Operation strategies using the Smart Grid Ready

interface in solar heat pump systems

1)Mitja Ortleb, 1)Josef Meiers, 2)Danny Jonas and 1)Georg Frey, Senior Member, IEEE
1) Chair of Automation and Energy Systems, Saarland University, D-66121 Saarbrücken
{mitja.ortleb | josef.meiers | georg.frey}@aut.uni-saarland.de
2) Ingenieurbüro Jonas, D-66740 Saarlouis

[email protected]

Abstract—The energy sector is faced with challenges since One way to create control strategies is to modify concepts
the increasing use of weather-depending renewables in power for continuously variable heat pumps and assign heat pump
generation. The resulting fluctuations must be balanced through demand to the SG Ready states, as done in this work for rule-
storage technologies and Demand Side Management (DSM) based timer controls. Another possibility applied here is to use
methods. Heat pumps are generally recognized as shiftable loads a reference load profile to generate a control signal based on
for DSM. More and more heat pump manufacturers in consumption. Furthermore, a control based on the electricity
2022 IEEE 1st Industrial Electronics Society Annual On-Line Conference (ONCON) | 979-8-3503-9806-9/22/$31.00 ©2022 IEEE | DOI: 10.1109/ONCON56984.2022.10126762

Germany are using the Smart Grid (SG) Ready interface, which price is possible.
enables grid operators on the one hand and system operators on
the other hand to control heat pumps for the purpose of DSM In the following, different control strategies will be
aiming at either grid power balancing or to increase the self- investigated and compared with respect to grid efficiency,
consumption rate of the residential energy system. The selection self-consumption and comfort for the system operator.
of a suitable control strategy is crucial here. Through intelligent
control, strong fluctuations of the power grid can be prevented. II. SIMULATION MODELS AND PARAMETERS
For this work, different control strategies for SG Ready enabled
solar and heat pump systems are implemented in a simulation A. Simulation software
framework and compared from a grid operator and a system The simulation of the building and all system components
owner perspective. is done via the TRNSYS 18 simulation environment. The
residential energy system model including building, solar, and
Keywords—Smart Grid Ready, simulation, control, heat pump heat pump system is based on the work presented in [10–12]
using models of the SHP-SIMLIB [13]. This system is coupled
I. INTRODUCTION
via the open-source software Building Controls Virtual Test
Photovoltaic (PV) systems in combination with heat Bed (BCVTB) version 1.6.0 with Matlab R2020b where the
pumps and a battery storage are a promising concept for the control strategies for the heat pump are implemented. Matlab
energy transition of residential buildings. The power produced transfers the control signal in each simulation step via BCVTB
by the PV system can be temporarily stored in the battery if it to TRNSYS and at the same time receives data needed for the
is not consumed directly. If this is not possible, the operation control (cf. Fig. 1). The control signal is compatible with the
of the heat pump can be forced to avoid feed-in into the SG Ready-interface of the binary state-controlled heat pump.
electricity grid. The fluctuating feed-in from renewable
energies creates problems for the grid operator, as the volatile
generation must be taken into account in grid planning and
operation to ensure stability [1].
Heat pumps, as variable consumers, can mitigate such
peaks by increasing their demand. The produced heat can be
stored in a buffer storage tank and used later for space heating
(SH) or domestic hot water (DHW) supply. For integration
into a smart grid, the Bundesverband Wärmepumpe e.V. Fig. 1. Simulation scheme.
(German Heat Pump Association) awards the SG Ready label
to smart-grid-capable heat pumps in Germany. Via the SG B. Boundary conditions
Ready interface users or grid operators can adjust four The International Energy Agency (IEA) Solar Heating and
different pre-defined operation modes (OM) [2]. According to Cooling Programme (SHC) Task 44 / Heat Pump Programme
the association, more than 1900 heat pump types are already (HPP) Annex 38 defined boundary conditions to provide a
equipped with this label, which makes it the standard in simple basis for the comparison of solar thermal and heat
Germany [3]. Some work has already been done on various pump systems (STHP), which are also used for this work [14,
control strategies for heat pumps in combination with PV 15]. It contains weather data for different locations [16]. Three
systems. However, almost all of them are dealing with binary- different buildings are defined, which represent varying heat
state (ON/OFF) or continuously variable heat pumps without loads and thermal insulation. For this work the SFH45
SG Ready interface [4–6]. There is hardly any work on the building is used, which represents a renovated building with
control of SG Ready heat pumps in this combination. Tjaden good thermal quality of the building envelope and a heat load
et al. presented such a relatively rudimentary control concept of 45 kWh/m2a for a floor area of 140 m2 [15]. The weather
[7]. In addition, Fischer [8] has studied SG Ready-capable data used are from the climate of Strasbourg [16]. The DHW
heat pumps. A control strategy that has already been applied profile which is also defined in the boundary conditions with
and introduced on the market is integrated in the E3/DC home an average consumption of 140 l per day at 45 °C corresponds
power plant S10, which works rule-based with weather to an energy consumption of 5.845 kWh per day given a cold
forecasts [9]. water temperature of 10 °C [14].

