Humanitarian Efforts for Improving Air Quality Using Solar Power

Full text

Abstract — Indoor air pollution is a major health issue
worldwide, responsible for 4.3 million deaths per year,
predominantly in developing nations and rural areas. In these
areas, families often cook indoors using wood, leaves, manure,
and any other locally-available biomass. Cleaner burning
cookstoves are occasionally used, but in many rural locations
they have experienced a low acceptance rate. The Breeze
Project was created to explore alternative means to ameliorate
the risk of indoor air pollution through solar-powered
ventilation systems. This paper provides the intermediate
results of this ongoing project as well as generalized lessons
learned over the course of three separate trips to our partner
community in Uganda. Our approach is characterized by first
understanding the problem first-hand by traveling to a specific
community, interfacing with community leaders and residents,
and collecting qualitative and quantitative observations.
Subsequently, we design a solution that identifies the perceived
needs of the community, remotely gather as much prior
feedback as possible from community leaders, and conduct a
second deployment of a small number of units. This serves as a
proof of concept, and qualitative and quantitative observations
can inform the researchers of whether the units both create
meaningful change and are well-received by the residents. This
deployment may also yield unexpected risks and opportunities
relevant to subsequent efforts. We follow this trip with a
second phase of design revision and prototype fabrication, this
time creating enough units to constitute a larger pilot. The
third pilot deployment should be focused on gathering feedback
on any design modifications, ideally observing any available
devices from the second deployment, and beginning to think of
how a successful solution could be manufactured locally or
regionally. In this paper we will demonstrate some of the
specific decisions we made during and prior to these
deployments in order to maximize their positive impact and the
potential for the project’s success in the future.
I. INTRODUCTION
According to the World Health Organization, close to 4
million people die prematurely every year because of indoor
air pollution and the inefficient use of solid fuels.
Additionally, exposure to household air pollution nearly
doubles the risk of childhood pneumonia and is a leading
cause of other conditions such as COPD, stroke, and lung
cancer [1]. Over the past few years, air quality has become a
more common topic of discussion. For many people,
however, bad air quality is associated with images of urban
centers, heavy traffic, and industrial areas. Cities like Beijing
and Pittsburgh immediately come to mind. Many are
unaware of the ubiquitous nature of air pollution and how
even in the most rural areas of the world, poor air quality
and fine particulate matter pose grave threats inside peoples’
homes.
While there are many categories of airborne pollutants
with significant health impacts, this paper focuses primarily
on fine particulate matter, or PM2.5. This includes any solid
or liquid particles with a diameter of less than 2.5 microns,
which is typically less than one thirtieth the size of an
average human hair and can infiltrate deep into the lungs and
bloodstream. [2]. The most common source of these particles
is combustion, and the less fuel-efficient the combustion
process is, the more particles are released[3]. In rural areas,
we observed the cook stoves for many families are made of
stone and brick, and where charcoal is unavailable, anything
that burns can become fuel. This can release dangerous
amounts of fine particulate matter, and because these homes
lack any substantial ventilation, these particles accumulate
for the duration of meal preparation.
1. Size comparisons for PM particles [2]
Indoor air quality problems in the developing world came to
our attention through an ad campaign by Toyota called
“Ideas For Good.” The public was tasked with thinking of
ideas to change the world using Toyota technology. After the
country voted on several hundred ideas, the top five were
Humanitarian Efforts for Improving Air Quality
using Solar Power
Joshua Schapiro, Michael D Taylor, Illah Nourbakhsh
Robotics Institute
Carnegie Mellon University
Pittsburgh, Pennsylvania 15213

flown to Carnegie Mellon University for the weekend to
have their ideas prototyped and tested. One project
repurposed the solar powered ventilation systems found in
the Toyota Prius to ventilate cooking areas of homes in
developing nations. Engineers from the Community
Robotics Education And Technology Empowerment
(CREATE) Lab at CMU led the prototyping efforts. They
pl ayed a pivotal rol e d esig ning , engin eeri ng a nd
experimenting on the different ideas, including the
ventilation system. After the weekend was finished, Toyota
encouraged the University to continue working on the
problems, and the ventilation system had the most promise
for positive impact. The money given to the University
allowed us to plan for an observation trip as well as a small
deployment of any prototypes.
