Content from this work may be used under the terms of theCreativeCommonsAttribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
1
Plus minus zero: carbon dioxide emissions of plus
energy buildings in operation under consideration of
hourly German carbon dioxide emission factors for
past, present and future
A Studniorz 1, D Wolf 3, N Kiessling 3, R Fahrich 2, C Banhardt 3and
G Tsatsaronis 1
1Technische Universit¨at Berlin, Chair of Energy Engineering and Environmental Protection,
Marchstraße 18, 10587 Berlin
2Technische Universit¨at Berlin, Hermann-Rietschel-Institut, Marchstr. 4, 10587 Berlin
3HPS home power solutions, Carl-Scheele-Straße 16, 12489 Berlin
E-mail: [email protected]
Abstract. The energy supply of private household buildings accounted for 16 % of the total
German CO2-emission in 2020. To fulfil the targets of a climate neutral building sector in 2045,
both, energy efficiency as well as on-site use of Renewable Energies in buildings are needed. One
concept of a climate neutral building is the so-called Efficiency House Plus, that features large
photovoltaic systems making it seemingly energy self-sufficient and CO2-negative by feeding
in more electric energy into the grid than needed for its operation on a yearly basis. In fact,
houses of this type are highly grid dependent especially during winter months due to their solely
electrically based energy supply and a missing long term energy storage. This paper analyses the
CO2-emission of Energy Efficiency Plus houses more in detail on a timely resolved basis for the
German electric supply system of the year 2013, 2021 and a perspective one 2030. An alternative
calculation approach for simplified normative evaluation of such buildings is proposed.
Keywords: single family house, CO2sustainability assessment, energy exchange, dynamic
building simulation, energy system analysis
1. Introduction
The climate protection target made within the Paris Agreement [1] to limit the increase in the
global average temperature possibly below two degrees Celsius and, if possible, to 1.5 degrees
Celsius, above the pre-industrial level were the basis of the Federal Climate Change Act [2].
This so-called “Bundesklimaschutzgesetz” (KSG) was revised in 2021 [3] This reduction goal
is targeting the emissions during the operation of buildings, which makes sence regarding the
building stock where operation energy is the main GHG emission contributor. However, a recent
study [4] with an analysis of more than 650 building life cycle assessment (LCA) results show
that the so-called embodied greenhouse gas (GHG) emissions are more and more dominating
the life cycle of new buildings. Embodied emissions stand for the emission of manufacturing and
processing of building materials. Thus it becomes obvious, that the whole life cycle of buildings
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
2
needs to be analyzed to assure their climate neutrality. Since embodied GHG emissions are
unavoidable, the decision whether a building is climate neutral or not, depends on the (negative)
GHG emissions during operation.
One building concept that seemingly achieves a CO2negative operation is the so-called
“Effizienzhaus Plus” (Efficiency House Plus (EHP)) [5]. EHPs feature large photovoltaic (PV)
systems, offering an annual amount of electric energy that exceeds the electric and heating
energy demand within a year. The remaining surplus of PV energy is fed into the electric grid.
The assumption regarding this exported PV energy is, that it substitutes grid electricity and its
corresponding CO2emission. While the approach is straighforward, the central question arises
about the corresponding CO2emissions of the grid electricity being substituted by the exported
PV energy.
In the absence of a detailed data basis, most CO2calculations apply a uniform and constant
CO2emission factor to account for both, electric energy drawn from and fed into the electric
grid, although recent studies [6, 7, 8] show that this approach can distort the actual proportion
of CO2emission and CO2relief of energy exchange between buidlings and the electric grid.
This paper reviews the influence on the CO2emission calculation in operation of EHPs when
using hourly CO2emission factors of the electric grid in comparison to the common static
approach.
2. Background and Method
To understand how EHPs reach a negative CO2emission in operation it is essential to clarify
their general working principle as well as the way of how energy and CO2balances are applied.
After that the dynamic CO2emissions of grid electricity in Germany are analysed, representing
one part to reveal the net CO2emissions of EHP. The last part needed to achieve this goal is a
rigorous building simulation model offering reliable data on how the grid usage of EHPs evolves
throughout a year.
