Multiple structure fires sparked by suspected gas line failure in towns north of Boston 7

[bimr]

Authorities ordered residents to leave their homes immediately after dozens of house fires broke out in a string of communities north of Boston Thursday evening.

The Massachusetts State Police were evacuating multiple neighborhoods and restricting access to the area.

"Residents in the affected towns of Lawrence/North Andover/Andover who have gas service from Columbia Gas should evacuate their homes immediately if they have not already done so," state police tweeted.

"Gas lines are currently being depressurized by the company it will take some time."

Link

 

[CWB1]

ax1e, I haven't worked on the distribution end so cannot speak to regulation there, however I did design gas drilling and production equipment for a few years and can attest there are a number of regulations governing gaseous systems' safety in the American oilpatch. Unfortunately I cannot point to specifics thanks to a lousy memory and an excellent regulatory team translating the legal-eze and doublechecking our work, however I do recall the location, sizing, and redundancy of PRDs and shutoffs being critical requirements to prevent catastrophic failure. Granted, the oilpatch is heavily regulated between the EPA, OSHA, and other agencies but its still difficult for me to fathom this distribution system being approved and implemented as described online.

 

[1503-44]

In my experience, US included, the existing regulations are not specific and it is common to see pressure reduction installations with the only form of safety being provided i.e. one, or more shut-in valves, i.e. no positive overpressure protection in the form of relief to atmosphere or tank. It is also routine for oversight agencies to approve a X2 shut-in valve system. I cannot explain it. The Lawrence disaster is the reason why I can't.

When designers are questioned on why, the typical responses are
1. Costs too much.
2. No room for a tank
3. No safe area nearby, no existing flare system
4. Its low pressure
5. I put in two dedundant shutoff valves

The first 4 are rather self-explicitory, however faulty they may be. As for the last, it appears to be a commonly held belief that redundancy somehow can increase the safety of an unsafe system to acceptable levels. Did you hear that right? Yes, similar to the "lightning never strikes twice in the same place" logic, that often comes with a lack of a deference to the Bayesian probabilities (formerly referred to as Murphy's Laws) that seem to be ever present in these types of "situations", but seldom, if ever, appear in supporting calculations documenting a system's ability to meet probabilistically acceptable safety levels. Here there were a number of factors that came together, poor original design, locating sensors far from protective device, no 100% positive pressure vent, lack of proper as-built drawings, insufficient experience of rework team, age of system and maybe a few others too, that increased the risk of this accident's probability.

Bayesian probability might be better suited to predict this, in that if one shutoff valve fails, it wouldn't be a great leap to expect that another one just like it would also fail, thus having two would not be providing much of any higher level protection at all. So the lesson is, if you do believe in the "two is better" philosophy, at least make the two different; use a shutoff valve and a relief valve, which points us back "common sense engineering experience" and the Alberta requirement.

 

[davefitz]

I am surprised at the various comments that suggests that there are no fixed regulations of natural gas pipelines and their distribution to consmumer piping in the US. Each and every state has defined statutes that in turn reference federal regulaions, ASME codes, and NFPA codes that must be met, and in turn each of these codes rigorously define the required protection against overpressure . See US 49CFR part 192, or asme B31 series , or NFPA 54 nat'l fuel gas code, for example.

I did not read the Massachusetts final report, but it seems that there was inadequate supervision of the construction processes that led to the lack of overpressure relief during the system upgrade. Its possible that both the original system and the final upgraded system would have met code requirements, but that the process of trasnforming the system from the 100+ yr old legacy system to the modern higher pressure distribution system was flawed and not correctly managed.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick

 

[1503-44]

"the process of trasnforming the system from the 100+ yr old legacy system to the modern higher pressure distribution system was flawed and not correctly managed". Totally correct.

Nobody is saying that there are no applicable codes and regulations. There are regulations, but they are too inspecific on this topic.

CFR Title 49 Part 192 is the Federal regulation that governs "TRANSPORTATION OF NATURAL AND OTHER GAS BY PIPELINE: MINIMUM FEDERAL SAFETY STANDARDS". In 49 Part 192 Section 195 through 197 you can see exactly what they say about overpressure protection.

https://ecfr.io/Title-49/pt49.3.192#se49.3.192_1195

Note that in SS 195 there is nothing more required than to provide suitable "pressure relieving or pressure limiting devices". SS 197 is slightly more specific, and in fact actually mention a redundancy requirement. If you are at less than 60 psi[g], all you need is a suitable regulator. If you are over 60 psi[g], you have some options, see Paragraph c1, where two regulators and a "safety device" inbetween, limiting pressure delivered by the upstream regulator, must be provided.
Now the rub there is ...
"This device [inbetween] may be either a relief valve or an automatic shutoff that shuts, if the pressure on the inlet of the service regulator exceeds the set pressure (60 p.s.i. (414 kPa) gage or less), and remains closed until manually reset."
In paragraph c2, a "service regulator and a monitoring regulator" are allowed. Two very similar devices.
Not until you get to paragraph c3 do you find the last option,
"A service regulator with a relief valve vented to the outside atmosphere,"
IMO, if you are going to require something, require the safest solution. Why is it the third option? Order of preference? Not mine!
The last two options in fact, IMO are the better ones of the 4 listed, but I still believe 3 can't be beat.

