Formation Water Processing 1

[engo911192]

Hey there!
I have to design a formation water process, in which the oil content should be reduced from 1000ppm to below 200. Since it seems to be impossible to measure the droplet size, the residence time is based on experimental results (additional question: is it possible or reasonable to calculate the residence time without knowing the exact droplet size distribution?!?).
A pre-requirement is the use of tanks for the separation process. My problem now is, that I don't really know how to design the tank in order to guarantee the residence time not only theoretical but also in practice. Is there a correlation factor between theoretical residence time and the time I can achieve in such a tank in practise?!? How exactly do I design such a tank, how about fixtures and fittings?!?

I'd appreciate any helpful tip.

Thanks in advance.

 

[zdas04]

There is a question you have to ask in front of the residence time question--is this a batch or continuous process? In other words are you trucking the fluids or pipelining them?

The device most often used for oil/water separation at the ppm level is called a "gun barrel" for reasons that seem to be lost in the mists of time. A properly designed gun barrel remains full to the oil and water overflow lines at all times. If a liter of fluid comes into the process than a liter of fluid must immediately leave the process. This implies that the inflow should be continuous and relatively slow (I like to keep it to about 1 gun barrel volume inflow in 24 hours).

If you are trucking to a gun barrel then you hit it with 80 bbl of cold fluid in a 10-15 minute period and blast a mixture of oil and water out of both the oil side and the gas side. I never design truck-unloading capability directly into a gun barrel. If I'm trucking water I put a HEATED pre-treat tank that is about twice the size of the gun barrel in front of it. Incoming fluid is discharged into the bottom of the tank and warm fluid leaves out the top through a choke set to my desired gun barrel flow rate. The big trick is to design the pre-treat tank with enough room above the overflow to accommodate anticipated truck arrival rated (i.e., if you expect one 80 bbl truck per hour and you are doing 400 bbl/day then you need to use a 1000 bbl pre-treat tank with the overfow set with about 200 bbl above the outflow, more trucks/hour requires a bigger pre-treat tank or multiple pre-treat tanks).

It is very important to set the pre-treat tank on a pedestal that puts the outflow line above the top of the gun barrel so that you don't have to pump the fluid (pumping creates a lot of shear forces that make the separation significantly more difficult).

I've designed systems that were anticipated to be 50% trucked and 50% piped and ran the piped water into the pre-treat tank along with the trucked water. The result was outstanding as long as I kept heat on the pre-treat tank and the gun barrel.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[engo911192]

Thanks for your detailed response.
It's a continuous process... The main part of oil and gas are sepparated in a first step already. There are approx. 1000m³/h od produced water left for which I have to reduce the oil content from 1000ppm to below 200ppm. Regarding the separation behavior the oil seems to be not to bad. (Static) Experiments showed that phase separation takes around 3 hours to yield the necessary oil content in the water phase.
Since I'd like to use a continuous process the question for me now is how to correlate the separation time of the static experiments (3 hours) to the tank volume in continuous operation. I guess it's not going to be 1000m³/h, 3 hours --> resulting tank (or filling) volume 3000m³ in order to reach an effective residence time of 3 hours?!? Or is it possible... due to using a gun barrel or any other flow-optimizing installations?!?

Thanks again for any advices...

 

[zdas04]

Your quantities are really big. The design that I use would require nineteen 1,000 bbl gun barrels in parallel. You are talking about serious money at that point. I've used powered centrifugal clarifiers (mostly from DeLaval) to get ppm water out of oil, I'd talk to someone like that to see if they have a version with the dam set differently to get ppm oil out of water.

30 years ago you would have dumped all this water into a pond and skimmed the oil off periodically. In North America the Migratory Bird Act put an end to that [pretty sloppy] activity. Other parts of the world have similar restrictions to open bodies of water with oil slicks.

I don't think you are going to be successful with tanks with this volume so I'd be looking really hard at throwing energy at it.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[shahyar]

The oil content from 200 to 8000 mg/L can be separated from water by gravity, or API separator. You can design a long channel or skim tank with internal baffles.
If oil content is less than 200 mg/L, gravity separator doesn't work. The available options, depends on the flowrate, are: IGF, Oil Removal filter, or adsorption by clay.
If you choose to use a non-baffled skim tank(conventional tank) the real oil separation residence time is less than one fifth of the tank residence time.
Regarding droplet size distribution, actually it is very rare that we can have access to droplet size distribution. So, you are not alone.
Hope this help

 

[MortenA]

The 200 ppm is not a problem in itself - although i think that an API separator would normally be for much smaller quantities?

The emission requirements in the north sea is less than 40 ppm and lower values will soon be introduced. Hydro cyclones eg can lower the oil content to 40 ppm. But on the top of my head i cant remember what the expected inlet oil contamination is for a oily water hydro cyclone. a 1000 ppm dose not sound totally off.

Best regards
Morten

 

[TD2K]

Just on what zdas04 posted, I was on an assignment in Indonesia with a client that was looking at using centrifuges to break some stable oil in water emulsions. The test results were very promising from what I remember but I don't recall the numbers, I'm pretty sure it was an Alfa Laval centrifuge.

 

[zdas04]

The Alpha Laval (if it is the one I'm thinking of) started life processing milk and then adapted to a bunch of other stuff. Nice thing about that technology is that it spins a lot slower (less shear) than the DeLaval stuff that started out as a device to remove water from oil. It has been a few decades since I had all of these factiods straight so I might have gotten them muddled some. I'm pretty sure that we had a 10,000 rpm DeLaval and a 400 rpm Alpha something, but it has been too long to be sure if it was Alpha Laval.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[engo911192]

thanks again for all the comments on my questions.
I know that tanks are - from the engineers point of view - not compulsory the method of choice in this case. Unforunately the operators insist on using just separation tanks, if it works somehow. Trying hard to bring them around...
I've one more question regarding the oil/water separation resp. produced water treatment... is there any literature on this topic - going beyond basic principles?!? (maybe going more into details regarding design and newest separation technologies)

 

[zdas04]

If there are any books I can't find them (and I've looked). SPE has had a couple of workshops on this subject lately, but papers are not mandatory at Workshops like they are at conferences so I'm not sure if anyone put papers in the SPE e-Library, but that is where I'd start.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[DuncanM]

Wish I'd seen this thread when it was fresh - hopefully you will be notified of this reply, OP.

If you have sufficient pressure drop available then a hydrocyclone vessel (or two) would be ideal for your flowrates; and would exceed your 200ppm target with ease (as long as there's no odd chemistry).

For example: We have a 30" vessel with 60 hydrocyclones fitted (rental equipment) which can treat up to 300m3/hr of water. Depending on oil & water chemistry, in an ideal case it could treat water with 3% oil down to sub-30ppm discharge.

We also have equipment suitable for sizing oil droplets; which is usually used alongside a test hydrocyclone to prove the technology with the installation's fluids.

If you're looking for some basics on current technology the search phrase "PWSM SLB" will help. You may also find contact details that way. If you've any specific questions I can do my best to answer here also.

The major installed base in the North Sea is hydrocyclones, CFU and IGFU. There are other technologies in development but I don't believe they are public as yet.