Please help with a gas pipeline problem we are having at my company. There is something stuck in a pipeline but the operators have no clue where. My question is, can I determine where the obstruction is in the system by recording the rise in pressure over time from a fixed measurement point upstream of the block? We can monitor static pressure, differential pressure, flow rate in mcf/d, temp and total volume sent past the sales meter. Ideally I would like to bring an equation to my masters to get a few crumbs from the table. Thanks.
>gas pipeline problem we are having at my company
>operators/technicians have no clue
>ask 4chan for advice
you guys are running a real bang up operation
>>8021482
kek
try ideal gas law PV=nRT
if you know P, n and T, you can find V, which basically gives you the length of tube you need.
If this doesn't work as a first approximation, you'll need to use van der waals equation I guess
https://en.wikipedia.org/wiki/Van_der_Waals_equation
same thing, find V.
>>8021482
Have you tried knocking on the side of the pipe until it sounds different somewhere?
Are you located in the US?
>>8021512
I was about to write this.
>>8021512
>knocking on thousands of kilometers of pipes
mathematicians
>>8021482
It should be very obvious bfrom a sudden change of pressure between compressor stations. You should not need to do any math, just read the instrumentation.
Once youve narrowed it down to between two points, use a pig you fucking savages.
>>8021482
>sci
>homework and work related problems
Thanks for the tips and keks. There are like thirteen pigs in the line because the part of Montana where the pipeline is located is know for inbreeding.
>>8021650
what
what does that mean?
There are 13 pigs in the pipe?
Yep. They just kept shipping them .
>>8021657
P.I.G.
Pipeline
Inspection
Gage
Device you push through a pipeline to determine damage/clear debris.
>>8021659
Oh.
pls dont hurt the pigs or force them thru the pipe
thermal camera. there will be a notable cold spot in the pipe where the constriction is.
you must be referring to some fancy space age pigs, the ones I've seen are like foam cylinders they use to push liquids downstream. Also, these guys are cheap, I don't think they have pipe laser robots.
>>8021672
In the natural gas industry much more sophisticated models are used, with calipers and electrochemical sensors so as to pinpoint the exact spot where deformation occurs. This is a huge benefit ass you can dig a 10 foot trench to repair pipe instead of a 2000 foot one.
I have seen models produced by GE and Rosen, which can have 3 foot diameters and be in trains 15 feet long. These are usually owned by a pigging company & in a near-constant state of renting out to various contractors.
I forgot to mention the pipeline is buried. Otherwise you could almost see through the poly pipe to find the plug.
>>8021685
that looks wild. The pipe is like 6-8" diameter depending on the location.
>>8021688
Im genuinely curious, this isnt me trying to insult you or your company. How does your company stay afloat if they dont know how to take care of a routine problem like this? How long have they been around?
>>8021755
do you know how many start up oil and gas companies there are? any faggot with a well has his own "energy management firm".
>determine the speed of sound through the whatever is in the pipe
>send some kind of shockwave/ loud noise through the medium
>record how long it takes for the echo to come back
>divide by two and multiply by the speed of sound
That's how far away the blockage is.
>>8021814
brb, patent office.
>>8021814
p.s.: give me a job
[email protected]
>>8021482
find out where obstruction is by measured values - no chance.
but when obstruction triggers relevant p/V changes check up major repair locations of last years and these dilation bends after it.
>>8021822
do you not have an internship this summer anon?
>>8021828
Nope. I guess I'll be doing research in computational structural optimization over the summer instead. It's super informal and unpaid, though. :/
>>8021829
you doing topology optimization? What sort?
>>8021829
you didn't apply to many places did you
>>8021831
Discrete combinatoric. I'm dividing the problem into a bunch of cells that can either be solid or void, then finding the optimal combination of cells. I came up with this pretty cool algorithm to do it too. It's a black box method that operates on sequences of bits, so I can apply it to all sorts of stuff.
>>8021833
My GPA is kind of bad because I'm more interested in working on personal projects that use more advanced material than what we learn in class.
>>8021838
>My GPA is kind of bad
first, thats not what i asked. second, nobody cares. they care about what you can do, not what your GPA is.
>>8021814
this optimism is worth a new meme.
>>8021841
>they care about what you can do
Well, apparently not. Hopefully they will after I publish some papers this summer, assuming everything goes according to plan.
>>8021842
Of course there will be noise, but that's nothing some audio analysis can't solve. Just run the data through this.
http://www.audacityteam.org/
>>8021851
you are going to half to take into account the damping effect of flanges and bends. you'd have to accurately model the whole pipe to know what your signal should look like.
>>8021852
Gradual bends shouldn't be too much of a problem for a longitudinal wave, though elbows might. Maybe begin the analysis at a relatively straight run of pipe. There shouldn't be that many sharp corners in a long range pipeline.
>>8021857
point is, you are still going to have to pull out schematics and probably do some full scale analysis. it won't be as easy as going out, slapping a vibration sensor on it, and whacking it with a hammer.
