- This topic has 30 replies, 6 voices, and was last updated 10 years, 4 months ago by Anonymous.
3 December 2007 at 6:33 pm #33656
Since my orignial post I heard that Chris Sanders helped to co-author BR497 Conventions for Calculating Linear Thermal Transmittance and Temperature Factors. As Chris and I were in contact a while ago so I emailed him about this positive/negative psi subject. He clarified matters in the form of calcs; see the attached.
Thought that it would be worthwhile to post his feedback here for others to see. Should help to avoid people having the same dilemas as me.
Mark22 December 2007 at 1:08 am #33657
Here are new thermal bridge programmes from Building Physics, they are called Heat 2 and Heat 3 (2D and 3D.) They are both cheaper than TRISCO and SISCO.
P.S. Heat 2 has was developed for and used on a PH Primary School in Riedberg. (I think that it was PHI people that wrote the software but not sure. PHI definitely seem to have played a part in its development.)18 December 2008 at 12:48 pm #33658
Some more thermal bridging software and tools:
Mark5 February 2009 at 1:23 pm #33659AnonymousInactive
I have recently been looking into the thermal bridging issue and the THERM software. After looking at the excellent work done by Roderick Williams, I wonder why the THERM software is needed. Surely THERM is just a U-value calculator i.e. a 2D finite temperature calculation. Within a standard U-value calculation, the mesh size if sized at the material depth, as the temperature drop through the material is assumed to be linear. Therefore by using a “normal” u-value calculation through various different “slices” of the thermal bridge and weighting them according to their height, effectively enables the thermal bridge calculation to be resolved and then applied on a project. Obviously the limitation is with 2D, however THERM does not carry out 3D calculations either. I may be (and probably are!) missing something fundemental, however I believe you can work through Rodericks worksheets using a standard U-value calculator for each of the slices. Is this right?5 February 2009 at 7:26 pm #33660
As I see it there are perhaps 2D conditions that could, more or less (i.e. within a reasonable level of accuracy), be assessed shorthand through the use of fractional areas, however there are many complex details that need to be assessed using tools such as THERM. Typical examples are corners (external, internal, eaves, ridge, wall/ground, basements, stepped party walls etc), and structural openings (head jamb, sill, threshold) where not only are fractional contributions play a part in heat loss but the geometric aspects of the building element also influence the thermal performance (positively and negatively).
For example, one boring Sunday I undertook a theoretical study considering the location of *where* you put a window within a wall. I found that if located at the front of the insulation the heat loss is “high”, if central it is “low,” if at the back of the insulation the heat loss is “high” again (forming a U-shaped curve). If the window encroaches beyond the insulation then heat losses begin to sore. I also learn that the adjacency of one material to another (or within a matterial) can have a suprising influence heat loss; another subtle geometric consideration. If you then factor in fractional areas then things begin to get very complicated very fast, consequently no shorthand calculation can really quantify these issues.
So using THERM or another modelling tool is a means to an end. After you've built the model and extracted the heat loss data then you can determine the psi-value with a nice shorthand calculation….but only once you have the heat loss data. I think that the apparent simplicity of the psi-calc is beguiling you into thinking that thermal modeling is not required in low energy buildings. It is.
Hope this helps,
Mark6 February 2009 at 7:30 pm #33661
Mark is correct but I'd just like to make the point that it is possible to design even to Passivhaus without ever needing 2D heat flow models or psi value calculations.
The PH convention of using external areas means that as long as you use good details (see CLP and various sources of PH details) you can ignore most psi values.
Situations where I think Therm and commercial variants are useful are:
1. For geeks like Mark and I wanting to try and understand how heat flows thru building junctions.
2. Where refurbishment or other constraints prevent clean thermal 'bridge free' details being used and a fix needs to be found.
3. When struggling to meet the 15 kWh/m2 target for Certification under PH – in this situation some psi values can be demonstrated to be negative and so credit claimed (although the design remains the same).
4. For exploring innovative solutions such as optimising the installation of uninsulated triple glazing to achieve low heat loss with less expense as demonstrated by Mark and Chris Herring.
5. For making pretty pictures for reports!
Nick7 February 2009 at 3:18 pm #33662
As Nathan enquired into the relevance of THERM, rather than thermal bridging calcs, I limited myself to that topic. I agree with you that you could design buildings using CLP/PH detail and avoid the need for psi-value calculations. However, whilst working on a PH scheme I have found that if it surprising how often, due to non-energy performance requirements (your “other constraints”), that CLP and PH details are not suitable (and sometimes even subtle changes can have surprising results). This is not to discredit the details, I am simply noting that client driven non-energy performance requirements can lead to some very different solutions that require bespoke assessment. (It is after all a little unfair to expect the CLP details to cover all situations…. even the well funded Accredited and EST details do not do this.)
Examples of details that have required bespoke assessment include:
a) Details resulting from the topography (terraced houses with terraced (that is stepped) levels at floor and roof (here the detail required to achieve zero bridging will pretty much change at each different level change…. We tried to standardise level changes where possible.)
b) When seeking to achieve level access and paving/patios adjacent to the perimeter of the properties (whilst also addressing NHBC requirements to have the sole plate 75-1500mm above ground level!)