In the USA, there are only options. A man with two watches never knows the time. The relief valve venting to safe location is not even first on the option list. 1 and 2, not so safe. Option 4 might have worked, but still no venting possible if that valve didn't close. I believe that if the USA requlation was worded exactly as is the Alberta regulation, the Lawrence disaster would have been avoided.

If anyone knows of any specific laws, codes or regulations, international, country specific, US federal, state, city or otherwise, addressing this topic, I would be very pleased if they could post the reference here. Thanks.

 

[davefitz]

AX1E;
Each is entitled to their opinion, but providing a full capacity relief valve in the tens of thousands of applications of consumer gas pressure letdown stations would be distributing a massive fire risk . I think the vast majority of cases use the combination of a fast shut off valve + dual press red valves in series + a "leakage" relief valve + pressure switches to avoid the major fire risk, and that design has proven to be effective and prudent . Simply imagining the safe venting of such large relief valves in urban settings boggles the mind.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick

 

[1503-44]

"relief valves in urban settings boggles the mind". I'll add that to my list of excuses.

What is a "leakage" relief valve? Why would that not require a safe location?

Large meter stations delivering gas from interstate pipelines, especially those that feed Class 3 and 4 locations, are where they are most needed. I don't believe that there are so many of those stations that it would make the requirement prohibitive. Furthermore these stations are usually located on the fringes of the cities that they serve, because as you recognise, they do need to create a controlled space to have a safe vent location to atmosphere. As such they would be entirely feasible to install. The billion spent on repairs here would have bought a lot of them. It turned out to be a false economy, no doubt the result of the desire to please shareholders willing to risk somebody elses life, limb and property. Pipelines need to "change the chip". That is the reason why just about everyone is fighting pipelines these days. They just don't get it. The public does. They have had enough.

And this happened before Lawrence.

https://www.metrowestdailynews.com/...e-pulls-back-on-gas-gate-project-in-hopkinton

That link only works on my tablet. I'll upload a pdf.

 

[1503-44]

Here it is,

 

[davefitz]

After years of use, virtually every throttling valve eventually develops leakage, usually minute amounts. If the downsteam low pressure system does not have any relief valves, then even a small throttle valve leak will overpressure the downstream system under zero demand situations. Therefore a "leakage " relief valve may be used , perhaps sized for 1% casualty flow to address leakage thru the 2 press red valves (in series).The venting requirements of a 1% casualty flow releif valve is much easier to implement than a 100% casualty flow relief vent.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick

 

[1503-44]

OK. While it should work for small leakages across a damaged valve, it wouldn't be much help during a regulator, or snap action valve failure scenairo.

 

[1503-44]

Update 29 July 2019
Columbia Gas settles Massachusetts blast lawsuits for 143 million

 

[davefitz]

The code-required system design would imply that 3 simultaneous failures to occur in order to overpressure the downstream system, which is an unlikely scenario.. Once again:
-2 independent pressure regulators in series, plus a
-fast acting ( 1 sec stroke period) stop valve, hardwired to a
-pressure switch with required calibration schedule
-usually a leage relief valve is considered as well.

This avoids the fire hazard of a full capacity relief valve

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick

 

[1503-44]

Yes we can do the calculations. However today's experts recognise (Nissam Nicholas Taleb, author of "The Black Swan: The Impact of the Highly Improbable" and four other works on the subject) that there is a big problem with calculating the probability of "extremely low probability" events, stating that data on events which happen so infrequently are equally as rare, therefore we con't accurately describe how frequently they really can happen, unless you have managed to obtain infinite data. Apparent knowledge of probability is porportional to the number of times an event has appeared in the data. As its appearence decreases, error margins increase. That's how "extremely low probability events" still manage to happen. What to do? The experts warn that, even though it is acceptable to use the theoretical equations when consequences are limited to acceptable outcomes, there is extreme danger inherent in using them when predicting probability of high consequence events. Use the equations to justify one shut-off valve for a single connection to your house??? Use the equations for the $5 black jack table (and you'll still lose 5% on the average BTW). But do NOT use them when you've got a whole city of people to protect. Make a risk-consequence matrix. Redbox the unacceptble results. Pick the acceptable mitigating solution. Basically my take is, if you can do anything to avoid unacceptable events, DO IT!. Making it impossible is just .... so much safer. Unfortunately it is easy to see only in hindsight.

 

[davefitz]

good luck with that.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick