Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in

Explosion at liquefied natural gas plant in Freeport Texas 7

Status
Not open for further replies.
Replies continue below

Recommended for you

Sort by date Sort by votes
Title 49 Part 193 has words that should result in a requirement similar to OSHA Process Safety Management standard. This should have resulted in a rigorous discussion about safety valve isolation administrative controls, and requirements in the plant operating manual.
It will be interesting to see how the investigation develops.
 
Upvote 0 Downvote
The report refers to vacuum insulated pipe as follows:
The 18-inch vacuum insulated pipe consists of an inner and an outer stainless-steel pipe with covering shroud. The inner pipe carries the cryogenic liquid. It is wrapped with multiple layers of super insulation made of alternating layers of a heat barrier and a nonconductive spacer. This creates a space between the two pipes that is “pumped down” using a vacuum pump to create a static vacuum shield. The vacuum shield protects cryogenic liquid from heat loss due to conduction, convection, and radiation.
Click to expand...

So an inner insulated pipe which contains the liquid inside, and an outer pipe, and a vacuum space between them which is pumped down by a vacuum pump.

Are there alarms when the vacuum is not sufficient to maintain insulation?

If the relief valve were not isolated, would it be connected to the inner liquid chamber? (rather than the intermediate vacuum chamber)

By the way if I use 18" outer diameter to calculate the volume of a 300' cyclinder, I get pi*300ft*(1.5/2)^2=530 ft^3 or (with 600:1 volume ratio) 318,000 SCF. But the report states 120,000 SCF leaked which is only 38% of that 318,000scf. I realize some of that volume is occupied by the intermediate vacuum chamber, but I don't think it would be 62% of the volume (leaving only 38% for the liquid), would it? Is it reasonable to conclude the pipe was initially only partially filled with liquid?


=====================================
(2B)+(2B)' ?
 
Upvote 0 Downvote
The pipe contained a mixture of gas and liquid. As the temperature rises the liquid boils, gasifies and raises the pressure, in this case too much.

The relief has to be connected to the inner pipe. Everything is there. Nothing should normally be in the annular space. You would want to limit pressure there, which would also limit the pressure entering the annular space, if it did happen to find a way in there.

I doubt that there is any alarm on loss of vacuum. (maybe) That's what the relief valve is, in a sense.

This is a stupid mistake that caused a lot of damage. Super lucky nobody died.


A black swan to a turkey is a white swan to the butcher ... and to Boeing.
 
Upvote 0 Downvote
NG Phase diagram
They probably were somewhere in the mid to lower central mixed phase region.

Screenshot_20220705-220123_Brave_e4jipb.jpg




A black swan to a turkey is a white swan to the butcher ... and to Boeing.
 
Upvote 0 Downvote
I would expect the vacuum space to have rupture discs. Once you loose vacuum and have an overpressure in the vacuum space, a relief valve that reseats does not help anything, as any repair of leaks involving the inner pipe and other internals involves cutting the vacuum containing outer pipe open.
 
Upvote 0 Downvote
Meaning a disk, or relief valve connected to the annular space (?).

Burst disks are do messy with inflamables. Never used one.

A black swan to a turkey is a white swan to the butcher ... and to Boeing.
 
Upvote 0 Downvote
Not having practical experience with cryogenic flammables (other then propane) my opinions might not pass 1503-44's evaluation. The idea about using rupture discs on the vacuum space, comes from the fact that it is really hard to get relief valve seats completely leak proof as need to maintain higher levels of vacuumed. Of course even if the plan is for the relief on the vacuum space to never operate, it still needs to be aimed in a direction that would be away from important equipment and occupied spaces. I would be surprised if this relief was taken to a flair stack, due to the hopefully low likelihood of it needing to operate.

One other thought. In this environment how are relief valves protected from getting stuck with ice from condensation? One side being exposed to possibly cryogenic conditions, and the other to hot humid coastal air.
 
