Author Topic: Probabalistic risk assessment vs. good practice  (Read 34458 times)

Guest

  • Guest
Probabalistic risk assessment vs. good practice
« on: December 13, 2004, 01:45:19 PM »
Risk assessment

For discussion:  does the introduction of purely risk-based fire legislation mean that an approach to fire safety along the lines “I probably won’t have a fire, therefore I need do nothing” or even “…I am spending too much and can get away with much less” legitimate?  

Consider the following – an organisation uses established risk assessment methodology to assess the risk from fire for its premises.  Based upon past incidents of fire; statistics from similar properties and the use of the premises (good training, strong management ethos and relatively well-controlled fire load, perhaps), they are able to make a quantitative assessment that the risk of having a major fire is demonstrably very low.  They also regularly practice evacuation and are able to reliably and fairly rapidly clear their building(s).

They conclude from this that they can dispense with much of their fire-resisting structure, all sprinklers and first aid fire-fighting equipment.  They conclude that they do not need to follow British Standards (e.g. BS 5588) or other relevant guidance (including that from the Regulators), because they control risk to a sufficient degree that compliance would be unduly onerous – the cost of doing so (and maintaining the kit) would be disproportionate to the risk reduction achieved.  

They are similarly satisfied that property / asset risks are sufficiently controlled and they self-insure.

If all the law requires is that risks are as low as reasonably practicable, it is certainly possible that the fire risk in many relatively low-risk premises is so low that there could be a strong argument that existing fire protection is unnecessary.  If the guidance on what is represented by ‘as low as reasonably practicable’ does not consider how ‘good practice’ and risk assessment methodology interact, we shall shortly be facing the situation where some larger organisations start to question the ’value’ of their existing fire protection (let alone improvements in fire protection).

As you may infer – this is not a theoretical discussion – there is at least one major organisation who is actively considering the above approach.  The organisation is big enough so that the property / asset risk can be ‘absorbed’ without undue risk.

We would be very interested in the views of those on the forum on the whole issue of ‘risk assessment’ vs. ‘good practice’.  Existing guidance in pd 7974-7 appears to be silent on this issue.

Chris Houston

  • Guest
Probabalistic risk assessment vs. good practice
« Reply #1 on: December 13, 2004, 03:34:45 PM »
I think this whole argument is based on a minunderstanding of the concept of "risk."

In most environments the probability of fire is quite low, but "risk" is defined "as the combination of the probability of an event and its consequences" (ISO/IEC Guide 73)

Therefore, the consequences of a fire must be multiplied by the probability.  If the result (risk) is still low, i.e. low probability and low consequences, then fine  - do away with the sprinklers, but if the conseqences are high (i.e. include possibility of death, or large expence) then risk control is needed.

So the real problem is a minunderstanding of risk, due to a lack of competant risk management professionals.

Probability x Consequences = Risk

Guest

  • Guest
Probabalistic risk assessment vs. good practice
« Reply #2 on: December 14, 2004, 01:54:07 PM »
Absolutely not - the risk assessment methodology used in the above is industry-standard and fully considers the hazards, probability and consequences on an 'event'.  If probability is low, the value of the 'consequences' part of the equation becomes largely immaterial.  The method does not, however, take any heed of ACOPs, BSs, ADB & other guidance - the assessment claims that risk is low so there is no need to.

The issue is not just sprinklers - it relates to most of the fire protection - if the 'purist' risk assessment methodology is followed there will be little fire protection left.

It is a fact that the probability of having a major fire in any one commercial premises (over its design life) will be very small indeed.  Even if you do have a fire, generally speaking the probability of death or serious injury (in an office building, for example) tends not to be high.

Once we have 'pure' risk-based legislation, the route will be open for any employer to reduce/remove their existing / future fire protection measures if they believe that they can demonstrate that the risk reduction they provide is disproportionate to the cost of provision / maintenance.  We could end up with little protection left.  

As a fire engineer, deciding not to provide well-established and proven 'safe' levels of protection, as enshrined in British Standards etc, on the 'assumption' that you won't have a major fire (dress it up as 'risk assessment' if you like) fills me with concern!  I would welcome the views of others!

Offline Colin Newman

  • Full Member
  • ***
  • Posts: 114
    • Healthfire
Probabalistic risk assessment vs. good practice
« Reply #3 on: December 16, 2004, 04:54:21 PM »
In the above scenario, the likelihood of a fire occurring is deemed as low.  What is the basis of that assessment?  Is it historical data?  If so, was the data set collected from premises with reduced fire precautions?  

