SVM-DTC can suppress torque ripple by at least 27% and
42%, respectively, and suppress flux ripple at least 25%
and 50%, respectively. The steady-state performances of the
PMSM driven by NDTC in the same setting operation condi-
tion are nearly the same as SVM-DTC.
V. CONCLUSION
This paper proposes a novel PMSM DTC scheme for the
industrial robot servo control system. The errors in the PMSM
are analyzed independently, and compensated by the intro-
duced torque active vector and flux linkage active vector,
respectively. To obtain the satisfied operation performance
of the industrial robot, the torque active vector and the flux
linkage active vector will be selected from the redesigned
switching table and applied the appropriate time based on the
duty ratio values. The proposed discrete calculation method
for the active factor in this paper which ensures the fast
dynamic response in NDTC can be obtained easily. And the
steady-state performance of the PMSM driven by NDTC can
be improved to a certain level based on appropriate dividing
criterion of active factors discrete values.
Experimental results clearly indicate that the PMSM driven
by NDTC scheme exhibits excellent control of torque and
flux linkage with lower steady-state ripples, and faster
transient response performance when compared to CDTC,
DDTC, and SVM-DTC.
REFERENCES
[1] M. A. M. Cheema, J. E. Fletcher, D. Xiao, and M. F. Rahman, ‘‘A direct
thrust control scheme for linear permanent magnet synchronous motor
based on online duty ratio control,’’ IEEE Trans. Power Electron., vol. 31,
no. 6, pp. 4416–4428, Jun. 2016.
[2] A. H. Abosh, Z. Zhu, and Y. Ren, ‘‘Reduction of torque and flux ripples
in space vector modulation-based direct torque control of asymmetric
permanent magnet synchronous machine,’’ IEEE Trans. Power Electron.,
vol. 32, no. 4, pp. 2976–2986, Apr. 2017.
[3] L. Tang, L. Zhong, M. F. Rahman, and Y. Hu, ‘‘A novel direct torque
controlled interior permanent magnet synchronous machine drive with low
ripple in flux and torque and fixed switching frequency,’’ IEEE Trans.
Power Electron., vol. 19, no. 2, pp. 346–354, Mar. 2004.
[4] A. Shinohara, Y. Inoue, S. Morimoto, and M. Sanada, ‘‘Direct calculation
method of reference flux linkage for maximum torque per ampere control
in DTC-based IPMSM drives,’’ IEEE Trans. Power Electron., vol. 32, no. 3,
pp. 2114–2122, Mar. 2016.
[5] I. M. Alsofyani, N. R. N. Idris, and K.-B. Lee, ‘‘Dynamic hysteresis
torque band for improving the performance of lookup-table-based DTC
of induction machines,’’ IEEE Trans. Power Electron., vol. 33, no. 9,
pp. 7959–7970, Sep. 2018.
[6] T. Yuan et al., ‘‘Duty ratio modulation strategy to minimize torque and
flux linkage ripples in IPMSM DTC systems,’’ IEEE Access, vol. 5,
pp. 14323–14332, Aug. 2017.
[7] L. Jin, S. Li, H. M. La, X. Zhang, and B. Hu, ‘‘Dynamic task alloca-
tion in multi-robot coordination for moving target tracking: A distributed
approach,’’ Automatica, vol. 100, pp. 75–81, Feb. 2019.
[8] Y. N. Tatte, M. V. Aware, J. K. Pandit, and R. Nemade, ‘‘Performance
improvement of three-level five-phase inverter-fed DTC-controlled five-
phase induction motor during low-speed operation,’’ IEEE Trans. Ind.
Appl., vol. 54, no. 3, pp. 2349–2357, May/Jun. 2018.
7586
[9] S. Payami, R. K. Behera, and A. Iqbal, ‘‘DTC of three-level NPC
inverter fed five-phase induction motor drive with novel neutral point
voltage balancing scheme,’’ IEEE Trans. Power Electron., vol. 33, no. 2,
pp. 1487–1500, Feb. 2018.
[10] C. Xia, J. Zhao, Y. Yan, and T. Shi, ‘‘A novel direct torque and flux
control method of matrix converter-fed PMSM drives,’’ IEEE Trans. Power
Electron., vol. 29, no. 10, pp. 5417–5430, Oct. 2014.
[11] Y. Yan, J. Zhao, C. Xia, and T. Shi, ‘‘Direct torque control of matrix
converter-fed permanent magnet synchronous motor drives based on mas-
ter and slave vectors,’’ IET Power Electron., vol. 8, no. 2, pp. 288–296,
Feb. 2015.
[12] D. Mohan, X. Zhang, and G. H. B. Foo, ‘‘A simple duty cycle control
strategy to reduce torque ripples and improve low-speed performance of
a three-level inverter fed DTC IPMSM drive,’’ IEEE Trans. Ind. Electron.,
vol. 64, no. 4, pp. 2709–2721, Apr. 2017.
[13] F. Niu, B. Wang, A. S. Babel, K. Li, and E. G. Strangas, ‘‘Compara-
tive evaluation of direct torque control strategies for permanent magnet
synchronous machines,’’ IEEE Trans. Power Electron., vol. 31, no. 2,
pp. 1408–1424, Feb. 2016.
[14] Y. Zhang, J. Zhu, W. Xu, and Y. Guo, ‘‘A simple method to reduce torque
ripple in direct torque-controlled permanent-magnet synchronous motor
by using vectors with variable amplitude and angle,’’ IEEE Trans. Ind.
Electron., vol. 58, no. 7, pp. 2848–2859, Jul. 2011.
