Pipeline Batching (Line Reversal) Questions

[Bluckles]

We are currently in the process of reversing the flow of one of our hydrocarbon pipelines (from diluted bitumen to diluent). I'm looking into the various ways this could be accomplished.

The first, which seems the cleanest, is to use N2 to push the diluted bitumen out, run cleaning and drying pigs, and then fill with diluent.

However, I always want to investigate simply using a fluid to push the diluted bitumen out. Are there guidelines for determining the interface volume between two liquids? I have all properties of the pipeline and fluids.

I have discussed it with others in my company, but we're all new to a batching process this long (460 km), big (24") and between such different fluids. So while looking elsewhere I thought I'd post here and see if anyone is familiar with this or knew of any resources I should use.

My questions are:
- How can I calculate the volume of contaminated fluid (knowing pipeline and fluid properties, and things like pressure and flow rate) ?
- Would batching pigs reduce this interface volume? How can that be calculated?
- Should cleaning agents or any special cleaning pigs be used?

Any help or direction is appreciated.

 

[BigInch]

Do you really mean you will reverse the pipeline, filling it with diluent and pushing back the bitumen/diluent towards the wells (or mining fields)?

2nd question. Why N2? Why not just use diluent alone?

My bitumen/diluent pipeline was actually two pipelines, one bringing only diluent from a coastal refinery to the well field production plant, where we mixed it with bitumen heated it, and then feed it into the bitumen/diluent to go back to the coastal refinery area where some of the diluent was separated, refeed back to the pipeline to the well production facility and some was loaded into tankers. We in effect ran a recirculating system.

The way to change the fluid is first to get out the old (or finally do a) start-up/shut-down study. That should have already been accomplished, if you indeed have an operating bitumen pipeline at this time. If that hasn't been done yet, you need one right now. Since you can't (or probably shouldn't) let the bitumen/diluent cool off in the pipeline, in the start-up/shut-down study you should have defined various operating scenarios based on variable time intervals between shutdown, cooling and reaching certain ambient temperatures during the shutdown time, and the various higher viscosities during that cool off period, and then define or assure yourselves of how the line could be restarted using various arrangements of your available booster and mainline pumps and standby units.

The typical easy way to shutdown a bitumen/diluent line for a long term repair or a turn-around of the diluent extraction refinery is to just keep increasing the diluent concentration until the bitumen/diluent concentration reaches 100% diluent, so I'm not sure why you don't just do that, but OK, maybe you've got reasons.

When increasing the diluent concentration, it is usually possible to reduce the flowrate as the diluent concentration rises and maintain more or less constant discharge pressures off the mainline pumps at lower levels, but if you just want to start slugging it with diluent, that will (probably) also work, but you need to be sure that pressures higher than the pipeline allowable pressure are not developed in the process. With an interface in the pipeline, you will not have the benefit of uniformly decreasing viscosity along the pipeline, as you will have the downstram segment of bitumen/diluent with high viscosity and the upstream side with diluent at low viscosity.

If you are indeed talking about REVERSING FLOW and backing up diluent/bitumen to the pipeline inlet and finishing with 100% diluent, that just might change the entire hydraulic conditions enough to make it totally unfeasable. You could be talking about a lower density diluent pushing the higher density bitumen/diluent mixture which could create high pressures at the inlet or other various sections of pipe containing 100% diluent and also depending on the elevation profile and the flow head losses. If this is the case, basically you are looking at doing an entire detailed pipeline review to find all pressures at all points along the line in the reverse condition.

Batching pigs probably won't help and might do more mixing than what would be created with liquid/liquid interface alone. (There are formulas and nomographs to determine interface generation volumes- Austin-Palfry, to name one). The use of batching pigs is not accepted by many companies who have found the trouble of injecting and retreiving the pigs is not worth the effort only to generate higher interface volumes. Basically, if you can maintain turbulent flow regimes, interface generation will be minimized, but that is also without great differences in viscosity between the two fluids. For that reason, the method I mentioned previously of increasing or reducing concentration until the desired concentration is reached is the preferred method.