If it had been collected from premises with approriate fire preacuations as per the appropriate standards then using this data set to argue a reduction in the very precautions that had influenced the data would not be a valid argument.  Such a data set could not be directly releted to premises without the level of preacuations that existed in the premises for which they were collected.

The argument being proposed is is akin to stating that your trousers have not fallen down for at least a decade, so now you can remove your belt! :oops:

Chris Houston

  • Guest
Probabalistic risk assessment vs. good practice
« Reply #4 on: December 16, 2004, 08:11:58 PM »
Exactly.

There is an argument that risk assessments should consider the "gross" and "net" risks.  That is, the risk as it occurs without the control measures and then with them.

The gross risk of a fire without the risk control measures (ranging from compartmentation, alarm, fire escapes, etc) surely could result in significant property damage and death in the worse case scenario.

Therefore, to most organisations, irrespective of their risk culture, death and loss of the building, is most likely to be considered catastrohpic.

I could concur that the risk of a fire in most workplaces may well be low, however, a "low" probability of a catastrophic event cannot be acceptable.  The control measures must reduce the probability to "impossible" or "unlikely ever to happen" before such a risk can be deemed as being properly controlled.  

I would not work in a place where the net risk of death was just "low".

Offline p.b.morgan

  • Newbie
  • *
  • Posts: 29
Probabalistic risk assessment vs. good practice
« Reply #5 on: December 21, 2004, 11:27:59 AM »
While risk assessment depends on identifying hazards what about the hard to identify ones like arson and terrorism? Commercial sector buildings are recognised as targets more and more post 9/11.

How does this equate with the 'low' probability cited by the large organisation our guest is involved with?
Penny

Chris Houston

  • Guest
Probabalistic risk assessment vs. good practice
« Reply #6 on: December 21, 2004, 12:12:20 PM »
I think those most at risk of terrorism are probably very aware of it - however for the rest of the organisations in the UK, the risk control measures for dealing with terrorism risks are often the same as those that ought to be deployed to prevent break-ins, theft, tresspass and other risks that have been faced for many years.

Any risk to staff health or safety has to be assessed and controlled under existing health and safety legislation.

Arson should be considered as part of the fire safety risk assessment.  It's a daily event in the UK and should not be considered a rarity.  In some sectors it is the most likely cause of fire.  Unlike an accidental fire, arson attacks can have more than one seat of fire and may be aided by an accelerant.

"The Risk Management Standard", AIRMIC, ALARM and IRM 2002 is available from the IRM website http://www.theirm.org/

Offline Fishy

  • Hero Member
  • *****
  • Posts: 777
Probabalistic risk assessment vs. good practice
« Reply #7 on: May 09, 2005, 04:28:14 PM »
It'll be interesting to see how the RRO deals with this - will it be acceptable for an organisation to assume that it controls the risk of a significant fire occurring so well that it doesn't need to have detection, suppression, fire resisting construction etc, because the cost/benefit of doing so is so low?

pd

  • Guest
Probabalistic risk assessment vs. good practice
« Reply #8 on: May 09, 2005, 06:56:06 PM »
The ODPM has been banging on about the concept of carrying out a FRS by reducing the risk to a minimum before applying general fire precautions to control the remaining risk. All very sound but it does open the door to the argument that if the risk has been so reduced that it is neglible then, as said above, the need for any protective measures becomes redundant.
This is a dangerous route, made more so by expecting 'nominally competent' people to make this judgement. Professor Everton has been trying to make this plea, largely falling on deaf ears, for months.
The result of poor fire precautions is demonstrated nearly every week by some fire disaster in a largely unregulated part of the world.

Offline eddy orr

  • Newbie
  • *
  • Posts: 19
Probabalistic risk assessment vs. good practice
« Reply #9 on: May 11, 2005, 01:34:33 PM »
It is not necessarily a bad thing that companies are reassessing fire risk, if done correctly it should be a good thing.  

However to do so on a high level, looking at the occurrence of major fires and not taking into consideration minor fires is deeply flawed.  Further to then not adhere to the fire safety standards (ie flammability, surface flame spread, fire withstand and containment etc) that are in place to mitigate the consequence from minor fires, then removing the means to extinguish those minor fires is wrong.

