The Great Thermal Mass Myth................

[Jeremy Harris]

The term "thermal mass" comes up time and time again on building related forums and discussions, yet as a parameter it has one notable feature - it does not really exist.

There is no such thing as "thermal mass" and never has been.

Mass is a simple physical property, in simple terms it's approximately how much a given volume of something weighs at the surface of the earth. This, in turn, depends on the density of the material.

For example, here are some densities for some common building related materials, in terms of the weight at the Earth's surface for a 1m square cube of the stuff (1m²?

Brick ~ 2000kg/m²

Concrete~ 2400kg/m²

Plaster and plasterboard ~ 2700kg/m²

Water ~ 1000kg/m²

Structural softwood ~ 550kg/m²

Typical hardwood ~ 700kg/m²

Granite ~ 2700kg/m²

On its own the mass of a given volume of material isn't that useful for working out how much heat it would take to either raise the temperature of the stuff, or for it to give off heat as it cools down. What we need to know is the specific heat of the material, expressed as the amount of heat energy (called sensible heat, which can be measured in Joules, J) needed to change the temperature of a certain mass (lets say 1m³ to match the data above) by 1 deg C (or more correctly a deg K, but it's the same thing for this purpose).

So let's list the same materials as above, with the amount of heat energy we need to put into increase the temperature of 1kg of it by 1 deg ?

Brick ~ 840 J/deg C

Concrete ~ 880 J/deg C

Plaster and plasterboard ~ 1080 J/deg C

Water ~ 4200 J/deg C

Wood ~ 1700 J/deg C (This is an average value, as the true range is dependent on variety, with a wide range, from 1200J/degC/kg to around 2300J/deg C/kg)

Granite ~ 790 J/deg C

So, if you want to create a house with the highest "thermal mass" (i.e. Heat capacity per unit mass, if that's a reasonable way of trying to define this unknown term), then here is a list of materials, with the highest heat capacity for 1 kg at the top, and lowest at the bottom:

Water
Wood
Plaster or plasterboard
Concrete
Brick
Granite

You may well spot a few odd things here. The first is that you cannot build a house with water (but you can include water as a heat distribution or storage system). The second is that concrete, brick and stone aren't great materials in terms of storing heat for a given mass.

Surprising, isn't it? Even more so when building professionals keep harping on about the virtues of so-called "thermal mass".

[ToughButterCup]

Thanks for this. I will link to this from my own blog.

People choose the wrong terminology all the time. And sometimes it causes confusion. But mostly, it doesn't, especially when -as most of us do- we go for a 'percentage' of what someone means. I think when people use the mis-applied term thermal mass , they might mean - it-keeps-its-warmth.

If you wanted to tell someone that your house holds on to its warmth well (or conversely stays warm for ages when it's freezing outside) which term would you use?

[Jeremy Harris]

I think the most useful parameter overall, in terms of perceived comfort, is the thermal time constant.  This is the time it takes for the house internal temperature to change in response to any change in the external temperature, in essence.  Say your house was at 21 deg C and during the night the outside temperature dropped to 0 deg C.  If the house has a long thermal time constant then it will be only a fraction below 21 deg C in the morning, with no heating, as the combination of the rate of heat loss, the heat capacity of the internal components of the house and the thermal conductivity of those internal components, will combine to release heat from the internal structure at a rate that is close to the heat loss rate, so the internal temperature doesn't drop much.

Exactly the same thing happens the other way around in summer, when it may be hotter outside than it is inside.  The ability of the materials inside the house to conduct and store heat will tend to keep the house internal temperature from rising during the hot part of the day.

The main things to watch are that the insulation used has a long decrement delay (there's a good explanation of that here: http://www.greenspec.co.uk/building-design/decrement-delay/ ) and that the materials inside the house have a high heat capacity and a reasonably good thermal conductivity, so that they can not only store heat, but that heat can flow into, and out of, them at a rate that is fast enough to maintain an even temperature.

In general, a house that has a long thermal time constant may have virtually no diurnal or seasonal temperature variation, and so be at a comfortable temperature all year around.

