In the United States we are experiencing a reversal of some of our safety regulations. We are also witnessing a decline in the size of some of the regulatory and investigative agencies. Examples are:
Some of these decisions can be justified on the following grounds.
But most of these regulations were written in response to a specific event, such as Deepwater Horizon/Macondo or Bhopal. There has to be a concern that the likelihood of a recurrence of one of these events is increasing.
There are indeed many problems to do with writing detailed regulations. Companies can be tempted to "game" the system, as Richard Rosera notes. Also, no regulator can have the expertise to develop rules for the vast number of hazardous chemicals manufactured and used by industry. And, as Mr. Kumana says, detailed regulations can provide a means for eluding accountability.
All of which is why the philosophy behind process safety management is to make it non-prescriptive and performance-based. In effect, the regulator is saying to industry, "You know more than we do about these chemicals, do what it takes to achieve safe operations".
The issue of automation and robotics also encourages a non-prescriptive approach. Technology is moving so fast - particularly in subsea operations - that no regulator can keep up.
The catch with a performance-based approach is that we are forced to address the vexed topic of "acceptable risk". That's a subject that probably deserves its own discussion.
I think Mr. Pitblado misses a key point in Mr. Bullough's note. Performance based rules are always going to be at the mercy of the group(s) evaluating both the hazards and the effectiveness of the mitigative measures. Large organizations with strong technical support and critical interests at stake will probably usually do a good job. Smaller organizations with less depth, less support and organizations evaluating tangential and less important areas may not tend to evaluate either correctly. Almost all US codes have the option of allowing the organization to prove that a different approach is equally safe. The onerous of that proof is on the organization and the authority having jurisdiction is right to treat any of these with some prudent skepticism. The real world, as Mr. Bullough notes, is often much more dangerous and difficult to analyze correctly than it immediately appears.
My experience with overly burdensome regulations is usually not in the prescriptive requirements but in the documentation that most organizations find necessary to ensure compliance. And that documentation is likely to only increase in a performance based approach. So perhaps there are some real downsides to that as well.
Pilot plant and laboratory consulting, safety, design,reviews, and training
Richard P Palluzi LLC
72 Summit Drive
Basking Ridge, NJ 07920
Following on from Robin Pitblado's comment, one of the most serious incidents that I was ever involved in had an initiating event that could never realistically have been forecast by even the most thorough hazards analysis team. It was something totally out of left field.
But the consequences of this initiating event were made much more severe by failures in the safeguards. For example, some of the relief valve headers were plugged with solid materials. This problem is something that a PSM program could be expected to identify.
The basic idea behind Nicholas Taleb's book Black Swans is that we cannot and should not try to predict the likelihood of highly improbable events. Instead we should accept that such events - "Black Swans" - can occur, and we should build our defenses on that assumption. He is talking about financial systems, but his ideas apply to the process industries. In other words, we should spend more time looking at the right hand side of bow-tie diagrams (or building Event Trees as well as Fault Trees).