1

SAMPLE EXAM CIEM5312

Emerging Technologies for
Transportation Infrastructure

Date :-
Time : 3 hours
Location : -

EXAM
 2

Question-P1 (12 pts) (only one answer is correct)

Q1-1. Which mechanism allows piezoelectric pavements to
generate electricity effectively? (2 pts)

A) Direct sunlight exposure through transparent overlays

B) Conversion of heat generated from tire friction

C) Conversion of mechanical pressure into electrical energy

D) Kinetic energy from wind created by passing vehicles

Answer: C

________________________________________

Q1-2. What is a significant environmental advantage of plastic

roads compared to traditional asphalt roads? (2 pts)

A) They can be built with biodegradable materials

B) They are resistant to high temperatures, minimizing maintenance

C) They reduce waste by incorporating recycled plastics into the

structure

D) They cool the road surface and surrounding area

Answer: C

________________________________________

Q1-3. In truck platooning technology, how is fuel efficiency

improved for each vehicle in the platoon? (2 pts)

A) Trucks are connected by physical coupling to reduce drag

B) Trucks maintain close proximity, reducing aerodynamic resistance

C) Each truck operates autonomously, optimizing fuel for each

section

D) Airborne deflectors on each truck reduce wind impact

Answer: B
 3

Q1-4. What is a primary technical feature of snow-melting

pavements that prevents ice and snow accumulation? (2 pts)

A) Embedded sensors that activate anti-icing chemicals

B) Heated pipes or electrical heating cables installed beneath the

surface

C) Micro-porous materials that absorb and trap heat

D) Reflective surfaces that amplify sunlight

Answer: B

________________________________________

Q1-5. Why might low-noise pavements require specific surface

treatments to maintain their noise-reducing properties? (2 pts)

A) To ensure they retain heat for snow and ice prevention

B) Because debris can clog the porous surface, reducing noise

absorption

C) Because the surface texture wears down quickly under heavy

traffic

D) To maintain structural flexibility during temperature changes

Answer: B

________________________________________

Q1-6. Which pavement characteristic primarily enhances the

noise reduction performance in low-noise roads? (2 pts)

A) The use of rubberized asphalt for added flexibility

B) Porous material that absorbs sound waves from vehicle tires

C) Reflective coatings that deflect sound waves away from nearby

areas

D) High-friction surface textures that minimize tire slippage

Answer: B
4
 5

Question-P2 (6 pts)
Q2-1. Determine the healing efficiency of the asphalt mixture using
the SCB bending and healing test (see Fig. 1) and the peak load
values provided in Table 1 for each healing cycle. Calculate the
healing index using the formula given below. Based on the
knowledge acquired in this course, identify the potential healing
method(s) used in the tests shown in Table 1, and provide a rationale
for your conclusions. (4 pts)

Fig. 1 Semi-circular bending and healing test programme

Healing index equation:

Cx
HI = × 100 %
C1
Where:
HI is the healing index (%),
C1 is the initial peak load (N);
Cx is the peak load measured from the x testing cycle
 6

Table 1 Asphalt mixture healed with healing method 1

C1 C2 C3 C4 C5

Peak
2500 1750 1720 1590 1300
load (N)
Healing
100
Index(%)

Answer:

Using the provided formula, the healing index can be calculated

for each healing cycle in the test and is presented in Table 1.

Table 1. Asphalt mixture healed with healing method 1

C1 C2 C3 C4 C5

Peak
2350 1750 1620 1550 1530
load (N)

Healing
100 74 69 66 65
Index(%)

Based on the potential healing methods applicable to this test,

capsule healing typically achieves no more than 20% strength
recovery and cannot be repeated multiple times. In contrast,
induction or microwave healing methods can achieve over 60%
strength recovery and remain effective for more than five cycles.
Therefore, the likely healing method used in this test could be
induction or microwave healing. It is also possible that a hybrid
healing approach (e.g., capsules combined with induction) was
employed.

Question 2-2: (only one answer is correct) (0.5 pts)

 7

Principle of Energy Balance and Boundary Conditions: If a

working refrigerator is placed in a perfectly insulated, closed
room, is left open, and is plugged into an external power source,
what will happen to the room’s temperature over time?

A) The room will cool down gradually.

B) The room will warm up gradually.

C) The room will remain at the same temperature.

D) The room’s temperature will fluctuate randomly.