 

 

[reddal]

On 20/05/2016 at 16:08, JSHarris said:

So, if you want to create a house with the highest "thermal mass" (i.e. Heat capacity per unit mass, if that's a reasonable way of trying to define this unknown term), then here is a list of materials, with the highest heat capacity for 1 kg at the top, and lowest at the bottom:

Water
Wood
Plaster or plasterboard
Concrete
Brick
Granite

You may well spot a few odd things here. The first is that you cannot build a house with water (but you can include water as a heat distribution or storage system). The second is that concrete, brick and stone aren't great materials in terms of storing heat for a given mass.

Surprising, isn't it? Even more so when building professionals keep harping on about the virtues of so-called "thermal mass".

 

Hi,

I agree with your point about the term 'Thermal mass' being much misunderstood and misused.

Its interesting to think about alternative measures of how good a material is at providing heat capacity. i.e. instead of Heat capacity per unit mass - we could look at Heat capacity per unit volume. I think ordering by that would give :

So wood does worse under this measure and concrete/granite a little better.

We can go further and estimate the cost of installing a m3 of each material (installing a m3 of water is going to involve tanks or pipes etc so isn't cheap!). I made up some numbers for the costs - which may be off - then used these to calculate the Heat capacity per unit cost. Ordering by that gives :

So in terms of heat capacity per £ spent - Concrete is the winner by a distance - and wood is very bad.

Maybe my estimates and/or calculations are wrong?

So we have 3 different measures - each of which captures a different aspect of what people mean when they say 'thermal mass' :

• Heat capacity per unit mass
• Heat capacity per unit volume
• Heat capacity per unit cost

Different materials look better or worse depending on which measure you use. 

Not sure if that is useful or just adds to the confusion - I just thought it was interesting :).

- reddal

[Jeremy Harris]

As I mentioned above, though, heat capacity on its own isn't a great deal of help, unless the material is sufficiently thermally conductive to allow heat to transfer in and out of it at a rate that is consistent with being able to regulate the temperature of the interior of the house.  This is one reason why the general rule of thumb on the impact of added heat capacity suggests that it is only the internal surface layer that has a really significant effect, and that anything that is deeper than around 100mm from the surface has almost no effect at all.

What is clear is that if you have a house where the insulation is on the outside (which is generally a good thing in terms of increasing the thermal time constant) then that insulation not only has to be effective in terms of having a high thermal resistance, but it also has to have a relatively high heat capacity, as the combination of the two increases the decrement delay.

[ToughButterCup]

2 hours ago, JSHarris said:

[...] but it also has to have a relatively high heat capacity, as the combination of the two increases the decrement delay.

Finally, 'click'. 

I might now be able to explain this issue to someone else now.... a proper test of understanding. I'll start with my architect : a 'thermal mass' sinner.

[Jeremy Harris]

There's an easy test you can do with insulation, all it needs is for you to touch it.  If it quickly warms up when you hold your hand on it, it almost certainly has a low heat capacity (think EPS, for example).  If, on the other hand, the insulation takes some time with your hand resting on it before it starts to feel warm, then the chances are it has a high heat capacity (or it may just be a lousy insulator!).

It's noticeable that wood fibre and cellulose don't "feel" like good insulation when you're handling them, they feel cooler and don't heat up very quickly.  They are pretty good insulation materials, but they also have a moderately high heat capacity (a lot higher than, say, EPS).

 

[tonyshouse]

I like living in a very dense high mass house, stable temperatures = comfortable, very slow to change temperature.

There is a problem with less dense materials as more volume is needed to store the same ammount of heat, (water excluded)

Would it be possible to define your thermal time constant please?

[Jeremy Harris]

I've never seen a formal definition of it Tony, but I think most people feel that houses that don't fluctuate in temperature rapidly seem more comfortable.

If I was to try and come up with a definition, given the range of diurnal temperature variation we typically get in the UK, then I think Id say it's the time taken for the internal temperature to change by 1 deg C for an external temperature step change of 10 deg C.

