Simulation Software for Passive Solar Design

[Mark Siddall]

I have begun to look into the subject of for environmental modelling software capable of calculating, in detail, passive solar designs. In the course of this investigation I have heard that software, calculating using the CIBSE admittance procedure, is flawed, as it cannot model thermal mass properly.

As I understand it the failure of the CIBSE admittance procedure has something to do with “each day being that same as the day before” i.e. the model has no way of feeding historical data into the its calculation in order to predict the buildings true behaviour. An example would be when using night purging techniques in the summer so as to exploit the coolth of the thermal mass to prevent over heating. For this to be simulated correctly the calculation model has to remember what the remaining thermal capacity of the material was in the last calculation, when the thermal capacity is reached and the subsequent impact upon room temperature.

This failure to “remember” the past calculations means that, as far as the calcs go, the scheme will be shown, in terms of comfort standards, to fail i.e. over heat. A further consequence of this is that by default the thermal model suggests that additional cooling loads need to be introduced. This may be achieved through increasing the thermal mass (which this will result in the designs failure in reality as there will be an excess of thermal mass which will in turn lead to additional heating demand) or by introducing mechanical plant (as this kit is not actually required it costs the client more money and potentially has a negative impact upon the carbon footprint.)

Given the above, it would seem that the subject of building simulation is fraught with complexity insofar as there are a number of simulation techniques are available; some of which are more successful than others. Which technique is the most reliable for a passive solar building? Who produces software that uses this technique? Who has the best algorithms? And ultimately which software offers the best value performance?

The whole field of simulation seems to be dreadfully confusing for the uninitiated. Any help would be appreciated.

Mark

[Andy Simmonds]

Mark
Have you contacted the Welsh School of Architecture, who have a track record in energy work, I think it is Phil Jones who runs the research department...

also is it worth contacting Ed Mazria (leading passive solar architect and campaigner) via www.architecture2030.com

I suspect David Olivier will post here in the near future too!

Again I suspect one issue underlying this is the relationship of modelling software and feedback from proper monitoring of performance - how informed is the one by the other in each model.

[Michael George]

Hi Mark

There are two applicaions that I am aware of that do not use the admitance procedure. These are TAS by EDSL and Apache by IES.  I have been using TAS for a few years now but do not have any experience with IES. Darren Bentley at EDSL is very knowledgable in terms of the empirical data and alogrithms used. Go to http://212.23.11.237/

[Mark Siddall]

That confirms some of what I have heard but I wanted to hear it from a third party. Most engineers in the North East use IES (at least the ones I’ve worked with). That these guys use IES and still conclude that thermal mass doesn’t work is worrying. I wonder how accurate the software is?

In my searches I have also come across TRNSYS and DEROB-LTH, in terms of accuracy does anybody any ideas how these compare to IES and TAS?


The Software Rumour Mill:
It would seem that there are a number of rumours circulating about software packages at the moment both IES and iSBEM. I’ve outlined them below. Can anyone pass comment upon the validity of them? (The best way of addressing this would be referances to any published data.)

Rumour 1: Has anybody heard the rumour that, due to the fact that iSBEM (and SAP for that matter) is a steady state model, it can’t actually calculate carbon footprint with the appropriate level of accuracy when compare to simulation?
Apparently this has something to do with the carbon calcs in iSBEM being based upon a system of weightings which are allocated to the energy consumption (rather than actual calculation). The result of this is that from one revision to the next, as and when the weightings are adjusted, you could pass Part L one month and fail the next all due to the fact that the weighting allocated to electrical operations has been shifted.

Rumour 2: Add to this the fact that I have heard that IES uses iSBEM to undertake its carbon calcs (even though I get the impression that IES is simulation software, see above). The consequence of this is that despite the fact that you have simulation software you can fail B.Regs in a much more readily than a package that uses fully simulated carbon calcs

If these rumours hold any water how can we be sure that building projects are being designed with any accuracy and will behave as predicted? The answer is that we can’t. In both cases the route of the problem would seem to be iSBEM.
Surely if one of, if not both, the governments principle design tools (iSBEM and SAP) are flawed then the whole idea of reducing carbon emissions is bordering on a farce. I hope someone can confirm that these rumours are just a rumours and nothing more.

