Module 3 notes

Course Name: AI Techniques for Electric and Hybrid Vehicles

(Professional Elective)

Course Code: 18EE743 Year of Study: 2023-24

Class: VII Semester Faculty Name: Dr. Venkatesha K
Course Name: AI Techniques for Electric and Hybrid Vehicles (Professional Elective)

Module 3: Optimization Techniques used in Active Magnetic Bearing AMB system for EVs

Introduction: The primary source of pollution are combustion of liquid fuels in

Industries or in transportation in which percentage of pollution is more from
transportation.

The brilliant & effective option for controlling pollution is use of EVs in place of
conventional one. In Evs, active magnetic Bearing [AMB] is widely used, using which
high speed with efficient performance can be achieved.
The major issue is from the control perspective as AMB are highly non-linear and
unstable, and the classical control approach alone can’t give efficient results. So the
AI based control approach such as
(1) ANN based control
(2) Fuzzy logic control
(3) Particle swarm optimization control (PSOC) etc. along with classical controller
can increase the reliability and performance of the overall system.
As the AMBs are used in most of the EV applications, their control like speed control
& torque control can be smoothly achieved by the use of AI based control techniques
1. Compare micro, mild, full hybrid, PHEV, renewable energy and battery electric vehicles
on the basis of their source and propulsion systems.

There are many types of EVs classified on the basis of their source and propulsion
systems. Table-1 shows classification of various EVs. Advancement in research took
the EV to a new level where, not one but two or more sources are used to make the
propulsion. These types of EVs are defined as hybrid electric vehicles [HEVs]

Note: Drive Train: group of components in a vehicle that generate power and deliver it
to the road surface, water or air.
Hybrid drive train: vehicle that has two or more power train connected for propulsion
are called hybrid drive trains]
2. List the advantages of conventional vehicles over electric vehicles

Advantages & disadvantages of Conventional vehicles (ICEV) over Electric Vehicle (EV)
Advantages:
(1) The requirement of initial torque is more for any kind of vehicle and
conventional vehicles are good in this.
(2) Heavy vehicles such as cranes, trucks, lifters etc. still use conventional engines
as they need more power and high initial torque.
(3) Conventional vehicle engine life is better than the battery life because
continuous charging & discharging of battery results in small battery life.
(4) EVs have limited driving range with a high initial cost. Conventional vehicles
during range is extendable as more petrol bunks are available across country.

Disadvantages

(1) Pollution from conventional vehicles is too much as compared to Full electric
vehicle or Battery Electric Vehicle (FEV/BEV) as they are mostly powered from
Renewable Energy Sources (RES)
(2) Reduction in carbon emission & nitrogen oxide (NO) gas with the use of EVs.
Whereas conventional vehicle produces 2.3 Kg for every 1 litre of petrol
consumption.
3. What are magnetic bearings? Compare active magnetic bearing with passive magnetic
bearing and also list the applications of Active magnetic bearing
Magnetic Bearing:
Conventional Bearing: A wheel bearing is a crucial part of the wheel assembly that
connects the wheel & the axle. It comprises inner metallic ring, outer metallic ring &
steel balls or roller balls in between them. It enables the wheel to rotate smoothly with
a minimum of friction.

Fig. 1 Ball Bearing [Low and medium duty] Fig.2 roller bearing [heavy duty]
A magnetic bearing is one in which there is no contact between the stationary part &
moving part of the bearing by the use of concept magnetic levitation.
Depending upon the excitation used, magnetic bearing is classified into two types
(1) Active Magnetic bearing (AMB)
(2) Passive Magnetic bearing (PMB)
Active magnetic bearing is one in which the required magnetic field is generated by
supply current in the coil of magnet & in passive magnetic bearing, the required
magnetic force is generated by a permanent magnet.
 Comparison between AMB and PMB
(1) AMB with electromagnets lasts for longer period & as long as power is fed to the
electromagnets till then behaves as an Active magnetic Bearing. But continuous use
of passive magnetic bearing will degrade the performance of magnetic properties
and overall performance of the system.
(2) Supply is always required for AMB for proper bearing action but in passive
magnetic bearing, permanent magnet will generate the required force.
(3) Heat, temperature, moist and other atmospheric factors affect the performance of
the passive magnetic bearing but active magnetic bearing totally depends on the
current flow in the coil.
(4) Passive magnetic bearing has limitations as far as magnetic field is considered. But
active magnetic bearing depends on the magnitude of the current flow in the coil.
(5) The speed/torque using PMB is lesser as compared to AMB.
(6) Closed loop operation is possible only with AMB, but PMB controlling will always be
a problem at a closed control loop operation.
Therefore AMB is preferred over PMB for bearing operations.