Authorized licensed use limited to: Ricerca sul Sistema Energetico RSE S.p.A.. Downloaded on December 19,2024 at 11:43:15 UTC from IEEE Xplore. Restrictions apply.
 C. Energy supply system (heat and electricity) designed to increase self-consumption and to shift increased
Main components are the air to water heat pump with SG heat pump operation power demand to times of increased PV
Ready interface, the buffer tank for the heat supply of the SH surpluses. OM 1 is only reached when a utility blocking is
and DHW circuits, the PV system, and the battery. activated. In the following simulations, however, this blocking
is disregarded due to its negligible utilization.
The heat pump has a coefficient of performance of 3.6
(A2/W35). The nominal heating power for this operating point When the generated PV power PPV is less than or equal to
is 6.36 kW. It is implemented with a minimum on/off time of the household electric load PLoad applied at that time
15 minutes in order to prevent high frequency switching. It is multiplied by the factor fon/off, the controller switches to the
connected to the system via a 1.0 m3 buffer tank. The heat is normal OM 2. With the characteristic values of the simulated
supplied either to the heating circuit for SH or for DHW. If PV and heat pump system, Tjaden et al. recommend a factor
there is a simultaneous demand for both, DHW is prioritized. of approximately fon/off = 70 % [7]. If the grid feed-in power or
the PV power surplus PSurplus exceeds the nominal power of
The PV system has a nominal power of 2.25 kWp. the ASHP multiplied by fon/off, the controller switches from
The power feed-in limit in Germany is 70 % of the nominal OM 2 to intensified OM 3. If PSurplus exceeds the maximum
power by law (§ 9 para. 2 no. 3 EEG [17]). The surface area feed-in limit PThreshold and there would be curtailment losses,
covered by the PV modules is 15 m2 and the efficiency at STC the controller changes to OM 4 to increase self-consumption.
conditions of the solar cells is 0.1497. The direct current
generated by the PV system is converted into alternating
current by an inverter with an efficiency of 96 % These are
standard values for flate-plate selective collectors of the IEA
Reference Heating System of IEA SHC Task 32 [18]. The
generated electrical power can either be consumed directly or
used for charging the battery.
The battery has a nominal capacity of 5.5 kWh. The
battery can only be charged up to 90 % of the nominal capacity
and discharged down to 5 % to avoid degradation effects. It is
connected to an inverter with an efficiency of 0.988
converting from AC to DC and 0.987 converting from DC to
AC. The battery has a charge/discharge efficiency of 0.949.
D. Implementation of the SG Ready Operation Modes Fig. 2. Flowchart SCO
The four SG Ready Operation Modes (OMs) and their
corresponding implementation are given in the following: B. Standard Load Profile Based Control (SLP)
OM 1: OFF state: The heat pump is switched-off. This operation strategy was developed by D. Lellinger
[19] to generate a control signal based on a standard load
OM 2: NORMAL state: The heat pump runs in normal profile. For this work, the BDEW standard load profile (SLP)
mode. At a temperature of 53 °C in the DHW zone of the from 1996/1997 is used. First, a daily profile averaged over a
storage tank the heat pump switches on to supply this zone year is created. Then the relative deviation of the average
with heat until it reaches 55 °C. In the SH zone of the tank, the quarter-hourly value from the average annual value is
heat pump supplies heat when the temperature difference calculated. The lower 10 % of the load values are assigned to
between the SH zone sensor and the setpoint is less than 0 K OM 4, the upper 15 % to OM 1. Halfway between these the
until the difference exceeds 3 K. The setpoint is calculated limit of OM 2 and OM 3 is set. Fig. 3 shows the daily profile
using a weather-compensated flow temperature control, which averaged over the year with the calculated limits.
depends mainly on the outdoor temperature and the design
heat load of the building and location.
OM 3: BOOST state: The heat pump controller is
recommended to switch-on: The storage tank setpoint
temperatures are increased by 5 K.
OM 4 SUPERHEAT state: Actuation signal to the heat
pump controller. Both setpoint temperatures are increased to
the setpoint of the SC zone plus 5 K.
III. OPERATION STRATEGIES
Four operation strategies will be presented and investigated:
• Self-Consumption Optimization (SCO)
• Standard Load Profile Based Control (SLP)
• Rule Based Timer Control (RBTC) Fig. 3. Averaged daily curve of the BDEW SLP
• Price and Rule Based Control (PRBC)
C. Rule Based Timer Control (RBTC)
A. Self-Consumption Optimization (SCO)
This operating strategy was developed by Fischer and
This operation strategy is based on a strategy developed attempts to increase or decrease heat pump operation at certain
by Jakobi et al. [19]. It is rule-based (cf. Fig. 2) and was times [6]. In Fischer's work, the hysteresis for the storage

