At our recent webinar on smoke and environmental ventilation of multi-storey buildings using shafts, we had a large attendance with many questions asked. Here you can see my answers to these questions, slightly edited for clarity.
There is also a recording of the webinar available. Watching the recorded version will still enable you to claim CPD points.
What was the design fire size used developing the Colt Shaft?
We used exactly the same fire size as BRE used for the BRE shaft which was a 2.5 megawatt fire. In terms of the key differences between the original Colt CFD model and the BRE CFD model, there were a couple of things which BRE didn’t report in their development report for the BRE shaft, not from any sleight of hand but simply because they weren’t considered as important in terms of the conclusions. Basically they allowed a leakage area from the Fire Room and obviously that then decreases the pressure slightly in the Fire Room and also allows the fire to develop better. When we originally tried to model without that opening we found that, first of all we couldn’t get the 2.5 megawatt fire size because we did not have enough ventilation and secondly that what smoke there was, was then pushed into the lobby together with flames because the lobby door suddenly provided a lot more ventilation to a restricted fire.
If lift doors are not smoke sealed, how does this affect the negative pressure in the lobby?
One of the advantages of going for a pressure differential sensor within the lobby is that the fan will modulate to match the leakage of the lobby. So if you have a very, very tightly sealed lobby, under the doors closed conditions the fan speed will drop to a very low level. If you have got more leakage, if you have got a lot of lift doors in the lobby for instance, then that will be taken account of, and the fan will just run at a slightly higher speed, so you end up with the same pressure but with a different fan speed achieving that pressure. So there’s no effect at all on the pressure - this just affects the fan speed needed to maintain the pressure.
What does “adiabatic” mean?
Adiabatic in this case is evaporative cooling, where we are basically cooling by moving the incoming air through a dampened screen so that we can make use of the conversion of water into vapour and make use of the fact that the latent heat has been removed so as to cool the air coming in.
Does the ventilation system have to factor in wind loads and wind shear on high rise buildings?
We don’t factor this in with a typical Colt shaft system. Normally the only reason that you are using a shaft system is that the corridor and the stairs are well within the building, so they are not actually exposed to the external wall. The only issue that we are really interested in is the location of the exhaust point to make sure that this won’t be adversely affected by winds, but normally that’s mounted up on the roof in a low pressure area so it’s not a problem.
There is certainly more of a concern in some quarters over pressurisation systems and whether those are affected more by stack effects and wind effects, but there is no guidance on that at the moment.
Is there a defined flow rate with a natural ventilator? Or does this depend on certain factors and which ones?
With a natural system then the answer is “no”. Our Regulations simply define the areas required for the ventilators, and the assumption is that if you meet this then it is deemed that you have satisfied the requirements. Certainly if you are looking at just using a 1.5m2 AOV there will be wind conditions where you will get very poor flow conditions but that is a weakness of our Regulations. It’s a pity that simple AOV’s are still permitted, but of course they are cheap and they are simple.
What was the time needed for the CFD simulation for the Colt Shaft?
We effectively have two different times: we have a timescale of the length of time that we are simulating and there is also how long does it take to run a simulation.
We ran the simulations for about 20 minutes, that was to give time for the fire to develop and for people to escape from the lobby and to allow 15 minutes for the Fire Service to arrive and start opening doors and fighting the fire and then a further 5 minutes to see what happens if we leave the door open and if we reclose the door. A 20 minute period is quite typical for these sorts of models. If you look at any guidance, roughly 20 minutes is typical.
If you are looking at physically how long it took to run the models, then typically these models would take somebody a couple of days to build and then it could be anything from 2 days to a week of computer time to run the model.
Missed the webinar?
Paul Compton is Technical Director for Colt, experienced in smoke control, HVAC, solar shading and louvre systems.