Our new build barely seems to change temperature at all, even in really cold weather it's unusual to see the internal temperature drop by more than about 0.3 deg C overnight, with no heating. I've not plotted it out yet (must get around to putting the logger in one day) but I'm guessing that there is a barely perceptible diurnal temperature variation internally, and that the time constant using the definition above is probably greater than 36 hours, maybe much longer.  It's hard to tell, as there is almost always some solar gain, even on the bleakest of days, something I have found a bit surprising.

 

[Lurkalot]

If I understand this (I probably don't) then the thermal delay of insulation is based on the amount of heat the material can store, combined with the rate the heat is transmitted through it. I'm probably getting too technical for myself here, but does this affect the way u-values are calculated?

For example if a material's heat transfer is measured when it is cold, for the initial time period some of the heat will be absorbed into the material and not transferred making the u-value appear lower than it would be if measured after heat has been applied for a few hours and the material cannot absorb any more at that temperature. Is this accounted for when testing? Am I completely misunderstanding this whole thing?

[tonyshouse]

U values are measured at steady state

buildings exist in the real world where temperatures, wind speeds, heating regimes and solar radiation vary all the time.

[SteamyTea]

I was pondering this yesterday (actually be thinking about it for a decade).

The really hard thing to understand about thermodynamics is fivefold.  You have power, energy, mass, temperature and distance.  So 5 dimensions, and we are not very good at thinking in 5D, most if us struggle with 3D and we live in that universe, stick on a 4th (time) and our brains start to seize up.

So back to basics.  2D

Imaging an insulating material that allows no energy to be transferred and has no dimensions.  You can pump in as much energy, as fast as you like, into one side of it, and there is no change the other side of it.  Quite easy to imagine this (even though it is totally impossible, but then I like 6 impossible things before breakfast).

Now imagine a dimensionless material that has the ability to absorb no energy at all so no changing temperature, so you can pop this into the sun and nothing changes (my second impossible thing).

Now put the two imaginary materials together, so you have new material that neither transfers or changes at all, no matter that you do to it.  What will happen?  Well nothing, it will just sit there.  We could give it some other dimensions i.e. length, width, height, mass, and still nothing would happen because any energy absorbed, will instantly be transferred though the material and come out the other side.

So let us change the material a little bit.  We add something that slows down this process, so in effect we change it from a perfect conductor to a less perfect conductor, so now it has a real thermal conductance.  This changes the way we calculate it, rather than thinking of energy passing though instantaneously, we now have to think of power.  Power is energy per unit time (time is just another dimension).  But as our material still absorbs all energy without changing at all i.e. no temperature rise or change in shape, we now have something that can suck in everything, just a bit slower, in effect it stretches time, which affects power (my third impossible thing of the morning).

So we are still no closer to really understanding normal material thermodynamics, so we shall change the material slight to allow it to not just absorb energy unchanged, but to heat up a bit.

We know this as heat capacity.  So for any level of energy we pump in, we get a temperature increase, we usually measure his in one of two ways, by volume or by mass.  But at the moment, out material does not have either of those properties, it is still imaginary remember.

But no matter, we can still observe what will happen if we increase the energy level now that it has the real properties of conductance and absorption, basically it will take a bit of time to warm up.  This is now in the real world and something we observe all the time.  The difference between say a kettle warming up and out imaginary material is now only the dimensions of distance and mass.

So let us think about distance for a moment, while remembering that we are still working in 2D only.  Keeping the intrinsic thermal properties of the material the same, we will give it length only to start with, another impossible thing, my forth.

What will happen is that if we add a packet of energy at one end of the material, it will slowly, or quickly, travel to the other end.  The thermal conductance sets how long this will take for any given length (W.m^-1, as W is energy [J] per unit time

).  So shorten the length, it gets to the end faster, lengthen it, it takes longer. That is easy to understand, so not an impossible thing at all.

At the same time our material is absorbing some of the energy at increasing in temperature (J.K^-1).  This is known as the conservation of energy, there is the same amount of energy, it has just changed form (to thermal, I never specified what form it was in at the start).  So for every increment along the length, there is less energy to move, so the next bit of the material warms up less.

