Frost damage conditions and internal insulation

[Browne Dan]

Dear WUFI team,

I am using WUFI Pro 4.2 1D to establish how different types of internal insulation affect the likelihood of frost damage to external masonry. This is the subject of my masters thesis and I would be grateful for advice on inputs and how to read outputs.

I am familiar with the process of frost damage to masonry as I am sure you are, but I would just like to clarify my assumptions in case they are wrong:

Freeze/thaw damage occurs in some types of masonry (brick and stone) and is dependant on many factors and does not necessarily occur at 0oC . For damage to occur masonry has to be sufficiently wet and the liquid water within the pores has to freeze.

Factors affecting frost resistance and the actual freezing temperature of liquid water in the pores include: soluble salt content and pore size, shape and distribution which are both beyond the scope of my research in WUFI. However I am hoping to use WUFI to establish the water content of masonry at zero degree events and how this changes with the addition of different types of internal insulation in the four wind-driven rain exposure zones across the UK. Even though water within pores may not actually freeze at 0oC, I am assuming the counting the number of 0oC crossings will show the worst-case scenario.


Assumptions

A 1. It is my understanding that if a brick is susceptible to frost damage then over 90% of the available pore space has to be full of water for damage to occur as water expands approximately 10% on freezing.

A 2. For the experiment I have chosen to use ‘Brick, historical’ which has a porosity of 0.31 m3/m3 or 310kg/ m3. Therefore I’m assuming that for damage to occur the water content has to be above 279 kg/m3 and freeze.

A 3. RH of 100% (which frequently occurs inside the wall) does not imply that interstitial condensation has completely filled the pores, the graph of water content and RH in the materials database for brick, historical shows that at 100% RH the lowest water content is 20-25kg/m3.


Questions

Result graphs are difficult for me to assess exactly what is happening within the wall.

Q 1. I would like to consider the effect of solar radiation on freeze thaw cycles so should the ‘short-wave radiation absorptivity’ and ’long-wave radiation emissivity’ be ‘on’ and if so, can we assume the approximate values as for ‘red brick’?

Q 2. One problem is that I’m running the simulation over 5+ years or until the moisture content of the wall stops rising, therefore ‘zooming in on specific’ freeze thaw events and trying to establish the water content as the wall temperature crosses zero is impossible, how best can I see these results?

Q 3. Is it possible to see a graph showing the temperature gradient and water content of the wall when the external temperature drops below 0oC? Or do I run the film and count zero crossings and note the water content?

Q 4. I am also using the WUFI U-value calculator to see the different values of the walls but I don’t understand the results graphs. It would be good to know a new average yearly or just winter U-value with the wall’s new moisture content after the addition of internal insulation. My graph shows red diamonds relating to WUFI results and black, blue and green lines. I can guess but what do they actually mean?

Q 5. Also I can’t understand why the porosity of Pavatex diffutherm and EPS is similar, 0.883 and 0.95 respectively? Surely difutherm is much more porous?

Q 6. Is it possible for me to establish with WUFI how the increased water content of the wall is achieved after internally insulating. Is it just that the drying out potential of the wall is reduced or has the internal vapour pressure changed as a result of internally insulating?

Q7. What is a sd-value?

I hope my questions make sense; physics, WUFI and heat and moisture transfer are all new to me. I will be very grateful for any help you have to offer.

Many thanks

Dan

[Thomas]

Freeze/thaw damage occurs in some types of masonry (brick and stone) and is dependant on many factors and does not necessarily occur at 0oC . For damage to occur masonry has to be sufficiently wet and the liquid water within the pores has to freeze.

Factors affecting frost resistance and the actual freezing temperature of liquid water in the pores include: soluble salt content and pore size, shape and distribution which are both beyond the scope of my research in WUFI. However I am hoping to use WUFI to establish the water content of masonry at zero degree events and how this changes with the addition of different types of internal insulation in the four wind-driven rain exposure zones across the UK. Even though water within pores may not actually freeze at 0oC, I am assuming the counting the number of 0oC crossings will show the worst-case scenario.