Apart from not providing necessary fire extinguishing equipment, it must be assumed that all minor fires will grow unimpeded, until there is an external intervention by the FRS, therefore any arguments based on the level of risk must treat minor fires as major fires in this case.
Hope this helps

Offline Ken Taylor

  • Sr. Member
  • ****
  • Posts: 414
Probabalistic risk assessment vs. good practice
« Reply #10 on: May 18, 2005, 01:33:01 PM »
To follow on from Chris's points, we need to consider the risk from fire as well as the risk of fire. That's what the compartments, sprinklers, etc are there for. Unless you are dealing with a guaranteed fire-free zone (empty concrete box?) you need to control the risk to occupants and others who may be affected. Does anyone imagine that those in control of the building will escape litigation after a death or injury from fire by arguing upon the basis of such an assessment?

Offline Fishy

  • Hero Member
  • *****
  • Posts: 777
Probabalistic risk assessment vs. good practice
« Reply #11 on: December 29, 2006, 09:35:09 AM »
Thought I'd resurrect this one...

Does the RR(FS)O have any impact on this thread?  Are we any closer to the Responsible Person's risk assessment being based almost entirely on the principle that it is unlikely that they will have a fire... and if not, why not?

Offline kurnal

  • Hero Member
  • *****
  • Posts: 6489
    • http://www.peakland-fire-safety.co.uk
Probabalistic risk assessment vs. good practice
« Reply #12 on: December 29, 2006, 10:01:33 AM »
Yes we are. Following the RRO guidance the first steps of the RA are to identify sources of ignition and fuels that may burn, and the risk that the two may come into contact.
If you have neither, then theres no risk and there is no need to incorporate any risk control measures.

Example - large shed where concrete slabs are manufactured, cast and stored, almost entirely by human  effort supplemented by a diesel fork lift.
The only risk control measure is that bins for empty bags are removed on a daily basis and taken out of building. No fire alarm, no FFE, only most basic of emergency procedures.

Offline Fishy

  • Hero Member
  • *****
  • Posts: 777
Probabalistic risk assessment vs. good practice
« Reply #13 on: December 29, 2006, 10:18:05 AM »
'No risk' is fairly straightforward - but what about risks that could be regarded as 'broadly acceptable', according to the HSE guidance?  If you look at the UK fire statistics, next to no-one dies in an office fire each year, despite the fact that there must be tens, if not hundreds of thousands of them in the Country.  Why, then, should we make any great effort to mitigate a risk which is already demonstrably acceptable?  As a manager of such a premises, can I not conclude that much of the existing fire precautions (which cost me money to manage and maintain) are unnecessary, because it is so unlikely that I will have a big fire, and even then there is a low probability that anyone would die?

I have my own views (which aren't entirely in agreement with the above); it's intended to provoke discussion, because there are Responsible Persons out there thinking along the above lines...

...and please let's keep the Insurers out of the picture at the moment - we all know they have a legitimate interest, but it's compliance with the law that's of primary concern.

Offline PhilB

  • Hero Member
  • *****
  • Posts: 963
Probabalistic risk assessment vs. good practice
« Reply #14 on: December 29, 2006, 11:01:56 AM »
Some good discussion points here chaps!!!

I have always argued that the likelihood of a fire happening is not really that important. You have to work on the assumption that a fire could happen. It may happen only once in 400 years but that once could be tommorow.

Yes you can have a building that has no ignition sources, and no fuel. Perhaps a concrete shell with no electricty supply and no furniture. As soon as people resort to the building you introduce ignition sources and fuel.

Therefore in my opinion there is a reasonable likelihood of fire occuring in all buildings. What you can do is reduce the likely development.

What worries me is that some people, including my dear friend Mr Todd, believe that if you reduce the likelihood of fire occuring you can reduce the preventive and protective measures required. This is a load of gonads in my opinion.

If you reduce the potential for development you can reduce the preventive and protective measures but only so far so as to maintain minimum standards for life safety.

And as far as the Fire Safety Order is concerned the responsible person must take general fire precautuions to ensure the safety of relevant persons.

This includes people in the vicinity i.e. your neighbours. Therefore the old arguement that all persons can walk away from my building so I need take no further action has gone. You must reduce the risk & mitigate the effects to ensure that neighbouring buildings will not be effected.