Author: Wolfgang Feist in January 2024

As far as the energy transition in Germany is concerned, there is widespread agreement among experts that the heat supply for space heating and hot water should no longer be covered by fossil fuels in the long term. However, which heat generators should be used instead is still the subject of heated debate. The alternatives under discussion include the switch to local and district heating, the substitution of oil and natural gas by alternative fuels on the primary energy side with renewable energy sources (so called E-fuels, e.g. hydrogen) and the increased use of biomass. However, a closer look reveals that none of these approaches mentioned so far will be able to provide an additional contribution of more than 10% of current consumption within around two decades - this is not the subject of this paper, but is easy to understand. What remains is the transition to fully electrical generation of heating and hot water ¹⁾ . This is a technology that has been used in individual buildings for decades and has made great progress in recent years thanks to the further development of heat pumps. However, a rough calculation shows (also on the basis of the figures documented in this paper) that direct electric heat generation is out of the question (operating costs are too high, but the total load on the electricity grid is also far too high). Heat pumps will therefore have to be the main source of heat in buildings; we estimate their share to be around 70% in the long term. Against this background, the question arises as to what additional electrical load will occur in the grid due to the operation of this number of heat pumps and how this can be sensibly covered by renewable energy wherever possible. This article deals with and quantifies both of these issues using published load curves for electricity and natural gas.

The highest 7-day average value for gas consumption by household (HH) and commercial customers (KV) to date was 3084 GWh/day, which occurred in the 6th week of 2018 (final energy). This is not an outlier; there have been other weekly averages at a comparable level in the past 6 years (Fig. 1).

_{Table 1 Space heating and hot water in DE in 2018; breakdown by final energy source (source: [Energy data], AGEB ) [7-Tages-Durchschnittsverbrauch = 7-day average consumption; Kernheizzeit = core of the heating period]}

The average final energy consumption of these sectors from natural gas was therefore 128.54 GW. In the sectors indicated, natural gas is used almost exclusively for heating and hot water. Gas stoves, which are used by around 6% of households in Germany, only consume around 3.3 GWh/day, or approx. 0.1% of this; by contrast, gas consumption in the industrial customer sector also includes a significant proportion of heating energy applications ²⁾ as can be seen from the annual trend in this sector; this is in the region of at least another 500 GWh/day in the highest load week. We do not include these two corrections (stoves and industrial sector) in the following analysis, which means that our calculations still slightly underestimate the total gas consumption for heating and hot water preparation.

According to the publication 'Energy data' [Energy data], the German energy demand for space heating and hot water preparation in 2018 was divided between the final energy sources as shown in Table 1. Natural gas accounts for a share of f_Gas = 48.9%. In the years 2019 to 2021, this breakdown remained almost unchanged. It can be assumed that this breakdown remains sufficiently accurate for a 7-day period during times of high heat load ³⁾ .

_{Table 1 Space heating and hot water in DE in 2018; breakdown by final energy source (source: [Energy data], AGEB ) .}

Existing natural gas boilers in Germany have average annual boiler efficiencies for heating and DHW heating in winter of around ηK = 89.2% (own calculation based on [Wolff 2004]) ⁴⁾ . The output supplied to the building heat distribution systems in the form of hot heating water for space heating and domestic hot water heating is therefore at least
$\displaystyle {P_{H \& W} = \frac {P_{Gas}\cdot \eta_K } { f_{Gas} }\;\;\;\;\; }$

If we use the values for the gas supply in week 6 of 2018, the gas share of this supply and the specified boiler efficiency, the ⁵⁾ highest contemporary 7-day average heat output requirement $P_{H \& W;7}$ for heating and hot water preparation in Germany is at least

$P_{H \& W;7}$ = 128,54 GW $\cdot$ 0,892 / 0,489 = 234,4 GW.

This entire output does not have to be replaced by heat pumps in future: An increasing proportion is also planned for the district heating supply ⁶⁾ and some of the existing biomass boilers, but also electricity heating systems, active solar systems and some oil and gas boilers will continue to operate. If we assume that heat pumps will cover 70% of the future ⁷⁾ space heating and hot water requirements, this results in a heat requirement $P_{H \& W;HP;7}$ for the heat pumps in the load week of a total of

$P_{H \& W;HP;7}$ = 234,4 GW $\cdot $ 70% = 164,1 GW.

The majority of these heat pumps will be outdoor air heat pumps ⁸⁾. In the winter load period, we estimate the average coefficient of performance for heating and hot water preparation of the heat pumps at COP7 = 3. This is an optimistic estimate also for future systems: the average outdoor temperature in these 7-day maximum heat load periods is below 0°C; under these conditions, commercially available heat pump systems today have COP values well below 3; it should also be noted here that we are talking about the use of heat pumps in existing buildings in their current state, i.e. without any significant improvement in thermal insulation. We deal with a scenario including possible improvements of the building fabric in section 6. This means that the additional heat pumps installed to replace oil and gas boilers result in an average additional electrical power demand Pel;HP in the German electricity grid of at least

$\displaystyle { P_{el;HP} = \frac {P_{H \& W;HP;7}}{COP_7} = \frac {164,1\; \text{GW}}{3} = 54,7 \text{ GW}_{el} }$.

This output occurs on average during the whole 7-day period - it should not be confused with the maximum load of the heat pumps (which is higher, but can be shifted by a few hours within the 7-day period). As this is a 7-day average value, this output cannot be shifted to another period, as the time constants of buildings that have not undergone energy-efficient refurbishment are at best sufficient for storage for around one day - but then the corresponding energy must be provided again the following day, so that the peak output on this following day is even higher. Electrical energy storage ⁹⁾ - which today comprise 0.04 TWh - of this magnitude (at least greater than 4 TWh) will not be available in 30 years' time either. The Pel;HP;7 output is therefore at least additional to the 7-day Pel;7 output demanded in the grid during this period and must be generated by additional ready-for-use power plants over this period.

If we evaluate the load data for the entire electricity grid in Germany for 2018 according to [E-Charts 2024], the highest 7-day average output to date is

$P_{el;7}$ = 72,4 GW.

See Fig. 2: The previous 15-minute peak load itself is around 15 GW higher at 87.1 GW. Outputs of this average magnitude occur regularly in the grid

_{See Fig. 2: The previous 15-minute peak load itself is around 15 GW higher at 87.1 GW. Outputs of this average magnitude occur regularly in the grid.}

The additional capacity to be made available for the heat pumps during the peak load periods for at least a 7-day average is therefore around 76% of the current maximum 7-day average load.

As we have already shown above, this additional capacity must actually be permanently ready-for-use during the relevant periods.

As the peak load period from December 5 to 20, 2022 shows (see Fig. 3), the maximum space heating load can also coincide with a winter lull.

_{Fig. 3: The heating load period from 5 to 17.12.2022 coincided with a continuous lull (electricity generation from wind power negligible). Source [E-Charts 2024]}

The relevant electrical output must therefore also be available in full from backup power plants (so-called “peakers”). These power plants must therefore be augmented by the value of Pel;HP . In a fully sustainable energy system, the associated primary energy must then be generated from seasonally stored energy sources; however, in a transitional phase or in exceptional situations, it is also conceivable that this could still be covered by natural gas, as this is only a limited part of the annual work. This is possible in terms of technology and energy quantities, but significantly increases the cost of supplying the heat pumps. ¹⁰⁾