Disadvantages of AMB
(1) Larger in size (2) Bulky (3) costly
Applications
(1) Transportation (2) Flywheel energy storage system (FESS)
(3) High speed tools (4) Watt-hour meter (5) Artificial hearts
(6) Centrifugal compressor
4. Discuss the basic requirement of EV design
Basic requirement of EV design

(1) High inertial torque & power density

(2) High efficiency at different torque & speed range
(3) High torque capability to climb hills
(4) Wide torque and speed range
(5) Long range
(6) Even for constant speed & torque operation, it should be efficient
To achieve the above stated requirement of Evs, AMB is best suited for bearing
operations.
5. With the help of basic and proposed block diagram of active magnetic bearing, explain the
various components used in closed loop system
Basic components of an Active Magnetic Bearing (AMB)

Fig.1 Basic block diagram of an AMB with basic components

Open loop AMB operation is non-linear & highly unstable in nature. To obtain proper
bearing operation the very first and basic step is levitation of the rotor at a particular
airgap & maintain it without any vibrations or oscillations.
The block diagram of an AMB with basic components is shown in Fig. 1
(1) Electromagnetic actuator: it has a magnetic core which is I, E U shaped iron core
with a copper winding in it as shown. Iron core is used because it provides low core
losses, low hysteresis and high permeability. For winding, copper is used, which is
stronger than aluminium.
 Fig.2 Proposed Closed loop AMB model

When excitation is given to these windings, the actuator becomes an electromagnet and
an increment in excitation increase the magnetically attractive power of this
electromagnet.
(2) Rotor: The attractive force of the electromagnet actuator acts on the rotor, that
compensates the gravity acting on it and results in levitation of the rotor. For the
proposed model of AMB, the rotor is made of ferromagnetic material and shaped like a
ball.
𝜇𝑟 ≫ 1 ⇒ 𝐹𝑒𝑟𝑟𝑜𝑚𝑎𝑔𝑛𝑒𝑡𝑖𝑐 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑒𝑔. 𝐼𝑟𝑜𝑛, 𝑐𝑜𝑏𝑎𝑙𝑡, 𝑛𝑖𝑐𝑘𝑒𝑙, 𝑓𝑒𝑟𝑟𝑖𝑡𝑒𝑠, 𝑠𝑖𝑙𝑖𝑐𝑜𝑛 𝑠𝑡𝑒𝑒𝑙
𝜇𝑟 > 1 ⇒ 𝑝𝑎𝑟𝑎𝑚𝑎𝑔𝑛𝑒𝑡𝑖𝑐 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑒𝑔. 𝑀𝑎𝑔𝑛𝑒𝑠𝑖𝑢𝑚, 𝑚𝑜𝑙𝑦𝑏𝑑𝑒𝑛𝑢𝑚
𝜇𝑟 < 1 ⇒ 𝑑𝑖𝑎𝑜𝑚𝑎𝑔𝑛𝑒𝑡𝑖𝑐 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑒𝑔. 𝑐𝑜𝑝𝑝𝑒𝑟, 𝑠𝑖𝑙𝑣𝑒𝑟 𝑎𝑙𝑢𝑚𝑖𝑛𝑖𝑢𝑚 𝑔𝑜𝑙𝑑
(3) Controller: is requirement for smooth & proper operation of AMB. The purpose
of the controller is to properly control the position of the rotor. Closed loop AMB
model has two controller. One control the position of the rotor which is labelled as
position controller. & second is to control the current in the coil of electromagnetic
actuator labelled as current controller.
a. Position controller: The position signal coming from position sensor is
compared with the reference position & error signal is given as input to position
controller. The position controller produces output depends upon the
calibration & setup value of controlling variables, which is further fed for
comparison from current sensor signal.
Position controller can be a classical controller like PI, PID, lead lag controller
etc. and in modern controller, they can be FLC, GA based PID controller.
b. Current controller: The current in the coil of electromagnet actuator is measured
by current sensor and that signal is compared with the output of the position
controller, the compared output is given as input to current controller which
generates output depending upon design & controlling variables, further the
output is fed to the power amplifier. Current controller is designed using PI
controller as they are suitable for faster & better transient state performance.
(4) Sensors
(a) Position sensor: It senses the real time position of the rotor & generates a signal
proportional to it which further sends for comparison to the reference position
signal.
Position sensor used could be (i) Induction type (ii) capacitance type
(iii) Laser type (iv) IR type
(b) Current sensor: senses the current in the coil and generates a signal analogous to
it, that signal is later sent to comparison with the output of position controller
eg. Hall effect sensor

