Simulation Software for Passive Solar Design

[David OLIVIER]

The Passive House Institute say that PHPP (calibrated against results of 100s of dynamic thermal simulations) can predict space heating energy in small buildings such as houses (and flats) to +/- 2 kWh/m2yr.

Large non-domestic buildings may become too complicated for this relatively simple approach but it seems that PHPP has been used successfully for many schools and offices.

I suspect the computer-modelled buildings that didn't work well contained classic design errors such as long highly-glazed facades facing south-west and low thermal capacity but densely-occupied and highly-glazed buildings and perhaps natural ventilation pushed beyond its abilities. Good results may be more likely if designers familiarise themselves with the LT method (developed by Nick Baker et al, Cambridge 80s onwards).

David.

[Mark Siddall]

David,
I completely agree on your recommendation that designers use the LT Method. The LT method is a good guide for the first stab at a non-residential design. It's the refinement of such as scheme that thermal simulation is key.

Interested to hear what you say about the PHPP being used in office and school designs.

Mark

[Mark Siddall]

Nick,
First of all I have similar concerns to you, though having said this some simulation is better than none, especially on large buildings.

Like most things at some point you have to use informed judgment and in many cases this judgment call is made by CIBSE. The main thing that CIBSE seek to do is establish something of a level playing field by establishing weather data sets and simulation test procedures.

Simulation software can, within reason, be fairly accurate. One of the problems is getting accurate real life data (its the old cause and effect thing whereby the act of taking a reading changes the initial conditions and thus introduces a degree of inaccuracy.) TAS, one of the better simulation packages has some data on its accuracy compared to real life data, go to http://www.edsl.zenwebhosting.com/
(then hold the cursor over "TAS software" then over "Tas Building Designer" and then click on "Empirical Validation" and "Simulation and Reality" to get a little of what your looking for.)

The Weather Front: -
The whole thing with simulation and weather files is that on the whole the weather data is fine for estimating the heat loads etc. the real challenge is considering over heating in the summer. For this, as I understand it, you normally take the worst case year in the last 100 years and design to work within the specified temperature parameters.

There are programmes out there Meteonorm for one that can be used to simulate the weather for a given location based upon the weather at three known geographical locations. Weather prediction is not an accurate science and therefore building simulation has to work within these parameters. Thankfully we have a reasonably good weather history that can be used to inform simulation.

I'm sure that Dr E will have far more to say when he next checks in...and no doubt correct my errors.

Mark

[Nick Grant]

Thanks both

The TAS validation seems to be for a simple box with a window and I don't understand how it exactly predicted the solar radiation for 7 days, perhaps I misunderstood that graph.

My scepticism of PHPP accuracy has to do with human rather than building variables. The idea seems to be to design and build something that you have some hope of modelling. ie thermal bridge free, controlled MVHR, airtight. I can believe the predictions - allowing for humans leaving windows open then using blow heaters (which won't show up in the heating energy monitored) or living at 23C etc.

Thanks for lead on LT method, I'll explore further.

Nick

[Mark Siddall]

Nick
I don't think you can really account for the human variable (people always find a way of cocking things up). All you can do is refer to an historic behavioral model for a given set of conditions i.e. a PH or AECB Gold home.

One reason why the PH is considered to be a success is that the buildings, when measured and compared to the normalised model, performed in accordance with predictions i.e. the rate of heat loss was as expected. This is not to say that occupants consumed the predicted quantity of energy sometimes this was more, other times less (some like it warmer, others colder). On average though the statistical agreement is that the predicted energy consumption of 15kWh in rough accordance with the reality.

I believe that this statistical agreement is due, in part, to the use of an all-air heating system and the impact that this has upon behavioral patterns (though I'd love to see a report confirming/disproving this theory). My contention is that, if in a PH you try to open the windows (when using an all-air heating system, as used in the normal PH) you'll get cold quite easily , this is due to the limited energy supply available from an all-air system. It is on this basis that over time the lesson learned by the occupant is "don't open the windows when the MVHR system is switched on."

Conversely under "normal conditions" where the occupant has rads or UF a greater range of energy supply is available to "fight" the ventilation losses caused by opening the window. This is achieved by cranking up the heat from the emitter.

This said once an occupant is living within the design/required-behavioral PH model there is still nothing to stop the occupant of a PH turning up the heat and living at 23C as they did at Lindas. (A new but more tightly controlled variable.)