ANS. (B). Since energy continuously enters the room through

the refrigerator’s power source and cannot escape as the room
is perfectly insulated, the total energy in the room will increase
causing the temperature to rise over time.

Question 2-3: (only one answer is correct) (0.5 pts)

Role of Thermostats in Molecular Dynamics Simulations: In an

MD simulation, a thermostat is applied to control the
temperature of the system. Which of the following best
describes the role of a thermostat?

A) It applies random forces to particles to increase temperature.

B) It adjusts particle velocities to maintain a target temperature.

C) It adjusts the potential energy of the particles throughout the

simulation.

D) It prevents particles from interacting across periodic boundaries.

ANS. (B). A thermostat in Molecular Dynamics is used to control

the system’s temperature by modifying the velocities of
particles. Since temperature is proportional to the average
kinetic energy, adjusting particle velocities allows the system to
reach and maintain a specified temperature. This is essential for
simulating conditions where temperature control is needed, but
the thermostat does not affect potential energy or the handling
of boundary conditions.
 8

Question 2-4: (only one answer is correct) (0.5 pts)

Concept of Equilibrium and Simulation Timescales: When

simulating a system of particles, you observe that some
properties change rapidly at the beginning but then stabilize
over time. What does this likely indicate?

A) The system has reached equilibrium.

B) The simulation is experiencing numerical errors.

C) The system needs a shorter time step.

D) Particle interactions are only occurring at the beginning.

ANS. (A). In Molecular Dynamics, it is common to see properties

like temperature or energy change quickly at the start as the
system moves from an initial, possibly non-equilibrium state,
toward equilibrium. Once these properties stabilize, it generally
indicates that the system has reached equilibrium, meaning that
macroscopic properties have stopped fluctuating significantly
and the system's behavior becomes steady over time.

Question 2-5: (only one answer is correct) (0.5 pts)

Speed, Kinetic Energy, and Temperature Relationship: In an MD

simulation, you decide to double the speed of all particles. What
is the resulting effect on the system temperature?

A) Temperature remains the same, as it is independent of velocity

and kinetic energy.

B) Temperature doubles.

C) Temperature halves.

D) Temperature quadruples.

ANS. (D). Temperature in an MD simulation is proportionally

related to the average kinetic energy of particles, as hinted,
 9

which in turn depends on the square of their velocity. If the

speed of all particles is doubled, their kinetic energy increases
by a factor of four, and consequently, the temperature also
1 2
quadruples. Remember, T ∝ Ek = 2 m V .

Question 2-6: (only one answer is correct) (0.5 pts)

Potential and Kinetic Energy Relationship: In a Molecular

Dynamics simulation, particles move around. How does the sum
of kinetic and potential energy behave if no external forces or
thermostats are applied?

A) The sum of kinetic and potential energy will continuously increase.

B) The total energy will oscillate due to particle motion.

C) The total energy will remain constant.

D) The sum of kinetic and potential energy will approach zero over
time.

ANS. (C). In a closed Molecular Dynamics system with no

external forces or thermostats, the total energy, which is the
sum of kinetic and potential energy, remains constant over time.
This reflects the conservation of energy in an isolated system.
While kinetic and potential energy may individually fluctuate as
particles move and interact, their sum stays constant,
maintaining the total energy throughout the simulation.
 10

Question 2-7: (only one answer is correct) (0.5 pts)

Principle of Energy Conservation and Elastic Collisions: A

particle is released from a height of 5 meters and allowed to fall
under gravity. It bounces back up in perfectly elastic collisions
with the ground. After 99 bounces, what height will the particle
reach?

A) Lower than 5 meters.

B) Higher than 5 meters.

C) Equal to 5 meters.

D) The particle reaches equilibrium and stops bouncing.

ANS. (C). With perfectly elastic collisions, energy is fully

conserved in each bounce, meaning there is no loss of energy to
heat, sound, deformation, etc. Therefore, after each bounce, the
particle will reach the same height from which it originally fell.
Thus, after 99 bounces, it will continue to reach the initial height
of 5 meters, as its potential energy is restored fully on each
bounce.
 11

Question-P3 (6 pts)
Name three key factors to be considered in the performance-
based tests of asphalt binder. For each factor, briefly explain its
influence.

ANS:

2 pts for each of the following factors:

• Time: Asphalt binder exhibits viscoelastic behavior,

meaning it deforms differently based on load duration. Short
loads cause elastic behavior, while long loads result in viscous
flow and permanent deformation.