2.1. Efficiency Plus Buildings and their CO2emission rating
As per definition EHPs reach negative values of primary and final energy demand. EHPs offer
low space heating demand that is supplied most of the times by heat pumps. In addition,
they contain a large PV system, that generates more electrical energy per year than needed for
electricity and heating requirements. This setup is favourable in general, since more PV capacity
and sector coupling is needed to enhance the energy transition in Germany.
However, energy supply and demand are not always such that a direct supply can be achieved
solely by the local PV system at all times. First EHPs reached temporal shares of 30-50 % in
which the energy demand was supplied by the PV-System only, whereas latest example show
values of up to 70-80 % incorporating batteries as short-term energy storages.
Thus, EHPs are fundamentally grid dependent due to two reasons. On the one hand, to achieve
reliable energy supply for the remaining part of a year, on the other hand to be able to make use
of the PV surplus energy by feeding it into the electric grid. From this point of view it becomes
clear that the ‘plus’ in energy of EHPs is time dependent. While during summer they produce
an energy surplus, EHPs strongly rely on grid electricity in winter times.
The approach used in [5] to estimate the CO2emissions of EHPs follows DIN 18599-1, so that
CO2emissions of imported from and exported to the grid energy can be summed up into one net
CO2emission. The DIN 18599-1 allows the same equations for CO2emission calculation as for the
primary and final energy demand calculation. As concrete values for accounting CO2emissions of
both, electricity drawn from and fed into the electric grid, a CO2emission factor of 550 g CO2
/kWh
is given, representing the annual average value of electricity within the German electric grid in
2014. With an average specific PV energy surplus of 20 kWh
/m2·aand following the declaration
of 550 g CO2
/kWh as the constant CO2emission factor, EHPs save 12 kg CO2
/m2·a.
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
3
2.2. Hourly based CO2emission factor of the German electricity mix
The actual German power plant system covers the national electricity demand under constantly
changing operation of fossil and renewable energy sources. Hourly data on power plant operation
are available in several publicly accessible databases (e.g. Fraunhofer ISE Energy-Charts,
Agorameter). Technology-specific CO2emission factors (e.g. [9]), as shown in Table 1, allow
the absolute emissions of the German power plant system to be calculated from the hourly-
resolved power plant input, as well as a specific CO2emission factor of the current electricity mix
based on the electricity demand covered. The CO2emission factor can be used to characterize
the CO2intensity of electricity production in the face of fluctuating electricity demand. This
hourly CO2emission factor is shown for the years 2013 and 2021 and a future scenario 2030
in Figure 1 (a). The diagram in Figure 1 (b) contains the same data points but shown as a
carpet plot to better illustrate daily and weekly changes of the CO2emission factor. High and
low values of the CO2emission factor can be assigned using the adjacent colorbar.
Clearly recognizable is the ’photovoltaic ellipse’ with the main axis February-October and minor
axis 08:00-16:00 at the end of June. Likewise, isolated vertical yellowish and green bars (’wind
turbine towers’) can be recognized mainly in March and May, which are caused by constant
highwind power feed in times lasting for days. Nevertheless, it must be noted that in summer
nights with little wind, the CO2emission factor reaches high values, because almost all electricity
demand must be covered by fossil energy sources. And finally fossil power plant supply big shares
of the electric energy demand during November, December and January. Within Figure 1 (a)
Figure 1. Hourly based CO2emission factor of the German power plant system for the years
2013, 2021 and extrapolated scenario 2030; with data from Agorameter
two phenomena can be described. On the one hand, the average level of the CO2emission factor
continues to decrease from 2013 (Ø 573 g CO2
/kWh) through the year 2021 (Ø 422 g CO2
/kWh) to
the future scenario 2030 (Ø 233 g CO2
/kWh). On the other hand, the range of values of the
CO2emission factor increases from 2013 to 2021/2030. Both phenomena can be explained
by the growing share of renewable energies. In 2013, the share of renewables in electricity
generation was about 10 %, while in 2021 it was already 42 % and is assumed to be 68 % in the
2030 future scenario [10]. With an increasing share of renewable energies, specific CO2emissions
decrease as fossil power plants are used less. On the other hand, the range of values of the
CO2emission factor increases, since renewable energies no longer only support the electricity
supply in an additive manner (2013), but take over relevant shares (in 2021) or even completely
