In a previous article (https://edenergy.be/the-myth-of-flexibility-production/?lang=en), we drew conclusions about the optimal production park and the measures to take regarding grid investments and research development, aiming to limit the societal and individual costs of electrifying two major needs: dwelling heating and individual transport.

Although the conclusions were straightforward, nuances were expressed concerning the figures. Indeed, using historical consumption patterns raises questions in a world where the evolution of needs is as significant as those related to the energy transition.

As announced, I made calculations to estimate the consumption for dwellings heating when they all will be equipped with heat pumps.

The first thing to do is to assess the current existing related consumption across the different fuels.  The pie chart (source = Belgian FPS economy) below gives the distribution of the use of the different fuels for the dwellings heating, in terms of number of households:

So, 66% of the households are heating their dwelling with natural gas.  This heating includes the preparation of warm water, although this does not represent more than around 5% of the total. Heat pumps have been separated from the direct (or accumulation) electric heating, that is a good point to avoid a double count for them.

In Belgium there were 11,76M people (grouped in 5,16M dwellings) in 2024 and their annual energy consumption across the different uses is:

Thus heating (including preparation of warm water) represented 7.320 kWh/year and person, thus a total 86 TWh/year.

Applying this value on the pie chart data gives:

These figures are consumptions, to translate them into heating energy needs, we have to apply the yield of the used heating appliances and the ratio GCV (gross caloric value) lower/higher for each fuel. With the following distribution we have (left hand natural gas, right hand fuel oil):

Electricity (direct or accumulation) have a yield of nearly 100%.  In a previous article we have shown that the effective yield of a heat pump (air/water type) is of 142% across the year (far below the announced COP of 3 to 5).  For the remaining (representing together 5% of the total) we can use: coal – 50%, propane-butane – 85%, biomass – 75%, solar thermal 100%.

So that the total heating needs is actually of 75.059 GWh/year (thus 75 TWh/year):

We can now distribute this yearly total need across the year using a standard dwelling heating consumption profile.  But here we have a problem: since the go live of MIG6 (in 2022), the SLP’s for gas (which made the distinction between gas consumption for residential and other usage) have been replaced by a unique RLP0N gas (reflecting the whole gas consumption without distinction on its usage), so that the last distribution we can use is the one from 2021.  Grouped by month this gives the following in MWh (1TWh = 1.000.000 MWh):

Assuming that this total need must be produced by air/water heatpump the related electricity would be of this amount divided by 142%, thus nearly 50 TWh/year.  But, since the yield of the air/water heatpump depends on the atmospheric temperature, we have to use it to determine the consumption, we can’t apply the weighted mean of 142% on each monthly total.

Assuming the heatpumps are provided with a vessel typically containing the heating (buffer) to satisfy the needs of one day, we can use the daily average temperature to assess the daily consumption.  Reusing the formula resulting from a previous article (see: https://edenergy.be/heat-pumps-paradigm-or-fallacy/?lang=en) we have:

r = K . (Te – Tb)² / (Ti – Te)

with: K = 0,0333, Te = atmospheric temperature, Tb = heatpump fluid boiling temperature (-20°C), Ti = inside (dwelling) temperature (20°C).  So that we are able to translate the heating needs into electricity consumption at day level.

For the atmospheric temperatures, each month has been selected in the years from 2006 till 2019, where its cumulated degree days was the closest to the normal value for this month (e.g: 2019-01, 2009-02, …, 2014-05, …, 2017-12).

Since the air-water heatpump has got a one-day demand buffer, the daily electricity consumption has been calculated as the daily sum of the hourly heating needs divided by the ‘r’ for this hour.  Then, assuming a flat distribution for the moments where the heatpumps refill their buffer, this daily consumption has been distributed along the hours, according to the ratio of the hourly ‘1/r’, divided by the sum of the ‘1/r’ along the day.  In the table below, the P is the electricity consumption by month, for the heating needs (see the table above with the monthly total):

The global COP of the heatpump is thus of 1,37 (75.059 GWh heat produced, with an electricity consumption of 54.909 GWh), in line with the calculation of previous article.  Not surprisingly lay the consumptions for the coldest months (i.e: January and February) above the heating needs for these months.  The yield of an air-water heatpump is lower than 1 when the outside temperature dives somewhere under 3°C.

Now we have all the necessary elements to calculate the optimal combination of productions types (solar, wind onshore, wind offshore and nuclear), to examine to which flexibility needs they lead.

First, we have to emphasize that the electrification of the dwelling heating will lead to an extra 64% of the total electricity demand that exists without (or prior to) this electrification.  An enormous increase that will also have a huge impact on the necessary grid investments, to enable its reinforcement to cope with this additional consumption.

Since 2/3^rd of this extra consumption occurs in 3 months (December, January & February), production assets that must be kept running all the year (like nuclear) can’t lead to a good fit.

The figures are crazy, that’s right, it is because the need of flexible (steerable) production assets is huge due to the strong seasonal difference in heating needs and the poor effective COP of the air-water heatpumps.  These figures, without any other further calculation make already the evidence that a molecule will be the mandatory route, as mentioned in previous publications.

I need to repeat that, however this sounds like ‘the fossil fuels strike back’, it is definitely not the case.  Renewable will make the molecules (e.g. via electrolysis followed by synthesis with [captured] CO2) during production surpluses.  And during scarcities, the molecules will deliver power (e.g. in fuel cells, or burned, with capture of CO2) in easily manageable units, like nano to small CHP’s so that the remaining energy (thermal losses) could be recovered for heating purposes.

Nevertheless, I made the calculation of the necessary electricity storage for each of the 5 cases. They range from 427 GWh to 677 GWh, based on autonomy of maximum 1 day in average circumstances (does not in extreme ones), totally unrealistic, even in 20 years.

You will also notice that when the nuclear power reaches 6.000 MW (6 GW), the optimal solar power (11,3 GW) is very close to the one already installed (10,8 GW).  The maximum of nuclear power that could be installed after complete electrification of the dwelling heating, is thus arround 6 GW as of today.  This figure will decrease over the years, due to the increase of solar panel installed power (according to European guidelines), so that by 2030 it would become suboptimal to have more than 3 to 4 GW of nuclear power, and probably 0 GW by 2050. Furthermore, we already know that due to its lack of flexibility and the very strong seasonal character of the demand, nuclear power can’t ever be a good match.  Without speaking about the physical impossibility to treat the waste, that will require a million year care, nor the sky high prices associated to the new nuclear power (Flamanville: 19 billions € for 1,6 GW – 6 times of an equivalent offshore wind turbine park cost, leading to a necessary market price of 130€/MWh to reach rentability).

The conclusions of the previous publications remain valid and are reinforced:

• heating needs will become very expensive in grid investment if the heat pump would be generalized, it is much better and cheaper to invest in dwelling insulation and promote hybrid (gas fired) boilers (0,3 – 1 kW electric and 1,5 kW thermal CHP + 5 kW extra boiler) with CO2 capture and injecting back into the grid pipes
• target 0 MW of nuclear power (by the end of the decade) and in the meanwhile limit the nuclear power installed to 2 GW a fast as possible, in parallel stimulate nano to small CHP (roughly from 1kW till 10 MW electric), with CO2 capture and invest in renewable fuel development
• flexibility, although it remains being a part of the solution, is not an option to bridge solely the gaps and shave the peaks (that will last over several days till 2 weeks), it can help but this help is very limited

Any other construction will lead to very important cost increase at societal level, in particular for all residential customers, and establish a permanent climate of anxiety and stress on the grid users (as well households as industrials and commercials).