Dear Mr. Browne,

yes, the water in the finer capillaries may freeze at much lower temperatures than 0 deg C, depending on the pore size distribution. WUFI does take into account that at temperatures below 0 deg C only part of the water may be frozen. This is possible because the moisture storage function contains some information about the pore size distribution (please refer to Dr. Künzel's thesis for details if interested).

However, the number of zero-crossings is usually taken as a proxy for the frost damage potential, and this seems to be a reasonable approach as long as no detailed knowledge about specific damage-inducing freezing temperatures is available. Also, at least the final part of the frost damage will certainly occur due to the action of freezing water in macroscopic pores and gaps (which may have been created earlier by accumulated damage on the level of the finer pores). In these macroscopic pores the water will freeze close to 0 deg C.

A 1. It is my understanding that if a brick is susceptible to frost damage then over 90% of the available pore space has to be full of water for damage to occur as water expands approximately 10% on freezing.

A 2. For the experiment I have chosen to use ‘Brick, historical’ which has a porosity of 0.31 m3/m3 or 310kg/ m3. Therefore I’m assuming that for damage to occur the water content has to be above 279 kg/m3 and freeze.

As far as I know, there is no generally applicable critical water content above which frost damage may occur. For example, our experience indicates that natural stone is usually quite resistant to frost damage, even at moisture contents above free saturation, while bricks are much more susceptible. So the critical limit must be determined for each material individually.
Sedlbauer and Künzel use 12 mass-% (i.e. 91% of free saturation and 79% of maximum saturation) as the critical limit in their WUFI study of frost damage on lime silica brick walls; Straube and Schumacher say in their WUFI investigation: "The dangerous moisture content is often in the range of 75 to 94% of the free water saturation. Given no other information we often choose to use 90% since it is conservative and one of the more common thresholds for brick. The same threshold can often be used for natural stone."

A general statement that (because of a 10% expansion of freezing water) frost damage will occur if 90% or more of the pore volume are filled would assume that each capillary is 90% filled, but this is not true in general. Usually, the smaller pores are completely filled, the larger pores are more or less empty. I guess the susceptibility of a material to frost damage would probably depend on the structural strength of the walls of the filled (or almost filled) pores, which in turn would depend on the pore size distribution and the pore structure.

A 3. RH of 100% (which frequently occurs inside the wall) does not imply that interstitial condensation has completely filled the pores, the graph of water content and RH in the materials database for brick, historical shows that at 100% RH the lowest water content is 20-25kg/m3.

This only seems so in the graph because the last part of the curve is very steep and seems to cover a whole range of moisture contents. However, if you look at the accompanying numerical table, you will see that an RH of 100% corresponds to a definite moisture content of 230 kg/m³.

We call the moisture content at 100% RH the "free saturation" and usually identify it with the moisture content the material absorbs in an imbibition experiment. Due to air pockets remaining in dead-end pores, this will be less than the "maximum saturation" where all pores are completely filled (e.g. because of condensation) and which is determined by the porosity. For details, see the topic "Reference | Material Data | Moisture Storage Function" in the online help.

Result graphs are difficult for me to assess exactly what is happening within the wall.

You can export the result data to an ASCII file using the menu item "Outputs | ASCII export" and view or analyse the results with a program of your choice (e.g. Excel etc.).

Q 1. I would like to consider the effect of solar radiation on freeze thaw cycles so should the ‘short-wave radiation absorptivity’ and ’long-wave radiation emissivity’ be ‘on’ and if so, can we assume the approximate values as for ‘red brick’?

The effect of solar radiation will probably be very important for your investigation, so the short-wave absorptivity should be set appropriately. You may use the default value for red brick as long as you have no reason to choose something else. If in doubt, perform a few test calculations with slightly varying values and see whether this affects the results of your analyses. If the effect is within what you'd accept as the uncertainty of your results, you may use the standard value. If the effect is beyond what you deem acceptable, you must put more thought into your choice or explicitly investigate cases within a certain range of this parameter. The same strategy also applies to all other input parameters.

