6.1 Case Studies

Introduction

These case studies have been created by a variety of practitioners and address areas where there are gaps in our understanding of the impact of retrofit measures. They are essentially ‘live’ retrofit research projects as monitoring is in most cases ongoing. The CLR team’s interpretation of results may change over time as new information is added and our understanding deepens. Each case study is available as a self contained PDF for ease of updating. New case studies will be added over time as funding permits.

For a detailed explanation of the kinds of graphs used in these case studies, refer back to Lesson 5.16, however this section also contains miscelleneous pointers on interpretation of graphs for easy access which will be updated over time as required.

Making sense of the results – interpretation

Significant amounts of moisture can be transferred via capillary action – either related to rain or groundwater. The presence of ground water (typically manifesting as rising or penetrating damp) is usually assessed by comparing sensors within 1m off the ground floor with those higher up. Rain (and groundwater) contribution to WME levels is assessed by:

1) The sorptivity of materials exposed to rain has been tested by Karsten Tube tests (8a/b), a moisture or microwave meter survey (1a/b) or there is membrane in the assembly and a low air leakage result which combine to minimise significant effects by reducing or blocking vapour flows from the inside (8a Poly membrane)
2) When we see low WME’s and where rain impact is likely to be low because of new EWI: in (2a,3), a rainscreen (7), a cavity (10), brick cream treatment (1a/b, 11). Occasionally there can be leaks in these which show up in the WME (2a, 10, 11).
3) By inference. Comparison of WME inside wall to external sensor is possible if we have it (8,10,6a) and/or comparison with rain data (2a,8). This may seem a good test but it is not conclusive because precipitation peaks in winter – as would hygroscopic adsorption of vapour and it can be hard to separate these effects. As a result there are two case studies where we are uncertain about the contribution of rain (4, 6a).

For this vapour to wet materials without condensation occurring there has to be a hygroscopic material present such as timber based materials, brick or mortar etc. Examples of timber based hygroscopic materials are: woodfibre (6a,11), sawn timber and cellulose (6e), wood and Gutex fibreboard (7), or wood (1c). We know wood based products can adsorb moisture directly from the air and form an equilibrium because this is already well documented. For 3, 10 we see a hygroscopic curve in brick but in 2a there is little effect because RH’s are never above 70%. (data from sensors in studies 4 and 8 lie on different hygroscopic curves: this could simply be variations in sorptivity or the influence of salt build up affecting the sorptivity of some of the materials e.g. brickwork below a rising damp front or behind ‘remedial plaster’ or or other damp proofing measures.)

Vapour is likely to be responsible where the Glaser calculation using internal and external readings shows a seasonal trend and hygroscopic materials are present. If a lower RH is measured than predicted using the Glaser calculation it can potentially be explained by the ‘surplus’ moisture being adsorbed and held in hygroscopic materials – in effect lowering RH in that area: hygroscopic adsorption in winter simultaneously brings down RH and increases the WME of the adsorbing material.

We see this effect in almost all case studies. One exception is case study 10. In IWI case study 8 some walls use a polyethylene vapour barrier that prevents this mechanism (8a Poly membrane), in contrast to other walls where an intelligent membrane has been used where a greater degree of water vapour enters the assembly from the interior.

Glaser is accurate in EWI cases: 2a, some sensors in 4, some sensors in 8.
Glaser is close in one case: (3).
Glaser is inaccurate in these cases: woodfibre (6,11), wood and cellulose (6e), wood and Gutex fibreboard (7), or wood (1c).
Glaser is inaccurate in these cases where hygroscopic effects related to masonry are assumed to be the explanation: (1a/b, 10).
Glaser in inaccurate for many other situations being monitored e.g. (8) which appear to be influenced by rising damp.

If you have any suggestions about how these interpretations could be improved, please do get in touch with tim@aecb.net.

‘Ambient’ readings

Some case studies (e.g. 1a, 2, 3, 4, 5, 7,11) are based on arrangements of sensors that are set up to try to measure as close to ‘true ambient’ conditions as possible. For external sensors this may mean hanging the sensor in a ventilated but rain proof cowl that also shields from solar radiation.