(in 2030), whereby alternating large shares of renewable and fossil power plants are used and
explain the large differences in the CO2emission factor. Thus, in 2021, there were times when
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
4
Table 1. CO2emission factors and shares of fossil power plant types for the future scenario 2030
Power plant type CO2emission t CO2
/MWh Cap (GW) Share(%)
Gas 0.37 40.5 80
Coal 1.09 3 6
Lignite 0.82 3 6
Others 1.50 4 8
Fossil power plant system 2030 0.53 50.5 100
the electricity demand was still almost exclusively covered by fossil fuels (ca. 600 g CO2
/kWh)
and times when a CO2emission factor of only 150 g CO2
/kWh could be achieved with the help
of renewable energy, which was not possible in 2013. For 2030, the CO2emission factor drops
to a value of 0 g CO2
/kWh at some times when all electricity demand is met by renewables. If
the supply of renewable energies is insufficient, the residual load in 2030 will still be covered
by a backup system consisting of conventional power plants. The supposed composition of this
backup capacity follows [10] and is shown in Table 1. Based on this breakdown, a weighted
emission factor for the backup power plants was determined (0.53 t CO2
/MWh).
2.3. Minute based heat and electricity supply of a single-family house
The linking element missing to utilize the hourly based CO2emission factors to calculate more
appropriate net CO2emissions is a precise, consistent, and reliable knowledge about the timely
resolved electric power demand of a single family house. This can be achieved by a rigorous
building model offering a minute-by-minute resolved thermal and electrical energy demand and
supply simulation of an EHP including energy exchange with the electric grid. In the following
the developed model named “piceaSim” including the energy system setup, namely models of
PV, inverter, battery, heat pump, thermal storage, underfloor heating system, and finally the
building model will be described briefly. Furthermore, model reliability is proven by a cross
validation for the building model with two other simulation software.
Modelling and simulation of piceaSim was performed in MATLAB/Simulink. PVLib [11] was
used to calculate incident surface radiation and power output, extended by a thermal time con-
stant model to reflect the module’s transient cell temperatures [12, 13] . Validation with the
commercially available software PV*SOL proved high accuracy at minutely resolution, resulting
in a coefficient of determination R2=0.99. Conversion efficiency curves for PV-to-battery,
PV-to-load and battery-to-load of a current domestic hybrid inverter were obtained from a se-
ries of laboratory measurements. The battery state of charge model applied ampere counting
method considering dynamic charge-acceptance. The battery voltage was determined through
a 2-D lookup table derived from laboratory measurements. The heat pump was modelled fol-
lowing a widely adopted performance map (e.g. manufacturers EN 14511 European standard
test data) based grey box approach including a lumped parameter thermal mass to account for
non-stationary behaviour such as cycling losses resulting from condenser or evaporator heat up
and cool down [14, 15]. Validation of these equation-fit models has been carried out in numer-
ous studies proving their well-suitedness for building energy simulations [16, 17]. The stratified
thermal energy storage is divided into two sections: The top part supplying domestic hot water
through an external heat exchanger (central freshwater station), the bottom part acting as a
buffer and hydraulic separator for the underfloor heating system. An adaptation of [18] using the
buoyancy algorithm from [19] was implemented, a description of the freshwater station model
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
5
can be found in [20]. The hydronic layout represents a common scheme for modern single-family
homes [21, 22, 23]. Heat pump and tank were sized according to European standard EN 15450
and recommendations of the national heat pump association [24]. The 5R1C model proposed
by ISO 13790 was extended for a second thermal capacitance comprising air mass. In addi-
tion, underfloor heating slabs were discretized more detailed as an electrical analogy (“Beuken”
model) [25] and coupled to the building zone following an approach presented and validated in
[26, 27, 28]. Heat delivery from carrier medium was reduced from a 3 dimensional to simplified
1 dimensional heat flow [29, 30, 31].The efficiency of ventilation heat recovery was calculated
utilizing the NTU method [32]. Measured historical weather datasets for the city of Potsdam in
10-minute resolution were acquired from DWD. Generation of minutely load profiles considering
season, weekday and cloud amount for household electricity and domestic hot water demand
was performed for each weather dataset according to national guideline VDI 4655:2021.
