Energy Modeling of Danish Housing Stock Using System
Dynamics
Reza Fazeli!, Brynhildur Davidsdottir?
1 School of Engineering and Natural Sciences, University of Iceland, Email: rfazeli@ hi.is
? School of Engineering and Natural Sciences, University of Iceland, Email: bdavids@hi.is
Abstract
Implementation of energy efficiency measures will play a crucial role in future energy
consumption in buildings. In this study, we developed a dynamic building stock model, which
captures the effects of building aging on total energy consumption for space heating demand.
We also studied the behavioral dynamics by focusing on the key processes affecting the
implementation of retrofitting measures. Calibrating the model to Danish housing stock, five
policy scenarios are evaluated in terms of total energy consumption and total costs.
Implementation of cost-effective measures resulted in 11.0% reduction in energy consumption
for space heating by the end of century. It was also found that implementation of mandatory
measures that are cost-ineffective can increase the energy saving by 3.3%, while incentives can
only affect the timing of retrofitting and not the amount of saved energy. Restricting the rebound
effect was found to be an effective policy (23.4% reduction in 2100). Total energy saving has
significantly increased to 48% in 2100, when tight building regulation was introduced in 2020.
We also compared five policy scenarios based on their total costs. Significant increase of costs
in incentive- based policy is driven by “free rider” effect. Therefore, the estimated costs should
be interpreted cautiously.
1, Introduction
The motivation for a shift in our current energy path and reduction of emissions of greenhouse
gases is substantial. Due to slow building stock tumover the majority of opportunities to
improve energy efficiency over the next several decades is in the existing building stock, that
is constrained by old equipment and aging infrastructure. Energy efficient and low-carbon
technologies will play a crucial role in the energy revolution needed to make this change
happen. According to the IEA (IEA, 2013a), the buildings sector is the largest energy-
consuming sector, accounting for over one-third of final energy consumption globally and an
equally important source of carbon dioxide (CO2) emissions.
In the Nordic countries, the buildings sector used 1527 petajoules (PJ) of energy in 2010, or
about 33% of total energy use, which is similar to the worldwide share of energy use despite
cold climates. The Nordic countries have progressively reduced the role of fossil fuels in the
buildings sector as well as increased the energy efficiency of buildings, by implementing
various polices including financial incentives, awareness campaigns, energy certificate
systems, a system for certifying qualified experts in addition to implementing strict building
codes (IEA, 2013b). Because building stock turnover in the Nordic countries is slow (on the
order of 1% per year), the majority of opportunities to improve efficiency over the next several
decades will be in existing building stock, most of which is constrained by old equipment, aging
infrastructure, and inadequate operations resources. However the potential is significant, and
the challenge is to “unlock” that vast potential and realize the benefits of a built environment
that is comfortable, efficient, and cost-effective.
Some of the key obstacles to studying the above mentioned inertia in detail are the analytical
difficulties considering the interrelated processes of dwelling age, new construction, demolition
and retrofitting. Particularly, the retrofitting process is directly related to the social dynamics in
the system, which is rarely studied at a system level (Santin et al., 2009). This study presents a
dynamic simulation model that captures the impacts of physical processes such as aging, as
well as social processes such as familiarity of households with retrofitting. The paper is
structured as follows. In the following section, we describe the scope and structure of the
simulation model as well as the calibration with Danish data. Then we assess the impacts of
five policy scenarios on energy consumption and greenhouse gas emissions. The contribution
of this study is the creation of a simulation model that enables the assessment of long-term
impacts of different retrofitting policies in terms of energy savings and emissions accounting
in the context of building stock tumover in Denmark. The generic model could though be
applied to any housing stock in the Nordic countries.