(5) Power Amplifier [PA]: PAs are used to amplify the input current fed to the coil. In
PA, the switching devices used is mostly MOSFET & IGBT, which are costly. Hence
a single switch PA is used to make the whole system cost-effective.
The working can be explained by dividing this whole block diagram into two closed
loops.
(i) Outer closed loop tries to make the position of the rotor at the reference position
value with position controller
(ii) Inner closed loop comprising current controller, power amplifier and current
sensor in a feedback path. The inner closed loop tries to make the current
adjustment in the electromagnetic coil such that the gravitational force acting on
the ball is totally compensated by the attractive force of electromagnet at the
reference position. That is Fg=Fem -> (1)
6. With the help of block diagram, explain flywheel energy storage system using Active
Magnetic bearing for EVs
Active magnetic bearing for flywheel energy storage system in EV system

Most of the vehicular applications require electrical energy sources (Battery) & storage
system (FESS) in which rotational operations are performed by the active magnetic
bearings. Losses which occur with the use of conventional bearing are one of the reason
for reduced efficiency of EVs, but with the help of AMB, that efficiency will be improved.
Energy stores in a flywheel energy storage system (FESS) is in the form of kinetic energy
depends on a turning mass rotating by an electric machine. As the kinetic energy is directly
proportional to the square of the rotor speed, more the speed of the rotor, will generate
1
more energy (𝐾𝐸 = 𝑚𝑣 2 𝑗𝑜𝑢𝑙𝑒𝑠)
2
A schematic diagram of a FESS is shown in figure in which EVs works as a motor
during charging mechanical energy to FESS & as a generator along with FESS during
discharging mechanical energy from FESS.
The casing is provided to create a vacuum, and the output of the energy will be either
directly fed to the machine or after converting it into DC its fed to batteries.

An important component of FESS is the electric machine, which should be such that,
it will fulfill the robust requirements: high efficiency, higher power density and a
wide speed range. So, the implementation of AMB in FESS improves overall system
performance.
7. Explain the various blocks of flow diagram of Fuzzy Logic Controller and hence discuss the
proposed AMB with FLC as a position controller for EVs
Control strategies of AMB for EV systems
(1) Fuzzy logic controller
The complete FLC working can be classified as
1. Fuzzification
2. Fuzzy Inference process
3. Defuzzification
First the applied crisp data input is fuzzified in fuzzy subset by fuzzification process.
Depending upon the designed rule base, inference process is applied to fuzzified input.
Later the output of the inference process is converted into crisp or Analog type data by
Defuzzification process.
1. Fuzzification: The process of generating a membership value () for a fuzzy variable
using a membership function.

Fig. Flow diagram of Fuzzy Logic Control

(FLC)
(2) Fuzzy Inference Process: The rules are designed using IF-THEN statements.
Depending upon these rules, output is generated which is further sent to Defuzzification.
(3) Defuzzification: The process of transforming a fuzzy output of a fuzzy system into a
crisp output. The available methods are
(i) Centroid of Sum (ii) Mean of Maxima (MoM)
The fuzzy logic tool available in MATLAB which can be used for designing the controller
for the proposed AMB model.
Input variables and output variables can be defined from the MFs which are triangular,
trapezoidal, sigmoidal etc.