Below: a DIY ventilated rain and solar shield for a sensor made from plant pots:

However most CLR case studies show that we consider the conditions close to the external weathering surface (for example a brick wall) to be more useful – and so sensors have often been arranged to measure the WME of the external brick or stonework and the RH and T close to the surface, whilst protecting the senor against rain and direct solar radiation (without shading the brick surface itself from rain or sun) for example:

Another example of a rain and solar cover for an external wall sensor is shown below where dense polystyrene was used to create the shield:

Miscellaneous useful pointers for case study graphs

If you are unclear about interpretations of any of the graphs used in Modules 5 & 6 please post your comments on the forum so that we can update this section as required.

Hygroscopic effects sections

In these sections you will often see a graph showing measured WME v measured RH. For hygroscopic materials these curves have similar characteristics to emc curves and are sometines overlaid with typical emc curves for timbers. In this example from an EWI case study where EWI has been in place since 2009 and conditions were monitored during 2014 -16 vertical red lines have been added to illustrate the RH and WME target limits suggested by the CLR team to help ensure moisture-robust assemblies:

In the graph above measured results show conditions in this assembly’s area of concern are safely below the CLR WME and RH limit, except in a few instances that the case study identifies as related to intermittent leaks during intense storms and a particular wind direction.

Diffusion Flow Rate graph

Example: look at the area of graph circled below and the red line within . This is below the axis, so the direction of this diffusion flow is ‘Outwards’. The line is a red colour which denotes it is (in this case) ‘from‘ the interior. For red lines above the axis the flow is ‘Inwards’ ‘to’ the interior.

‘Direction and magnitude of diffusion flows’ diagram

Data for the permeability of the assembly materials can be used in conjunction with measured data from embedded sensors to calculate the likely diffusion flows away from, or towards, a particular sensor location. In the case studies the diagram as illustrated below is used to summarise the total vapour-only flows in the assembly being monitored. How much vapour flows (the magnitude) is illustrated by the size of the arrow and its associated value. The direction of the arrows shows the direction of flows relative to the interface and also indicates the respective sources of vapour (i.e. the moisture held in either the interior or exterior air). These diagrams do NOT show any capillary flows reaching the chosen sensor location: however a very high diffusion flow rate out of an assembly will suggest that additional moisture loads are arriving via capillary flow and leaving via evaporation and diffusion.

Above: diagram illustrating a net loss (via diffusion) of 2.5 litres of vapour per m2 of wall over the study period from the interface. At the moment we cannot infer the exact magnitude of capillary flows through the wall, but it may be possible in the future pending further development of our monitoring and analysis techniques.

In future versions of these case studies it may become useful to produce flow summary diagrams for each year, as changes in annual flow rates and drying direction will help identify trends over several years.

Low diffusion flows combined with low WME overall suggest that an assembly is mainly responding to atmospheric moisture rather than liquid water (typically rising damp or rain ingress).

Rainwater ingress

WME graphs: if WME increases suddenly and reduces rapidly then this suggests a liquid water (rain related) leak into the assembly – it is possible to correlate these with weather data, some case studies do show correlation with severe rain storms. RH levels will obviously also follow such events.

Table of case studies:

Some of the case studies listed below are still in preparation but will be added when ready, and further case studies will be added as more become available. (Only those already completed have links to a pdf case study document in the list below the table).

The graph immediately below provides an interesting overview of a selection of sensors across different projects positioned at the interface between insulation and masonry, after SWI. The sensors’ names generally indicate insulation type and the key allows easy comparison between the changes in WME of the masonry over time. Obviously not all projects have been started at the same time, so the curves have been arranged for clarity rather than representing true start and end dates. Detail on each project including masonry wall types, assembly build-ups and U-values etc. are available in the case studies. Use the case studies to understand the CLR team’s ‘interim conclusions’ concerning the performance to date of each assembly.

*Projects not attributed to specific years but earliest project being monitored started in 2012.