• Temperature: High temperatures soften the binder, leading

to increased deformation, while low temperatures stiffen it,
causing it to crack under stress.

• Aging: Over time, exposure to oxygen and UV light causes

oxidative aging, which hardens the binder, increases stiffness,
and reduces flexibility, making the pavement more susceptible
to cracking.
 12

Question-P4 (8 pts)
Understanding the behavior of waves in various media is crucial
for many engineering applications, including pavement
engineering. The governing partial differential equation (PDE) for
wave propagation is expressed as:
2 2
∂ u 1 ∂u
2
= 2 2
∂ x c ∂t

Where,
u ( x , t ) represents the wave function, c is the wave speed, x is the
spatial coordinate, and t is time.

Based on the given information below, formulate loss functions of

a Physics-Informed Neural Network (PINN) to approximate the
solution to the above PDE.

Let the boundary conditions be defined as:

u ( 0 , t ) =f 1 ( t ) for t ∈[0 ,T ]

u ( L ,t )=f 2 (t ) for t ∈[0 , T ]

u ( x , 0 )=g ( x ) for x ∈[0 , L]

Where,
f 1 ( t ) : Sinusoidal function

f 2 ( t ) :Cosine function

g ( x ) : Initial displacement function

Q4-1: Using the given PDE, fill in the missing expression inside the
bracket in the physics loss to represent the residual (the error) of
the PDE at the points ( x i ,t j ). (2 pts)
N
1
L p hysics =λ 1
N
∑ ()2
i=1

Consider:
 13

u pinn is the predicted wave function by the neural network.

λ 1 is regularization parameter.

N is the number of training points for x and t .

Answer:

( ( ))
2 2
1
N
∂ u pinn ( x i ; t j ) 1 ∂ u pinn ( x i ; t j )
L p hysics =λ 1
N
∑ ∂x
2
− 2
c ∂t
2
2
i=1

Q4-2: Based on the given boundary conditions, fill the missing

terms in the following boundary loss expressions. (4 pts)

- For the first boundary condition:

M
1
Lb oundary ( 0 , t )= ∑ ()2
M J=1

- For the second boundary condition:

M
1
Lb oundary ( L , t ) = ∑ ( )2
M j=1

- For the initial condition:

K
1
Li nitial ( x ,0 )= ∑ ()2
K i=1

-Loss of (boundary + initial):

Lb oundary +initail =λ2 ¿

Consider:
K is the number of spatial points sampled for the initial condition.

M is the number of time points sampled for the boundary

conditions.
T surface ( t ) is the known surface temperature condition such as 40°C
at t = 0, decreasing to 30°C after 1 hour
 14

λ 2 and λ 3 are regularization parameters.

Answer:

- For the first boundary condition:

M
1
Lb oundary ( 0 , t )= ∑ ( u (0 ; t j)−f 1 (t j )) 2
M J=1

- For the second boundary condition:

M
1
Lb oundary ( L , t ) = ∑ ( u(L ; t j)−f 2 (t j ) ) 2
M j=1

- For the initial condition:

K
1
Li nitial ( x ,0 )= ∑ ( u (ti ; 0)−g( x i )) 2
K i=1

-Loss of (boundary + initial):

Lb oundary +initail =λ2 (Lb oundary ( 0 ,t ) + Lb oundary ( L , t ) ) + λ3 (Li nitial ( L ,t ) )

Q 4-3: Based on all the loss functions to train the PINN, write down
an expression for total loss. (1 pt)

Lt otal =¿

Answer:
Lt otal =L p hysics + Lb oundary

Q 4-4. What is the primary difference between a traditional neural

network and a physics-informed neural network (PINN)? (1 pt)

a) PINNs require more data than traditional neural networks.

b) Traditional neural networks integrate physical laws into their
learning process.
c) PINNs include physical laws (like PDEs) in the loss function,
 15

whereas traditional NNs are purely data-driven.

d) PINNs require manual calculation of gradients, while traditional
NNs do not.
e) PINNs are used only in engineering, while traditional NNs are used
in general tasks.

Answer: c) PINNs include physical laws (PDEs) in the loss

function, whereas traditional NNs are purely data-driven.
 16

Question-P5 (6 pts)
Q5-1 As part of your responsibility, you have to focus on two house
appliances: a washing machine and a vacuum cleaner.