[15] Y. Ren, Z. Q. Zhu, and J. Liu, ‘‘Direct torque control of permanent-magnet
synchronous machine drives with a simple duty ratio regulator,’’ IEEE
Trans. Ind. Electron., vol. 61, no. 10, pp. 5249–5258, Oct. 2014.
[16] F. Niu, K. Li, and Y. Wang, ‘‘Direct torque control for permanent-magnet
synchronous machines based on duty ratio modulation,’’ IEEE Trans. Ind.
Electron., vol. 62, no. 10, pp. 6160–6170, Oct. 2015.
[17] D. Mohan, X. Zhang, and G. H. B. Foo, ‘‘Generalized DTC strategy for
multilevel inverter fed IPMSMs with constant inverter switching frequency
and reduced torque ripples,’’ IEEE Trans. Energy Convers., vol. 32, no. 3,
pp. 1031–1041, Sep. 2017.
[18] Z. Zhang, Y. Zhao, W. Qiao, and L. Qu, ‘‘A space-vector-modulated
sensorless direct-torque control for direct-drive PMSG wind turbines,’’
IEEE Trans. Ind. Appl., vol. 50, no. 4, pp. 2331–2341, Jul./Aug. 2014.
[19] X. Zhang and G. H. B. Foo, ‘‘Overmodulation of constant-switching-
frequency-based DTC for reluctance synchronous motors incorporating
field-weakening operation,’’ IEEE Trans. Ind. Electron., vol. 66, no. 1,
pp. 37–47, Jan. 2019.
[20] A. Berzoy, J. Rengifo, and O. Mohammed, ‘‘Fuzzy predictive DTC of
induction machines with reduced torque ripple and high-performance
operation,’’ IEEE Trans. Power Electron., vol. 33, no. 3, pp. 2580–2587,
Mar. 2018.
[21] Y. Zhou and G. Chen, ‘‘Predictive DTC strategy with fault-tolerant function
for six-phase and three-phase PMSM series-connected drive system,’’
IEEE Trans. Ind. Electron., vol. 65, no. 11, pp. 9101–9112, Nov. 2018.
[22] Y.-S. Choi, H. H. Choi, and J.-W. Jung, ‘‘Feedback linearization direct
torque control with reduced torque and flux ripples for IPMSM drives,’’
IEEE Trans. Power Electron., vol. 31, no. 5, pp. 3728–3737, May 2016.
[23] Z. Zhang, C. Wei, W. Qiao, and L. Qu, ‘‘Adaptive saturation controller-
based direct torque control for permanent-magnet synchronous machines,’’
IEEE Trans. Power Electron., vol. 31, no. 10, pp. 7112–7122, Oct. 2016.
[24] Y. Zhang and J. Zhu, ‘‘A novel duty cycle control strategy to reduce both
torque and flux ripples for DTC of permanent magnet synchronous motor
drives with switching frequency reduction,’’ IEEE Trans. Power Electron.,
vol. 26, no. 10, pp. 3055–3067, Oct. 2011.
[25] L. Jin, S. Li, B. Hu, and J. Yu, ‘‘Noise-suppressing neural algorithm
for solving time-varying system of linear equations: A control-based
approach,’’ IEEE Trans. Ind. Informat., vol. 15, no. 1, pp. 236–246,
Jan. 2019.
TIANQING YUAN received the B.S. degree
in automation from the Civil Aviation Univer-
sity of China, Tianjin, China, in 2011, and the
M.S. degree in electrical engineering from the
Shenyang University of Technology, Shenyang,
China, in 2016. He is currently pursuing the Ph.D.
degree with Northeastern University, Shenyang.
His research interest includes permanent mag-
net motor drives, power electronics, and high per-
formance control strategy.
VOLUME 7, 2019
T. Yuan et al.: High-Precision Servo Control of Industrial Robot Driven by PMSM-DTC Utilizing Composite Active Vectors
DAZHI WANG received the B.S. degree in
automation from the Southeastern University of
China, Jiangsu, China, in 1985, the M.S. degree
in automation from Shenyang Ligong Univer-
sity, Liaoning, China, in 1992, and the Ph.D.
degree in control theory and control engineering
from Northeastern University, Shenyang, China,
in 2003.
He is currently a Full Professor and the Head
of the School of Information Science and Engi-
neering, Institute of Power System and Drives, Northeastern University.
His current research interests include renewable energy generation system,
power electronics, power quality control, and motor drives.
XINGHUA WANG received the B.S. degree in
social sport from Beijing Sport University, Beijing,
China, in 2014, and the M.S. degree in leisure
and sport management from Hong Kong Baptist
University, Hong Kong, in 2015. She is currently
pursuing the Ph.D. degree in education economy
and management with Northeastern University,
Shenyang, China.
She is currently a Lecturer of the P.E. Depart-
ment, Northeastern University. Her research inter-
ests include sport management, sport marketing, and public sport service.
VOLUME 7, 2019
XINGYU WANG received the B.S. degree in elec-
trical engineering and automation and the M.S.
degree in electrical engineering from Liaoning
Technical University, Liaoning, China, in 2014 and
2017, respectively. He is currently pursuing the
Ph.D. degree in power electronics and power
drives with Northeastern University, Liaoning,
China.
His research interests include sliding electrical
contact and basic theory of electrical equipment.
ZHENAO SUN received the B.S. degree from Bei-
jing Jiaotong University, Beijing, China, in 2011,
and the Ph.D. degree from the University of Chi-
nese Academy of Sciences, Beijing, in 2016. He is
currently with Northeastern University, Shenyang,
China.
His current research interest includes renewable
energy power generation technology.
7587