Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

EE 677

Smart power grids for a sustainable future


Monsoon 2023

Prabir Barooah

Project Ideas

1
Topic 1: Load ow analysis (LFA) of IITG micro-grid to
identify solar impact on voltage stability

Project summary: The goal is to identify if that luctuating solar generation in


the IITG micro-grid is causing large voltage luctuations. This will be done by
performing Load low analysis (LFA) of the campus’s distribution network,
with existing data (with solar), and then the counterfactual no-solar situation
(assuming all demand is imported from APDCL).

Data and software needed to the analysis are already available. But basic
familiarity with load low is useful. Multiple cases for the campus network has
to be set up via data preparation, solved by LFA code, and the voltages
examined to assess if solar generation is causing large voltage luctuations.

Resources needed to do the project:


(1) demand and renewable generation: raw data is available, has to be
converted irst so that it can be read by the LFA software.
(2) LFA software: available (a Matlab script developed by Dr. Praveen Tripathy)
(3) Topology of IITG distribution network: available.
f
f
fl
f
f
f
f
IIT Guwahati Electrical Network
(IITGEN)

33/11 kV Sub-Station
1×5.0 MVA, 33/11 kV Tr.-1

M1
ESS-9 ESS-1

ESS-6
2×0.63 MVA, 11/0.415 kV Tr.

2×0.50 MVA, 11/0.415 kV Tr. 2×0.50 MVA, 11/0.415 kV Tr.

ESS-8 ESS-2

2×0.75 MVA, 11/0.415 kV Tr. 2×0.75 MVA, 11/0.415 kV Tr.

ESS-7
ESS-5
2×0.63 MVA, 11/0.415 kV Tr.

ESS-3A
2×0.50 MVA, 11/0.415 kV Tr.
2×1.25 MVA, 11/0.415 kV Tr.

AMF
DG

0.75 MVA
AMF
DG
1.25 MVA
Topic 2: Development of open-source software for load
ow analysis for a distribution network with loops

• Existing code by Dr. Tripathi is designed for radial distribution networks

• Load ow analysis is simply solving a nonlinear equation f(x; u, theta) = 0


for the unknown x, where x is a tall vector of bus voltages, and the known
input u contains real and reactive power injections, and parameter \theta
contains information of the distribution network.

• f(x) = 0 is typically solved by Newton-Raphson, but due to high R/X ratio in


a distribution network, the Jacobian sometimes becomes singular and the
iterations do not converge. There are many ways to xing this problem,
such as adding eps*I etc. These ideas will have to be tried in this project
to see which one works, if any.

• In python/matlab, shared via gitlab. Tested on a few cases studies to verify


the code. Good documentation.
fl
fl
fi
Topic 3: Rooftop solar thermal water heaters for a
sustainable India: IIT Guwahati case study

• Goal: do a techno-economic analysis of solar thermal water heaters for the IIT Guwahati
campus. (Techno: what size, where to install etc. Economic: cost, and bene t from not
buying power from APDCL)

• Steps:

• Estimate water heater power and energy demand as. function of time over a year.
Multiple scenarios based on projected increase in campus population. (Talk to IPM,
visit hostels and analyze fac/staff quarters)

• Design (In a simulation) solar thermal water heaters for hostels and faculty/staff
housing, based on commercially available solar thermal water heaters.

• Estimate capital cost of these retro ts, cost savings/loss over the current scenario
(total and levelized), and carbon savings.

• Try various scenarios: all hardware imported, domestic industry half-mature or fully
mature, multiple grades of availability of trained labor for installation, etc.
fi
fi
Topic 4: Absorption refrigeration can reduce peak electricity
demand, but is it feasible for households: a simulation study

• Background: Absorption refrigeration converts heat to cool directly,


without needing mechanical work (so, no compressor). So it only
needs a heat source, but consumes little electricity. It is a mature
technology and is used in thermal power plants and re neries etc.
where there is a lot of waste heat. But with increasingly hot
summers, it might be possible to use it even in large buildings.

• Project goal: take a simple model of an absorption refrigeration


from a textbook/paper, simulate for various weather conditions in
Assam’s summer and estimate the potential for how much cooling
it can provide with a typical roof area. Use these numerical
calculations to assess whether it can provide any meaningful
amount of cooling at all.
fi
Topic 5: Is building pre-cooling effective as (virtual) energy
storage
• Context: There is/will be a lot of solar energy at 2 pm but not at 6 pm, or a lot of wind
energy at 3 am but not at 11 am. Instead of storing that energy in an expensive battery to
be used when needed, one can cool the interior of a building below the comfortable
temperature when the building is not occupied. That cools the structure down, and since
the building structure has high thermal mass (or thermal capacitance), it stays colder much
longer, helping to keep the interior space cold even when it is warmer inside. This
technology was demonstrated in the 1980s, as “structure precooking” at night, in the USA.
But it wastes energy since the building leaks its “cool”. Now there is a renewed interest
since green solar/wind energy can be cheap.
• Project goal: to do a simulation with a building’s model - with Indian weather data (say.,
Guwahati) - to check if pre cooling can ever be useful in this climate.

• Steps: set up a dynamic simulation of a building with air conditioning and time varying
outdoor weather, using Matlab or Python. Run various cases, and check.

• Of particular interest: check the claims made in recent papers (pick a speci c one)

fi
Topic 6: Techno economic analysis of thermal energy
storage for IITG

• Background: Suppose you use a lot of energy for space cooling: x kWh of thermal
energy over a day. And you want to go green by using locally generated solar PV energy.

• Option 1: (conventional) store the electricity generated during the day in batteries
and using that to do space cooling both day and night. Expensive due to batteries.