Mark

[Michael George]

Hi again Mark,

I am not aware of any publications but do have an opinion on the rumours you mention.

Isbem seems very similar to the SAP software in that it models findings via tabulated data such as wall areas and u-values. Neither perform calculations based on empirical weather data.

When using TAS however you need to actually ‘draw’ the geometry of the building in 3D. This includes solar shading, as it is possible to model the shading effects of anything you can draw.  You then apply building elements and zones to which detailed construction materials and internal condition data are subsequently applied.  Internal conditions include accurate [W/m2] figures for incidental gains such as occupancy, lighting and equipment. Heating, air conditioning, air infiltration and ventilation is also modelled as is the opening of apertures such as windows. All of these parameters are modelled using schedules with any values you like set at any time you like. All of this is then simulated using an empirical weather data file. Results are outputted for any day [or days] you like.  Solar shading and solar gain show clear and  significant impacts on results.

I do not know if IES uses Isbem in any way but I am pretty sure that TAS does not.  I will know for sure when the new version is released but I understand it is still being written so that it remains accredited for use with the Part L.

Regarding thermal mass not working, I am very interested in this. I have not intentionally modelled this yet but it would be easy to do. Do you have more details of the IES findings?

[Mark Siddall]

Michael,
For further details on the calculation procedures available regarding thermal mass I would recommend you look in the CIBSE Guides (I think its Guide A but not sure.) This explains that the CIBSE Admittance Method can not calculate thermal mass and that full simulation is required......hence my hunt for improved software.
Other instances of thermal mass "not working" (when someone is using IES say) could be attributable to operative behavior/inputs. So, at this time, I can't comment on individual software packages.
Since the original post I have had a look at a couple of software packages. For a range of reasons, not just quality of simulation, TAS is looking like our favorite option at the moment.

Mark
 

[Michael George]

Thanks Mark , I will have a go [when I have time!]at modelling different constructions which  reflect different amounts of thermal mass. I'll post the result  here. If you ahave any specific constructions in mind please let me know. You can e-mail me at mdjgeorg at glam dot ac dot uk

I also came accross this which may interest you as it indicates a link between Tas and isbem.

 ODPM validation details here, [prepared by Ove Arup]
http://212.23.11.237/Downloads/060504%20-%20ARUP%20REPT%20APPROVING%20TAS.pdf

[fostertom]

PaulinMontreal recommends Hot2000 http://www.greenbuildingforum.co.uk/forum/index2.php?DATEIN=tpc_reyahkhcw_1144599747
I don't know if it's up what's required here.

[David OLIVIER]

I know people who've used NBSLD / SERI-RES extensively. Reliable if user-unfriendly. The minimum temperature without heating is lower in a lowmass than in a high-mass bulding.

It's known from experience and measurements that high- and low-mass buildings behave differently. Insulation and airtightness also make a low-mass building more thermally stable. Don't these engineers acknowledge the results of scientific experiments?!

SAP-2005, which is used to model most dwellings, takes no account of thermal capacity - BREDEM iis a static monthly (or annual) degree-day model. Work in Canada shows that a high-mass dwelling is able to utilise passive solar gains (from large south facing windows) much more effectively than a low-mass dwelling (both were close to Gold Standard insulation and airtightness) On this basis, a low-mass dwelling uses 500-2000 kWh/yr (depending on size) more space heating energy than a high-mass one.

David.

[Mark Siddall]

David,
Your Canadian case studies, are they available as publish data? If so could you point me in the right direction, books, reports, websites etc.
Thanks,
Mark

[Dr Eric Roberts]

Simulation software: There are a huge number of codes out there do a search for the following paper:

"Contrasting the capabilities of building energy performance simulation programs" by Drury Crawley.

This will give some insight into some of the leading ones. All have different levels of functionality and usability.

