ukchinaenergyTLEE

Modelling retrofit changes
to the housing stock:
simulating decision making
and capturing geographic variation
Dr Timothy Lee
@MyBCU
www.facebook.com/birminghamcityuniversity
Introduction
• Challenging 2050 target
– 80% reduction in CO2 emissions
• Housing accounts for c.28% of current
emissions
• Therefore changes needed to existing
housing if target is to be met
• Stock models projecting uptake of energy
saving measures can aid planning towards
the 2050 target
@MyBCU
www.facebook.com/birminghamcityuniversity
Shortcomings of
current modelling
• No geographic variation in model stocks
– Eg: City centre flats have different retrofit
options to off-gas rural properties
• Uptake rates of retrofit measures are
imposed
– No consideration of the decision making of the
dwelling owners who will actually drive the
uptake of measures and therefore the energy and
CO2 savings
@MyBCU
www.facebook.com/birminghamcityuniversity
Geographically based
housing stock model
• Develop a test model for the North East of
England (1.1 million homes)
• Model dwelling stock variation across the region
Output area (OA) (c.150 homes)
-> Lower level super output area (LSOA) (c.700)
-> Medium level super output area (MSOA) (c. 3500)
-> Local authority (LA) (c. 80,000)
@MyBCU
www.facebook.com/birminghamcityuniversity
Housing data
• English Housing Survey data on 935
dwellings in the North East
– Provides sufficient data for calculation of energy
demand
– Decile IMD rating and 4 level ruralness scale
• Census Data - Output Area (OA)
–
–
–
–
Built form/detachment
IMD Decile
Ruralness
Heating fuel type
@MyBCU
www.facebook.com/birminghamcityuniversity
Geographic energy data
• DECC:
– Experimental data for gas and electricity use at
LSOA level
– Data for gas and electricity use at MSOA level
and LA level
– 1657 LSOAs in North East in 2011 Census
– 341 MSOAs in North East in 2011 Census
– 12 LAs in North East
@MyBCU
www.facebook.com/birminghamcityuniversity
Geographic model results
@MyBCU
www.facebook.com/birminghamcityuniversity
Understanding and
Modelling Decision Making
• Retrofit measures only installed when
owners decide to install
• Owner-occupiers
–
–
–
–
Energy Saving Trust (2009)
Element Energy (2008)
Discrete choice survey data
Multiple Criteria Decision Making Model
• Agent based model
@MyBCU
www.facebook.com/birminghamcityuniversity
Understanding and
Modelling Decision Making
• Simple additive weighting
Available technologies:
– Fabric Improvements, Heating System Upgrades,
Renewable Energy Technologies
Considers:
– Price, Saving, Maintenance, Disruption, Subsidy,
Taxation, Recommendation
@MyBCU
www.facebook.com/birminghamcityuniversity
Adoption curves
7000
6000
5000
Cond ×
Combi ×
Regular ×
Oil ♦
Electric ×
Solid ♦
Community ♦
GSHP ×
ASHP ×
4000
Total installed
3000
2000
1000
2015
@MyBCU
2025
Year
www.facebook.com/birminghamcityuniversity
2035
2045
Next steps
• Improve behavioural understanding
• Expand to all tenure types
• Integrate agents with geographic
model
@MyBCU
www.facebook.com/birminghamcityuniversity
Thank you
Any questions?
[email protected]
@MyBCU
www.facebook.com/birminghamcityuniversity