The climate in Uganda proved to be the ideal target for
this type of work. Situated on the equator, this land-locked
country is very tropical, consisting of two dry and two wet
seasons a year. If it were a desert climate, with very little
rain, the residents could potentially cook outdoors year
round. Because of the wet seasons in the area, residents are
forced to cook inside. Lack of resources and technology
often lead to poor ventilation with bio-mass cooking fuel.
Which leads to the chronic respiratory issues we see in the
area[3].
The villages we are working with are in the Makukuulu
Parish of rural Uganda. The community is a 4 hour drive
from the capital city of Kampala. We work closely with
Kyempapu [4], a non-profit community center. They are
highly respected by the community members and have been
extraordinary hosts during our visits.
II. PHASE 1: OBSERVATION
The first goal of this trip was to find locations greatly
affected by indoor air pollution. Through the CREATE Lab’s
global network of partners, established over many years of
working with communities on prior projects, we found
contacts in Uganda, and Rwanda. They welcomed us to visit
and observe their communities. Because of the language
bar r i e r s in t h ese c o u n t r i e s, th e majority of th e
communication with the families during our trips was done
using our contact as a translator. They would tell the families
why we have come to their home and relay any questions
they may have.
In August 2012, two engineers spent several days in both
Uganda and Rwanda. The contact in Uganda runs a small
non-profit in the rural villages of the Makukuulu parish. The
families typically have small cooking buildings, separate
from their sleeping quarters. There is no infrastructure for
electricity or running water. The kitchens were typically
made of brick, with either a thatch or metal roof. Several
bricks may be missing from the area above the fire to create
some air movement and ventilation, though we found the
airflow to be insufficient to remove most of the particulate
matter. This left black soot covering all surfaces of the
kitchen. Biomass is the primary source of fuel, such as left
over plant material from their crops (especially corn cobs).
2. Typical building for cooking in the Makukuulu Parish
Our contact in Rwanda was in the city of Kigali. The
homes in and around the city were comparatively more
modern, including some with running water and electricity.
Most families used coal as a fuel source, which typically
burns cleaner and more completely.
Out of the two locations, we surmised that the villages we
visited in Uganda would benefit most significantly from an
effective ventilation solution, especially given the lack of
easily obtainable coal needed for most cleaner-burning
cookstoves. Of equal importance, we felt that the level of
open communication and enthusiasm we observed from the
community would engender a more collaborative approach
to the technical development.
While each family differed, we observed a general set of
family member roles in cooking. Kitchen responsibilities
such as food preparation, cooking, and cleaning almost
exclusively fell to women, though children were often
involved as well (either assisting with tasks or being present
as a simultaneous responsibility of the cook). Matches were
expensive and difficult to obtain, with the villages located
several hours from a city center. This meant the children
would often be tasked with keeping the fire burning 24 hours
a day. Additionally, since the homes often did not have
electricity, when the children did their homework it would be
next to the fire. Both of these cases lead to additional
elevated exposure to the harmful smoke.
Since the fires were often kept low, kerosene lamps and
candles were also used for lighting sources while cooking.
This sometimes lead to wax or oil contaminating the food
without their knowledge. With limited food and time to
recook, the families would typically choose to eat the food to
avoid waste and hunger.
These observations, while limited in scope to the
communities and homes we visited, allowed us to
confidently proceed with the first prototyping phase to assist
these families.

III. PHASE 2: INITIAL PROTOTYPES
After returning from the Phase 1 trip, we decided to build
and deploy six prototype units in the homes we visited as an
initial proof of concept. Once in the field, we wanted to
gauge acceptance of the units, justify any additional effort
we put towards the project, and measure any improvements
of air quality. We would also solicit feedback on the devices
throughout their functional lifespan in order to improve the
next prototypes in durability and effectiveness.
Since the initial prototypes were built by hand, it was
important that they were easy to assemble and did not have
too many custom parts. The first prototypes consisted of a
fan unit and a separate power supply.