The 5R2C building model in MATLAB as part of piceaSim was cross validated with other ex-
isting simulation models to ensure its prediction quality for the building heat demand. For
validation, the architectural design of a prototype single-family house was used and three build-
ing models were implemented in MATLAB (piceaSim), Modelica (BuildingsLibrary) and the
industry simulation software IDA ICE for comparison. The models implemented in piceaSim
and Modelica were single zone building models, while the IDA ICE model was implemented as
a multizonal model to assess comparability for the single zone model to multizonal building be-
haviour. All parameters of the three building models were harmonized including the dimensions
of outer walls, windows, floors and roof, as well as the thermal properties of the correspond-
ing building components. For simulation the same weather and solar radiation data (TRY 2015)
were chosen, and air infiltration rates, air handling units, thermal bridge corrections and internal
heat gains were aligned. The resulting monthly heat demands are shown in Figure 2 (b). The
Figure 2. Cross validation of three building models by: (a) dynamic building heat load shown
for November; (b) monthly heat demand (both based on climate data time series TRY 2015)
calculated monthly heat demand for the piceaSim building model is comparable to the results of
the examined Modelica and IDA ICE models. The monthly result positions itself right between
the results of the other two models for the winter months. The cumulated yearly heat demand
results in 3228 kWhth for IDA ICE, 3664 kWhth for Modelica and 3464 kWhth for piceasim. For
the dynamic heat demand simulation, the piceaSim and Modelica models provide very similar
results over the whole year. In Figure 2 (a) the simulated building heat demand for all three
models is shown for the second half of November as an example. Since the IDA ICE model is im-
plemented as a multizonal model, heat demand is dampened due to the heat exchange between
rooms and added heat storage in partition walls. Another difference is the minimum heating
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
6
demand. Rooms with limited exposure to solar gains need to be heated in the multizonal model,
while the single zone models applies solar gains to the whole building volume, dropping the heat
demand to zero for the times solar gains overtake heat demand.
The results for the dynamic behaviour and monthly aggregated results for heat demand suggest
that the building simulation model in piceaSim provides very similar results compared to the
simulation models from Modelica. The single zone models perform similarly in monthly heat
demand simulation and dynamic behaviour. The results are comparable to industry simulation
software IDA ICE, although the heat demand of the multizonal model tends to be lower and
shows a behaviour for certain zones that are not represented in the single zone models. Still the
single zone model of piceaSim allows simulation within reasonable error ranges to estimate the
building heat demand. The building model used within in the simulations of this contribution
has a specific head demand of 38 kWhth
/m2·a.
3. Simulation
EPHs can be characterized by their average electricity surplus of around 20 kWhel
/m2·a[5].
Taking this average value into account a parameter study with PV-capacities varying between
8/10/12/14/16 kWpeak (180 ◦azimuth – South, tilt angle 30 ◦) and an annual household
electricity demand of 2500/3000/3500 kWhel still without thermal supply was perfomed to cover
a wider range of possible system setups including the one with an electricity surplus of around
20 kWh
/m2·a. Furthermore, all combinations mentioned were simulated for an EPH with and
without a battery as well as for the two weather data time series of 2013 and 2021 (weather
data time series of future-scenario 2030 equals 2021).
The simulated system behaviour ist presented in Figure 3 (a) for three days of summer and in
Figure 3 (b) for winter. It can be seen, that during summer almost no grid demand is existent,
although the household load is high. Due to high PV gains all electric demands are supplied
and in addition large amounts of PV surplus energy can be fed into the grid. This is different
in winter, where only small PV gains can be achieved and an additional heating demand must
be supplied. This leads to almost continuous need for grid power supply. For the same time
sections chosen to illustrate the system behaviour the timely varying CO2emission factor is
shown in Figure 3 (c) for summer and Figure 3 (d) for winter respectively. On basis of these
simulations the interconnection between times of grid demand (blue) and feed-in (yellow) to the
dynamic CO2emission factor are evaluated precisely throughout the year.
