Effect of wind on natural smoke ventilators

[GB]

SCA guidance provides design parameters when analysing smoke control systems particularly useful when using a CFD model.

It states that the effect of wind when using natural ventilation can be adverse and should be considered.

There are no design parameters within SCA guidance or the relevant BRE report the guidance refers to guide on how to consider the effect of wind on the AOV..

Is anyone aware of relevant design parameters to consider the effects of the wind in a CFD model reviewing the performance of AOV's in a corridor & adjoining stair?

[Phoenix]

What design parameters do you want?  Just try a couple of different reasonably expected wind conditions.  Many modellers use no wind as their wind condition.  That's seldom realistic.

[GB]

I have looked at the prevailing wind conditions which is from the West with an average speed of 1 - 3m/s.

We have placed the AOV's on the East & North elevations.

We have therefore considered the effect of the most common wind condition - but how do you assume a design parameter that is reasonable?

We can with any natural ventilation system model a wind effect that will stop the clearance of smoke - for example an Easterly or Northerly wind at say 5m/s reasonable?

If the AHJ wants you to use mechanical extraction instead of natural ventilation - then design parameters can be specified which are unreasonable in order to prove failure of the natural vent system!

[Phoenix]

It is not unreasonable to expect that winds from the North or East at about 5m/s (moderate breeze?) might occur with some level of regularity so if the system doesn't work under such conditions then it has significantly reduced reliability.

If it is for life safety then the reduced reliability is a real issue.  There are a number of avenues that could be explored for natural ventilation but they will all involve some redesign of the building and, ultimately, the use of fans might be the only viable solution that fits in with the design of the building.

[Fishy]

Not my thing, but I note that Chapter 10.5.5 of CIBSE Guide E (Wind overpressures) refers you to BS CP3 Chapter V, which has been replaced with BS 6399-2: 1997 (Loading for buildings: Part 2: Code of practice for wind loads)...

[Phoenix]

When analysing wind loads on a building it is necessary to consider winds acting on each face.  So it is with smoke control.

[kurnal]

True enough Phoenix but isn't it all smoke and mirrors in any case and any calculations in support of mechanical smoke extract systems  will be based on assumptions that may or may not apply such as the fire size, location of the fire, leakage past the door, how long the door is open, ventilation within the compartment etc. Although we have some guidance on this (BRE) I often wonder how realistic this is? I have seen even this guidance stretched to ridiculous degrees to justify extended travel distances by fire engineers, including reduction in fire sizes and even  basing leakage rates on fire doors with taped thresholds (in the event the abject workmanship on that particular  project the workmanship was such that the threshold gap was 25mm in most cases but nobody was interested in making the developer do anything about it).
 

[Phoenix]

Hi kurnal,

Yes, you're right. All the fancy calculations and fire modelling amount to....
....well, that's the question.  Are they reasonable estimates or are they just nonsense?

I try to have a little faith and think that if they are used conservatively then they can present best estimates and can tend to indicate that a building might be safe.  As you have pointed out, they are not always used conservatively.

If we chuck away all the fire engineering guidance and modelling techniques we have because we don't trust the people using them, then we have nothing, so I think we have no choice but to continue to have a little faith...

There are people who know that the modelling techniques are merely best guesses and who therefore use them and handle them with caution, openness and honesty, and there are others who don't care about the accuracy of the predictions and manipulate them to suit their needs.  Telling the difference between the two approaches is often very difficult for an enforcing authority.

Look at weather forecasting.  They have some of the biggest computers in the world with the most complex modelling techniques and they can't get yesterday right.  Predicting fire and smoke behaviour is not so fraught with difficulties as weather prediction but it is still subject to many variable inputs, many of which are difficult to determine.  I don't think we'll ever get it right, I don't think it's possible, but at least some of the good guesses can get nearer to the truth.  I think we should, for the time being, adhere to what we know is reasonably foreseeable when making predictions and temper the outcomes of our predictions with some common sense.

[Mike Buckley]

It goes back to the scientific method, create a model based on the best knowledge available then review the model based on the results and then adjust the model if necessary.

Fine the initial model may have been a best guess but it is better then nothing and can then be adapted.

[GB]

On the same thread....BS 7974-7 Section 7.1.12 discusses reliability of smoke control systems (which I assume is both mechanical & natural?)

It states that there is very little reliable data available however estimates that reliability is around 85 to 90% which I presume is subjective opinion or an industry assumption.

Does anyone know of a source of reliable independent data (i.e not data sponsored by..... a party either for or against smoke control??)

[Fishy]

On the same thread....BS 7974-7 Section 7.1.12 discusses reliability of smoke control systems (which I assume is both mechanical & natural?)

It states that there is very little reliable data available however estimates that reliability is around 85 to 90% which I presume is subjective opinion or an industry assumption.

Does anyone know of a source of reliable independent data (i.e not data sponsored by..... a party either for or against smoke control??)

Nothing that I'm aware of - in any case PD 7974-7 (it isn't a full BS) is, in my opinion, a pretty academic document that's close to being un-useable for life safety related analyses.  Probability of Failure on Demand (PFD) is data you might get for certain systems used in process or nuclear safety, but not typically for commercial off the shelf 'conventional' fire safety kit.

I've only ever seen probabilistic risk assessment used successfully to compare different fire life safety approaches, not to justify a single approach from first principles.  CIBSE Guide E gives some cautionary guidance on this. Application to loss prevention is a different matter, as you might be able to ascertain the value of the assets you're considering whether to protect or not fairly accurately, but I were ever tempted to use the figures in the PD I'd do it cautiously because many of them are decades out of date.

[Phoenix]

I agree completely, Fishy.  I have used 7974-7 for some property protection matters on large scale projects, basically cost benefit analysis, but it has very little relevance for life safety matters or for application to individual buildings.  As for the reliability of the data, who knows?  It's very limited and it's not easy to ensure that the data you take from it is entirely compatible with what you need.  

Probabilities do have a part to play in a building's fire risk assessment when considering the prioritisation of remedial works but only in a qualitative way.

GB, that 85% - 90% you quote is pretty meaningless.  For a full picture of this we would need definitions of success and failure and we would need a large sample.  I don't believe we have a large sample of how smoke control systems have performed in fire.  As for the definitions, we would strictly only need a definition for success or failure as the two would be mutually exclusive but the definition would not be simple.  Each building has different fire safety objectives - life safety? business continuity? heritage? property protection? all of these? other? (often they don't know what their objectives are!) and success or failure can only be judged against the objectives of the system.

In short, qualitative judgement is often the most pragmatic approach unless the data is fully reliable and relevant.

Ask, what's "reasonably foreseeable?"