Authorized licensed use limited to: Ricerca sul Sistema Energetico RSE S.p.A.. Downloaded on December 19,2024 at 11:43:15 UTC from IEEE Xplore. Restrictions apply.
 temperature is increased by 5 K at times when operation B. Self-Consumption Rate
should be avoided and decreased by 5 K at favorable times. The self-consumption rate SCR is the ratio of self-used
Since Fischer does not use a SG ready heat pump, the strategy energy to the generated PV energy EPV [24]. In this case the
needs some adjustment. In times of high solar irradiation self-used energy results from EPV2Load and the battery charging
(according to Fischer between 11 am and 2 pm), the hysteresis energy. A self-consumption rate of 100 % would mean that
is to be lowered by 5 K. In our case, this is to be done by the entire PV energy is used and nothing is fed into the grid.
setting the intensified OM 3. In times of medium irradiation, This is also financially worthwhile, since the remuneration
OM 2 is set. In times of low irradiation Fischer increases the price for electricity fed into the grid is lower than the purchase
hysteresis by 5 K to avoid amplified operation. Unfortunately, price. While the feed-in tariff has been falling continuously
there is no OM with attenuated heat pump operation besides since 2014, the purchase price has been rising almost
OM 2. Thus OM 2 is also set in this case. This control is only continuously [25, 26]. The SCR within this definition also
activated when the outside temperature of the building is contains the losses of the battery. Thus, high battery losses can
above 10 °C. Below this OM 2 remains activated to avoid result in high SCR values. The SCR is defined with:
intensified OM 3 in times with little irradiation.
𝐸𝐸𝑃𝑃𝑃𝑃2𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 +𝐸𝐸𝑃𝑃𝑃𝑃2𝐵𝐵𝐵𝐵𝐵𝐵
𝑆𝑆𝑆𝑆𝑆𝑆 = . (2)
D. Price and Rule Based Control (PRBC) 𝐸𝐸𝑃𝑃𝑃𝑃

The purpose of this rule-based strategy is to keep the costs

low for the system operator. In times when the electricity price C. Percentage of Curtailment Losses
is high, the heat pump operation is reduced and in times of low By prescribing a maximum feed-in limit, high PV power
electricity prices, the operation is increased and the buffer outputs can lead to curtailment losses Ect. In order to represent
storage is charged. For this, the day-ahead prices of Germany these losses, the quantity l, the percentage of curtailment
/ Austria / Luxembourg for the year 2016 are evaluated. Day- losses, is introduced [4]. These should be kept as small as
ahead trading is the trading of electricity for the following day, possible and can be calculated with:
which is traded on EPEX Spot [20], a spot market of the 𝐸𝐸𝑐𝑐𝑐𝑐
European Power Exchange. Even though trading on power 𝑙𝑙 = . (3)
𝐸𝐸𝑃𝑃𝑃𝑃
exchange only accounts for about 20 % of the total trading
volume, exchange prices are considered important indicators D. Seasonal Performance Factor
of wholesale prices [21]. The one year data set used is from
The Seasonal Performance Factor SPFHP,pen provides the
the entsoe Transparency Platform. It was cleaned using [22,
ratio of the amounts of heat consumed for SH QSH and hot
23] and interpolated to 15-minute values.
water production QDHW by the load of the whole heating
The algorithm determines the maximum and minimum system Eel,HP+,pen [27]. In case of violation of the comfort
day-ahead prices separately for each day. The range between criteria defined by Task 44 of the IEA SHC program (hot
the maximum and the minimum is divided and assigned to the water tap temperature and room temperature of the building),
four SG Ready OMs. For this work, four different penalty functions are applied [28]. They are added to the
modifications (mods.) with different limits were simulated electrical load of the heating system and represent the missing
and evaluated. Due to limited space in the following only the heating energy needed to reach the comfort criteria. The
extreme values mod. 1 and mod. 4 are included. electrical load of the heating system includes the electrical
energy consumption of the HP, the HP control, the HP storage
In mod. 1, the lower 5 % of the day-ahead prices on the charging pump, the DHW pump, the heating circulation
corresponding day are assigned to OM 4. Above the 5 % limit pump, and the additional controller. The heat quantity is
up to the 10 % limit is the range assigned to OM 3. Starting measured at the connection point of the buffer storage tank to
from this limit up to 85 % is the range for OM 2. The upper the SH and DHW circuit. The seasonal performance factor is
15 % belongs to OM 1. For mod. 4 these limits are set to 30 a measure of the efficiency of the entire heating system:
%, 50 %, and 70 % respectively. After calculating the limits,
the day-ahead price on the corresponding date and time is 𝑆𝑆𝑆𝑆𝐹𝐹𝐻𝐻𝐻𝐻+,𝑝𝑝𝑝𝑝𝑝𝑝 =
𝑄𝑄𝐷𝐷𝐷𝐷𝐷𝐷 +𝑄𝑄𝑆𝑆𝑆𝑆
. (4)
retrieved and compared with the calculated limits of the 𝐸𝐸𝑒𝑒𝑒𝑒,𝐻𝐻𝐻𝐻+,𝑝𝑝𝑝𝑝𝑝𝑝