• Option 2: use the solar-generated electricity to produce extra chilled water during the
day and store it in a big tank. Use that to do air conditioning both during the day and
night. Storing chilled water in a tank is less expensive that storing electronics in a
rechargeable battery.

• Project goal: Estimate existing demand for air conditioning energy and project onto the
next twenty years, for a few distinct scenarios. Determine the capital and operating cost
of two two energy storage options.

• Steps: determine an appropriate size of a chilled water tank, estimate the reduction in
energy import from APDCL over a year due to using the tank to store chilled water, and
then estimate the cost savings
Topic 7: Effect of increased AC use on the Indian power
grid: how will peak demand likely to change?

• Starting point:

• Assume a distribution of
AC’s by region, for various
years. Compute hourly
electricity demand from
AC’s. Add to the baseline
demand

• Repeat for various


scenarios for economic
growth rate, climate
change/consumer
preference
Topic 8: “India cooling future” analysis

If India’s energy consumption reaches, say, by 2050, Hong Kong’s level of 80 kWh/d/p,
what fraction of that is likely to be due to space cooling?

> What is the expected carbon footprint of that energy use (GtCO2/yr and GtCO2/yr/p?)
> What is the expected peak demand (kW, not kWh) and the time of that peak demand,
due to space cooling?

How is that carbon footprint likely to vary depending on the scenario:


(a) “Business as usual” scenario, of using lots of coal for electricity.
(b)50% of electricity from wind and solar, rest coal and gas but without increasing the
share from nuclear
(c) Same as (b) but also increasing the share of nuclear so that coal and gas’s share
becomes less than 10%?
Topic 9: Data analysis from smart meter data

Smart meter: A smart, at the minimum, collects demand data (kW) every few minutes (5 to 15 to 30,
depending on the model etc.). Or, energy demand in that time duration (kWh for the last 15 mines). It
also measures other variables such as voltage etc., again depending on the model and other factors.
This data is periodically transmitted to the utility (DISCO). At present, DISCOMs in India use smart meter
data only to do billing. But additional information can be gleaned, such as:
1. Is a consumer consuming a lot more power or energy compared to what they are supposed to (e.g. a
5 kW connection, but peak demand is 20 kW!)
2. Is an equipment about to fail?
3. ….

This project will involve analyzing data from a large collection of smart meters, with all meters from one
area (neighborhood, city, state etc.), to detect features that can help the DISCOMs or the consumers.

PROBLEM: I have put in a request to APDCL to get access to smart meter data from Guwahati circle, but
not received anything yet!
So, you may have to use publicly available datasets from other countries (there are some; search
online)
Topic 9,10: How much energy storage will India need

Consider various scenarios of renewable generation for India till 2050 (X1,Y1 GWp solar,
wind, or X2,Y2 GWp wind, solar, etc.)

For each scenario, how much energy storage is needed to be able to meet peak demand of
the Indian grid with 99.99\% reliability. (Assume all of India is just one big consumer, no
need to consider geographic spread of demand and generation)

1. Read and recreate the method in the Prayas+LBNL report


2. Try the method in a technical paper, and compare with the Prayas report

Or, do 1 and 2 as separate projects by separate teams


Grid-Scale Battery Storage:
Costs, Value, and Regulatory Framework in India

Webinar jointly hosted by


Lawrence Berkeley National Laboratory and Prayas Energy Group

July 8, 2020

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 36, NO. 3, MAY 2021 2473
1

A Probabilistic Method for Reserve Sizing in Power


Grids With High Renewable Penetration
Kendall Parker , Graduate Student Member, IEEE, and Prabir Barooah , Member, IEEE

Abstract—Power systems employ reserves to ensure there is ad- criterion that would be useful for infrastructure development”.
equate generation and storage to meet demand. With the increased Four years later, the authors in [3] conclude that “...thus far,
use of volatile and uncertain sources of electricity such as solar and contributions to the theoretical analysis of storage sizing had
wind, the need for reserves is becoming more critical to ensure
reliable supply of electricity. Due to the high cost of reserves, been limited”. In a review of methods for energy storage system
methods are required to determine minimal reserves needed. This (ESS) siting and sizing from 2016 [4], results show that only two
work proposes a probabilistic data-driven method to determine the works, [5] and [6], utilize analytical methods without complex
minimum reserve size (both power and energy) to satisfy a spec- mathematical optimization tools to size energy storage with
Topic 11: Energy markets in India: lessons from USA and
Europe?

Precise project goal is not clear to me as yet, but will require a massive literature review
Design and analysis of solar farming

• Not this time


Design and techno-economic analysis of solar powered
cold storage with battery and absorption refrigeration

• Not this time!


Feasibility study of phase change materials as
energy storage for military human waste digester

• Not this time


Quiz or homework questions:
as of 2021, 8% of Indian families own a car. In USA, it is X% and in Vietnam it is Y%. If USA fraction of Indian families own a petr
will be the energy and carbon footprint (per day per person) in India?

> How will it increase India’s oil export? (Don’t just give a number, express it in appropriate units to aid discussion)

> What about extra emissions due to road congestion? Look at road capacity of Indian roads now, and what will be the average
all those extra cars?
>If instead we have the same vehicles but replace them with EVs, what will be the energy and carbon footprint?
Scenario A: current mix of coal and others
Scenario B: Far higher solar and PV
Scenario C: Far higher nuclear

> If instead we use public transportation, what will be the energy and carbon footprint
> x kms/day in buses, petrol based
> y km/day in trains
> z km/day in airplanes
Will time-of-use pricing for EV charging help the
Indian grid reduce peak demand (Indian context)

• Lots of reading, and interviews with EV owners will be


need. So, not this time

Resources: Evaluating Time of Use rates for Electric Vehicle Charging for Distribution
Companies in India, by Chandana Sasidharan; Bhawna Tyagi; Shyamasis Das

You might also like