Design of Robust Control for a Motor in Electric Vehicles

Abstract

In this study, robust control methods are employed to control the permanent magnet synchronous motor. Features such as good performance despite the vehicle's load torque disturbance, measurement noise, system parameter changes and high‐frequency uncertainties of structured and unstructured types make this approach more effective. The control objective is controlling and maintaining the vehicle speed and motor torque in driver's desired references. In this study, a μ controller and a reduced order μ controller are designed for the first time due to simplicity and easy implementation of the optimal H∞ controller. Simulation studies are carried out in this study to evaluate the control performance of the designed control systems. Comparison studies show that the proposed controllers have several advantages compared to the proportional–integral–derivative controller including reference speed and torque tracking, mean square error, disturbance rejection and noise attenuation.

Full text

IET Electrical Systems in Transportation
Research Article
Design of robust control for a motor in
electric vehicles
ISSN 2042-9738
Received on 12th November 2018
Revised 19th June 2019
Accepted on 15th August 2019
E-First on 5th September 2019
doi: 10.1049/iet-est.2018.5084
www.ietdl.org
Tahere Pourseif1 , Mina Mohajeri1
1Electrical and Engineering Department, Shahid Beheshti University, Tehran, Iran
E-mail: t.pourseif@gmail.com
Abstract: In this study, robust control methods are employed to control the permanent magnet synchronous motor. Features
such as good performance despite the vehicle's load torque disturbance, measurement noise, system parameter changes and
high-frequency uncertainties of structured and unstructured types make this approach more effective. The control objective is
controlling and maintaining the vehicle speed and motor torque in driver's desired references. In this study, a μ controller and a
reduced order μ controller are designed for the first time due to simplicity and easy implementation of the optimal H∞ controller.
Simulation studies are carried out in this study to evaluate the control performance of the designed control systems.
Comparison studies show that the proposed controllers have several advantages compared to the proportional–integral–
derivative controller including reference speed and torque tracking, mean square error, disturbance rejection and noise
attenuation.
1 Introduction
Replacing gasoline cars with electric cars has gained importance in
recent years due to increasing environmental pollutants and global
warming. One of the main components of a vehicle, which requires
precise, quick and accurate control, is the electric motor.
Development of electric vehicles (EVs) requires high power
density drive and high performance energy management. Different
electric motor systems have been used to propel these vehicles. The
most popular traction motors for this application are induction
motors (IMs) and permanent magnet synchronous motors
(PMSMs) [1–3]. PMSM drive plays a significant role in speed and
torque control applications due to recent developments in
permanent magnet materials, semiconductor power devices and
control theories. PMSMs are widely used in high-performance
applications such as EV, hybrid EV (HEV), industrial robots and
machine tools because of their high-power density, compact size,
high air gap flux density, high torque capability, high-torque to
inertia ratio and high efficiency [2–4]. Furthermore, they are used
because of lack of rotor windings, low no-load current under
nominal speed [5–7], low maintenance and high reliability in
comparison to other types of motors especially IMs. The overall
performance of a controller for PMSM drive depends not only on
the quickness and the precision of the system response, but also on
the robustness of the control strategy. PMSM can obtain higher
power density, efficiency and smaller volume in comparison to IM.
Therefore, PMSMs have gradually become popular in high
performance drives such as traction of EV. Both types of PMSM
including surface-mounted permanent magnet and interior
permanent magnet are used in EV application [8]. A comparison
between these two types of PMSM has been made in [9].
The standard proportional–integral–derivative (PID) controllers
are usually used in order to control the speed of electrical motors
because of their simplicity and practicality [10, 11]. Due to
physical considerations, some limitations applied in electric drives.
These constraints deteriorate the controller's performance which is
referred to the integrator windup increase system overshoot and
settling time. In addition, this control scheme requires accurate
information regarding the operation conditions and system
parameters. The control performance cannot be maintained when
system parameters change or disturbance exists. Some researchers
have used combined control methods such as fuzzy PID scheme
[12] for real-time adjustment of controller parameters. It should be
mentioned that these methods adopt the conventional fuzzy
approach, which does not require any mathematical model and can
achieve a suitable control performance. However, the most
significant problem is the difficulty in analysing the stability of the
controller. To resolve these problems, some other control methods
are proposed in [13]. In [14, 15], Takagi–Sugeno (T–S) fuzzy
models are used to design a speed controller for PMSM drive
systems. However, the system parameter uncertainties as well as
the external noises are not addressed in [14]. Furthermore, the
robustness of the uncertain PMSM is verified using the T–S fuzzy
model, but they only reflect simulation results and the effects of the
unknown external noises are not considered. In [15], the effects of
the parameter uncertainties and the unknown external noises are
considered and the T–S fuzzy controller is designed. Therefore, the
T–S fuzzy control system is robust but its algorithm is quite
complicated and time consuming.
Using adaptive control method and combining it with other
controllers such as conventional PID, one can adapt controller
parameters with the changes in the process condition and
consequently, desired performance was achieved in [16–18]. In
[17], extended Kalman filter has been applied to estimate the whole
state vector of the plant, however, this requires intensive
computing. Adaptive neural network-based methods are desirable
when the system parameter changing rate is slow. There are some
parameters in PMSM such as stator resistance, inductance, rotor
flux and load torque which may change fast [18]. With fast
changing in parameters some of these methods become very
complicated and their implementation turns difficult [17–19].
Many researchers have introduced model predictive control
(MPC) as one of the most robust control techniques. It is based on
the optimisation of the cost function depending on the difference
between the output and the tracking trajectory. This method
improves insensitivity to parameter uncertainty and external
disturbances and handles the state and control constraints. In [20],
MPC is applied to PMSM which is fed by matrix converter. Here,
the load torque is considered as a disturbance and is estimated by
Kalman filter. In this technique, the determination of the optimal
control requires solving the online non-linear optimisation problem
as the main drawback of this method that limits its application.
Another disadvantage of this method is its model dependency.
Considering uncertainties in the PMSM system, load torque
disturbances and external noises, achieving robust performance
(RP) is a troublesome issue.
IET Electr. Syst. Transp., 2020, Vol. 10 Iss. 1, pp. 68-74
© The Institution of Engineering and Technology 2019
68