current day. Depending on which limits the price is located

between, the corresponding OM of the heat pump is set. E. Fractional CO2 Emission Savings
The fractional CO2 emission savings fsav,CO2 is a measure
IV. EVALUATION FIGURES of how many CO2 emissions can be saved compared to a
When evaluating the simulation with regard to the grid conventional heating system (in this case a gas heating
efficiency and system operator efficiency of the various system including electrical control and pumps) [29, 30]. The
operation strategies, various aspects must be taken into total efficiency of the reference system in this case
account. For this reason, the following evaluation figures are is ηref = 0.9. Eel,ref is the electricity consumption of the
introduced, which allow a differentiated evaluation. reference heating system in one year (Eel,ref = 204 kWh). The
A. Self-Sufficiency Factor quantities QSH,ref and QDHW,ref, which quantify the heat
The self-sufficiency factor SSF indicates how much of the consumed in the whole year, are already given by the
total energy demand ELoad is covered by self-consumed building and the DHW profile (QSH,ref = 6476 kWh/a,
energy. The self-consumed energy results from the directly QDHW,ref =2076 kWh/a). The quantity GWP stands for the
consumed energy from the PV generation EPV2Load and the Global Warming Potential of the respective final energy
battery discharge energy EBat2Load, since the battery is only source and defines its particular global warming potential in
charged by the PV system. Ideally, the degree of self- terms of CO2 (GWPel = 0.521 kgCO2eq/kWhfe,
sufficiency would be 100 %. The SSF can be defined with: GWPgas = 0.307 kgCO2eq/kWhfe). The electrical load of the
SSF = (EPV2Load + EBat2Load) / ELoad. (1) heating system Eel,HP+,pen includes the previously mentioned