For the more frequent periods with heat pump loads that are not in the doldrums, the electricity required by the heat pumps should come as directly as possible from renewable generation - wind power is particularly suitable for this, as the energy supply from PV is so low during the core heating period (CHP: November 23 to March 15) that it is completely absorbed by other applications. This means that additional wind turbines need to be installed. It can be seen that their average available output in winter should be increased by an amount equivalent to Pel;HP;CHP. This is the average electrical power drawn by the heat pumps during the whole core heating period. This results from the average gas demand in the CHP (almost exactly 2000 GWh/day or 83.3 GWGas) for heating and hot water applications, again divided by the gas share and the average coefficient of performance and multiplied by the efficiency of the gas boilers. If there is less than this output of around 36 GWel , then the peaker power plants introduced above must also make an increasingly higher contribution to the regular annual work for the heat pump power supply; this in turn leads to a higher demand for the fuel used for this, the production of which is associated with significant additional losses ¹¹⁾. The PEr method allows a more precise optimization of the power required for this, see [PEr 2013]. As the average winter availability of wind power in Germany is around 36% ¹²⁾, this results in an additional demand for wind turbines exclusively for heat pumps with an installed rated output $P_{Wind}$ of

$P_{Wind} = P_{el;HP;KHP}$ / 0,36 = 36 GW$_{el}$ / 0,36 = 100 GW_Inst.

This is 1.45 times the total installed wind power capacity in Germany at the end of 2023 ¹³⁾. This capacity is an addition due to the switch to heat pumps to the expansion capacity required for all other applications. If the expansion is spread over 25 years, an additional 4 GWInst of wind power capacity must be installed each year in addition for the heat-pumps. There would then be surplus production from these plants in times of lower heat demand - this would be used to generate storable energy (e.g. in the form of hydrogen or e-methane) according to the PEr concept. The exact demand for this branch of the future energy system can be determined using the PEr method ([Grove-Smith 2016], [PEr 2013]), which, under the boundary conditions given here, results in slightly higher values for the installation requirements of additional renewable electricity generators than in the approximation presented here.

Through the consistent application of energy-efficient building component refurbishment for all existing occasions, e.g. replacing windows anyway, the heating demand of buildings can be reduced by at least 50% within the conversion period of 25 years [Bastian 2022]. This would reduce the power requirement by a good 50%, i.e. to only around 27.4 GW, even at peak heating load times. The coefficients of performance of the heat pumps for generating this heat increase in addition by at least 15% due to the lower temperature level of the heat transfer, which reduces the electrical load again to around 23.8 GW. The cooling time constants of the buildings also increase due to the lower heat loss, and the temperatures drop less quickly and to a less low level with possibly lacking heating outputs: However, we do not quantify these latter potentials here. Both the required backup power Ppeak as well as the additional wind power capacity to be installed for the heat pumps will then remain within a feasible range: Ppeak = 24 GWel ; PWind = 44 GWInst or additional annual installations of around 1.75 GWInst /a. Even this additional installation requirement is still an ambitious target and comes as an addition to the expansion requirements planned to date; however, these figures appear achievable with a little effort ¹⁴⁾.

_{Fig. 4 Load profile in the electricity grid (2018 total); source [E-Charts 2024].}

We have already determined the total final energy consumption for space heating and hot water elsewhere (see Breakdown of final energy consumption in Germany). In 2019, this resulted in a final energy demand for space heating of 26.6% and for hot water of 5.3% of the total final energy demand, corresponding to

A total of around 32% of final energy consumption in Germany is used to generate space heating and hot water, a higher proportion than any other application in the energy balance, followed by transport ¹⁵⁾. This breakdown of annual consumption values is consistent with the breakdown of power requirements that we determined in sections 1 to 5. We would like to point this out again separately because some statistics that do not focus on energy use but on the provision of energy sources give a different impression, making space heating appear far less significant: This is due to the fact that, for example, CO2 emissions are allocated to power plants and heating plants without considering the intended use of the electricity or heat generated there. However, the reason for the operation of a power plant or heating plant is the electricity and heat demand of the buildings that are supplied from it. It is important to classify this correctly, as this is the only way to correctly estimate the conversion costs of the energy system.