For solar simulation there is simply no better tool than the daddy of all modular simulation tools TRNSYS. This program excels in solar simulation but is very difficult to use and you may need to have the ability to code your own software modules to gain the most from it.

All simulation tools are able to modelthe effect of thermal mass, if anyone is showing that they can't or there is no effect then it is with the user and the input that the problem lies.

IES does use the SBEM systems for a large part of its simulation (the actual heat transfer is by finite difference). This has inherant problems in that you are restricted to certain specified system types and the vague inputs required by the NCM. TAS does not use any SBEM inputs within its simulation (apart from in the Part L macro) and hence gives greater flexability.

The NCM (SBEM and SAP) and the building regulations is a flawed document and directs people to the easy option of creating energy intensive buildings in which it is easier to strip out services in order to pass rather than reward good low energy design from the outset.

Mark, I have been involved exclusively in building modelling and simulation for over a decade, if I can assist in answering any questions you may have please let me know.

[Mark Siddall]

Dr Eric,
Thanks for the insight,hearing this from a third party was what I was what I had been hoping for. (I didn't want trust the reps too much as they are mainly interested in selling the software.)
I'll hunt down the document that you have recomended.

Mark

[Nick Grant]

Hi Eric

As you have been using these models for over a decade you'll be able to answer the question I've been keen to ask since this thread started. What is the accuracy of the predictions of such models, ie the energy bills and summer temperatures?

It would be great to hear about some case studies with kWh/m2 data.

I am an amateur but have seen so many apparently computer modelled buildings that simply don't work.

In using dynamic simulation for UK designed and built buildings are we kidding ourselves?

As a numerical simpleton, I'm a fan of designing out as many elements that would need modelling. eg thermal bridges, uncontrolled ventilation, complex interactions between solar gain and mass etc.

This seems to be part of the Passiv Haus philosophy as well.

Nick

[Dr Eric Roberts]

Hi Nick

As with all things the accuracy of the answer depends soley on the accuracy of the information input. The programs themselves replicate the underlying fundamental physics of heat transfer, air flow etc, the problems arise of the ability of the user to input the data correctly.

As with all building simulation - we are trying to predict the furture from the past. In this I mean all our predictions are based on the use of historical weather data, they are merely indicators of how the building will perform under certain enviromental criteria.

There have been numerous studies over the past few decades of validating simulation against real life meaured data,  including some of my own. Actual measured data from operational buildings is very difficult to come by, but simulations are able to replicate this and again provided you give it all the correct information you can gain good predictions of energy and internal performance. But you need all the weather data for the site to do this correctly - this is nigh on impossible to obtain - so we use more generic data sets.

There is a lot of poor modelling conducted on a daily basis, there are numerous reasons for this which I will not go into here.

Dynamic simulation of all buildings should be an integeral part of the building design process to analyse the complex interactions that occur, from this surely more informed design decisions can be made?

I alwasy ask those who are questioning the use of simulation in any design the following.....

Would you consider flying on a plane or driving a car that hadn't gone through rigerous computer simulation and numerical testing?

Why should a building be any different?

It is the input of the data input and interpretation of the output you need to watch not the programs themselves.

Eric

[Nick Grant]

Thanks Eric

However my question wasn't so much about the mathematical abilities of computers as about the accuracy and usefulness of the whole process of dynamic simulation of real building (designed by architects, engineered by engineers and built by Roberts) from first principles.

I suspect a factory built aeroplane is easier to model than a typical semi. In addition the aeroplane has evolved and the empirical equations have been refined by trial and error. If an aircraft has double the predicted fuel consumption it won't make New York and yes I'd accept a ride in an old biplane if offered but that's not really the issue.

I really don't have an axe to grind on this, I'm just very curious. As I said, my hunch from ignorance is that most buildings as designed have too many varying variables (to clean up Peter Harper's lovely phrase) and we should put more effort into simplification.

Must say I'm concerned that results are so sensitive site weather files given the huge variation that happens from year to year, with or without climate change.

So can we predict performance to say 10% or do we always have to alter the inputs to make the answer come out right?

Nick