2. Model
To investigate the dynamics of the dwelling stock we have used the system dynamics (SD)
approach. SD is extensively applied in the study of dynamic systems by representing them as a
set of interrelated stocks, flows and feedback mechanisms and simulating their temporal
evolution (Forrester, 1969; Groesser and Ulli-Beer, 2007; Sterman, 2000). This study improved
the model developed by Y tcel, (2013). Using the developed system dynamics model, we aim
to explore the effectiveness of certain policy options that can alleviate the inertia of an existing
dwelling stock. In this study, the system scope is set to Danish housing stock and its energy-
related renovation, i.e. refurbishment that improves building energy performance (in contrast
to renovation without impacting on energy use, or simple maintenance). The buildings are
classified according to three construction periods (pre-1960, 1960-1980, and post-1980). The
temporal scope of the model is set to the period 1990-2100.
2.1. Feedback Loops
The final energy consumption of a household depends on five factors: building energy
requirement, household income, heating-degree-days, energy expenses and technological
progress. The causal loop diagram that depicts the relation between key factors and household
energy consumption is given in Figure 1.
Figure 1: Causal loop diagram of energy consumption in dwellings
The developed model incorporates two fundamental household behaviors driven by the
improvement of building energy performance and changes in energy expenses. These behaviors
are shown as balancing loop for efficiency improvement (B1) and balancing loop for energy
conservation (B2) in Figure 1. In this analysis, four balancing loops have been identified as key
challenges that limit the gain from retrofitting options. They influence the energy performance
of dwellings as well as the attractiveness of retrofitting measures:
e Efficiency improvement (B1)
With the increase in household energy consumption, and consequently household energy
expenses, households’ willingness for retrofitting increases. After implementing
retrofitting measures, the energy consumption is expected to decrease, as well as the ratio
of energy expenses to income per household. According to this balancing loop, the
willingness for retrofitting diminishes as the energy efficiency level of a dwelling
increases.
e Energy Conservation (B2)
With the increase in household energy expenses, the ratio of energy expenses to income
for households rises. The tendency of the households to renovate is related to two factors:
the perceived level of energy expenses to household income, and economic profitability of
retrofitting. After implementing retrofitting measures, the ratio of energy expenses to
income for households decreases. Therefore, the household will increase the intensity of
energy consuming activities as a result of increasing income (or decreasing cost of
consumption), which directly corresponds to the rebound effect broadly discussed in the
energy consumption literature.
¢ Saturation of retrofitting potential (B3)
Another factor that drives the retrofitting is the perceived economic profitability of
retrofitting. As more renovation is done in buildings, the building energy efficiency
improves which reduces the potential for further improvement, which consequently
reduces the potential money saving from retrofitting. Therefore the potential for efficiency
improvement saturates and the tendency/ability of households for renovation incrementally
declines.
e Saturation of market (B4)
The rate of retrofitting depends directly on the potential adaptors. So it will saturate as
retrofitting measures are implemented in the dwellings.
These balancing loops counteract the interest for retrofitting by reducing the willingness of
households to renovate as well as profitability of energy saving measures as the energy
efficiency level of a dwelling increases. However, there are three reinforcing loops in favor of
more renovation in buildings:
e Leaming through experience (R1)
It’s expected that with the implementation of retrofitting options in dwellings, the
cumulative experience increases, which subsequently decreases the investment cost of
options. The reduction in investment cost will increase the economic profitability of
retrofitting options.
e Development of retrofitting (R2)
In this model, to identify the expected economic gain from retrofitting, the expected saved
money from retrofitting was normalized by being divided by household energy expenses.
Therefore, with the reduction in energy expenses due to the renovation, the relative
economic gain from retrofitting increases, which increases the rate of retrofitting.
e Social Exposure (R3)
As more buildings are being renovated, households are getting more familiar with the
experience and due to the effect of word of mouth, they get familiar with the main purpose
and achievable benefits of building renovation. With the increase in familiarity of
households, the rate of retrofitting increases.
2.2. Coflow Structure
Sterman (2000) pointed out that system dynamic modelers often need to capture not only the
total quantity of material in a network of stock and flows, but also the attributes or
characteristics of the stocks. While the stock and flow network reflects the amount of material
in the fundamental stock, it does not reveal anything about the characteristics of that stock.