For a realistically implementable heat supply transition (= conversion of heating and hot water preparation to renewable energy), the following prerequisites result from the analysis presented here:

• There is no alternative to largely switching to electrical energy as the final energy source for heating systems. The basis for this are heat pumps - direct electric heating systems have far too high a power requirement.
• However, the resulting additional average power requirement in the electricity grid in winter is significant. This results in two further indispensable components for the success of this changeover
  1. Wind energy generation in particular must be ramped up to a much greater extent than previously planned.
  2. The heat output requirement of existing buildings must be noticeably reduced (by at least around 50% on average).

The stated prerequisites can realistically be implemented within a period of around 20 to 30 years - at the end of this period, CO2 emissions for space heating will be in the order of zero. However, implementation presupposes that all opportunities to switch to heat pumps (e.g. replacing a boiler anyway) and all occasions to improve thermal insulation (e.g. re-roofing, replacing windows, painting of facades) are used to retrofit the systems and components concerned with state-of-the-art energy efficiency.

[Bastian 2022] Zeno Bastian, Jürgen Schnieders, William Conner, Berthold Kaufmann, Laszlo Lepp, Zack Norwood, Andrew Simmonds, Ingo Theoboldt: Retrofit with Passive House components; Energy Efficiency 1/2022

[BNetz 2024] Bundesnetzagentur: Wöchentlicher Gasverbrauch, webbasierte Datensätze unter https://www.bundesnetzagentur.de/DE/Gasversorgung/aktuelle_gasversorgung/_svg/GasverbrauchSLP_woechentlich/Gasverbrauch_SLP_W_2023.html?nn=870134 am 14.01.2024 abgerufen

[Energiedaten] Hrsg. BMWi, http://www.bmwi.de/Navigation/DE/Themen/energiedaten.html

[E-Charts 2024] Web-basierte Daten zur Stromversorgung https://energy-charts.info/charts/energy/chart.htm?l=de&c=DE&interval=year&year=-1&source=total; abgerufen am 17.01.2024

[GIE/AGSI 2024] Gas Infrastructure Europe, AGSI, webverfügbare Daten unter https://agsi.gie.eu/, abgerufen am 17.01.2024

[Grove-Smith 2016] Grove-Smith, Jessica; Wolfgang Feist; Benjamin Krick: Balancing energy efficiency and renewable energies: An assessment concept for nearly zero-energy buildings. In: Bertoldi, P. JRC of European Commission (ed.): 9th International Conference Improving Energy Efficiency in Commercial Buildings and Smart Communities, 2016. p. 894-902. Link zum externen Artikel hier (In Englisch)

[ISE 2024] Öffentliche Nettostromerzeugung in Deutschland im Jahr 2023. Fraunhofer ISE. Abgerufen am 14. Januar 2024.

[PEr 2013] „Primärenergie Erneuerbar – PEr“. In: Passipedia (weitere Literaturquellen ebenda). Passivhaus Institut 2013-2023. https://passipedia.de/grundlagen/energiewirtschaft_und_oekologie/erneuerbare_primaerenergie_per

[Wolff 2004] Wolff, Dieter; Peter Teuber; Jörg Budde; Kati Jagnow: Felduntersuchung: Betriebsverhalten von Heizungsanlagen mit Gas-Brennwertkesseln, Fachhochschule Braunschweig Wolfenbüttel, April 2004

_{Fig. 5 Weekly average values of total gas consumption (incl. industry and power plant sector) in Germany, 2018-2021 as well as 2022 and 2023. A pronounced annual variation can also be seen here; this indicates that there are also significant proportions of space heating applications in the other sectors; in this paper, we refrain from quantifying these and adding them to those already shown because a differentiation, e.g. from other applications, requires detailed information on the respective individual cases.}

The original German article on Passipedia :Zunahme der elektrischen Last im Netz durch eine systematische Wärmepumpen-Strategie

Note: This information is also available in a video recording: “Load of heat pumps in the German electricity grid”. (in German)