If you do not want to account for night-time radiation cooling, the safest option is to set the long-wave emissivity to zero.
If you think radiation cooling is important for your investigation, you may set it to the default (or any other convenient) value, but then you must also use a weather file which contains data on the long-wave radiation exchange. This topic is somewhat complex, an explanation can be found in the section "Reference | Climate Data | Long-wave Radiation Exchange" in the online help.

I'd recommend to start with the emissivity set to zero until you feel comfortable enough with the program, and to see at a later point whether inclusion of these effects is relevant for your investigation (again, by doing test calculations with and without emission taken into account).

Q 2. One problem is that I’m running the simulation over 5+ years or until the moisture content of the wall stops rising, therefore ‘zooming in on specific’ freeze thaw events and trying to establish the water content as the wall temperature crosses zero is impossible, how best can I see these results?

Q 3. Is it possible to see a graph showing the temperature gradient and water content of the wall when the external temperature drops below 0oC? Or do I run the film and count zero crossings and note the water content?

Export the results to an ASCII file and analyse them using your favorite data analysis tool. Set a monitoring position at the location where you want to detect zero-crossings (at the surface, there's a monitoring position by default) and filter the file accordingly.

As to the moisture contents at zero-crossings, there are several options depending on whether you want to evaluate the water content at one specific location, as the average over a slightly extended region or as the average over the whole material layer.

You can export the RH together with the temperatures at the monitoring position and convert it to the local water content using the moisture storage function of the respective material.
If this is too cumbersome and/or you want the moisture content determined across a somewhat wider region, you can export the water content averaged over each layer. You may also define a separate "diagnostic" layer with the appropriate width around the location of interest if the whole material layer is too wide for your purposes.

If you want to use the film, note that you can use single-step forward and backward to find the times of interest (not during the calculation, but when viewing the film after the calculation). Hover the mouse over the displayed curves and read the respective values from the status bar. Right-click to export the curves to an ASCII file.

Q 4. I am also using the WUFI U-value calculator to see the different values of the walls but I don’t understand the results graphs. It would be good to know a new average yearly or just winter U-value with the wall’s new moisture content after the addition of internal insulation. My graph shows red diamonds relating to WUFI results and black, blue and green lines. I can guess but what do they actually mean?

The U-value calculator reads the layer thicknesses, the material parameters and the heat transfer resistances for the surfaces from the project file and computes various U-values from these data.

The curve labeled "dry" assumes that all materials are dry and thus uses the "dry" thermal conductivity from the respective material parameters.

The curve labeled "u80" assumes that all materials have the equilibrium moisture contents corresponding to an RH of 80%. It determines these moisture contents from the respective moisture storage functions and then computes the corresponding thermal conductivities and the resulting overall U-value.

The curve labeled "Typical Built-In Moisture" does the same but assumes that all materials have the moisture content that was entered in the "Typical Built-In Moisture" field of the material properties dialog (if an entry is missing there, the tool will complain with an error message and assume zero moisture content).

The red dots are effective U-values determined from the actual heat flows computed by WUFI, thus reflecting the changing influence of the actual variable water contents. This would be what you are interested in. The underlying formulas, based on monthly mean values of heat flux and temperature gradient, are given in the online help of the U-value tool.

Q 5. Also I can’t understand why the porosity of Pavatex diffutherm and EPS is similar, 0.883 and 0.95 respectively? Surely difutherm is much more porous?

The porosity of diffutherm is based on density measurements, the EPS porosity is just a guess. But since EPS is just a foam structure consisting mainly of air bubbles with very thin walls separating them, I don't quite see why a material consisting of wood fibres should necessarily be much more porous?

Q 6. Is it possible for me to establish with WUFI how the increased water content of the wall is achieved after internally insulating. Is it just that the drying out potential of the wall is reduced or has the internal vapour pressure changed as a result of internally insulating?