Coflows however can be used to keep track of the attributes of the items that are flowing
through the stock and flow structure. As a result, coflows are parallel structures that can be
“used to account for the attributes of items flowing through a stock and flow network”.
Average residence
time for ouflow
Average building
lifetime
— at Stock LA sey = .
Inflow + Outflow >! dwolings a
. “Construction — Demolition
erage -
Attribute —_ ‘Energy —
| ry } Intensity
+ \te ‘Total 2) tu . + 5
SS PD
Increase in Attribute | Decrease in > sy Energy *€ S
attribute attribute Increase in energy | Requirement | Decrease in energy
if requirement requirement
Marginal Ae
attribute per unit
= Marginal Energy
Intensity
Figure 2: Generic coflow structure
Figure 2 (Left, Sterman, 2000) illustrates the generic coflow structure. As each unit of the
fundamental “Stock” increases the quantity of that stock, a unit of the associated attribute is
added to “Total Attribute” stock. The “Marginal Attribute per Unit” is simply the number of
units of the attribute added to the “Total Attribute” stock, for each unit the fundamental “Stock”
is increased. As the number of units of the fundamental “Stock” is reduced through the outflow,
there is a corresponding decrease in the number of units of the “Total Attribute”. The number
of units by which the attribute is decreased is the product of the “Average Attribute” quantity
and the “Outflow” rate, where the “Average Attribute” quantity is the ’Total Attribute” quantity
divided by the quantity of fundamental Stock”.
5
The right side of Figure 2 presents an example of how we use this standard coflow structure to
model the energy use in dwelling stock. The fundamental stock is the number of “Dwellings”
in Denmark. The underlying attribute is the amount of “Energy Requirement” for space heating
and cooling where the “Increase in Energy Consumption” is the product of the “Construction”
of new buildings and “Marginal Energy Intensity”. The “Decrease in Energy Consumption” is
estimated by multiplying the rate of building “Demolition” by the average “Energy Intensity”.
2.3. Aging Structure
As was mentioned earlier, in this study, the dwelling stock is divided into three age groups:
¢ Old Buildings: buildings more than 30 years old
.
Mature Buildings: buildings between 10- 30 years old
New Buildings: building less than 10 years old
Figure 3 shows a stock-and-flow diagram of the key components of our dwelling stock model.
Stock-and-flow-diagrams are used to represent the structures of a system in close relation to the
equations that are actually simulated. With the three stocks and the aging rates, an aging chain
for dwelling stock was formed (the middle chain in Figure 3). The original energy requirement
and energy requirement (accounting for retrofitting impact) are the main coflows (the first and
third chains in Figure 3).
Original Energy | Original Energy | i Silos Ener val
a Requirement for Ri -
Tnerease in Buildings Increase in| Mature Buldings | !nereasein |__Buikings | Reduction in
OENBB OEMBs {OFOBs. ‘OERs
Original energy
intensity of NBB
Average original
if ecety OF ¢ energy intensity of OBs
Construction ft Vejagel Vintage 2 Vintage 3
Delay 5 \ “ etme y
° } |
Construction
Original energy
Mature Buildings - Old Buildings: |
pM [pwn |
Demolition
Buildings
Growth rate
Eneray Intensity Energy Intensity of Energy Intensity
Mature Buildings Or euldnge
ER of Building New Buildings ¥ Mature Buildings jnts for Old Building:
:
ER of
Demolition
5 ER of Aging 2
Retrofiting in
Marginal Eneray MBE:
Retrofitting in
MB
Intensity of NB
Retrofitting in
0B
Q )
o
Figure 3: Stock-and-flow-di.
