Merrimack Valley Gas Explosions Part 4

21 Feb 2019 at 20:10 // under Safety Chemical Engineering Process Control

In my final post on the Merrimack Valley Explosions¹ , I want to cover the recommendations that were issued in the Safety Recommendation Report. The report makes five urgent recommendations, one to the state of Massachusetts and four to NiSource, Inc., the parent company of the operating company in charge of the gas network.

The recommendation to the state was to:

Eliminate the professional engineer licensure exemption for public utility work and require a professional engineer’s seal on public utility engineering drawings. (P-18-005)

This is the equivalent of requiring, by law, that a chartered engineer signs off on the work. This is standard for many bits of work, but there are various exceptions to this requirement. So if this recommendation had been in place before the accident, an experienced and accredited professional engineer would have had to be involved. Would a Professional Engineer have managed to spot that the pressure sensing lines were on the section of line to be taken out? Note that the drawings that were used to plan the work did not have the pressure sensing lines on them. Those drawings were held by a different department.

Note that I am not against requiring sign off by accredited or professional engineer. Just that I am skeptical that it would have had much of an influence in this case.

The first recommendation to the company was:

Revise the engineering plan and constructability review process across all of your subsidiaries to ensure that all applicable departments review construction documents for accuracy, completeness, and correctness, and that the documents or plans be sealed by a professional engineer prior to commencing work. (P-18-006)

There is again the recommendation to require approval by a professional engineer, but this recommendation also includes the need for all applicable departments review the planned work. Except that is already the case. Footnote 6 of the report states:

The Columbia Gas Capital Design Job and Constructability Review Checklists allow the Engineering department to determine what departments outside of the Engineering and Construction departments review the project.

So what we have here is someone made a mistake of not getting the metering department involved, because they thought that they didn’t need to be involved. How can you ‘ensure that all applicable departments review construction documents’ when someone doesn’t notice that the metering department should be involved, because the drawings they have don’t show the pressure sensing lines on them?

Yes, ensuring the correct people are involved in reviewing the documents is the right thing to do. I just don’t think anyone within the company didn’t believe they were not already doing that. If you had made the recommendation to them the week before the accident, again, I am not convinced it would have had any effect. They would have turned around and said ‘we already do that’.

The second recommendation to the company was:

Review and ensure that all records and documentation of your natural gas systems are traceable, reliable, and complete. (P-18-007) (Urgent)

Ahh, the fun game of ensuring that documentation is up to date and accurate. This is a constant battle for operators of old equipment. That is not an excuse, just a point of note. It is something that is difficult to do. And again, in this case there were a set of drawings that were correct, just that they were held by a different department. Perhaps the drawings that the department planning the work had should have included the extra information on it, but that is not really what is being said. You can’t fit all information on every diagram. A P&ID² has a lot of information on the routing of pipes, valves and instruments, but it doesn’t show the actually layout. It is a great representation for a chemical or control engineer, but an inspection engineer would find it of little use. Similarly an isometric drawing of the line is useful for the inspection and mechanical engineers to point out where defects are etc, but a control engineer would find it very difficult to follow because they generally just look at one line at a time, not how the different pipes interact with each other.

The third recommendation was:

Apply management of change process to all changes to adequately identify system threats that could result in a common mode failure. (P-18-008) (Urgent)

In this case I do agree that it was wrong not to go through the company management of change process for this type of job. It seems that the company only used their management of change process for certain types of work and the replacement of this section of pipework was covered. However this is again a system that would have relied on humans checking the work. Management Of Change systems can look very gand and cleaver, particularly when people like to use the acronym MOC. But it doesn’t magically fix anything. All it is is a form that different people have to fill in. The same people that made the mistake in the first place. Would an MOC form have manage to catch the mistake when the checklist method that they did us did not?

The fourth and final recommendation to the company:

Develop and implement control procedures during modifications to gas mains to mitigate the risks identified during management of change operations. Gas main pressures should be continually monitored during these modifications and assets should be placed at critical locations to immediately shut down the system if abnormal operations are detected. (P-18-009) (Urgent)

This is probably the only recommendation that I agree with and believe would have definitely avoided the accident occurring. I am surprised that something like this did not exist in the first place, but then everything is obvious with hindsight. Again though, as this is hindsight, we should ask the question: would this risk have definitely been identified? Would the folk have definitely considered monitoring the rest of the network or would they just have focused on the small section they had been working on?

As you can probably tell, I am not particularly comfortable or happy with the immediate recommendations. Perhaps when the final report is produced, there will be other recommendations that provide a more comprehensive way to reduce the chances of a repeat event, rather than just relying on people making the correct decisions with the information they have.

1. Well, my final post for now. I may revisit the topic if more information comes to light. ↩
2. Piping and Instrumentation Diagram ↩

Go Top