As shown exemplarily in Figure 3 the grid demand takes place in time sections of high
CO2emission factors whereas feed-in mostly takes place during times of low CO2emission factors.
To characterize a system setup, an average value of the CO2emission factor was calculated for
grid demand and feed-in considering the actual CO2emission factor weighted by the amount of
energy drawn from or supplied to the electric grid throughout the year. The results of all system
setups simulated will be shown in the following.
4. Results
Simulations were performed for multiple combinations by varying the four parameters of installed
PV-capacity, electricity demand, availability of a battery as well as different datasets for weather
and CO2emission factor. As a representative case the results of 12 kWpeak PV and 3500 kWhel
household electricity demand are selected and will be shown grouped by the three weather-
and CO2-datasets used and the two scenarios of EHP with (PV + battery) and without battery
(PVOnly). Simulation results of all other scenarios are represented by an additional range of
values illustrated by spread bars around the representative case. The presentation of results can
be divided into relative and absolute key indicators. As the central relative value, the weighted
average CO2emission factor was calculated for times of grid use and feed-in times. As absolute
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
7
Figure 3. Simulated EHP system behaviour and timely interlinked CO2emission factors for
(a,c) 3 days of summer and (b,d) 3 days of winter (EHP - PVBat - 12 kWpeak PV - 3500 kWhel
electricity demand) - 11 kWhinst battery
values of CO2emission for grid use and feed-in were calculated to estimate the net CO2emission.
The relation between the CO2emission factor for grid use and feed-in times for the three reviewed
datasets 2013, 2021 and 2030 is shown in Figure 4 for EHPs with and without battery.
It has to be emphasized that for both scenarios, PVOnly and PV + battery identical weather
Figure 4. Weighted average CO2emission factor calculated by dynamic simulation for grid
demand and feed-in times of EPH without (PVOnly) and with battery (PV+battery) for different
years
and CO2emission time series data is used.
Two main trends can be observed in both setups with and without battery. Firstly, a growing
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
8
Figure 5. Comparison of monthly electricity demand and CO2emission factor for variation of
PV capacity for EHP without and with battery show almost constant CO2emission factor for
electricity drawn form the grid in 2030
gap between the CO2emission factor of grid demand and feed-in times is clearly visible.
The difference accounts for scenario without/with battery up to 94/110 g CO2
/kWhel in 2013,
122/155 g CO2
/kWhel in 2021 and 168/202 g CO2
/kWhel in 2030. Moreover, the level of CO2emission
factor for both, grid use and feed-in, decreases over the years due to the decarbonization of
the electricity sector. To be highlighted is the only minor deviation of the simulated system
configurations regarding the relative CO2emission factors of EHPs with and without battery.
The central finding is, that the CO2emission factor of energy drawn from the grid is almost
equal throughout the year, implying a comparable share of fossil power plants in the grid during
summer nighttime and winter season (an insight matching Figure 1 (b)). Nevertheless, it can
be seen, that grid use mainly occurs during the winter months (1,2,11,12). The share of grid
power demand within these four months accounts up to 58-60 % for EHPs without and to 91-96 %
for EHPs with battery. EHPs with battery show lower grid use in summer months whereas grid
use during winter remains high.
The results of using the static method (DIN 18599) in comparison to the dynamic method
(introduced here) to account for the net CO2emissions of EHPs are presented in Figure 6.
Here, the absolute net CO2emissions in operation, calculated by the two different methods, are
presented in absolute numbers [kg CO2per year] for the datasets 2013, 2021 and 2030 for EHP
with and without battery. Negative values represent a relief of the global climate. Regarding the
static method (red bars) negative values are achieved by any EHP in every year. In contrast,
CO2-savings calculated with the dynamic method (green bars) reach significant lower values
throughout all datasets and system configurations simulated.