‘am of the building sector
2.4. Retrofitting
Retrofitting with the aim of energy savings is typically made as an integrated part of other
renovations, and considering the wear and tear of e.g. roofs or windows. However, the rate of
retrofitting has so far been slow in Denmark, as the majority of the houses still lack sufficient
energy efficiency, e.g. insulation. This is partly because aspirational types of renovations (in
the kitchen and bathroom) are prioritized over energy saving renovations (Gram-Hanssen,
2014). In this study, the impact of four factors that contribute to decisions concerning energy
retrofitting of existing dwellings have been studied:
e Perceived economic attractiveness of retrofitting: Retrofitting is considered to be
economically attractive when the normalized saving index (the ratio between the
lifetime energy avoided cost of a retrofitting option and energy expenses) exceeds a
certain expected level, which could vary by households.
e Familiarity with retrofitting: Familiarity is a key factor that directly affects the retrofit
decision-making process. Word-of-mouth effect played an important role in the
implementation of retrofitting options. The larger the stock of retrofitted buildings, the
greater the exposure to and knowledge of that experience among potential adopters,
increasing the chances that households will consider and renovate their houses. The
familiarity can also be improved by informative policies and marketing.
e Willingness to renovate: Willingness to renovate represents more than simple
familiarity. Many people are aware of the achievable gain through retrofitting, but do
not take them seriously in their decision. In this study, it was assumed that the
willingness to renovate increases when the ratio of saving from energy improvement to
income level goes above the motivation threshold.
¢ Potential adaptors: Rate of retrofitting depends directly on the potential adaptors. The
larger the stock of retrofitted buildings, the fewer the potential adapters, reducing the
rate of retrofitting.
3. Case Study: Danish dwelling stock
To study the energy performance of Danish dwelling stock, the developed system dynamics
model needs to be calibrated with the historical data. In this study, 1990 was selected as the
base year and the data collected from Danish Energy Authorities (2014) between 1990 and 2012
was used to calibrate the system dynamics model. Three types of dwellings are studied: Single
Family Detached House (SFDH), Single Family Terraced Houses (SFTH) and Multi-dwelling
Houses (MDH). The distribution of dwelling stock based on construction year 1990 and by type
7
of the dwelling is shown in Figure 4. Dwellings built before 1960 account for 70.0% of total
stock and considering their low energy performance, the focus of energy improvement plans
should be on retrofitting these old buildings.
u SFDH Pre 1960 a SFDH 1960-1980 = SFDH post 1980
u SFTH Pre 1960 uw SFTH 1960-1980 @SFTH post 1980
um MDH Pre 1960 = MDH 1960-1980 = MDH post 1980
Figure 4: Distribution of Danish dwelling stock by type and construction year 1990 — source:
statistics Denmark
As explained in the previous section, the dwelling stock is divided into three groups according
to their age. The evolution of Danish dwelling stock by type of buildings is shown in Figure 5.
1800000
1600000
1400000
1200000
1000000
800000
600000
400000
200000
0
QO cd cb > oh cd cP co ce D O
PSY S?_ oF 8? © gh) SP M_gS
SST SSS SST ST SST
+ DO oo O
SPW PF ws
Qo
© ad ad
$ Py’
os
mDetached houses Terraced houses = Multi-dwelling houses
Figure 5: Total number of Danish stock by type — Source: Statistics Denmark
3.1. Evolution of Building Regulations in Denmark
Denmark is one of the first EU countries to set up their national Zero Energy Building (nZEB)
definition and roadmap to 2020. Based on the Building Regulations 2010 (BR10) (The Danish
Ministry of Economic and Business A ffairs, 2010), the total demand of residential buildings
for energy supply for heating, ventilation, cooling and domestic hot water per m? of heated floor
8
area must not exceed 52.5 kWh/m’/year plus 1650 kWh/year divided by the heated floor area.
The minimum energy performance requirements from set building regulations will gradually
become stricter, starting from the actual standard, BR10, with an interim milestone in 2015 and
a final target in 2020 (table 1).
