Machine Learning for improved monitoring and forecasting of Mumbai
flood
Professor: Subimal Ghosh
UID: CE01
Study on Wave Energy Devices
Professor: Srineash V K
UID: CE02
Evaluation of prestress loss in UHPC bridge girders
Professor: Najeeb Shariff Mohammad
UID: CE03
Machine Learning for Hydroclimatic Clustering
Professor: Arpita Mondal
UID: CE04
Driving simulator based automated driver warning system using
statistical and AI/ML techniques
Professor: Tom V Mathew
UID: CE05
CE01
Description: The project aims to gather flood data from
multiple sources, and prepare real time Mumbai flood map. The
next step is to go for high resolution rainfall forecasting and
converting to flood. A website will be developed to disseminate
the information to the Mumbaikars.
Number of students: 3
Year of study: Students entering 3rd year
CPI eligibility criteria: 8.5
Prerequisites: Sound knowledge of statistics and
knowledge in ML
Duration: 1 year
Learning outcome:
Weekly time commitment:6 hrs
General expectations:
Assignment:
Instructions for assignment:
CE02
Description:The rapid depletion of non-renewable energy
sources coupled with the increase in the energy demand
necessitates the requirement for the usage of alternate energy
from renewable resources. Oceans have a huge energy potential
which could be utilized effectively using wave energy devices.
Therefore, it is essential to have studies involving wave energy
devices/converters to understand their performance for different
wave conditions. The implementation of wave energy devices
requires fundamental understanding with respect to the
performance and the design. This is the focus of the project.
The aim of the project will be to investigate the hydrodynamics
of the Wave Energy Devices (WED) and to examine their
performance for different wave conditions. There are different
types of wave energy devices or converters such as Oscillating
Water Column, Surge type wave energy converters, Attenuators,
Point absorbers. The student is welcome to choose his/her own
WED and work on it.
Number of students: 1
Year of study: Students entering 4th/5th year
CPI eligibility criteria: 6.5
Prerequisites: NO
Duration: 2 months.
Learning outcome:
Weekly time commitment: 8h
General expectations:
Assignment:
Instructions for assignment:
CE03
Description:Loss of prestress is a major concern in
prestressed concrete bridge girders. Although, a few models
exist to evaluate the prestress loss, occurring due to creep,
shrinkage and relaxation, in normal strength concrete, there is
very little data on the behaviour of UHPC in terms of long-term
behaviour. In the present study, a brief literature review on
the existing models for UHPC will be carried out and the
time-dependent behaviour will be simulated using the
state-of-the-art hygro-thermo-chemico-mechanical models.
Number of students: 2
Year of study: Students entering 4th/5th year
CPI eligibility criteria: >8
Prerequisites: Background in coding is recommended
Duration: 4 months
Learning outcome: The student would get a comprehensive
understanding of how UHPC bridge girders age and the process to
evaluate the residual strength PSC girders affected due to
ageing and time-dependent degradation.
Weekly time commitment:20 h
General expectations: Commitment towards the work
Assignment:
Instructions for assignment:
CE04
Description:Rainfall, floods and droughts are physical
processes that have been observed to organize in clusters both
in space and time. In the absence of good quality and quantity
of observed data, engineers and hydrologists borrow similar data
within a cluster for design, planning and management. In this
project, the students will learn and apply unsupervised machine
learning techniques for hydroclimatic clustering.
Number of students: 2
Year of study: Students entering 2nd year, Students
entering 3rd year, Students entering 4th/5th year
CPI eligibility criteria: >8
Prerequisites: Background in probability and statistics,
coding experience, particularly in R or Python.
Description:Driver warning technologies hold the
potential to reduce traffic crashes and save precious human
lives. Scientists and engineers hypothesize that the
measurements obtained from advanced driving simulators can
predict equivalent measurements in the real world that lead to a
better understanding of the driver behaviour in critical driving
situations. The aim of the project is to study the variations in
driver state using a combination of driver behavioural and
driver physiological signals using sensor technologies in
driving simulator. Driver warning can be developed through
statistical analysis and AI techniques.
Number of students: 2
Year of study: Students entering 4th/5th year
CPI eligibility criteria: >8
Prerequisites: Data analysis tools
Duration: 6-8 months
Learning outcome: Statistical data analysis and modeling,
exposure to driver simulators and sensors