The fan unit was built using stacked components made of
laser cut acrylic, assembled using screws. This allowed for a
strong, three-dimensional shape while being lightweight and
easy to assemble (and replicate with the necessary
fabrication equipment). In the center was a 120mm computer
fan, surrounded by 6 white LEDs to provide illumination of
the kitchen and cooking area. Switches were used to control
the fan and lights separately.
A loop in the housing was intended to provide an easy
installation method using rope and/or wall hooks. In some
homes we envisioned a rudimentary pulley system with a
counterbalance to help reposition the device using its large
handles.
The fan unit was equipped with a USB plug for power.
This allowed for it to be easily tested and to run directly off
a USB battery pack. Both the fan and LEDs run directly off
5 volts with small in series resistors to limit the current
through the LEDs.
On the back of the unit was affixed a 4-inch diameter hose
adapter. This allowed us to attach a 4-inch aluminum drier
hose to the back of the fan and run the other end outside.
Since the air is being evacuated outside, no filtration was
installed.
We sourced a solar powered USB battery charger as an
integrated power solution. This allowed the user to leave it
on the roof to charge with a USB cable routed into the
kitchen for power to the fan.
To pick the fan, we attempted to find a quiet, and easy to
source computer fan. We then discovered by lowering the
voltage to the fan from 12 volts to 5 volts, we can consume
less overall power, yet still pull air from a small confined
area. From there, the enclosure was designed to house the
fan, allow the exhaust pipe to attach to the rear and hold the
LEDs. We then performed multiple tests outside in a
makeshift shed with a small fire. We concluded that the fan
and lights would provide enough suction and light for the
intended purposes.
3. Six initial prototypes brought to Uganda
In June of 2013, we brought six prototypes to the villages.
Unfortunately due to unexpected travel complications, only
approximately 48 hours was spent in the village. The first
day was spent meeting the six families and installing the
units. On the second day, we revisited the families to see if
everything was working properly, to gather as much
feedback as possible, and answer any questions the families
might have after their first use.
We relied heavily on our local contact for translation,
without whom the language barrier would have been
insurmountable. Communication was still challenging due
to the unfamiliarity of our contact with the technical aspects
of the device. Nominally, additional time in the villages and
with our contact may have alleviated this issue.
Another challenge we faced was the batteries themselves.
The units were not designed to be waterproof or significantly
water-resistant. They were installed in transparent bags to
provide some level of protection, but the seal was not
perfect, and the charging efficiency of the units was reduced.
The batteries also had to be manually switched on. After the
users finished for the evening and turned the fan and LEDs
off, the battery would enter sleep mode and turn off. In order
to turn the system back on, the user needed to press a button
on the battery to turn it on. These issues prevented any
permanent installation of the battery, as the user would need
to access it to turn it on and bring it indoors during inclement
weather. The problems with the batteries added to the
difficulty of teaching the users and exposed aspects of the
design that would need to be greatly improved.
This first pilot allowed us to assess the technical
effectiveness of the device in reducing indoor air pollution.
Using an HHPC-6+ handheld particulate counter, air quality
was measured during cooking and five minutes after turning
the fan on. The results are in figure 4. Anecdotally, when we
arrived at one particular home to inspect the unit on the
second day, the family was cooking, but the fan was not on.
We entered the kitchen to help turn the fan on but
experienced significant difficulty breathing with our

unaccustomed lungs. After several minutes with the fan on,
we were able to hold a conversation with the family inside of
the kitchen without coughing. We concluded from our
measurements and observations that a relatively small
amount of airflow could significantly improve indoor air
conditions if installed and used appropriately.
4. Comparison of air quality when using the ventilation system
An unpredicted outcome we observed was the positive
impact of the light. We expected that it would help with their
cooking, but in the short time we were in the village, we
observed unanticipated events such as a woman creating
crafts while cooking. Crafts are often a form of income for
the women we met. The light allowed them to multitask
during the less active phases of cooking (when, for example,
a pot needs to be occasionally stirred).