• The washing machine works for 10 minutes.

• The vacuum cleaner works for 15 minutes.

• The simulation lasts for 60 minutes.

Part A (2 pts):

Determine when the washing machine and vacuum cleaner are

operating simultaneously during the allotted 60 minutes.

Part B (2 pts):

How would you adjust the intervals between turning two devices to
lessen conflict, assuming that you are attempting to minimize the
amount of time that two devices are turned on?

Answer:

Part A: Times when both appliances are running simultaneously

• The washing machine operates in 10-minute intervals: 0-10,

20-30, 40-50 minutes, and so on.

• The vacuum cleaner operates in 15-minute intervals: 0-15,

30-45 minutes, and so on.

Simultaneous running times:

• 0-10 minutes: Both the washing machine and vacuum

cleaner are running.

• 30-40 minutes: Both are running during this period.

Thus, during the 60-minute period, the two appliances are

running simultaneously from:

• 0-10 minutes
 17

• 30-40 minutes

Part B: Reducing the overlap between appliances

Changing one of the appliances' start times or their operating

intervals will help to minimize overlap.

For instance, five minutes later, you can turn on the vacuum.
The vacuum cleaner will operate from 5 to 20 minutes, 35 to 50
minutes, etc., if it starts at 00:05 rather than 00:00. This
minimizes the time when both are running at the same time.

Alternatively, you could adjust the intervals:

• Change the vacuum cleaner’s interval to 20 minutes (rather

than 15). This would reduce the overlap, as it would run less
frequently. This would result in fewer simultaneous operating
periods and maximize the time when only one appliance is
running.

Q5-2 Describe and draw typical layers of digital twin architecture. (2

pts)

Answer: on slide number 72

 18

Question-P6 (6 pts)
Q6-1 What is the force field in the context of molecular dynamics
simulations? What are the main components of a force field? (3 pts)

Answer: A force field refers to a set of equations used to

describe how particles (typically atoms or molecules) interact
with each other. These interactions are modeled using a
combination of potential energy terms. Common components of
a force field include bond stretching, angle bending, torsional
(dihedral) angles, and non-bonded Interactions, namely
Lennard-Jones potential and Electrostatic (Coulombic) potential.

Q6-2 What are the three main components of a neural network? (1

pt)

Answer: Input Layer - Hidden Layers - Output Layer

Q6-3 What is the main difference between traditional neural networks

and PINNs? (1 pt)

Answer: The main difference is that PINNs incorporate physical

laws and constraints into their learning process, while
traditional neural networks rely solely on data-driven learning.

Q6-4 Example: What are the three main equations used in the
asphalt aging model presented in the lecture? (1 pt)

Answer: The three main equations are:

Temperature Equation

Modified Oxygen Pressure Equation

Carbonyl Area Equation

 19

Question-P7 (6 pts)
Recycling is one of the main end-of-life (EoL) strategies for asphalt
pavements. What are the main challenges in asphalt recycling? List
at least 4 challenges.

Answer:

• Contamination of recycled materials can compromise the

quality of the recycled asphalt pavements.

• Variability in the quality/composition of the binder

reclaimed from asphalt pavements

• Hardness and level of ‘recovered binder activation’, in

other words, how much of the old binder is participating in the
new binder instead of just acting as ‘black rock’.

• Recyclability of asphalt with legacy materials, such as

asphalt with coal tar or asbestos.

• Lack of material specifications

• Uncertainties in the long-term performance

 20

Question-R1 (6 pts)
Draw a simple switch set in a diverging direction. Name the main
elements and point them on the drawing.

Answer:
 21

Question-R2 (6 pts)
Q2-1 What is IoT, and what is its role in railway systems? (2 pts)

Answer: The Internet of Things (IoT) refers to physical objects

with sensors connected to the Internet. IoT connects the
physical and digital worlds and allows real-time monitoring and
decision-making, among many other applications. Basically, IoT
is an enable that is the basis of multiple applications, for
example, predictive maintenance where the health condition
data from components go to the cloud, gets then processed, and
decision-making is conducted to prevent malfunctions, reduce
repair time, and ensure the safety and availability of the railway
operations. While algorithms/methods that provide the advice
are independent of IoT or the use of Internet, IoT is an enabler
for efficiency, for instance, you don’t need to locally store data
from your sensors located remotely in moving trains or
locations difficult to access, but directly get them from the
cloud.

