لا يكلف مقارنة بما يحقق من اهداف

@saddek

Shlonik?

There are many types of risk analysis. I have good experience of risk analysis for major hazards in oil, gas and chemical plant; in fact for 102 such plants including complete refineries, ammonia plants, petrochemical complexes and pharmaceuticals plants. Follow up studies over a period of 36 years give experience of how many post analysis accidents occurred covering over 4000 process unit years

There are many steps to a good process plant risk analysis, The key ones are hazard and operability analysis, quantitative risk analysis, equipment design review and either pre-commissioning inspection or mechanical integrity inspection.

It should be recognized that no risk analysis is worthwhile unless it gives recommendations for risk reduction. I have never studied a plant which did not need risk reduction. Also, risk reduction recommendations are worthless unless they are acted upon, and this generally requires considerable help to companies in implementing safety engineering.

Hazard and operability analysis generally requires from two to five days per process unit for a team of one facilitator, one scribe and a typical team of 3 to 5 persons from the company involved, i.e. between 14 and 35 man days. Typical cost is between US$14,000 US$ US$70,000 per process unit, although some companies allow for larger numbers of participants as an aid to communication to the design or operating teams. The US$1 -2000 per day cost covers salaries, consultant fees, company overheads for experienced employees and direct costs of meeting accommodation.

Quantitative risk analysis involves looking up/calculating the frequency and size of releases of hazardous substances, the effect of safety systems, the consequences of releases in terms of fires, explosions and toxic effects, the impact on people and possibly on plant equipment, and finally presenting the results. From personal experience, there are very few risk reduction recommendations arising from QRA, but when there are the impact is very large.

The effort required for QRA varies enormously, depending on the degree of automation of the calculations. A complete refinery can be analysed by a single analyst in about 4 weeks with the best automation, wit several tens of thousands of accident scenarios calculated. With more primitive software and a manual interface with design data the analysis can take a two or three man team one year.

Detailed safety inspection is usually carried out by a 4 to 8 man team of operations engineers with extensive experience, and usually takes about 1 day for a process unit such as a crude oil distillation unit. Smaller teams will take longer for the inspection. Inspection usually identifies far mor problems than QRA, but most are of smaller impact.

Determining the value of a risk analysis can be difficult because in principle you have to wait a long time to determine if your calculations of accident frequency are correct or alternatively you have to carry out many analyses. The BP company has done just that, as has the present author.

For new plant with risk analyses made during the design stage the value is predominately the early and increased production obtained by solving problems before the plant is built. The first full scale risk analysis carried out by the author’s group in 1978 for a small fine chemicals plant allowed production to commence after just one week of commissioning, compared with an expected time of 8 weeks. In a very large plant, designed and constructed over many years, the HAZOP and HAZID studies carried out, together with the safety engineering improvements to the design cost about $7 million, but the plant was brought into production about 9 months ahead of schedule, with a benefit from early production of about $700 million, a benefit to cost ratio of about 100. (Not all the benefit came from the HAZOP and QRA, some came from design review).

Some of the potential benefit for risk analysis can be found from “I told you so” accidents, where the company management ignored or failed to implement recommendations. In one case a platform manager refused to implement a proper permit to work system. Six years after the risk analysis was completed and recommendation made a maintenance team made a mistake in failing to shut down properly before starting maintenance. Gas escaped, exploded and killed 11. The cost of plant damage was $400 million.

Some of the risk analyses have had gratifying results. After a risk analysis was carried out for a chlorine storage unit a recommendation was made to turn a worker transport waiting room into a safety refuge. Twelve years later a large escape of chlorine occurred with a hazardous plume of several kilometres length passing over the waiting area. No one was hurt. In all there were twelve such “near miss” accident preventions from the 102 risk analyses

all are true to the level that cost is incurred .

Thanks

Any form of risk mitigation costs money. I have seen many times senior managers "willing to take the risks". Sometimes that pays off, sometimes it fails spectacularly. Such as allowing helicopters and passenger aircraft in the same airspace. A risk was taken to launch the space shuttle Challenger, when data available prior to lauch showed a 12% chance of what happened, happening.

The culture of the entitiy doing the work plays a big role - are they risk averse or risk takers? What are the consquences if they mitigate the risk (don't launch the Challenger and wait for warmer weather) versus take the risk. Obviously 20-20 hingsight tells us the answer.

It is not easy before taking the risk. Yes, even the act of risk analysis costs money. You can do probabilisitic risk assessments based upon the expected costs of failure (such as FMEA). But it all comes down to the management decisions. And sometimes the phrase "better lucky than good" applies.

Sorry for the non-answer but the issue of risk mitgation is as much psychological / cultural as it is a technical issue.