So now we have two ways of describing our material, the conductance and the absorption and you can see that they are linked by 2 dimensions, power and length (W.m^-1.).  So hopefully you can see that, at this point in this essay, mass makes no difference.

So now let us add this mythical material property, mass.  Mass is an easy way to describe something on Earth.  We don't talk about how many grains of sand we want, or how many atoms are in a person, but we could.  So mass is really just an easy method of describing one property of a material.  In some ways it would be much easier to think of relative density to a hydrogen atom (or silicon).  Oddly mass is the one SI unit that is not defined yet (and is odd because we use the kilogram rather than the gram).

Now the real trouble starts as materials, even our imaginary material, do not have consistent thermal properties.  Take a kilogram of water at 4°C and it has different thermal properties to a kilogram of iron at 4°C.  The water will absorb more energy for the same temperature increase.  This is all do do with the internal structure of pure materials and makes it impossible to just use mass alone as an indicator of all thermal properties (the fifth impossible thing, though not one of mine).

So you can see why we dislike using mass as a descriptor for thermal properties and how it affects the temperature within a building.  To describe what is going on we need to use two equations, thermal conductivity [W.m^-1.K^-1] and heat capacity [either specific J.kg^-1.K^-1, or volumetric J.m^-3.K^-1].

Now it is possible to combine these two for any given material, within a fixed temperature range, by just multiplying them together.  It is also much easier to do this using volumetric heat capacity as no density conversion is needed, but either can be used.

I = W.m^-1.K^-1. J.m^-3.K^-1.

This needs a bit of jiggling about to get to something that is understandable to end up with

I =  J.s^-1.m^-1.K^-1.J.m^-3.K^-1 which looks like an impossible formula (almost my 6th of the morning), but is really similar to acceleration and can be thought of in the same way.  But may be easier to rearrange it to:

I = J.m^-2.K^-1.s^-1/2

So you can see that rather than think of mass being the overriding factor, it does not even feature in a formula that fully describe the thermal properties of buildings.

Now that really is my 6th impossible thing before breakfast.

 

(I do have a 7th impossible thing, get the effing forum colours sorted out, it has taken me an hour to type up and I have a migraine now, all it needs is a pale grey or pale blue background where you type, that cannot be hard to sort out).

 

Edited June 14, 2016 by SteamyTea

[SteamyTea]

Why have I got a trike though in the last bit.  I can't get rid of it either.

Seem to have sorted that using Word.

I may have to edit the above later when I can see again.

Edited June 14, 2016 by SteamyTea

[mafaldina]

(I do have a 7th impossible thing, get the effing forum colours sorted out, it has taken me an hour to type up and I have a migraine now, all it needs is a pale grey or pale blue background where you type, that cannot be hard to sort out).

Try turning your screen brightness down a tad.

[SteamyTea]

Yes right, never thought of that, it is at the lowest setting.

[Jeremy Harris]

That's a more eloquent working through of my thinking when I very deliberately chose the provocative original thread title, ST.

The challenge, as always, is how to clearly convey concepts that may be a bit difficult for some to understand (primarily those without a background in physics in this case) in a way that most can wrap their head around.

BTW, there are still things going on to look at how the screen colours might be changed - from what I can gather it may not be particularly easy and needs some themes purchased and then set up so they are user-selectable.

[mafaldina]

If you use Chrome download the Care your Eyes extension. A lot of dyslexic students use this to get the best background colour for their individual needs.

 

[SteamyTea]

I think we should keep this to thermal properties and not my eyesight.

There are 1200 words there if anyone wants to edit it.

Edited June 14, 2016 by SteamyTea

[TerryE]

On 22/05/2016 at 20:09, tonyshouse said:

Would it be possible to define your thermal time constant please?

I was a mathematician by training, so I am one of those rare wierdoes who is comfortable thinking in equations. So if you are like me then this Wikipedia description is useful: Thermal time constant.  The physics of this is dictated by what is called the heat equation, and this is classed as what is known as linear time invariant systems.  This formula has the dimensions of time, that is you can measure it and quote it in seconds, hours of whatever time unit takes your fancy, and it is a measure of how sluggish your system is to respond to step changes in external conditions.  Dropping the funny formula symbols, this article includes a summary: In other words, the time constant says that larger masses and larger heat capacities lead to slower changes in temperature, while larger surface areas and better heat transfer lead to faster temperature changes.

I would say: the longer the time constant the more sluggish is the response to changes in external temperature.   The time constant for the walls of my house with its external stone skin and twinwall filled with cellulosic filler is a few days, and because of this I don't really have to worry about diurnal temperature variations in designing my heating system

Edited June 3, 2018 by TerryE
Typo in heating system

[SteamyTea]

In other words, the time constant says that larger masses and larger heat capacities lead to slower changes in temperature, while larger surface areas and better heat transfer lead to faster temperature changes.

That is one way to describe it for homogeneous materials.

I think the main thing is to not assume that the mass of the material is the overriding property, it is the combination of the two and the shape.

Now, for a laugh, let us do a composite material with constantly variable temperatures either side.  Who wants to pick the partial difference equation, and who wants the geometric model.  I will use the statistical model.

Actually I won't, I am going to sit on a boat in Aylesbury.

[Jeremy Harris]

There's been some work done (I think by the BRE, but it might be someone else) that shows that for normal masonry building materials thermal conductivity limits the effect on the thermal time constant in practice.  The general rule of thumb is that anything more than 100mm below the surface doesn't have any significant impact, because heat flow is so slow by the time you get down to 100mm that it can't have much of an effect on regulating the temperature of the house.

This means that with insulation outside the thermal store created by the structure, there's not any significant merit in increasing the thickness of the internal structure above about 100mm, assuming masonry construction.

 

[TerryE]

5 hours ago, SteamyTea said:

Who wants to pick the partial difference equation, and who wants the geometric model.  I will use the statistical model.

Come on Nick, why the cynicism? If you remember I did a post on the eBuild forum where I did just that for my wall profile and did some time simulations showing the time response to various external temperature profiles, and I showed the cross-sectional heat profiles. The equation is the standard heat equation. OK, it was a 1D model, but that was good enough.  It should that for the standard twinwall with an external stone skin, the time constant was so long compared to 24 hrs that I really didn't have to worry about diurnal effects or modelling them when doing my Heat balance calcs. A simple steady state approximation is perfectly adequate.

At that point I stopped doing time varying thermal models - no point.

[tonyshouse]

JSH was saying, "The general rule of thumb is that anything more than 100mm below the surface doesn't have any significant impact, because heat flow is so slow by the time you get down to 100mm that it can't have much of an effect on regulating the temperature of the house."

this may well be true for poorly insulated high heat demand buildings but the better insulated a building is the less true this becomes.

 

[Jeremy Harris]

The thermal conductivity of a material doesn't change just because you add insulation to one side of it.  The key thing here is how long it takes for heat to move from deep in the material to the surface and how quickly the surface can dissipate that heat into the house.  The latter depends primarily on the temperature differential between the air in the house and the material; the former depends on the temperature differential and the thermal conductivity of the material.

Once you reach the point where the time taken for heat to travel from deep in the material to the surface is greater than the diurnal fluctuation in temperature then it's effect becomes small, especially when you consider that deep inside an internal, well-insulated, wall, roof or floor structure the temperature is only going to have a tiny temperature differential compared to the house, and so heat movement will inherently be a lot slower than in a less well insulated house, where the air in the house may well have quite a wide diurnal variation. 

[TerryE]

I think there are separate but related discussions about use of "thermal mass" (and yes, that meant to tease J) for external bulking and interior bulking.  In my and Jermey's house the bulk of the thermal capacity is on the exterior shell to the living environment, that is in the slab and the the filling in the the Larson strut frame.  This acts a huge high-stop filter on external variations.  Any internal bulking is intended to top up heat losses.  The is a lumped system, and any internal heat capacity and time constants need to be matched to exterior ones.  IMO the main benefit of interior "thermal mass" is to act as an absorber of internal heat variations -- the odd 4 visitors, vacuuming, etc.  But designing a total system to work effectively is going to be complicated.