As a consequence of the 13 details I examined:
5 of the details assessed were completely unavoidable i.e. they were site specific or were influenced by non-energy considerations.
1 was a result of poorly/undocumented detailing to assist with airtight construction (hybrid balloon frame)
3 helped to verify how the PH windows should be installed so that we actually achieve our goal (construction tolerances could be determined, and I learned about some subtle things about the installation details that are less obvious than the PH details sometimes suggest.)
4 were strictly developed to enable us to mine the design for energy performance. This was a result of changes to the design, namely requirement for laminated glazing to achieve Secured-by-Design, degraded energy performance (we lost 0.1W/m2K on the Ug). Mining the psi-values has helped to ensure that we keep the built costs down over the project as a whole….
In the end by assessing all the details we have achieved roughly a 35% improvement upon EST Best Practice (~70% improvement upon Accredited details). The SAP/Code Assessor on the project was staggered by how much improving the little appreciated Y-value could have such a significant impact upon performance.
Having learned the lessons we should be able to reduce the psi-calcs on future schemes. I agree that building refurb is one area where psi-calcs will pay for themselves many times over. ….I also agree that geeky interest helped spur me into undertaking the psi-calcs. 😉
Mark9 February 2009 at 1:27 pm #33663Tom FosterParticipant
4. For exploring innovative solutions such as optimising the installation of uninsulated triple glazing to achieve low heat loss with less expense as demonstrated by Mark and Chris Herring
Sounds intruiguing! What's this?9 February 2009 at 6:39 pm #33664
Have a look through the AECB Windows thread in the “Materials and products”
The concepts are discussed repeatedly.
Also have a look at “thermal bridging around windows with wrap around external insulation”
Mark9 February 2009 at 6:43 pm #33665
Thanks Mark, was looking for the threads for Tom when I spotted you had posted!!!
Absolutely agree with all your points re Therm, just wanted to make the point for general browsers that a cookbook approach is possible, especially if it is possible to keep things simple, which as you point out, is always a struggle!
Nick11 February 2009 at 12:54 pm #33666AnonymousInactive
Many thanks for your reply. My concern is that THERM is only a 2D approach not 3D – therefore ground floor corners etc require a 3D software. A 2D approach looks at the construction in slices (inside to outside), therefore by breaking the bridging element down into slices, a simple hand calculation would provide the same detail as THERM. Unless THERM does carry out a 3D approach… I probably need to look into THERM in more detail. Actually, maybe the hand calculation is just a 1D analysis.11 February 2009 at 6:32 pm #33667
Yep, hand calc is 1D!
Corners largely get dealt with (approximately) by using external measurements and I have personally only seen 3D used for columns going through a floor.
If you look at some of the graphical output from Therm you will see the heat flow is not 1D!
Typically when calculating psi values, Therm is used to work out the 1D heat loss thru the plain bits either side of the detail – eg I beam walls. Is already drawn so quick to do and ensures calculation is the same as for the 2D bit. This won't make much sense unless you have doe psi value calcs and I am too rusty to explain clearly!
Nick12 February 2009 at 8:35 pm #33668
As Nick says long hand is 1D and THERM is 2D (the “2D” bridges that I mentioned earlier are fractional bridges in 1D heat flow; they have actually 2D or even 3D affect upon the energy flow).
Yes a 3D corner, say two walls and a floor, requires 3D analysis for the most accurate of calculations. Luckily in this situation 2D analysis is sufficiently accurate for the most part and provided that you have the 2D analysis the additional heat loss/impact upon surface temps is marginal.
If you think that you can calculate 2D heat loss using your 1D methodology then here is the challenge. Take a 90 deg corner, internal dims 1m x 1m, one 300mm wall has a U-value of 0.1W/m2K the other is 250mm and has a U-value of 0.12W/m2K. Just to keep things simple there are no physical thermal bridges in either of the wall element. Using your methodology calculate the internal psi-value without using THERM.
I'll declare right now that I can not calculate the psi-value using a 1D method. Hats off if you can do it.
P.S. If you do manage please post the result and the methodology on the forum. I'd be very interested to see how you managed. (I'm not a physicist so would not have a clue where to start.)31 March 2009 at 9:54 am #33669AnonymousInactive
Sorry I had not seen your last post.
I was thinking (incorrectly) that the hand u-value calc was a 2D analysis, which it obviously is not. I have recently been looking at the THERM program in more detail using the suggestions above.30 September 2009 at 6:58 am #33670
I recently observed that Internal psi-values result in slight geometric underestimation of the heat loss. This results from the fact that linear psi-values consolidate geometric and structural heat loss into a singular value.
In the case of internal psi-values the linear length of the psi-value is determined by the length of the internal datum, the result, say at a walls external corner, is a small geometric square of wall that is not accounted for. Thus when two internal psi-values meet, wall/floor junction, there is a small cubic volume of unaddressed heat loss. Conversely the use of external dimensions will account for this cube and may result in a small over estimation due to the overlapping linear geometries (erring on the side of safety).
All in all, at a building scale, this additional heat loss may be relatively trivial (depending upon geometric complexity) but this discrepancy would show up when 2D and 3D analysis of a wall/floor junction are compared. I've not managed to check this with a 3D tool. Has anyone else noticed/been thinking about this?
- You must be logged in to reply to this topic.