Upvote 0 Downvote
I think that there probably is no vacuum protection, alarm, relief or burst disks attached to the annular space. The best way to keep a vacuum is to weld every joint and not to have any connections to it. Least that's how we do it on encased, insulated, offshore pipelines where the annular space is not a vacuum and contains only an air gap, or insulation, either of which must remain dry. The annular space and casing pipe is sealed between "flange" fittings at each pipe joint. The flange there being a flange, but with a weld neck on each side. That may also be the design strategy applied here, where the vacuum is maintained joint to joint, rather than in a continuous annular space from beginning to end of the pipeline itself. I think there is an LNG import terminal in Peru that might be doing that, brining the LNG in from a body, rather than loading arms along side a dock. I do not know the specifics of LNG pipe design. I have designed an ammonia pipeline (-29C) with insulation and an HDPE jacket pipe that was enclosing a continuous annular space, so there are various ways the cat can be skinned. I just don't know what was done in this plant.

The relief valve in question must have been connected to the LNG pipe. That's the pipe that must have over-pressure protection (according to all pipe design codes I have worked with).

A loss of vacuum would be noted by an unexplained increase in the temperature profile, gasification of the LNG and subsequent pressure rise, then relief valve activation at over-pressure, if it wasn't isolated, just as this system was supposed to work. An independent alarm on vacuum loss might be nice, but doubtfully specifically required by code.

Natural gas pipelines can also freeze during rapid blowdown. The Joule-Thomson effect reduces the gas temperature at the outlet, condensing and freezing the valve. The blowdown rate is limited to avoid freezing. Surely the relief valve for LNG must be pretty heavily insulated.

A black swan to a turkey is a white swan to the butcher ... and to Boeing.
 
Upvote 0 Downvote
I was thinking about the long periods where the relief valve just sits doing nothing. Ensuring that ice can not form inside the relief valve while it is watching an inspec system (hopefully it's entire service life) seems like a tall order.

As stated in most codes, "This Code is not a Design Guide". As we all have found in our practice, we need to go beyond the applicable codes to build safe systems.

PSM should be part of the design process, as an aid in designing out hazards. Administrative controls while often necessary, are not sufficient for safety.
Screenshot_from_2022-07-06_16-58-42_tls0ln.png

[URL unfurl="true" said:
http://websites.umich.edu/~safeche/swiss_cheese.html[/URL]]The main purpose of the Swiss Cheese model is to visualize how a hazard may be able to pass through the vulnerability of the many different measures in place.
Click to expand...
 
Upvote 0 Downvote
Oil & Gas Journal

Freeport LNG Development LP is targeting initial production from its 15-million tonne/year liquefaction plant on Quintana Island, Tex. in mid-December, subject to the meeting regulatory requirements.

As of Nov. 14, reconstruction work necessary to begin initial operations, including utilization of all three liquefaction trains, two LNG storage tanks and one dock, was about 90% complete, the company said in a release Nov. 18. All reconstruction work expected to be completed by end- November.

Proposed remedial work activities for a safe restart of initial operations have been submitted to the relevant regulatory agencies for review and approval, the company said.

Each of Freeport LNG's three liquefaction trains will be restarted separately. It is expected that about 2 bcfd of production will be achieved in January 2023. Full production utilizing both docks is expected to begin in March 2023.

Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."
 
Upvote 0 Downvote
No doubt they'll have a lot of eager customers in Europe. Maybe they'll be putting out full capacity before the end of winter (northern hemisphere)

=====================================
(2B)+(2B)' ?
 
Upvote 0 Downvote
nothing like losing an insurance claim to getting that plant up and running pronto.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick
 
Upvote 0 Downvote
As a California resident, boy am I glad that plant exploded last year. I can't imagine what our prices would be if we were competing with Europe.
 
Upvote 0 Downvote
Every time the US opens another export terminal, your gas price goes up.

Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

Murph 9000
  • Locked
  • Question
Natural gas explosion at the Realty Tower Building
2
Replies
20
Views
21
Murph 9000
Murph 9000
hacksaw
  • Locked
  • Question
Texas plant
Replies
4
Views
5
1503-44
1503-44
MJCronin
  • Locked
  • Question
Fatal explosion and fire at blending plant in Australia - October 12, 2023
Replies
0
Views
8
MJCronin
MJCronin
MJCronin
  • Locked
  • Question
Contractor killed in Robertson County, Texas power plant boiler explosion !!!! 3
Replies
6
Views
7
davefitz
davefitz
MJCronin
  • Locked
  • Question
"Boiler Tank" (sic) explodes in Mexico Tequilla Plant 9
2
Replies
38
Views
39
1503-44
1503-44
Back
Top