Table 1: Evolution of the primary energy performance requirements towards nZEB levels in Denmark
(COHERENO, 2013)
Residential buildings 52.5 + 1650/A* 30 + 1000/A
2
(housing sector and hotels) FiMuwh ie kWhim*/yr 20 kWhim'/yr
Minimum
requirement | Non-residential buildings
(offices, schools, hospitals,
others)
71.3 + 1650/A 41 + 1000/A
kWhim’/yr kWhim’/yr
25 kWhi/m*/yr
3.2.C alibration of the model
The developed model is used as a dynamic tool that can be used to explore different scenarios
to improve our understanding regarding the interactions of various processes. While doing so,
the model’s correspondence with the Danish dwelling sector with regard to the size of the
dwelling stock and socio-economic characteristics is maintained. Before the scenario analysis
phase, the model is tested for the validity of its structure. The historical data on the evolution
of building stock and energy consumption in dwellings from 1990-2012 was used to calibrate
the developed system dynamics model.
The model is initialized based on actual data corresponding to year 1990, and the 1990-2012
period (a period about which reliable data was accessible from Kragh and Wittchen, (2014) is
used for behavioral comparison purposes. Model-generated behavior for the stock of dwellings
per type compared with the actual data can be found in figures 6-8. As can be seen from the
plots, the model is able to capture the general trends.
1990 1994 1998 2002 2006 2010
~ 1240000 1240000
© 1200000 X 1200000
** 1160000 x 1160000
1120000 x 1120000
i 1080000 1080000
1040000 1040000
F 1000000 1000000
5 960000 ~, 960000
SEGSESESTSERRRERRRRRS
——SFDH Model xX SFDH Obs.
2011
201
Figure 6: Number of dwellings - Single-Family detached houses (model vs. observation)
1990 1994 1998 2002 2006 2010
250000 250000
z x
® 200000 x 200000
un
@, 150000 * 150000
=|
g 100000 100000
8 50000 50000
5 0 0
ot iso} oan
Z ZESRSRRSRRSSSSSSSSSSEEE
ee 3 NNNAN NANANAN
——SFTH Model x SFTH Obs.
Figure 7: Number of dwellings - Single-Family Terraced houses (model vs. observation)
1990 1994 1998 2002 2006 2010
x 100,000 100,000
90,000 ¥ x 90,000
5 80,000 x x 80,000
4 70,000 70,000
£ 60,000 60,000
50,000 50,000
40,000 40,000
3 30,000 30,000
20,000 20,000
10,000 10,000
ow ise} no b fo} 0
=a a4
Seq
SHURRERGGRRRREERRRERRAR
—MDH Model x MDH Obs.
Figure 8: Number of dwellings — Multi-Dwelling houses (model vs. observation)
The outcome of the developed model on energy consumption by type of dwellings is also
compared to the collected data in the validation period. Based on figures 9-11, the outcome of
the model is acceptable.
25,000
q
& 20,000
15,000
10,000
5,000
Energy Consumpti
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
——Energy Consumption in SFDHs Obs. Energy Consumption in SFDHs Model
Figure 9: Energy consumption for space heating in Single-family detached houses in GWh-
model vs observation
6,000
6 5,000
5
8 4,000
3,000
§ 2,000
i ”
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
—Energy Consumption in SFTHs Obs. ——Energy Consumption in SFTHs Model
Figure 10: Energy consumption for space heating in Single-family Terraced houses in GWh-
model vs observation
14,000
12,000
F 10,000 ———_eooe ee
8,000
6,000
4,000
2,000
0
Energy Consumption inGWh
b= se jel by co NO jel i our nN
SEGPPLESTT SERRE RR RRR RES
—Energy Consumption for MDHs Obs. ——Energy Consumption for MDHs Model
Figure 11: Energy consumption for space heating in Multi-dwelling houses in GWh- model vs
observation
11
3.3. Retrofitting Scenario
During the last decade, several standards and regulations regarding energy consumption of
buildings have emerged with the focus on new buildings, specifying increasing levels for
energy efficiency requirements. However, these standards provide less guidance on the
renovation of existing buildings that will have to face similar challenges in the near future.
Appendix 6 in the Danish Building Regulations 2010 (BR10) contains a summary of measures
that are often cost-effective to implement in relation to existing buildings undergoing
renovation. Based on that guideline, the primary energy demand exclusive of renewables in
kWh/m2 annually is reported in the SBI 2013 report for different building types, year of
construction and heat supply fulfilling the energy requirements in BR10 for existing buildings
undergoing major renovation. We have estimated the required energy saving in single-family
houses and multi-dwelling houses to fulfill the energy requirements in the Danish Building
Regulations (table 2).
Table 2: Retrofitting potential ing to building ion 2010 — own
Building type and age category Energy saving potential
Single-family house — Old Building 51.3%
Single-family house — Mature Building 36.4%
Multifamily house — Old Building 53.7%
Multifamily house — Mature Building 35.5%
The contribution of the present work is to study the long-term impacts of effectiveness of
various policies in terms of energy savings from buildings in Denmark, considering the
estimated energy saving potential in table 2.
4. Results & Discussions
After calibrating the model with the historical data, the model was used to simulate the Base
scenario. In the Base scenario, it is assumed that households do not invest in energy-related
renovations, and they do not conserve energy. Figure 7 demonstrates the change in the energy
intensity of the three dwelling type-groups, as well as the average of the whole stock (the yellow
line) in the Base scenario. Because of the construction of more efficient dwellings, and the
demolition of inefficient old ones, the energy intensity levels are on the decline.
—
-)
S
b/sqm
Bee
SES
sss
Energy Intensity in
-
aes
Seas
a
MDH === Aggregate Energy Intensity
Figure 12: Projected energy intensity for three dwelling type-groups, as well as the average of
the whole stock
Then, we were interested in analyzing the impact of policy scenarios in terms of energy
saving and required investment.
First of all, we studied the scenario in which the proposed energy requirement for
buildings becomes effective in 2020. Total energy consumption for space heating in base and
BR20 scenarios are compared in Figure 13.
50,000
45,000
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
‘Total Energy Consumption for
Space Heating in GWh
——BR20 ——Base
Figure 13: Total energy consumption for space heating in GWh
According to figure 13, the energy consumption for space heating in residential buildings
is expected to decrease by 2.5% in 2030 and by 22.5% at the end of century. This insignificant
reduction (0.3% per year) clearly indicates the importance of the extent of inertia caused by the
existing building stock against an energy transition in the residential sector in Denmark.
To affect the stock turnover, there are two possible policies that should be assessed in
terms of energy reduction potential; one is the scenario with a high construction rate (EU
average level) and the other one is the demolition policy with a high demolition rate (+40%
more than the historical trend). In order to see the potential effectiveness of such policies, the
construction rates and the demolishing rates of all dwelling types are varied after 2015. The
impacts of these two policies on space heating demand and energy intensity compared with the
base scenario is shown in Figure 14.
2030
Space heating demand Energy Intensity
3.0%
2.0%
:
0.0%
-1.0%
-2.0%
-3.0%
-4.0%
-5.0%
mHigh construction rate = High demolishing rate
Figure 14: Impact of high construction and demolition rate on energy consumption for space
heating and average energy intensity of building stock in 2030
Figure 14 shows that the changes in total space heating demand in 2030 are +2.4% and -
3.7% in high construction and high demolition scenarios, respectively. However, both policies
are effective in reducing the average energy intensity of building stock by more than 3% until
2030. This figure clearly shows that demolishing old buildings is significantly more effective
in reducing the space heating demand compared to the investment on construction of new
efficient buildings.
Next we run a set of scenarios to assess the potential of retrofitting. In the “R1” scenario the
implementation of retrofitting measures was possible from 2015. According to BR10, if the
annual saving from a retrofitting option multiplied by the lifetime, and divided by the
investment, is greater than 1.33, that measure is deemed to be cost-effective and will be
implemented. It basically means that applying that measure will pay for itself within 75% of its
expected lifetime. In the “R2” scenario, we let the model choose from non-cost effective
measure to reach the higher energy saving level (80%). Then, in the “R3” scenario, we included
the financial incentives of 100 $/sqm for households. Figure 15 demonstrates interesting
dynamics in these three retrofitting scenarios.
14
50,000
45,000
40,000
35,000
30,000
1990 2010 2030 2050 2070 2090
Time (Year)
Total heating ion : Base
Total heating ion :R1
Total heating jon :R2
Total heating on : R3
Figure 15: Comparison of total energy consumption for space heating in three retrofitting
scenarios and base scenario
Compared to the base scenario, implementation of cost-effective measures will reduce the
energy consumption for space heating by 10.7% by the end of the century, which is
unexpectedly low. This trend is due to the impacts of the balancing loops that were discussed
previously. Although the energy reduction was significant in the “R1” scenario at the beginning,
the reduction trend vanishes after 2030. This behavior can be justified by the increase in energy
consuming activities as a result of an improvement in the efficiency of buildings (the “rebound
effect”). In the “R3” scenario, the more aggressive retrofitting plan resulted in reducing the
total energy demand for space heating by 3.3% compared to the “R2” scenario. Providing the
incentives to households increases the rate of retrofitting after 2030, but the differences in
savings with “R2” disappears by the end of century. To analyze the overall impacts of rebound
effect, in a hypothetical scenario of “R4’”, we disconnected the link for the second balancing
loop (B2). Figure 16 depicts the differences in total energy consumption in retrofitting
scenarios.
30,000
45.000
40,000
35,000
30,000
1990 2010 2030 2050 2070 2090
Time (Year)
Total heating ion : Base
Total heating on : R3
Total heating ion :R4
Figure 16: Total energy consumption for space heating in three retrofitting scenarios
15
The differences in total energy consumption between the “R3” and “R4” scenarios represent
the increase in energy consuming activities motivated by rebound effect. Compared with the
scenario with rebound effect, the reduction in energy consumption for space heating has
increased to 6.1% and 9.2% in 2030 and 2100, respectively. In the integrated scenario “Il”, we
assumed that the proposed building regulation will be in place in 2020 and this condition was
added to the “R4” scenario.
2030 2100
0%
10% "ll
-20%
-30%
-40%
-50%
-60%
mBase wR1 =R2 #R3 wR4 wll
Figure 17: Energy consumption reduction by percentage in various scenarios
The significance of having tight building regulation in 2020 on reducing the energy
consumption is clear in figure 17. Although the differences in 2030 are not significant, we can
observe that the saving potential becomes much more evident by the end of the century.
At this point, a comparison of the total costs for retrofitting in different scenarios is critical to
have a better understanding of the effectiveness of policy scenarios. In this study, total costs
covers the investment cost for retrofitting to the households and the costs of incentive policies
to the policy makers.
Total Investment
80B
60B
40B
208
0
1990 2010 2030 2050 2070 2090
Time (Year)
otal investment: Base Tout tnvesiment
Toa lavestnat: RI Tua vestnet Ra
‘Toual Investment: RQ. Total Investment
Figure 18: Comparison of total cost for five policy scenarios in USD
According to Figure 18, the total investment costs for cost-effective measures in “R1” scenario
can reach 18.38 and 27.66 billion USD in 2030 and 2100, respectively. As expected with having
an incentive policy in the “R3” scenario, the total cost dramatically increases compared to “R2”.
But, considering the fact that the proposed retrofitting measures by building regulation are cost
effective, we believe that this huge investment of 75.05 billion USD until 2100 is driven by the
“free rider" effect. Free riders are people who would invest in retrofitting anyway, so they
shouldn’t receive the incentives. Therefore, it’s critical to keep in mind that the estimated total
costs should be interpreted cautiously. Besides, Figure 18 shows the reduction in total cost by
having the building regulation in place in 2020 by 2.7% in 2100.
5. Conclusions
As the building stock tumover is slow in the Nordic countries (on the order of 1% per year),
the majority of opportunities to improve efficiency over the next several decades will be in
existing building stock, most of which is constrained by old equipment, aging infrastructure,
and inadequate operations resources. One of the key obstacles to study the inertia in detail is
the analytical difficulty considering the interrelated processes of dwelling age, new
construction, demolition and retrofitting.