Q2-2 What is Blockchain, and what is its role in railway systems? (2

pts)

Answer: Blockchain is a technology that secures records of

transactions and makes it almost impossible to alter/temper the
data that is introduced. This allows the high security,
traceability, and reliability of data/transactions as they are all
verified. Fraud in ticketing/payments can be minimized. Railway
maintenance and operation involve various suppliers (of
components, energy) and blockchain could guarantee tracking
origin and handling of each component. Smart contracts, that
execute themselves and from where rules for verification and
payments of works can automatically be generated require that
the data/transactions are of high security.

Q2-3 Discuss barriers to implementing new technologies in railways

(2 pts)

Answer: We discussed in class the following. Cost: high initial

investment, budget constraints, and long-term financial planning
 22

when stakeholders do not want to commit resources without

guaranteed success. Standardisation and regulations:
environmental impact and compliance with existing safety
standards, implementation process can be slowed down when
designs undergo testing and approvals. Adoption resistance:
preference for familiar designs and processes, cultural barriers
if the innovation disrupts traditional practice, change
management requiring training communication, and adjustment
to organizational structures. Technical challenges: integration
with existing technologies, interoperability (working technology
in different regions and systems), complexity in implementation
due to complexity in maintenance and monitoring that require
new skills in the workers.
 23

Question-R3 (8 pts)
1) Explain the Doppler effect and give an example of it. (2 pts)

Answer: When there is a relative motion between the source and

the observer of a wave, the received frequency shifts from the
emitted frequency. The frequency shift depends on the relative
speed between the source and the observer. A typical example
is the difference in the frequency of the horn one hears between
a train is approaching and leaving.

2) Describe how the received frequency changes compared to the

emitted frequency when an observer is moving closer to the source of
a wave. What if the observer is moving away from the source? (2 pts)

Answer: When the observer is moving closer to the source, the

received frequency is higher than the emitted frequency. When
the observer is moving away from the source, the received
frequency is lower than the emitted frequency.

3) Explain why there is a double Doppler frequency shift when

measuring structural vibration with a laser Doppler vibrometer. (2 pts)

Answer: For the emitting beam from an LDV to the object, the
LDV is the source and the object the observer; the vibration
causes a Doppler frequency shift. For the reflective beam from
the object back to the LDV, the object is the source and the LDV
the observer; the vibration causes another Doppler frequency
shift, so the total frequency shift is doubled.

4) Explain the advantages and challenges of train-borne

monitoring compared to track-side monitoring (2 pts)

Answer: Advantages: Very efficient, non-invasive to traffic,

frequent monitoring, realistic loading conditions. Challenges:
More uncertainties/disturbances, more challenges in data
analysis.
 24

Question-R4 (6 pts)
What are the key advantages of purely data-based modeling
approaches?

Answer: 2 points for each of the following advantages:

• A fast surrogate/replacement model for computationally
heavy models
• An alternative when there is no model available but data is
sufficient
• Highly flexible to different modeling problems with various
input and output
 25

Question-R5 (6 pts)
What types of layers are commonly used in a digital twin architecture,
and what are they for? Name at least three types of layers.

Answer:
1 point for each of the listed layer types and 1 more point if the
description of the layer type is correct. Max 6 points.

Layer
Layer type What is the layer for?
number
Physical
object/system,
Physical layers, Data sensors, IoT devices
1
acquisition layer that
provide/communicate
data
Digital model of the
physical object,
simulations performed
Digital layer, Model layer,
2 based on the model(s),
Simulation layer
data processing and
information generation,
etc.
Data sense-making,
feature extraction,
Cyber layer, Data/Model
3 AI/ML algorithms,
integration layer
simulations based on
data, etc.
Interaction layer, Data transmission and
Transmission layer, communication
4
Synchronization layer, between the physical
Ingestion layer twin and digital twin
Data sense-making,
feature extraction,
Data processing layer,
5 AI/ML algorithms, etc.
Inference layer
(can be an additional
layer to number 3)
User interface,
Service layer,
6 actionable information,
Consumption layer
etc.
 26

Question-R6 (6 pts)
Discuss the following statements.
1) The standards define the level of confidence the RAMS are to
achieve in a railway infrastructure.
2) Resilience in railway infrastructure refers to a fast recovery during
external disruptive events while continuing with critical functions and
providing transportation services.
3) Embankment instability is a possible adverse effect of flooding on
railway infrastructure, and better drainage system design is a
countermeasure for addressing pre-, during-, and post-flooding
events.