The community was, as before, extremely welcoming and
patient with us. While some families were initially confused
by the purpose of the ventilation systems, they were
graciously accepting and appreciative of our efforts to work
with them. An unexpected side-effect of their overwhelming
politeness was that it was sometimes difficult to gather
critical feedback, in part because of the language barrier, and
in part because of what we perceived as a reluctance to
criticize or potentially hurt our feelings.
This barrier is still present, albeit to a lesser extent than
before. While we have not solved this difficult issue, we
intend to consciously frame future projects in a way that
more strongly emphasizes co-development and reduces the
apparent risk of critical feedback being taken personally.
With the aid of our community contacts, we were able to
obtain one of the original prototypes six months after
installation. The bearing in the fan was greatly clogged by
soot and debris, preventing the fan’s rotation. After a quick
cleaning, the fan was able to freely turn once again. This
demonstrated the importance of not only carefully selecting
components for durability, but also teaching local
community members how to troubleshoot and fix the units
when possible.
IV. PHASE 3: SECONDARY DEPLOYMENT
The next step after the initial six-unit proof of concept
deployment was a larger deployment. We wanted to build on
what we learned, and try to make it as sustainable as
possible. It was very important to us that we worked more
with the locals to install and maintain the units. This would
hopefully empower the community to maintain and ideally
improve the design without having to rely on outside aid.
The first major hurdle for this deployment was
fundraising. Given the recent success of crowdfunding, we
chose to create and publicize a campaign for this phase of
the project. After a year of redesign and testing, we launched
a fundraiser online looking for $10,000. Approximately half
of the budget was allocated to building 30 units, while the
other half was for travel, compensation for community
members who would maintain the units, and payment for
lodging and any resources consumed while in the
community.
We were able to raise about half of the targeted amount in
the first three months, with the remainder contributed by a
local foundation with connections to the CREATE Lab.
Using the lessons learned from the first prototype trip, we
hoped to improve the device in several aspects. One major
challenge was the size of the ventilation system. The original
prototypes were sturdy and robust, but we felt were too large
to transport to Uganda without spending several thousand
dollars in shipping cost. The largest component of the
original system was the exhaust tube. We replaced the
aluminum drying hose to a small diameter plastic tube,
which was also less expensive and easier to route in the
homes. Additionally, we hoped plastic tubing would be
easier to source while in country.
Another change to the design was the fan itself. The
newer fan has sealed bearings to keep the dust and dirt from
clogging them. Like the original prototypes, the new
prototypes have LEDs for lighting while cooking and
independent switches for controlling the LEDs and fan. This
feature received positive feedback during our first trip, and
the LEDs appeared to be a strong incentive for individuals to
first use the system.
We replaced the off-the-shelf solar battery with a lower-
cost custom solution. A solar panel is still used for charging,
but the newest design utilizes a stand-alone panel with
batteries and charging circuitry kept within the “remote” part
of the system within the home. This potentially allows the
system to be serviced and repaired, unlike the previous
sealed system purchased online. The system was easily
disassembled such that individual components could be
replaced. We predicted that the most common components
to replace would be the three rechargeable AA batteries and
the fan itself.
We also designed the system so that variable lengths of
electrical cables and air hoses could be used depending on
the needs of an individual house. The “remote” piece was to

be mounted by the door of the kitchen, so the system could
be turned on immediately upon entering. This housed the
batteries, charging circuitry, as well as the switches for the
fan and LEDs. The solar panel was originally designed for
charging phones. After affixing a USB cable shielded by
weatherproof epoxy, the panel would be secured to the roof,
with the other end of the cable plugging into the “remote”.
Using a USB cable for charging made it simpler to test both
the battery charging system and the power output of the
panels. The last part of the system is the “head” unit. This
includes the fan and LEDs. It is designed with mounting
holes to be secured above the fire. After the “head” unit and
“remote” are secured, the two are connected using
thermostat wiring or other readily available conductors. The
tubing is then cut to length and run outside from the “head”
unit.