>>8021851
in case company tries to solve prob one their own:
how big is change in pressure loss(we actually dont now in this thread).
estimate possible size of obstruction.
finding out if/which device fits to problem and if this event justifies invest.
order device.
waiting for device to arrive and of course pay for it.
built in device.
adjust device.
make tests, check reflected freq... ,... .
hoping measurement really helps figuring out where obstruction is. reality is a mad opponent.
in case company decides for solicitation:
doesnt matter which measurement method is used as long as firm solves problem(sound, PIG, mexican, magic).
thats why this is optimism. im sure OPs chief is looking for a solution now.
>>8021835
That's neat. Good luck beating conventional gradient based topology optimization. See "On the usefulness of non-gradient approaches in topology optimization," Ole Sigmund.
>>It's a black box method that operates on sequences of bits
if it's really blackbox, and not a genetic algorithm, try to see if you can beat genetic algorithms.
Sounds like it might be interesting to apply to linkages and truss structures though.
Pete engr student here. What is your gas flash comp or gravity? Also the operating conditions of the pipeline?
>>8021882
Well what I'm working on is more of a global search technique, so it's not quite comparable to gradient descent or evolutionary algorithms. Thanks to some clever mapping, I map the entire enumerated search space onto a number line, such that similar model combinations are located near each other, then form a statistical model for where in the search space the best solutions occur. I'm still debugging it, but it ought to achieve very good results very quickly.
>>8021549
But it's rigorous
>>8021549
>sucking thousands of kilometres of pipes
Engineers
>>8021482
Send out a sound pulse and measure time to reflection.
If the obstruction is practically a total blockage you can use the fill rate and pressure buildup to determine volume between compressor and block. Divide by cross sectional area and voila: you get the distance.
Ultrasonic transducers on the wall can measure propagation speed also in the fluid. This will change with fluid pressure. Measure at various places along the length (binary chopping, say) and locate where there is a transition.
>>8023409
> use the fill rate and pressure buildup to determine volume
I think this is what I am looking for. Its a natural gas pipeline with a total blockage.They wont buy shit so, can I get the information I need to solve this from a sales meter that calculates million cubic feet/day? Without buying new equipment I can tell what the static (line) pressure, differential pressure (through an orifice plate), flow rate, total volume, specific gravity (in this case .9789) and temp. I think the pipe diameter stays pretty constant in the underground sections but there are some 45's when it comes above ground for the pig launchers/blowdowns.
Can you do my work for me, Anon? What equation would I use to calculate flow rate and pressure buildup to determine volume?
>>8021755
I am equally astounded. I can't into thermodynamics because I am a born untermensch (I still tryhard) but I am leagues beyond the people running the region in simple reasoning. It's amazing to me they don't kill or disable more of their workers.
>>8021549
1. A gas company has an obstruction in a pipeline. Find the lateral surface area of the pipe in order to tell how much pipe the crew will need to knock on. Put your answer in terms of pi. (15 points)
>>8021822
>p.s.: give me a job
Oh, and if (against all odds!) this is patentable, you have the right to be named as an inventor on the application. It does not matter if you are intern or employed, paid or unpaid.
In some jurisdictions you have also a right to a "reasonable compensation" for the invention.
>>8021482
Are there pressure gauges stationed at different spots in the pipe, for regular checks/maintenance? Assuming you have the ability to close off gas flow through the pipe, you can read the pressure change simultaneously at different positions along the pipeline. If there's a significant obstruction like you say there is, the latent pressure in the pipe will equalize at significantly different rates on either end of the blockage. Find the two pressure gauges which seem to contain the length of pipe with the blockage, then you can use more exact methods like sonic testing.
Don't worry about the math, just pour a few gallons of this stuff down there and start pumping. I'll invoice you tomorrow when you're up and running.
Can you reverse the flow? That might pull it out the other end.
>>8021508
dude, you have no idea what you're talking about, do you?
>>8025087
Well, I have lots of ideas that are patentable, but seeing as how I have a net worth of -$20k I don't think I can afford to file. So right now, work > patents.
>>8025769
In some jurisdictions you might be obliged to inform your employer if you have made a patentable invention and then it is up to them to determine of they wish to patent it, keep it a trade secret or let it go. There are a few ifs and buts here but that is the simple explanation.
And keep in mind that oil tech does a lot of patenting and have a fair bit of patent conflicts. Since journalists consider oil to be "dirty" and "unglamorous" they prefer to talk about Apple patent conflicts. Still, the conflicts are simmering there.
>>8025118
This is what I would have thought was the standard system for pipelines. Should just start with whatever gauge along the flow is the first to show a reading that deviates from the norm and the normal one before it. The blockage would have to be between those two points. I would also have thought that you would be able to close such a section of the pipeline off and then service it.
>>8021482
>Please help with a gas pipeline problem I am having in my introductory fluid mechanics course.
Fixed that for you, dipshit.
