By Gregory Hale
There are times when mistakes happen time and time again and no one seems to learn from them. That has got to change and Carlos Barrera is trying to help.
“The ultimate goal of an incident investigation is to find the underlying reasons for the occurrence of the event in order to prevent its reoccurrence. These reasons can be divided in two categories: root-causes and causal factors,” said Barrera, managing engineer at Menlo Park, CA-based Exponent during his break out session Monday during the 16th Process Plant Safety Symposium at the AIChE Spring Meeting and 10th Global Congress on Process Safety in New Orleans. “There is an incident and the company forms a team to investigate. It could last a couple of days or it could take a long time. You want to find the root cause.”
Sometimes the investigation ends up being very mechanical and it just seems like people are checking boxes on a form. It should not be that way, Barrera said.
“Lessons learned should not be a slogan everyone learns until the next incident comes up,” he said. “An investigation is a much more complex scenario. If you want to eradicate the problem, you have to find the root cause.”
Incident investigation teams identify root causes, but don’t pay enough attention to the causal factors, leading to the generation of recommendations that sometimes take care of the symptom but don’t cure the disease, he said.
Kelly Keim couldn’t agree more.
The Exxon Mobil research and engineering global technology sponsor for process safety said his company goes to great lengths to truly understand the root cause of the problem.
“We really have to find the cause of the phenomenon,” he said
During his presentation, Keim talked about how they could use data analysis to determine common themes from all process safety events from a company’s global manufacturing operations across 50 sites. The process covers identification of all process safety events, selection criteria for those to undergo greater analysis, identification of opportunities and where additional investigation could add value.
Part of what Exxon does is they characterize the significance of the event by applying API RP 754 performance indicators for the refining and petrochemical industries and they also use their own device called the incident risk analysis tool. They also standardize the investigation methodology. That means they are going to specify what data they are going to collect and then they give each incident an individual identifier. They will then start to dig down and collect as much information as they can that shows the activity, equipment, direct cause, root cause and the management system.
Once they get the information they are able to place it into four categories, near miss, tier 1, tier 2 and tier 3.
A tier 1 incident gets a full investigation and a tier 2 incident will also most likely get a closer look. Tier 3 and a near miss will get an investigation, but not as intense as the top two.
“After collecting the data and information, they create graphs; now is the time for true data analysis,” Keim said.
Through their own analysis and intense investigations, Keim said they were able to cut down on major incidents by about 10 percent.