You might do a series of WUFI calculations; one with the realistic insulation material as the reference case, one with an artificial insulation with the same thickness but high thermal conductivity (so that the thermal insulation effect is almost zero but any vapor diffusion retarding properties remain) and one with an artificial insulation with the same thickness but very high or very low vapor permeability and added or removed vapor barriers (so that the thermal insulation properties remain but the vapor retarding effect is changed).

If the overall effect is simply the sum of the heat transport and the vapor transport properties of the insulation, this method should cleanly separate them. If they affect each other, more elaborate tests may be necessary.

Q7. What is a sd-value?

The sd-value measures the vapor permeance of a material layer by giving the thickness (in meters) which a stagnant layer of air with the same vapor permeance would have.

In WUFI's surface transfer coefficients dialog you can enter the sd-value of a vapor-retarding surface coating or layer (if any). A surface layer which has no thermal or hygric effects beyond simply impeding vapor diffusion may be modelled more conveniently in this way than explicitly including it in the component assembly.

The definition of the sd-value is explained in more detail in the online help topic "Reference | Material Data | Water Vapor Diffusion". Its use for modelling surface coatings is explained in "Reference | Surface Transfer | Surface Coatings".

Kind regards,
Thomas

[Browne Dan]

Dear Thomas,

Well what can I say? Thanks you so much for the quick and detailed reply, lots of information for me to get my head around.

I don’t suppose you fancy writing my thesis?

Thanks again

Regards

Dan

[Thomas]

I don’t suppose you fancy writing my thesis?

I'd like to, but I think I'll better leave this to the expert...

Regards,
Thomas

[Browne Dan]

Thomas,

It is very kind of you to have negotiated the writing of my thesis with Dr. Hartwig Kunzel!




Regards

Dan

[Nickolaj]

Hey there,

I fell over this topic and got interested as I'm sitting with similar questions regarding saturation levels, moisture content and the risk of frost damages to masonry clay walls for my master thesis. I'm some what familiar with some of the processes behind frost damages to walls, but there might come some very basic questions along the way.


Q1)
Earlier in the topic it is mentioned that the critical level of saturation for bricks is around 75-94%, but also that the critical level of saturation differs between materials. I've read some articles stating that for older/historical bricks the critical level of saturation can be as low as 25-30%. How is your opinion/experience on this statement, would it be worth to consider zero-crossings with saturation levels down to 25-30% as potential risks?

Q2)
Also, you say "the number of zero-crossings is usually taken as a proxy for the frost damage potential", do you have a reference for this? and is there any sort of guidelines regarding a “safe” number of zero-crossings to avoid frost damages to the wall or such?


Thank you for your help.

Nickolaj

[Manfred Kehrer]

Hello Nickolaj

Q1) Hard to say whether this approach is reason. 25% to 30% seems pretty low to me to create structural damage.
Q2) I am not aware of any references.

Hope this helps.

Manfred

[Nickolaj]

Hey Manfred,

Thank you for your fast reply.


Another question came to my mind when I read the topic again, regarding application of water repellent coating in Wufi.

I tried to test a coating type with the data from the manufacturer. I followed the Wufi tutorial on how to apply a coating to a typical brick wall, by consturcting the 10mm coated brick layer, and the following data was used when editing the brick material:

Vapour diffusion resistance coefficient (µ) = 178
Resistance to passage of vapour Sd (mt) = 0.267
Capillary action water absorption coefficient W24 [kg/(m² h0.5)] = 0.12

But I'm a bit uncertain about the input value for the Adhering Fraction of Rain.

Default value for Adhering Fraction of Rain is set to 0.7 for external walls, but what might be an approximate value when a water repellent coating has been applied to the wall? I assumed the coating might not necessarily eliminate all the driving rain onto the surface of the wall.

Thank you for your help

Kind Regards
Nickolaj

[Manfred Kehrer]

Just keep the absorption coefficient to 0.7

[Podmolik_J]

Yes, the value 0,7 means that 70 % of the water comes in contact with the surface of the wall, therefore coating won't affect this.