Authorized licensed use limited to: Ricerca sul Sistema Energetico RSE S.p.A.. Downloaded on December 19,2024 at 11:43:15 UTC from IEEE Xplore. Restrictions apply.
 penalty function when the comfort criteria is violated. The 25 °C. If the outdoor temperature is above 26 °C, the room
fractional CO2 emission savings are defined with: temperature may also be higher (1K inside for 3K outside).
𝐸𝐸𝑒𝑒𝑒𝑒,𝐻𝐻𝐻𝐻+,𝑝𝑝𝑝𝑝𝑝𝑝 𝐺𝐺𝐺𝐺𝑃𝑃𝑒𝑒𝑒𝑒 When evaluating this evaluation figure, it was noticeable that
𝑓𝑓𝑠𝑠𝑠𝑠𝑠𝑠,𝐶𝐶𝑂𝑂2 = 1 − 𝑄𝑄𝑆𝑆𝑆𝑆,𝑟𝑟𝑟𝑟𝑟𝑟+𝑄𝑄𝐷𝐷𝐷𝐷𝐷𝐷,𝑟𝑟𝑟𝑟𝑟𝑟 . (5)
𝐺𝐺𝐺𝐺𝑃𝑃𝑔𝑔𝑔𝑔𝑔𝑔 +𝐸𝐸𝑒𝑒𝑒𝑒,𝑟𝑟𝑟𝑟𝑟𝑟 𝐺𝐺𝐺𝐺𝑃𝑃𝑒𝑒𝑒𝑒 the time in which the room temperature exceeds the upper
𝜂𝜂𝑟𝑟𝑟𝑟𝑟𝑟
limit of the comfort range is almost exactly the same for all
strategies considered. This value is several times higher than
F. Relative Import Bill the time in which lower limits are violated. The high room
The Relative Import Bill RIB considers the load shift from temperature is due to fact that no active cooling is installed.
high price periods to low price periods [31]. The quantity For this reason, the figure ttemp< is introduced, which only takes
accounts for the reduction in the electricity bill. The electricity into account the temperature drops below the comfort range.
bill is calculated as on one day as a sum energy demand from
the grid over one timestep t, EGrid(t), multiplied by the I. Number of Heat Pump Starts
corresponding day-ahead electricity price p(t) in timestep t. The number of heat pump starts nstarts provides information
The possible reduction in the electricity bill is this sum minus about the lifespan of the heat pump [35]. Too frequent starting
the minimum possible electricity bill. This is the sum of the should be avoided.
grid purchase at timestep t multiplied by the minimum daily-
J. Summary
day-ahead price on the corresponding day d, pmin,day(d). The
difference between the electricity bill and the minimum Table I. provides an overview of the classification of the
possible bill is related to the difference between the maximum evaluation figures including valid range.
possible bill and the minimum possible bill. The maximum TABLE I. CLASSIFICATION OF THE EVALUATION FIGURES
possible bill is calculated with the maximum daily-day-ahead
price pmax,day(d) and RIB can be defined with: Figure Valid Range Operator Supportive Grid Supportive
SSF 0% - 100% High value -
∑ 𝐸𝐸𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 (𝑡𝑡)𝑝𝑝(𝑡𝑡)−∑ 𝐸𝐸𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 (𝑡𝑡)𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚,𝑑𝑑𝑑𝑑𝑑𝑑 (𝑑𝑑)
𝑅𝑅𝑅𝑅𝑅𝑅 = ∑ . (6) SCR 0% - 100% High value -
𝐸𝐸𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 (𝑡𝑡)𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚,𝑑𝑑𝑑𝑑𝑑𝑑 (𝑑𝑑)−∑ 𝐸𝐸𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 (𝑡𝑡)𝑝𝑝𝑚𝑚𝑚𝑚𝑚𝑚,𝑑𝑑𝑑𝑑𝑑𝑑 (𝑑𝑑)
l 0% - 100% Low value -
Desirable for the RIB would be a value as close to zero as SPFHP+,pen >0 High value -
possible. This would mean that the electricity was purchased fsav,CO2 a. 0% - 100% High value -
at the most favorable times. A favorable exchange price leads RIB 0% - 100% Low value Low value
to lower electricity procurement costs for the energy suppliers GSCabs >0 - <1
and thus to a more favorable end customer price. However, ttemp< ≥ 0 min Low value -
this relationship is not linear, as the levy of the German nstarts >0 Low value -
renewable energy law (EEG-Umlage) increases with a.
only applies if Eel,WP < Eref * GWPref / GWPel
decreasing exchange price [32]. In addition, other aspects
affect the cost calculation of the end customer price. Thus, RIB V. RESULTS
is relevant for the grid operator and customers with price
variable tariffs. Table II. shows the evaluation results for all strategies. In
order to better classify the values, the table has been colored.
G. Absolute Grid Coefficient The range between the worst and the best value in the
GSCabs is designed to provide information on how well the comparison is scaled from red to green.
end user’s electricity load profile matches the availability of
electricity [33]. It indicates whether electrical energy TABLE II. ABSOLUTE EVALUATION RESULTS
consumption ELoad is more likely to occur during periods when Evaluation value SCO SLP RBTC PRBC1 PRBC4
the day-ahead electricity price is higher than the average
SSF [%] 33.05 28.37 31.53 29.44 27.47
relative price p̅ or during periods when it is less. This allows
SCR [%] 98.04 89.19 93.43 90.97 92.04
an evaluation of the grid impact by a building from the utility’s
perspective. l [%] 0.007 0.117 0.068 0.109 0.091
SPFHP+,pen [-] 2.947 2.588 2.981 2.671 2.199
For a flexible load such as a heat pump, a GSCabs of 0.9 means fsav,CO2 [%] 51.95 45.38 52.47 47.31 36.70
that, on average, electricity is consumed when the day-ahead RIB [%] 45.30 35.87 45.22 38.01 39.28
price assumes 90 % of its mean value during the evaluation GSCabs [-] 1.019 0.958 1.021 0.960 0.940
period. A low GSCabs is desired because it means that ttemp< [h] 113.0 108.3 119.6 360.2 537.6
electricity is mainly consumed when the price of electricity is
nstarts [-] 2286 1147 2137 1391 1009
low, or when the demand for electricity is low. This makes
sense for both the end consumer and the utility. GSCabs can be A. SCO
defined with: The rule-based self-consumption optimization strategy by
𝐺𝐺𝐺𝐺𝐶𝐶𝑎𝑎𝑎𝑎𝑎𝑎 =
∑ 𝐸𝐸𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 (𝑡𝑡)𝑝𝑝(𝑡𝑡)
. (7) Jakobi et al. works as intended and ensures by far the highest
∑ 𝐸𝐸𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 (𝑡𝑡)𝑝𝑝̅ self-consumption. Both the self-consumption rate of 98.04 %
and the self-sufficiency factor of 33.05 % are the highest in
H. Comfort Criteria the comparison. Also, the percentage of curtailment losses of
The total time ttemp during which the comfort criteria are 0.007 % are much lower than the values reached by the other
violated is of great importance for the home user. As regulated strategies. The comfort criteria are violated for 113.0 h in the
in Task 44, the room temperature must not fall below 19.5 °C evaluation period of one year, which is correspondingly good.
[28]. The upper limit for the temperature can be found in DIN The SPF of 2.947 is among the better ones in the comparison.
Standard 1946 Part 2 [34]. Up to a maximum outside The CO2 emission savings of 51.94 % are also a good value.
temperature of 26 °C, the room temperature must not exceed The intensified operation is almost exclusively switched on