Answers:
1) Different standards suggest processes and methods to
demonstrate a certain level of RAMS. Although standards show
the generalities of the methods and frameworks, the final
decision on the levels of confidence for RAMS should be defined
by the infrastructure managers for their specific assets.
2) It is not only the ability to recover but also the ability to
adapt to external disruptive events. For instance, learning from
previous disasters can help us develop better systems that
anticipate, prepare in advance, and respond to those events.
Resilience can be quantified with different approaches, such as
the R4 Resilience Framework, which considers robustness,
redundancy, resourcefulness, and rapidity. In the statement, fast
recovery refers to Rapidity, which is only one aspect. Also,
resilience is a characteristic of the system that does not only
apply during external disruptive events. Integral design of
methods to improve resilience can tackle pre-, during, and post-
external disruptive events.
3) Flooding can also affect the electronic systems and
signaling, as well as the catenary, bridges, tunnels, etc., so not
only the embankment but the whole railway system. Depending
on the severity of the event, more components are affected.
Changes in the track design, such as better drainage systems,
refers typically to pre-flooding countermeasures that can be
conducted, as during flooding or immediate post-flooding is not
reasonable to plan such major design modifications (immediate
post-flooding refers to the learning period after the flooding,
which could lead into the need of design of better drainage as a
major conclusion). From Network rail examples, pre-event could
be flood defense systems, including barriers with membranes
that allow water to pass by just in one direction, cleaning the
 27

drainage systems from branches-leaves to prevent water runoff,

construction of pumping stations in flood-prone locations, and
installing tracks and signaling systems at higher levels.
Examples of during-flooding: monitoring weather and flood
warning signals from specialized agencies. Examples of post-
event could include all sorts of assessment methods for the
damage (by drones or inspections), logistics for recovery, and
replacement operations. Also, the learning process so to
evaluate actions to prevent such damage.
 28

Question-R7 (6 pts)
What are the coefficient of friction (2 points) and coefficient of
adhesion (2 points)? How can the rail coefficient of friction be
measured (2 points)?

Answer:

The coefficient of Friction (CoF) characterises friction level on a

surface. A high CoF indicates a rough surface and a low CoF
indicates a smooth/slippery surface. In practice, rail CoF varies
with weather conditions.

Coefficient of adhesion (CoA) is the ratio between wheel-rail

tangential force and normal load, which varies with train type
and traction/braking conditions.

Since CoF is the upper bound of CoA, CoF can be measured by

gradually increasing CoA (or traction/braking torque/force) until
the friction saturation is achieved when the CoA cannot further
increase with torque/force.
 29

Question-R8 (6 pts)
In the photos below you can see a Stadler FLIRT (conventional
diesel-electric) and a Siemens Mireo Plus B with lithium-titane
batteries (battery-electric). Describe the pros and cons of these
technologies, that is: a) Internal combustion engine with diesel, and
b) battery electric systems.

Answer:

Diesel engines have traditionally provided reliable, high-torque

power, with many examples of implementation, proving
extensive range and independence from electrified
infrastructure. Diesel engines are relatively straightforward to
implement with a large availability of fuel distribution networks.
However, they produce significant emissions, notably carbon
dioxide (CO₂) and nitrogen oxides (NOₓ), which contribute to air
pollution and climate change.

Battery electric trains offer low emissions, and they can even be
emission-free in their operation if green electricity is used. They
also have the highest reduction of WTW energy use. They are
quite silent, with reduced noise that is welcomed by the people
living in the surroundings of the railway infrastructure. Although
various tests are being conducted to improve their flexibility on
non-electrified routes, they still show a limited range,
suggesting the need for partial track electrification in addition to
vehicle retrofit and energy management systems. The charging
times of batteries are getting shorter with new technology, but
these still require additional infrastructure and logistics.
Batteries also add weight (heavier trains require more energy to
transport them) and take up space (although the new
technologies in batteries are making it possible to have them in
every time smaller sizes and higher efficiencies). Although the
 30

operation can be emission-free when seeing the whole life cycle

of the batteries, they definitely produce environmental
impacts/emissions related to their production and disposal.
 31

THE END