Gå til hovedinnhold

Seeking more relevant results

Risk analyses are meant to be a key tool in maintaining acceptable safety and contributing to continuous improvement. But the industry now needs to think along new lines in this area.

This article was originally published in
Safety - Status and Signals 2012-2013

To order a hard copy
(free of charge):

Send an e-mail to

This warning from the PSA reflects the fact that today’s use of such assessments often lacks important elements and provides little benefit.

Norway’s petroleum regulations require risk analyses to be conducted at every stage from design and development, via operation and modifications, to cessation and removal. Their main purpose is to serve as decision support, so that facilities are safely and robustly designed and activities are pursued in a safe manner.

But risk analyses carried out by the industry today contribute only limited new knowledge, according to Odd Thomassen and Vidar Kristensen in the PSA’s process integrity discipline.

With long experience from these assessments, they find it strange that the companies fail to set tougher standards for the outcome of the work done.

“We see that many of the analyses focus on repeating existing knowledge,” says Thomassen.

“They are largely used to systematise historical error and accident data, rather than applying the lessonslearnt to preventing new incidents.”

“Given the amount of knowledge they already possess, we think it’s odd that the companies don’t see a need to get a better return on the resources devoted to this area,” adds Kristensen.

“The industry now has detailed information on what’s required to build a facility in a safe and robust way,” he notes.

“It knows pretty well everything about designing a process plant to help prevent leaks or to limit their consequences if they happen.”

Kristensen points out that the industry is well informed about how facilities and barriers should be tested and maintained in order to perform their intended functions.

“So the companies have little need for the results from today’s risk analyses in order to design their installations in a good and robust way.

“Given that, the industry should be asking itself more often why it’s doing this kind of assessment and what it’s going to use them for.”

“We see not least that historical errors and/or accident frequencies are often used to assess risk in relation to most types of incident which could occur,” says Thomassen.

“When the companies conclude on that basis that the risk is acceptable, decisions are being taken on a slender and uncertain foundation.

“It’s more important to assess how far they have control over conditions we know are significant for incidents occurring – and the uncertainty related to these factors.”

In his view, the current approach suggests that many companies have failed to grasp what risk analyses are and are not, their basis, and how their results should be understood and applied.

Future risk analyses should be better suited to systematising and highlighting what the industry knows and what it does not, says Kristensen.

“It’s important that they bring uncertainties into the open, rather than concealing them. They must also reflect to a greater extent the relevant status of technical, organisation and operational conditions.”

The PSA has devoted much time to following up the industry’s understanding and use of risk analyses. This work is also a priority in 2013, in part through various supervisory activities.