Authorized licensed use limited to: Ricerca sul Sistema Energetico RSE S.p.A.. Downloaded on December 19,2024 at 11:43:15 UTC from IEEE Xplore. Restrictions apply.
 between 9 am and 6 pm. This results in a relatively high RIB losses among the PRBC strategies with 0.091 % due to the
of 45.30 %. But the overall lowest grid purchase of about intensified operation at midday. The efficiency of the heating
4680 kWh compensates the high-cost RIB from the system system suffers from the frequent switching to the intensified
operator's perspective. From the system operator's point of OMs. The seasonal performance factor for mod. 1 is at a
view, this strategy is thus one of the most sensible among medium level. The SPFHP+,pen of mod. 4 is 2.199, which is the
those compared. From the grid operator's point of view, lowest value in the whole comparison. The same result can be
however, this strategy is not very beneficial as shown by the observed for the potential CO2 savings. In particular, mod 4
RIB and the GSCabs of 1.019. again achieves the worst value with fsav,CO2 = 36.70 %. Mod. 4
has also the highest grid purchase of all the strategies. All
B. SLP mods. are also not convincing when it comes to violating the
The control according to deviations from the standard load comfort criteria. With ttemp< = 537.6 h, mod. 4 has the most
profile is a strategy that does not change over the year and thus frequent violations of the comfort criteria. Mod. 1 is also
cannot react to external influences, such as particularly strong significantly above all other strategies with ttemp< = 360.2 h.
solar radiation. The OMs are changed every day at fixed times. For this reason, the compared PRBC strategies should not be
This results in one of the worst self-consumption rates of acceptable from the system operator's point of view. In
89.19 % and a comparatively low self-sufficiency rate of addition, the self-consumption is relatively low. Only the low
28.37 %. The curtailment losses of 0.117 % are the highest in number of heat pump starts is beneficial for the system
the comparison. The room temperature falls below the operator. With 1009 starts, mod. 4 achieves the lowest value
comfort criteria for a total of 108.3 h. The efficiency of the in the comparison.
heating system is rather in the lower range with a SPF of
2.588. Only the low number of heat pump starts with 1147 and From the grid operator's point of view, the PRBC
the low RIB of 35.87 % indicate a system operator friendly strategies are the most acceptable and also recommendable.
control. The grid consumption is average with 5237 kWh. The GSCabs of 0.940 for mod. 4 is the lowest of all strategies.
From the system operator's point of view, this strategy makes The RIB is also low for all mods. and remains below the
little sense, especially because of the very low self- average. This can be explained by the fact that the OMs are
consumption. From the grid operator's point of view this directly dependent on the day-ahead electricity price.
strategy is highly recommendable, as can be seen by the low VI. CONCLUSION
RIB also the GSCabs is 0.958. Thus, the electricity is mainly
consumed in times when the demand is below average. In this work seven different control algorithms for SG
Ready heat pumps were simulated and compared. Some
C. RBTC strategies used external data, such as day-ahead electricity
The rule-based timer control depending on solar radiation prices or standard load profiles, others depended on real-time
according to Fischer switches, like the SLP control at certain variables from the simulation, and still others managed
times, but only between OM 2 and 3. The control is activated completely without additional data. All strategies have
only when the outdoor temperature is at least 10 °C. The self- different strengths and weaknesses. For different system
sufficiency factor of 31.53 % is the second highest in the operators and grid operators, different evaluation variables are
comparison. The self-consumption rate of 93.43 %, as well as of importance. Therefore, an objective classification of which
the percentage curtailment losses of 0.068 %, are also among strategy is best cannot be made easily.
the best. With fsav,CO2 = 52.47 % and SPFHP+,pen = 2.981, the
If each evaluation variable is equally important in the
best values in the comparison can be achieved. Temperature
evaluation, Table III. is helpful and gives an overview of the
undershoots occur in 119.6 h. The relatively low grid purchase
results. Each evaluation variable has been normalized to a
of 4764 kWh somewhat compensates for the poor RIB of
range of 0 to 100. The 100 is assigned to the most desirable
45.22 %. Overall, this strategy can thus be recommended for
outcome in this comparison and the 0 to the least desirable.
system operators. The GSCabs of 1.021 % is relatively high.
Similarly, the RIB of 45.22 % is at a high level. Thus, it is not TABLE III. NORMALIZED EVALUATION FIGURES
very beneficial to the grid.
Evaluation value SCO SLP RBTC PRBC 1 PRBC4
D. PRBC SSF [%] 100 16 73 35 0
The PRBC control changes its states depending on the SCR [%] 100 0 48 20 32
current value of the day-ahead electricity price. This results in l [%] 100 0 45 7 24
SPFHP+pen [%] 96 50 100 60 0
relatively frequent changes of the OMs
Fsav,CO2 [%] 97 55 100 67 0
For mod. 1, OM 2 is set most of the time. The intensified RIB [%] 0 100 1 77 64
OM 4 is mainly used early in the morning, since the day-ahead GSCabs [%] 2 78 0 75 100
ttemp< [%] 99 100 97 41 0
electricity price is usually lowest at this time. The self- nstarts [%] 0 89 12 70 100
sufficiency factor of 29.44 %, for mod. 1 is lower than the Average [%] 66 54 53 50 36
average value of all control strategies. With 27.47 % mod. 4
has the worst self-sufficiency factor of all control strategies. SCO reaches by far the best result with an average of 66 %.
This also results in the highest electricity consumption of all However, From the grid operator's point of view, this strategy
strategies. The self-consumption rate for all mods. is between performs poorly, as can be seen by the GSCabs and the high
90.97 % (mod. 1) and 92.04 % (mod. 4). The self- RIB. There is no grid operator friendly strategy in this
consumption increases with increasing duration of intensified comparison that also has good self-consumption. The PRBC4
operation. Mod. 4 shows good self-consumption. This is control with the best GSCabs, has the worst self-sufficiency
because of the high power consumption by intensified OM 3 level. To sum up, none of the presented control strategies is
and 4 mainly middays to the afternoon. The situation is similar beneficial for both the system operator and the grid operator.
for the curtailment losses. Mod. 4 has the lowest curtailment Thus, a trade-off which allows for a grid beneficial operation