Fig. Simulation of proposed AMB with FLC as a position controller for EVs

After creating the FLC, it can be used as a position controller for the proposed AMB
model as it is a modern controller so there is a no tuning of control variables is required
and a little disturbance in the system can be easily controlled by the FLC.
8. With the help of flow chart, working and implementation of ANN, explain ANN-PID
controller applied to AMB techniques. hence discuss the proposed AMB with FLC as a
position controller for EVs

Artificial neural network using MATLAB

(1) ANN is a tool using which parameters of controlling variables of any classical controller
can be calculated. The advantages of ANN is, they are easy to understand and
implementation by non experts.
(2) The tedious mathematical observations for calculating value of controlling variables in
the classical controller like Kp, Ki, Kd can quickly be calculated by this tool.

Fig. Various layers of ANN

 Xnet=x1w1+b1+x2w2+b2……
Fig. A layout of ANN tool of MATLAB

(1) In MATLAB, the ANN tool is available, which can be used to create an ANN based
classical controller. In this case, ANN based PID controller.
(2) ANN structure is designed by selecting number of neurons in the system layers
comprising of input layer, hidden layer and output layer.
(3) The training of neuron can be performed by using the required method of training
algorithm.
a. Leven berg – Marquardt
b. Bayesian Regularization
c. Scaled conjugate gradient
(4) Training can be performed up to several time until the error between the output &
targeted output is minimized. The number of iterations can be set according to it.
When the error between the output & targeted output has minimized the value of
controlling variables of PID controller will be found. ANN-PID controller can be applied
for AMB model for efficient performance of the system.
 yes

no

parameters

Fig. Flow chart of working and implementation of ANN

 The flow chart clearly shows that, it is required to set the ANN structure with
respect to number of nodes ‘n’, weighted coefficient of each layer and learning rate &
inertia factor.
 Number of iterations depends upon the number of nodes for up gradations.
 Obtain the input and output signals and calculate the error.
 Calculate the PID controller parameters using ANN
 Perform PID controller output calculations. Based on error, set ANN learning by
changing the weights, threshold values and content of nodes and go for next
iterations
9 Explain Particle Swarm Optimization technique. With the help of flow chart, explain
working and implementation of particle swarm optimization and hence discuss the block
diagram of PSO implementation in AMB for EV system.
Swarms are natural inspiration to devise problem solving method. They often exists in
nature. For instance, swarm(group) of birds/insects/fish/humans/flying robots.

Benefits of swarm
1. Predators may not attack swarms of animals as the group is larger than a single
animal
2. Move quicker in group with influence of all the members called as an individual
particle
3. Achieve the goal faster with the decentralized unity

Swarm Intelligence: is the collective behaviour of decentralized, self organized systems,

be it natural or artificial. This technique also involves social learning. Swarm
intelligence is seen in humans behaviour easily & it has multiple components.

1. Inertia: stick to own idea or working for self in random movement

2. Past Experience: based on previous experience
3. Social influence: based on neighbours, friends & social circle or network
Particle swarm optimization:
1. A meta heuristic based on flocking/swarm behavior
2. Solves difficult problem
3. Maintain a population of solution. Instead of using the word population as in Genetic
algorithm, it is called as swarm and likewise each member of a swarm is called as
particle (fish/human/birds etc.)
PSO used inertia (own), past experience, social influence & innovation to guide search for
better solution.
 Table: List of variables used in PSO

(Current position)

(Past
experience)
(global/social
influence)

Considering the list of variables used in PSO as shown in Table and after setting up the values
of the particles and velocity using the following equations, the position of individual updates as
(New position)

+
A flow diagram for PSO algorithm flow is shown in figure.

Fig. Flow chart of working and

implementation of PSO

In PSO, each single solution is a "bird" in the search

space. We call it "particle". All of the particles have
fitness values which are evaluated by the fitness
function to be optimized, and have velocities which
direct the flying of the particles
 Fig. Implementation of PSO in AMB for Evs system

Particle swarm optimization technique is used to calculate gain values of a

conventional controller like PID controller. The calculated value of the gains is fed
to the PID controller for controlling the AMB as shown in Figure.
Number of iterations increase the accuracy of the calculations, so almost 50-100
iterations are performed.