A remarkable deviation between the two methods appears within the datasets of 2030, where
net CO2emissions of EHP turn positive for some setups mirroring a net CO2emission despite
a surplus of PV energy supplied to the electric grid. This finding is in line with the increasing
gap of specific CO2emission over the years from 2012 to 2030 (compare Figure 4).
The spread bars shown in Figure 6 indicate the values of different system setups and illustrate
the strong influence of the PV-capacity and electricity demand on the results. All simulated
system setups shown reached a PV energy surplus which was fed into the grid. As a guideline
EHPs are said to reach 20 kWhel
/m2·aPV energy surplus on the average [5]. This information was
used for the parameter study with ranging results of PV energy surplus between 11-64 kWhel
/m2·a
for simulated EHP without battery and 6-59 kWhel
/m2·awith battery.
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
9
Figure 6. Comparison of calculated net CO2emissions for EHP in operation using the static
method (red, DIN 18599) and dynamic method (green, hourly CO2/,emission factor)
5. Critical Review
In this study we use timely resolved average CO2emission factors (AEF) for the German power
system in order to determine the net carbon emission of an EHP. This approach implies the
simplification that for each kWhel additionally fed into the grid, all power plants currently online
reduce their power contribution proportionally. In view of a fully functional wholesale day-ahead
power market, only the marginal power plant running at the highest marginal cost is supposed to
reduce its power instead. This relationship has been examined already in various studies defining
the term marginal emission factor (MEF) [33, 34, 35]. Making use of the MEF requires either
a fundamental model of the wholesale electricity market to identify the marginal power plant,
relying on a considerable number of input data, parameters and uncertain assumptions [34], or an
approximate model based on the marginal system response, which neglects must-run capacities
as well as (local) transmission system restrictions [33]. Development and/or application of such
a model is beyond the scope of this work. Furthermore, the MEF is determined based on a
purely economic model following the merit-order. Thereby fossil fuel must-run capacities as
well as (local) transmission system restrictions and measures are neglected. As a result, it
remains uncertain if the marginal power plant or other, even renewable energies capacities, will
be reduced in performance.
The effect of relying on AEF instead of MEF depends highly on the structure of the actual power
plant fleet bidding into the respective market. With the current German power plant fleet, open
cycle gas turbine often represent the marginal power plant. Their specific CO2emission of
approx. 370 g CO2
/kWh is rather in the range of the timely respective AEF in 2021 (compare
Figure 3 (c)). That is why in this case AEF and MEF deviate only little and bring comparable
results. For future scenarios with higher shares of renewable energy capacities, however, it can
happen that renewable power generation pose the marginal capacities. Additionally considering
fossil fueled must-run capacities, this leads to an overestimation of avoidable CO2emissions,
especially during summertime with both approaches. We therefore highly recommend to broaden
the presented methodology towards the concept of MEF. Here, MEF should be calculated with
an increased level of detail regarding the mentioned points above to especially account for future
scenarios with very high shares of renewable energy capacities.
6. Conclusion
In this paper the application of hourly CO2emission factors (dynamic method) for energy
surplus single-family houses (EHP) to estimate their CO2emission in operation, is introduced
and reviewed.
The two main findings when comparing the absolute CO2emissions calculated are, that, following
the dynamic method, CO2emissions turn out to be significantly higher than those calculated
with the static method. Even the case occurs, that negative CO2emissions are calculated with
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
10
the static method and positive CO2emissions are reached when using the dynamic method.
The second finding relates to the comparison of EHP with and without battery. While one could
expect EHP with battery to achieve lower CO2emissions than those without, the opposite turns
out to be valid. EHP without battery show lower CO2emissions in operation than those with
battery in 2013 and 2021, if the dynamic method is applied. In contrast, in the year 2030 EHP
without battery emit slightly more CO2than EHP with battery.
Regarding the relative CO2emission factors, it can be stated, that during feed-in times of EHPs
the CO2emission factor is significantly lower than the one during grid demand times throughout
all simulated system setups. The deviation between the CO2emission factor of feed-in and grid
power demand increases with increasing decarbonisation of the electric grid. This means that
for future already partially CO2neutral electric power systems the avoidable CO2emission by
PV feed-in tend to be zero since they coincide with already very high shares of renewables in the
grid. Grid power demand, instead, still can be significantly CO2affected. This especially holds
true for space heating related power demand during wintertime. Or in other words, sooner
or later all EHP with a significant electricity grid demand in wintertime turn out to be net
CO2emitters, if not the need for seasonal energy time shift is adressed.