Authorized licensed use limited to: Ricerca sul Sistema Energetico RSE S.p.A.. Downloaded on December 19,2024 at 11:43:15 UTC from IEEE Xplore. Restrictions apply.
 in compliance with the system operator’s interests and / HPP Annex 38: Report C1 Part B: Building and space heat load,”
compensation mechanisms should be developed and IEA SHC Task 44 / HPP Annex 38.
[16] Meteotest, Bern, Switzerland, Meteonorm 6.1.0.9: Global
introduced in future research. Other works suggest that model meteorological database for engineers, planners and education,
predictive control (MPC), schedule optimization, or other 2009.
optimization algorithms can also lead to good results [6, 36]. [17] § 9 EEG 2021 - Einzelnorm, 2022. Accessed: Jun. 20 2022. [Online].
Further improvements in grid utility could also be achieved Available: https://www.gesetze-im-internet.de/eeg_2014/__9.html
through utility blocking which is not considered in this work. [18] Heimrath, R., Haller, M., 2007. The reference heating system, the
template solar system of Task 32. A report of IEA Solar Heating and
Furthermore, the system parameters, i.e. buffer storage size, Cooling programme – Task 32 Advanced storage concepts for solar
could to be adapted to different control strategies to enhance and low energy buildings. Report A2 of Subtask A. IEA SHC Task
the compliance with comfort criteria. 32.
[19] M. Jakobi, T. Tjaden, V. Quaschning, U. Stöckli, and L. Meier, “Von
ACKNOWLEDGMENT der Simulation zur Realität: IEC 61499 konforme Regelanwendungen
für Solare Energiesysteme,” in PV-Symposium 2018, pp. 193–194.
This work was part of the research project SolWP-Hybrid [20] About EPEX SPOT | EPEX SPOT. [Online]. Available: https://
funded by the federal state of Saarland and the European www.epexspot.com/de/node/1762 (accessed: Jul. 21 2022).
Regional Development Fund (ERDF 2014-2020). The authors [21] SMARD | Großhandelspreise. [Online]. Available: https://
www.smard.de/page/home/wiki-article/446/562 (accessed: Jun. 20
are grateful for this support and thank the project partners, 2022).
Viessmann Heizsysteme GmbH, Sonnenkraft GmbH, and [22] K. Rheinberger and A. Ulmer, GitHub - klaus-rheinberger/DSM-
DualSun and Stadtwerke Saarlouis. data: Data set for the Special Issue "Demand Side Management of
Distributed and Uncertain Flexibilities" in the journal "Energies".
REFERENCES Accessed: Jun. 20 2022. [Online]. Available: https://github.com/
klaus-rheinberger/DSM-data
[23] Special Issue "Demand Side Management of Distributed and
[1] A. Spring et al., “Untersuchung der Korrelationen aus
Uncertain Flexibilities". [Online]. Available: https://www.mdpi.com/
Tageslastgängen und PV-Einspeisung zur Bestimmung der
journal/energies/special_issues/demand_side_management_
maximalen Netzbelastung,” in 28. Symposium Photovoltaische
distributed_uncertain_flexibilities (accessed: Jun. 20 2022).
Solarenergie. Accessed: 01.06.22. [Online]. Available: https://
[24] J. Weniger, J. Bergner, T. Tjaden, and V. Quaschning, Eds.,
mediatum.ub.tum.de/doc/1185433/1185433.pdf
Economics of Residential PV-Battery Systems in the Self
[2] SG Ready-Label | Bundesverband Wärmepumpe (BWP) e.V.
Consumption Age.
[Online]. Available: https://www.waermepumpe.de/normen-technik/
[25] Statista, EEG-Vergütung von Photovoltaikanlagen bis 2022.
sg-ready/ (accessed: Jun. 1 2022).
[Online]. Available: https://de.statista.com/statistik/daten/studie/
[3] SG Ready-Datenbank | Bundesverband Wärmepumpe (BWP) e.V.
173265/umfrage/durchschnittliche-eeg-verguetung-von-
[Online]. Available: https://www.waermepumpe.de/normen-technik/
photovoltaikanlagen-seit-2009/ (accessed: Jun. 20 2022).
sg-ready/sg-ready-datenbank/ (accessed: Jun. 1 2022).
[26] Statista, Strompreis für Haushaltskunden in Deutschland bis 2021.
[4] B. Siegel and J. Bergner, “Betriebsstrategien für PV-Speichersysteme
[Online]. Available: https://de.statista.com/statistik/daten/studie/
im Vergleich,” HTW-Berlin, Sep. 2015.
154908/umfrage/strompreise-fuer-haushaltskunden-seit-2006/
[5] B. Matthiß, J. Binder, and B. Schott, “Reduzierte Netzbelastung und
(accessed: Jun. 20 2022).
optimierter Eigenverbrauch von dezentralen PV-Speichersystemen