In this study, we develop a dynamic building stock model, which captures the impacts of
physical processes such as aging. We also illustrate the importance of behavioral dynamics by
focusing on the key processes affecting the implementation of retrofitting measures: word of
mouth, social exposure and the willingness of the consumer. Initially, we compared the impact
of a high construction rate with a high demolition rate.
It was found that although the total energy consumption has slightly increased in the case of
higher construction, they are both effective in reducing the average energy intensity of building
stock by 3% until 2030. In this analysis, four balancing loops have been identified as major
challenges that restrict the implementation of retrofitting measures, and also the benefit from
these efforts. Calibrating the model to Danish conditions, five policy scenarios are compared to
a base case focusing on energy consumption and total cost.
Primarily, we estimated that the implementation of cost-effective measures can reduce the
energy consumption for space heating by 11.0% by the end of century, which is surprisingly
low. Although the energy reduction was significant in the “R1” scenario at the beginning, the
reduction trend vanishes after 2030. Due to the insignificant saving, the “R2” scenario with
17
aggressive retrofitting measures, and a scenario with financial incentive “R3” were analyzed.
Total energy consumption for space heating is expected to fall by 3.3% and 3.2% in “R2” and
“R3” compared to “R1”. Then, a scenario “R4” was studied where the impacts of rebound effect
are limited. Compared with “R3”, the reduction in energy consumption for space heating in
“R4” has increased to 6.1% and 9.2% in 2030 and 2100, respectively. The significance of
having tight building regulation in 2020 on reducing the energy consumption was assessed in
“TI” scenario and the estimated energy saving in 2100 is 48.0%.
We have also compared the scenarios in terms of total costs. According to Figure 18, the total
investment costs for cost-effective measures in the “R1” scenario can reach 18.38 and 27.66
billion USD in 2030 and 2100, respectively. As expected with having incentives for households
in the “R3” scenario, the total cost dramatically increases compared to “R2”. But, considering
the fact that the proposed retrofitting measures by building regulation are cost effective (in the
“RI” scenario), we believe that the significant share of the 75.05 billion USD investment until
2100 is driven by “free rider" effect. Therefore, it’s critical to keep in mind that the estimated
total costs should he interpreted cautiously.
References
COHERENO, 2013. nZEB criteria for typical single-family home renovations in various
countries.
Danish Energy Authorities, 2014. Energy statistics 2014.
Forrester, J., 1969. Urban dynamics. MIT Press.
Gram-Hanssen, K., 2014. Existing buildings - Users, renovations and energy policy. Renew.
Energy 61, 136-140. doi:10.1016/j.renene.2013.05.004
Groesser, S., Ulli-Beer, S., 2007. The Structure and Dynamics of the Residential Built
Environment: What Mechanisms Determine the Development of the Building Stock?, in: 25th
International Conference of the System Dynamics Society.
IEA, 2013a. Transition to Sustainable Buildings - Strategies and opportunities to 2050. Paris,
France. doi:10.1787/9789264202955-en
IEA, 2013b. Nordic Energy Technology Perspectives. Paris, France.
18
Kragh, J., Wittchen, K.B., 2014. Development of two Danish building typologies for residential
buildings. Energy Build. 68, 79-86. doi:10.1016/j.enbuild.2013.04.028
Santin, 0.G., Itard, L., Visscher, H., 2009. The effect of occupancy and building characteristics
on energy use for space and water heating in Dutch residential stock. Energy Build. 41, 1223-
1232.
Statistics Denmark — StatBank.dk/BOL101"
Statistics Denmark — StatBank.dk/BOL103"
Sterman, J., 2000. Systems Thinking and Modeling for a Complex World. McGraw Hill.
The Danish Ministry of Economic and Business Affairs, 2010. Building Regulations.
Copenhagen, Denmark.
Yicel, G., 2013. Extent of inertia caused by the existing building stock against an energy
transition in the Netherlands. Energy Build. 56, 134-145. doi:10.1016/j.enbuild.2012.09.022