Incidentally I read a paper that suggested that this is one are that German engineers will be focusing upon in the future (statistical models based upon use patterns etc. Very complex, it will be interesting to see if they can get the whole thing to work, people being people and all).

With all this in mind I wonder how the AECB standards will be policed. Will they be actual measurements i.e. "forcing" an individual to live within a given footprint, or will they refer back to a normalised model as done with the PH?

Mark

[Nick Grant]

Mark

Obviously I agree that human influence can't be predicted but I'm questioning whether the dynamic models have any worthwhile accuracy given all the variables. ie if the inputs are very sensitive to user behaviour then a simpler model is justified.

I agree re limiting input but it seems many PH occupants use auxiliary heating which i think i mentioned on another thread with ref from something I read recently. Bit similar with how water use when cylinder is small, pressure low, shower curtain a poor fit and bathroom floor is shagpile. This combination will save more water than any water saver shower head.

Mark suggested I split this thread as we are getting onto PH/Gold heating:

http://www.aecb.net/forum/index.php?topic=949.0

Lets keep this to original subject

Nick

[Eric Parks]

Hello All - I've been following this thread with interest for a week or so but have only now found the time (new baby and part-time MSc currently occupying my spare time) to get involved. 

A few comments: 

Having recently completed the 'Computer Applications' unit on the UEL CAT Msc course, which introduced us to dynamic simulation in building design using IES, it's encouraging to see many of the same issues being discussed here.  Amongst other things, I can say that Eric Roberts comments on the 23rd of April were echoed by the course tutors throughout the week.  Several MSc student have previously prepared essays on the weather data issue as well and I am aware of a study done by ARUP and ZedFactory that looked at the potential effects of climate change on buildings. 

Using IES (ApacheSim) I am currently modelling the effects of different levels of airtightness on the AECB Silver Standard dwelling as described in the most recently published information:  80m2, , semi-d, passive solar design with overheating proectction etc.  One of my initial aims, in the absence of empirical, 'live' data, is to effectively calibrate the model to yield more realisitc results.  I would like to be able to gain a clearer understanding of how important airtightness is in approximate kWh/m2/yr terms - as given in the AECB Standards. 

After many unsusccessful attempts and entering my own user profiles (what a chew!)  for heating, casual gains, ventilation etc, I have used the standard NCM profiles supplied.  Rather suprisingly I managed to achieve results for the dwelling's heating in kWh/m2/yr (62) very close to the 58 predicted by the AECB.  I have also managed to achieve a similar 'success' (my 'real U-value' results - 146 / AECB's 'real U-value' results - 147 ) with a building regs compliant (non-passive solar design but same as previous model in every respect). 

I've not yet managed to achieve the AECB Gold levels of performance with the model, mainly because setting up a heat recovery profile, which is critical, would take more time than I have to invest at the moment and its not on the 'critical path'. 

I realise the shortcomings of SAP (and presumably SBEM) in terms of what they leave out - cooking, lighting and what they don't properly account for - thermal mass.  Would NCM based profiles suffer the same affliction?  Given that my main interest in the modelling above is the heating energy, it's not critical for the exercise, but I would like to know more about the NCM as the IES and BRE information on it so far seems thin on background explanation or even 'foreground' explanation - i.e. what it includes!

I have two questions: 

(David O.?) Was the research conducted to estblish the AECB standards based in some part on NCM data? 

If not, is it valid to use it to calibrate in the model in the exercise described above? 

One final note - In addition to stressing the importance of simulation validation (with measured data where possible), one of the main topics of discussion during the week was understanding the purpose and usefulness of buildng simulation not so much in terms of absolute values (this design will produce x kg/CO2/m2/yr) but more in relative terms (design A is a % improvement on design B). 

Any input or comments welcome. 

Eric P.

[Nick Grant]

Congratulations Eric!

Simulating to compare advantage rather than absolute seems sensible and this could justify modelling a more simplistic building element - room with window.

Nick

[Mark Siddall]

Came across these useful links for weather data/solar irradiation:

Monthly solar irradiation/weather data (free from NASA)
http://eosweb.larc.nasa.gov/sse/

Monthly solar irradiation (free from NRC, Canada)
http://www.retscreen.net/

Hourly solar irradiation/weather data (a commercial product estimating irradiation for anywhere in the world using stochastic model)
http://www.meteotest.ch/

Mark

[Mark Siddall]

New translated info on PHPP. Sets things straight. Sounds like a good little tool.

http://www.passivhaustagung.de/Passive_House_E/PHPP.html

Mark