[27] D. Jonas, D. Theis, J. Meiers, and G. Frey, “Model-Based Analysis of
durch modellprädiktive Betriebsführung von Speichern,” in OTTI-
Solar Thermal and Heat Pump Systems Using TRNSYS,” in
Konferenz „Zukünftige Stromnetze für Erneuerbare Energien“,
Proceedings of SWC2017/SHC2017, Abu Dhabi, 2016, pp. 1–12.
29.01.-30.01.2014, Berli.
[28] M. Haller, “System Simulation Results of the IEA SHC Task 44 /
[6] D. Fischer, J. Bernhardt, H. Madani, and C. Wittwer, “Comparison of
HPP Annex 38: Report C3,” IEA SCH Task 44 / HPP Annex 38.
control approaches for variable speed air source heat pumps
[29] J.-C. Hadorn, Ed., Solar and heat pump systems for residential
considering time variable electricity prices and PV,” Applied Energy,
buildings. Berlin: Ernst, 2015. [Online]. Available: http://
vol. 204, pp. 93–105, 2017, doi: 10.1016/j.apenergy.2017.06.110.
onlinelibrary.wiley.com/book/10.1002/9783433604830
[7] T. Tjaden, J. Weniger, and V. Quaschning, Eds., Richtige
[30] D. Jonas and G. Frey, “Model-based Analysis of the Performance and
Dimensionierung von Photovoltaik, Wärmepumpen und
the Environmental Impact of Solar Thermal and Heat Pump
Pufferspeichern.
Systems,”
[8] D. Fischer, “Integrating heat pumps into smart grids,” Dissertation,
[31] U. I. Dar, I. Sartori, L. Georges, and V. Novakovic, “Advanced
Kungliga tekniska högskolan.
control of heat pumps for improved flexibility of Net-ZEB towards
[9] E3/DC GmbH, “E3/DC Kurzanleitung,” Oct. 2016.
the grid,” Energy and Buildings, vol. 69, pp. 74–84, 2014, doi:
[10] D. Jonas, D. Theis, F. Felgner, and G. Frey, “A TRNSYS-based
10.1016/j.enbuild.2013.10.019.
simulation framework for the analysis of solar thermal and heat pump
[32] P. Götz, J. Henkel, and T. Lenck, “Zusammenhang von
systems,” Appl. Sol. Energy, vol. 53, no. 2, pp. 126–137, 2017, doi:
Strombörsenpreisen und Endkundenpreisen,” Energy Brainpool
10.3103/S0003701X17020074.
GmbH & Co. KG, 2013.
[11] D. Jonas, G. Frey, and D. Theis, “Simulation and performance
[33] K. Klein, R. Langner, D. Kalz, S. Herkel, and H.-M. Henning, “Grid
analysis of combined parallel solar thermal and ground or air source
support coefficients for electricity-based heating and cooling and
heat pump systems,” Solar Energy, vol. 150, pp. 500–511, 2017, doi:
field data analysis of present-day installations in Germany,” Applied
10.1016/j.solener.2017.04.070.
Energy, vol. 162, pp. 853–867, 2016, doi:
[12] D. Jonas, M. Lämmle, D. Theis, S. Schneider, and G. Frey,
10.1016/j.apenergy.2015.10.107.
“Performance modeling of PVT collectors: Implementation,
[34] Raumlufttechnik; Gesundheitstechnische Anforderungen (VDI-
validation and parameter identification approach using TRNSYS,”
Lüftungsregeln), 1946-2:1994-01, Deutsches Institut für Normung
Solar Energy, vol. 193, pp. 51–64, 2019, doi:
e.V., Berlin, Jan. 1994.
10.1016/j.solener.2019.09.047.
[35] P. Perrin, “Simulationsbasierte Analyse der Einflussfaktoren auf
[13] D. Jonas, SHP-SimLib. A library of solar and heat pump subsystem
Betriebszahlen von Wärmepumpenanlagen,” Dissertation, Fakultät
simulation models in TRNSYS, unreleased (Release in 2022/2023).
für Maschinenbau, Technischen Universität Carolo-Wilhelmina zu
[Online]. Available: https://github.com/DnJns/SHP-SimLib/
Braunschweig, 2012.
[14] M. Haller, R. Dott, J. Ruschenburg, F. Ochs, and J. Bony, “The
[36] T. Kemmler, R. Maier, C. Widmann, and B. Thomas, “Using a
reference framework for system simulations of the IEA SHC Task 44
Thermal Energy Storage to Provide Flexibility for Heat Pump
/ HPP Annex 38: Report C1 Part A: General simulation boundary
Optimization Control with Rapid Control Prototyping and SG Ready
conditions,” IEA SHC Task 44 / HPP Annex 38.
Standard,” in Proceedings of the 13th International Renewable
[15] R. Dott, M. Haller, J. Ruschenburg, F. Ochs, and J. Bony, “The
Energy Storage Conference 2019 (IRES 2019), Düsseldorf, Germany,
reference framework for system simulations of the IEA SHC Task 44
2019.

Authorized licensed use limited to: Ricerca sul Sistema Energetico RSE S.p.A.. Downloaded on December 19,2024 at 11:43:15 UTC from IEEE Xplore. Restrictions apply.