Best paper nominee, presented at Ubicomp 2014
Authors: Jacky Bourgeois, Janet van der Linden,
Gerd Kortuem, Blaine A. Price and Christopher Rimmer
Contact: jacky.bourgeois AT open.ac.uk
Abstract:
Domestic microgeneration is the onsite generation of low- and zero-carbon heat and electricity by private households to meet their own needs. In this paper we explore how an everyday household routine – that of doing laundry – can be augmented by digital technologies to help households with photovoltaic solar energy generation to make better use of self-generated energy. This paper presents an 8-month in-the-wild study that involved 18 UK households in longitudinal energy data collection, prototype deployment and participatory data analysis. Through a series of technology interventions mixing energy feedback, proactive suggestions and direct control the study uncovered opportunities, potential rewards and barriers for families to shift energy consuming household activities and highlights how digital technology can act as mediator between household laundry routines and energy demand-shifting behaviors. Finally, the study provides insights into how a “smart” energy-aware washing machine shapes organization of domestic life and how people “communicate” with their washing machine.
1. Conversation with my Washing Machine:
An in-the-wild Study of Demand Shifting with
Self-generated Energy
Jacky Bourgeois, Janet van der Linden, Gerd Kortuem,
Blaine A. Price and Christopher Rimmer
1
In collaboration with
2. 2
Electricity generation with solar panels
alters people’s relationship with energy
“Energy farmers”
3. Local Energy Generation is Complex
• Self-generated energy is used
locally or is exported to grid
• Additional energy is imported
from the grid if required
• Import costs are higher than
export payments received
• Generation incentive payments
vary by country
“optimizing” energy use in the
home is complicated
3
Solar Photovoltaic (PV)
Generation
Export
To the grid
Import
From the grid
Self-consumption
4. Local Energy Generation is Complex
“Energy Gap”: Consumption and
local generation are out of sync
• Generation and consumption
vary during day
• Generation and consumption
vary by weather and season
• Typically generation peaks
around midday, consumption
peaks in early evening
4
Electricity Profile of household #12 on 7 May 2013
(Consumption vs Generation)
5. Previous Research
• Most ubicomp and HCI energy
research has focused on
consumption and demand
reduction
• “Double-dividend of solar
generation” [Keirstead 2007]:
households adopt new energy
saving practices
• “Looking out of the window”
[Price et al 2013]: householders
estimate weather impact to shift
demand
5
6. What role can Ubicomp technology play
in enabling or supporting new
energy practices in households with
solar generation?
Specifically: demand shifting
6
7. Case Study: Doing Laundry with Washing Machine
7
Laundry practices and washing
machine use is good case study:
• Everyone needs to wash clothes
• Involves whole family
• Temporal constraints (deadlines)
• Environmental impact
• Emerging demand-shifting
practices
by Gloria Garcia
8. “In-the-Wild” Study with Households
Objective
• Understand household practices
• Explore design alternatives for in-home
technology
Scope
• 8 Months
• 18 households
• 64 participants
8
9. Study Methodology
• Home instrumentation
• Participatory energy data analysis
• Design and deployment
of technology interventions
• Qualitative studies:
• Home visits
• Interviews & focus groups
• Thematic analysis
9
10. Study Methodology: Energy Data
• 20M data points over 2 years
• Household electricity generation
• Household electricity import
• Household electricity export
• Washing machine use (timing
and electricity consumption)
• Other appliances (timing and
electricity consumption)
10
14. #1 Delayed Energy Feedback via Email
14
• Participants received email with
summary energy report few days
after they have used the washing
machine
• Report outlines:
• Predicted solar energy
generation for next 5 days
• Past daily generation and
washing machine use
• Idea: enables householders to
reflect on behavior and plan
future washing machine use
15. #1 Delayed Energy Feedback via Email: Findings
15
• Users did not engage with energy
reports, neither in a positive nor
negative way
• Interpretation:
• the gulf between email and
real family life is too large
• Planning of washing machine
use is not something that is
done on the computer
17. #2 Real-time Feedback via SMS Text Messages
• Participants received SMS a few
minutes after washing machine
use
• 'You ran your washing machine
at 15:45 today (3.7% green). You
could have achieved 43.6% by
starting it at 10:34.‘
• 'Congratulations! You ran your
washing machine at 13:48 today
(65% green). The expected
maximum for today was 71%.'
17
18. #2 Real-time SMS Feedback: Findings
• “Just saying ‘your washing used
63 percent of solar’, that’s in
itself is not really useful to us.”
• “unless you’re going to keep all
these text message and analyse
them, you are not going to get
that information.”
• “It’s like shooting in the dark!”
18
20. #3 Proactive Suggestions via SMS Text Messages
• Participants received a SMS
message at a time they had
chosen. This message:
• Suggests best time of day to
run washing machine during
the next 36 hours
• This involved predicting solar
energy generation for each hour
of a day and uses past weather
and generation data, and local
weather forecast
20
21. #3 Proactive Suggestions: Findings
• Very positive response from
participants
• Some participants followed
suggestions
• Even if participants did not
follow the suggestions they
appreciated that the
information was there for
them
21
22. #3 Proactive Suggestions: Findings
• Huge diversity across households
– where each family wanted to
receive their proactive message
at a different time
• Many requests for changes to
mobile phone numbers for the
messages, thus involving more
members of the household
22
24. #4 Embedded Control
• Display and interactive control near the
washing machine which was actually
controlling the machine and receiving
feedback (Zigbee)
• Shows best time to use washing machine
• User can select auto-start at best time
• User can select constraints for start and
end time
24
25. #4 Embedded Control: Findings
• Mostly positive reactions
• Actionable information at
right time and right place
• Participants suggested many
refinements:
• Start time should
continuously adapt to current
weather
• The system should pause the
washing machine when a
cloud passes
25
26. #4 Embedded Control: Findings
• New laundry practices:
• load machine in the morning,
set to auto-start, leave for
work
• Appropriation:
• Participants used Information
about best start time to
manually control other
appliance (dish washer)
26
27. Conclusion
1. Technology support for demand-shifting is viable and effective
• Supporting emerging practices, not behavior change
2. Engagement and utility increased from
• decontextualized information -> embedded contextual control
(i.e. email -> washing machine display)
• retroactive feedback -> proactive suggestions
3. Decisions about timing of washing machine use is negotiated
through “conversations with my washing machine“
4. Future work: from one appliance to many appliances
27
28. Conversation with my Washing Machine:
An in-the-wild Study of Demand Shifting with
Self-generated Energy
Jacky Bourgeois, Janet van der Linden, Gerd Kortuem,
Blaine A. Price and Christopher Rimmer
In collaboration with
Editor's Notes
Good Afternoon everyone
I m Jacky, Phd student at both the Open University in the UK and Université de rennes 1 in France
And I’m going to present you a study we’ve done with Janet van der Linden, Gerd Kortuem, Blaine Price and Christopher Rimmer
in collaboration with eon, a major energy company.
Increasingly people have solar panels on their roof and are effectively producers of electricity. Or electricity farmers.
Also, there is evidence that people with such solar panels think differently about energy.
However
… And that peak of consumption is mostly made of interactive consumption, appliances that requires user interventions.
So far Most ubicomp and HCI, energy research has focused on consumption and demand reduction
Only few have looked at local generation, yet there are opportunities for supporting emerging practices
Previous research has by and large assumed that by making people more aware they can change their behaviour. Whereas in this research we want to give people active tools to help them achieve these new practices.
Keirstead observed what he called the Double Dividend in the context of local generation. This means that people not only produce green energy but also reduce what they are consuming.
We also found in an earlier study that householders are creating simple manual ways, such as looking out of the window, to estimate weather impact because they want to shift their demand.
we chose to focus on one appliance – the washing machine.
It is a good example of an interactive appliance that has a central place in the household and the household routines.
Doing the laundry is a household activity that has many subtle routines and habits and that vary highly between households.
In that study, our objective was to…
We conducted this study over 8 months, with 18 participating households around Milton Keynes in the UK
All had solar panels on their roof for at least a year
The households were quite different, from 2 to 5 members, stay at home versus going to work, with and without children, retired etc
We monitored electricity in these houses,
we analysed the data with participants
Then we designed and deployed technology interventions based on emails, text messages and electronic tablets
We collected our qualitative data through home visits, interviews and focus group and we performed a thematic analysis
What did we monitor:
LOTS of DATA!!
We knew everything about these people. Minute by minute.
The generation coming from solar panels
The import from and export to the grid
Data from a specially constructed washing machine, using zigbee for communication and control. We knew every wash they did, what time, which type of cycle etc
And we also monitor other appliances through a variety of smart plugs
We had scheduled interviews at regular intervals, but also we had lot of contact with the householders in-between for help and technical assistance. Many home visits were made
There were thus many opportunities to gather additional and informal qualitative data.
So here is our intervention plan in 4 stages around different time and place
In the remaining of this presentation, I will detail each of these interventions
First, we explored delayed energy feedback via email
What was very interesting is that people did not comment at all about these emails.
We think that this is because people don’t read their email as when they are doing their laundry. Often in a separate room and at a very different time.
Then we moved onto real time feedback via text messages
So for our second intervention – we asked for participants phone numbers so that we could send them a message straight after each wash was finished.
The message would tell them how well they’d done. If they’d done well they were congratulated – but otherwise they might get a suggestion for what would have been a better time.
By and large people enjoyed receiving the text messages. A number of people said they enjoyed getting a message from their washing machine. It was funny and unusual to have a conversation with your washing machine
However, they felt that the message was coming too late. People were not able to change anything. Particularly if they had a poor score there was no opportunity to improve it as it was after the fact.
Ironically when we stopped sending these messages, people kept asking to have them back.
Then we moved onto a more proactive suggestion, BEFORE the washing machine load
When we asked participants when they’d like to receive these messages, we got a huge diversity of answer across households
For example, I want it at 7 in the morning because then it fits in when we’re getting ready for work and school.
I want it the night before – say around 8 in the evening
I want it every day.
Only twice a week
We also got many requests for changes to mobile phone number, as we were not sending messages to the right person in the household. They wanted to make sure the message would go to the persons who were mostly involved with doing the laundry, often not the same person who is mostly concerned with the household energy concerns. So this meant more people were becoming involved in participating in the study and in thinking about the tools needed to support them.
Finally, we went onto embedded control
We set up an electronic tablet near the washing machine
Showing the best time to use the washing machine
The user can select auto-start at best time and select constraints for start and end time
Then the application was actually controlling the washing machine and receiving feedback
We also observed the emergence of new laundry practices. People would say ”I load my machine in the morning, set to auto-start, then leave for work“
We also noticed that people appropriated the tool for their own purposes. So they would use the information for the washing time to manually switch on other high consumption appliances. For example, manually switch on the dishwasher. Even the hot tub!