Based on the previous findings and based on the fact that for the reviewed datasets the
relative deviation of the CO2emission factor for feed-in and grid power demand only show very
little sensitivity regarding different system setups, namely battery (non-)/existent PV capacity,
electric demand, one central recommendation can be made:
The introduction of CO2emission factors separately for feed-in and grid use for building GHG
emission evaluation within the regulatory framework seems to be highly beneficial. Both, the
ISO 16745 as well as DIN 18599 already provide the possibility to consider deviating CO2emission
factors. These new CO2emission factors could be determined using dynamic simulations.
However, for more accurate results, the hourly AEF has to be replaced by a more specific
hourly resolved CO2emission factor in the simulation. The emission factor has to correlate with
the distinctive power plant type that is avoided in case more renewable energy is fed into the
grid. Such hourly emission factor data is not available by now, as is described in the limitations.
The dynamic simulation would then only need to be performed once e.g. by a public authority
to determine the CO2emission factors applicable for any further standardized GHG emission
evaluations. Before being applied on a broader basis a crosscheck whether to use AEF or MEF
should be carried out carefully.
Acknowledgments
Parts of this work were developed within the research project FlexEhome:
The grid-supporting solar house with complete thermal and electric supply, funded by
the German Federal Ministry for Economic Affairs and Climate Action (project reference
number: 03EGB0025A). The funding is gratefully acknowledged. We would also like to thank
Agora Energiewende for open data access and especially Fabian Hein regarding CO2emission
factor calculation for future scenarios.
SBE-BERLIN-2022
IOP Conf. Series: Earth and Environmental Science 1078 (2022) 012048
IOP Publishing
doi:10.1088/1755-1315/1078/1/012048
11
References
[1] United Nations Framework Convention on Climate Change 2015 Paris agreement: Fccc/cp/2015/10/add.1
[2] Bundestag 18082021 Federal climate change act (bundes-klimaschutzgesetz): Ksg
[3] Bundestag 18082021 Erstes gesetz zur ¨anderung des bundes-klimaschutzgesetztes
[4] R¨ock M, Saade M R M, Balouktsi M, Rasmussen F N, Birgisdottir H, Frischknecht R, Habert G, L¨utzkendorf
T and Passer A 2020 Applied Energy 258 114107 ISSN 03062619
[5] German Federal Ministry of the Interior Building and Community What makes an efficiency house plus?
principles and examples of energy-generating buildings
[6] Vuarnoz D and Jusselme T 2018 Energy 161 573–582 ISSN 03605442
[7] Olkkonen V and Syri S 2016 Journal of Cleaner Production 126 515–525 ISSN 09596526
[8] Lausselet C and Brattebø H 2021 Int. J. Life Cycle Assess. 26 2263–2277
[9] Petra I, Thomas L and Gunter K Entwicklung der spezifischen kohlendioxid-emissionen des deutschen
strommix in den jahren 1990 bis 2020
[10] Prognos, ¨
Oko-Institut and Wuppertal Institut Towards a climate-neutral germany by 2045. how germany
can reach its climate targets before 2050
[11] Stein J S, Holmgren W F, Forbess J and Hansen C W 2016 Pvlib: Open source photovoltaic performance
modeling functions for matlab and python 43rd Photovoltaic Specialists Conf. PVSC ed IEEE pp 3425–30
[12] Niederl A 2017 Leistungssteigerung von photovoltaikanlagen durch modulk¨uhlung Innehalten und Ausblick:
Effektivit¨at und Effizienz f¨ur die Energiewende (Verlag der Technischen Universit¨at Graz) pp 53–54
[13] Kurz and Nawrowski 2019 Applied Sciences 91626
[14] Wetter M and Afjei T Trnsys type 401-kompressionsw¨armepumpe inklusiv frost-und taktverluste
[15] Baster E 2017 Air source heat pump modelling for dynamic building simulation tools based on standard test
data Ph.D. thesis University of Strathclyde
[16] Ralf D and Afjei Thomas e a Models of sub-components and validation for the iea shc task 44 / hpp annex
38 part c: Heat pump models
[17] H¨using F 2020 Modelling of inverter heat pumps in trnsys Proc. of the ISES EuroSun 2020 Conf. – 13th
Int. Conf. on Solar Energy for Buildings and Industry ed Charalambides A, Streicher W and Mugnier D
(Freiburg, Germany: International Solar Energy Society) pp 1–8 ISBN 978-3-9820408-2-0
[18] Dr¨uck H Multiport store - model for trnsys: Stratified fluid storage tank with four internal heat exchangers,
ten connections for direct charge and discharge and an internal electrical heater: Type 340 version 1.99f
[19] Kepplinger P, Huber G and Petrasch J 2015 Energy and Buildings 100 50–55 ISSN 03787788
[20] Tjaden T April 2013 Techno-¨okonomischer Vergleich von Solarthermieanlagen mit Photovoltaik-
W¨armepumpen-Systemen mittels dynamischer Simulation Masterarbeit HTW Berlin Berlin
[21] Dipasquale C and Fedrizzi R e a Technical report on system sizing and optimised control strategies
[22] Clauß J and Georges L 2019 Applied Energy 255 113847 ISSN 03062619
[23] Pollerberg C, Schmidt D L, M¨onnigmann M, L¨ohr Y, H¨orsting A, Bierwirth M, Wolf D, Kolbe C,
Wiebicke C and Abul-Ella Z Schlussbericht zum verbundvorhaben: Einsatz von latentw¨armespeichern
mit w¨armepumpen zum lastmanagement von stromnetzen - latenter stromspeicher
[24] Bundesverband W¨armepumpe eV Leitfaden hydraulik
[25] Feist W 1994 Thermische Geb¨audesimulation: Kritische Pr¨ufung unterschiedlicher Modellans¨atze 1st ed
(Heidelberg: C. F. M¨uller) ISBN 3788074868
[26] Wystrcil D and Kalz D 2013 Model-based optimization of control strategies for low-exergy space heating
systems using an environmental heat source Proceedings of BS2013 - 13th Conf. of Intern. Building
Performance Simulation Association ed Etienne Wurtz pp 2575–2583 ISBN 9781629939988
[27] Wystrcil D and Kalz D 2014 Pr¨adiktive regelung thermoaktiver bauteilsysteme unter ber¨ucksichtigung der
hydraulischen typologie. Human-centred building(s) ed van Treeck C and M¨uller D (Aachen: RWTH
Aachen University) pp 562–569 ISBN 978-3-00-047160-5
[28] Klein K, Hermann M and Herkel S 2016 Bautechnik 93 1–7 ISSN 09328351
[29] Gwerder M, Lehmann B, T¨odtli J, Dorer V and Renggli F 2008 Applied Energy 85 565–581 ISSN 03062619
[30] V˜osa K V, Ferrantelli A and Kurnitski J 2020 Journal of Building Engineering 31 101387 ISSN 23527102
[31] Solar Energy Laboratory Trnsys: Volume 6 multizone building modeling with type56 and trnbuild.
[32] Gl¨uck B 1990 W¨arme¨ubertragung: W¨armeabgabe von Raumheizfl¨achen und Rohren 2nd ed Bausteine der
Heizungstechnik Berechnung, Software (Berlin: Verl. f¨ur Bauwesen) ISBN 3-345-00515-8
[33] Braeuer F, Finck R and McKenna R 2020 Journal of Cleaner Production 266 121588 ISSN 09596526
[34] Seckinger N and Radgen P 2021 Energies 14 2527
[35] Regett A, Boing F, Conrad J, Fattler S and Kranner C 2018 Emission assessment of electricity: Mix vs.
marginal power plant method 2018 15th International Conference on the European Energy Market (EEM)
(IEEE) pp 1–5 ISBN 978-1-5386-1488-4
