Breaking down household natural gas use

We are a typical UK household with a connection to the national gas grid. This means that we use natural gas to heat our home, heat our water, and cook our food.

Although I know how much gas we use overall, I don't have a breakdown on how much gas we use for each activity, or even each appliance. Working this out is important to successfully prioritise activities to reduce gas consumption. I can look up national averages, but these aren't personalised and can be imprecise for individuals.

So how can I work out individual breakdowns for space heating, water heating, and cooking?

The trouble with gas meters

Our gas consumption is measured by a meter where the supply enters our house, but we don't have individual meters for individual appliances. If I want to work out electricity consumption for individual appliances I can buy plug in Watt-meters. The same is not true for gas appliances, all of which are plumbed directly.

What I do have though is a smart meter that can give me half hourly gas usage.

The first thing I tried was to download this data and see if I can tell apart individual appliances from the pattern of my gas use. Here's an example from the 27th February this year:

You can see the heating turn on in the morning, a burst of usage as the water in the radiators is brought up to temperature, and then lower usage throughout the day. You can also see a burst of use about 4pm when the hot water turns on to heat the cylinder.

What you can't tell though is how much usage was for hot water and you can't tell when, or indeed if, I was cooking with gas that day.

If I looked at days when the heating wasn't on, the picture was clearer, but still not perfect. If you look at this data from 9th April 2022, you can see that the heating wasn't on because there's no big block of usage. You can see that we cooked breakfast with some gas, heated water at 4pm again, and then cooked an evening meal about 6:30pm:

This is clearly better, but what if I cook when I'm also heating water? I'd have to keep a note of what time I cooked each day and then manually curate the data. This would be time-consuming and error prone. I needed a better way.

Alternate days for water heating

We don't have a combi-boiler that heats water on demand. Instead we have a boiler that heats an insulated tank of water that we then draw on when we need hot water. Although this may seem like the minutiae of domestic hot water supply, this detail means that I have another option.

Our tank is 210 L. This is quite a lot of hot water and certainly more than we use in a typical day. As such, we don't need to heat the water every day. In fact, we can easily go two days without re-heating the tank and can extend that to three days if we don't have a bath or use a lot of hot water for something else.

(Incidentally, this has a small added perk in that it reduces overall heat loss from the cylinder because the average water temperature is lower. That's a topic for another day though).

Realising this possibility, I set our boiler to only heat the tank on Tuesdays, Thursdays, and Saturdays. This gives us a maximum gap of three days and ensures we still have plenty of hot water. The critical bit is that it allows us to see how much gas we use on Mondays, Wednesdays, Fridays, and Sundays. This usage is just cooking:

Now you can clearly see days that are higher, with cooking and hot water heating, and days that are lower, with just cooking. These days follow a regular pattern which makes analysis easier. I ran with this pattern of hot water/cooking for the entire summer of 2022.

If I also include the winter data, I now have three types of day:

Type of day	What was happening that day?
"Heating"	Heating, but also maybe hot water and maybe cooking
"Hot water"	Hot water, and maybe cooking, but no heating
"Cooking"	Cooking only. No hot water or heating

From just the cooking days I can work out how much gas we typically use for cooking.

Then I can subtract this average cooking value from the "Hot water" days to work out how much we typically use heating water.

Finally, for heating I have two options. I can repeat the same process, subtracting how much we typically use for hot water and for cooking, or I can just look at the annual usage for our house and subtract hot water and cooking. I chose the latter for its simplicity and lower margin of error.

The final breakdown

In table form, for 2021, this is:

Gas use	Usage (kWh/year)	Usage (%)	Emissions (kg/year)
Heating	9,940 kWh	88%	2,030 kg/year
Hot water	1,050 kWh	9%	200 kg/year
Cooking	330 kWh	3%	70 kg/year
TOTAL	11,300 kWh	100%	2,300 kg/year

My personal takeaway here is that cooking and hot water are such small fractions of the total that almost all effort should be focussed on space heating. Switching space heating to something that emits less carbon dioxide will mean getting a heat pump and running it on electricity. This will also reduce emissions from hot water production. I will be exploring heat pump economics in future articles, especially considering how to retrofit one into an existing 1950s house.

Conversion of our gas cooker to an induction hob will also be explored, but it's clear that this is a relatively low priority target when it comes to carbon dioxide. There may be other arguments for switching, including improved internal air quality, or better cooking experience, but there's unlikely to be a strong carbon dioxide-based argument when there's still a gas boiler in the house.