5. Prototype for the second deployment
During the parts selection process, several in-home tests
were conducted. These mainly consisted of testing different
solar panel / battery / fan combinations to approximate the
operating time of the system from a full charge. The solar
panels provided approximately 5W to three 2000mAh
batteries. The fan consumed approximately 200 milliwatts,
and it was estimated that the fan and LEDs could run for 16
hours on a full charge. Assuming adequate sunlight, we
estimated that it would take less than a full day to charge the
batteries fully, hopefully providing for brief periods of
clouds or rain. The fan itself was also tested inside our
homes to check the flow rate. While the fan was
significantly smaller than the original 120mm computer fans,
the flow rate was similar.
After funding was secured, we began assembly and
preparation. Two engineers from the CREATE Lab
independently traveled to the community in February of
2018. These two trips were separated for logistical reasons,
but provided the advantage of a longer overall interaction
with the community.
Investigation during the first trip focused on parts
sourcing. In order for this and similar projects to make a
sustainable and empowering impact, components must be
sourced as locally as possible. We found that while the
hosing, fans, and other mechanical parts are inconsistently
available in different local marketplaces, solar panels and
general electrical supplies were significantly more prevalent
than observed on prior trips. Systems designed for a small
home can be purchased for approximately 50 USD in several
places around the city. While this is still too expensive for
many families, the momentum toward independent home
solar power is evident.
The second trip focused on installation, with four days
spent in the villages. This allowed the engineer to train 3
workers to replace components like the fan or batteries if
they fail, and these workers were left with a supply of
critical replacement components. They were also trained on
the usage of the ventilation units, how to build the cables,
and how to install the systems. After just two hours of
training, they were able to install 13 units in approximately
10 hours over two days. The newly trained community
members performed all of the installation work in addition to
training the home owners how to use the systems. The
installations were successful with no apparent issues
requiring intervention. After both trips by the engineers, the
remaining systems were installed by the community
members without supervision.
Unfortunately, testing of the efficacy of the ventilation
units in removing indoor air pollution was inconclusive.
Since returning, it has been reported by our community
contacts that the units have remained functioning four
months later without failure. While a lack of quantitative
evidence limits our conclusions, we hope to receive
additional feedback regarding any qualitative changes in
observable air pollution or quality of life, including whether
the devices continue to be regularly used.
V. FUTURE DIRECTIONS FOR BREEZE
Before conducting larger pilots, it would be beneficial to
run more tests in country to get quantifiable results on the
effectiveness of the fan and the solar charging system itself.
This should include using multiple systems with different
fan sizes, as well as different solar panel and battery sizes.
Locally-purchased solar panels and batteries should be
included and prioritized in this study. This will ensure we
have more concrete results and a path to sustainability.
Once a configuration of fan, battery and solar panel is
selected, local businesses and manufacturers should be
approached in order to form a partnership which could
develop into a locally-run industry if the demand and
available resources permit. Some of the existing solar kits
include lights or other accessories. Approaching these kit
manufacturers to make a fan accessory is one potential
possibility.
Other things to consider is how to get the local workers as
involved as possible. This could be done in the design
process itself, market research, and evaluation of the
prototype units in home. Local workers know how things
work in country, whether that is building things efficiently

with the parts available, speaking with families for market
research ensuring we get honest answers, finding target
pricing, and communicating effectively to families about the
struggles they endure using prototypes. This will also help
spread knowledge of the project as well as the problems we
aim to solve.
Once larger pilots with a final design are conducted, it
will show us how often maintenance is needed on these
types of systems. We will then know if service plans are
needed or if maintenance can be done by home owners.
VI. CONCLUSIONS
We note several lessons and challenges encountered
during these first phases of the project. First are the benefits
of an observation-first approach. When the project began, it
was unclear why there was not a local prevalence of cleaner
burning cook stoves, but quickly learned upon arrival that
the community’s distance from the city made it impractical
to acquire the necessary fuel. Additionally, cooking
practices are established and centralized in every culture,
and the community must be empowered with the decision of
whether or not to radically change these practices. We
approached the impact of air pollution using an additive
strategy to existing practices rather than attempting to
replace them. Our first observations also revealed that LEDs
would be a very beneficial and relatively low cost addition to
the system. The kitchens we visited were often dark with
very few windows, and even during the day it may be hard
to see. By adding the LEDs, the system has a secondary and
immediate benefit. While health improvements may not be
immediately apparent, the light source may provide a reason
to continue using the system.