Let me know if you need more detailed help. You can find my e-mail and phone numbers under contacts section in the "Myspace" web pages, or just Google "virtualpipeline", if you can't access Myspace from your company web.

http://virtualpipeline.spaces.msn.com

 

[Bluckles]

My original post probably wasn't clear enough. Yes, the diluted bitumen line will actually be reversed, purging the currrent dilbit that is in the line back towards the field. I wish I could be absolutely specific about the details, but we're in the very preliminary stages of looking at this. So I don't know if current variable pumps are simply going to be physically reversed or if new pumps will be brought in. No valves need to be reversed.

The system is essentially as you described it: Diluent is piped from Z to A, added to the bitumen at A, and then this dilueted bitumen is pumped back to Z, and separated. A new diluted bitumen line is necessary to increase capacity, and is being installed as we speak. Once it's completed, this new line will replace the old DilBit line. The old DilBit line will then become a diluent line, and the old diluent line will become obsolete. As a result, the old line which used to bring dilbit from A to Z, will now flow in reverse, from Z to A, bringing diluent.

I'm not going to be in charge of running the actual shutdown/startup process for the change; I am simply looking at the timeline and cost associated with this reversal.

They are considering nitrogen, rather than simply purging the old dilbit with the new diluent, because of the amount of diluent that will be "contaminated" by residual dilbit (Yes, even though they are going to be mixed eventually, I think the worry is ending up with consistantly different mixtured and having to deal with that). As for a lighter fluid pushing a heavier, I'm working on the batching hydraulics now. The current model I wrote only accepts a single fluid, so it obviously needs to be upgraded. The new dilbit pumps shut off for low pressure at around 500 kPa, so obviously I need to check if diluent pushing dilbit can achieve that.

So I'm back at my original problems (now that the situation is hopefully cleared up). Assuming it's hydraulically feasible, how do I find how much diluent would be contaminated by the current diluent/diutmen mixture? I've heard from several people (and now you) that batching pigs will not help in this case, so I suppose that is out, even though the fluids will be around 50 cSt versus 0.5 cSt.

So the "standard" solution is to simply start pumping diluent, let everything mix, and then once 100% diluent comes out the other side, assume problem solved? You wouldn't worry about residual bitumen contaminating the diluent (especially because the company were are piping for has requirements on the purity of diluent they receive)?

 

[BigInch]

I don't see the QC problem.

If you increase diluent concentration in the line to be reversed, you will still have control of the diluent mixture when you begin feeding to the downstream company again, just as you do now. As you are receiving increasing diluent concentration in the reversed pipeline and sending it to the mixing tank, simply decrease the amount of diluent coming from the old diluent line to compensate, as you blend in the bitumen coming from the field as always. Then start sending that same as always %blend down the new pipeline.

If you slug the old diluent/bitumen line with 100% diluent, just put any quantity of interface you generate into a temporary holding tank (a tank truck or two). I doubt you will make more than 1000 bbls under the worst circumstances Although I didn't check the numbers, I'd probably estimate around 1/2 bbl/mile of pipeline traveled, but in any case, so what? You can blend any quantity of interface with 100% bitumen to the right concentration before you send it down the new pipeline. No?

I think all you need to do is to check the hydraulics of the reversed line for higher pressures, although I don't understand how you happen to have enough pumps (to reverse flow) at the present outlet of the diluent/bitumen line. Just lucky or a very favorable elevation profile? Perhaps you will need some tmp units there?

http://virtualpipeline.spaces.msn.com

 

[Bluckles]

Hi, I'm still having some trouble with this. I'm looking for a simple order of magnitude model to calculate the mixing interface for this problem.

I tried using a model from a paper by Rachid, Araujo and Baptista, but it was FAR too involved for the estimation we need. To the point where I couldn't get anything out of it because the formulas were too complex and the explanation of terms was weak. Basically I need to know if we're looking at having 50m of mixing or 1km.

I have the fluid properties, pump data, elevation profile. I've even got a full hydraulic spreadsheet I wrote. But nothing that handles mixing concentrations.

Any suggestions? The main model I've heard referenced is the Austin-Palfrey model. Is this going to be overly involved or is it a simple calculation I will be able to do essentially on paper? If not, can anyone suggest a simple model that gives a mixing volume estimate that I would be able to find in a paper (purchased or otherwise) online?

Thanks so much.