Without our contact in the community, this work would
not have been possible. Our contacts leadership allowed us
to borrow the initial legitimacy and trust needed for us to
better understand and collaborate with the users. While in
Uganda, our contact also provided necessary translation and
organized transportation, lodging and food. Mutual trust and
continuous communication during and between our efforts
have been the foundation of this project.
Our contact’s guidance also enabled us to navigate
situations that we were unprepared for, especially since this
was our first experience attempting to assist with air quality
in a culture we were unfamiliar with. In addition to helping
us to better understand cultural norms and facilitating
communication, the community leader provided support as
we acclimated to a very different day-to-day life for the
relatively brief time we were present.
Travel and logistics can be frustrating in unfamiliar
circumstances where it is difficult to be self-sufficient,
especially when accustomed to the rapid, impersonal
transactions enabled by the internet and large commercial
establishments. We have found that this difficulty lessens
with patience and travel, regardless of where we have
visited.
Designing and testing the system can be difficult because
of distance, available resources, and climate. Nominally, we
would always co-design solutions which are manufactured
and serviced locally, but matching this ambition with the
necessary relationships and resources can be very difficult.
By far, the most important challenge (and one which we
are still learning to address) is to discard assumptions, be as
culturally aware and respectful as possible, and conscious of
the time and resources expended by the community to
accommodate the efforts of researchers and engineers. The
impact of projects like Breeze must be judged in relationship
to how these resources (social, monetary, or otherwise) could
otherwise be used to improve quality of life. Justification
requires meeting a verified (rather than hypothesized) need,
and maximizing impact may often require minimizing
involvement. The ultimate goal is empowerment and
independence, which means we must constantly prepare and
strive for complete autonomy of the community.
VII. OTHER PROJECTS
During Joshua’s most recent visit, he has become aware of
additional problems faced by rural Ugandan communities.
Two specific topics are waste management and the lack of
shoes for children.
With many of the villages hours away from the closest
recycling facilities, anything made out of plastic is left to
collect dust in piles on the side of the road, scattered in
fields, or burned (which releases toxic airborne pollutants).
It was observed that 10-20% of the school children in these
specific villages do not own shoes. Their parents often
cannot afford them as they are subsistence farmers. These
children may walk several miles a day to and from school, as
well as to get water for their families from distant and scarce
water sources. These children suffer from cuts and scrapes
by accidentally walking over broken glass or nails during
these long daily walks. Even worse, most of these children
suffer from soil-transmitted diseases including infestation by
chigoe fleas. These fleas infest the feet and make it
extremely painful to walk and play, potentially leaving the
children as social outcasts or limiting their ability to travel to
school.
Moved by these stories, Joshua has now begun a new
project called Eco-Shoes. The goal of Eco-Shoes is to take
the plastic waste in the villages and rework it into sandals for
the children using solar-powered injection molding
machines. This project is designed to clean up the plastic
waste, give children protection for their feet, and create jobs
for local craftspeople. One avenue of exploration is to set up
a payment system for the collection of plastic. If paid per
pound of plastic, anyone could contribute materials and
benefit financially. The craftspeople building the shoes
would be supported with equipment, training, and
compensation with the goal of creating sustainable and
independent businesses with positive social impact.

References
1. “Household Air Pollution and Health.” World Health
Organization, World Health Organization, www.who.int/en/
news-room/fact-sheets/detail/household-air-pollution-and-
health
2. “Particulate Matter (PM) Pollution.” EPA, EPA, https://
www.epa.gov/pm-pollution/particulate-matter-pm-
basics#PM!
3. Torres-Duque, C., et al. “Biomass Fuels and
R e s p i r a t o r y D i s e a s e s : A R e v i e w o f t h e
Evidence.” Proceedings of the American Thoracic Society,
vol. 5, no. 5, 2008, pp. 577–590., doi:10.1513/pats.
200707-100rp
4. Kitanda Sub-county. “KYEMPAPU.” Kyempapu,
kyempapu.org!