6 replies to this topic

[TCA Consulting]

Greetings,

I am a mechanical engineer working on troubleshooting an existing tank blanketing system. This is my first exposure to a tank blanketing system, so I am far from an expert. I have done some research into blanketing systems and probably know just enough to get into trouble at this point! In addition to various valve manufacturer guidelines, I have read Mr. Ostand's article, "Tank Blanketing: A Complete Guide..."

The system consists of 3 crude oil tanks blanketed with natural gas. Either Tank A is used (primary - 13,000 bbl capacity) or Tanks B and C are used if Tank A is out of service (3,000 bbl each). There are a pair of Shand and Jurs model 94261 valves used to pad and depad all three tanks. The valves are located at ground level. The connection points for the gas and sensing lines at the top of Tank A are approximately 35 feet above the valves. Tanks B and C are shorter, about 25 feet tall, but I do not know their location/elevation relative to Tank A. (Site visit scheduled for Sept 21).

The sensing lines from the two valves are connected together at ground level and a common sensing line is run to the top of Tank A. A second sensing line tees off at ground level and runs to Tanks B and C. The pad/depad lines are run in the same manner - a common line to Tank A and a common line to Tanks B and C. There is also a separate common pressure equalizing line connecting all three tanks. There are knock out tanks and blowdown valves at the bottom of the vertical piping to prevent condensation accumulation in the sensing lines.

Valve setpoints are +0.3" wc for the pad valve and -0.3"wc for the depad valve.The operators have measured an operating pressure range at the bottom of the sensing pipe of 0.2"wc to 0.8"wc. The maximum range measured is 0.0"wc to 1.8"wc. Tank MAWP unknown at this time, but will be determined. The tanks are equipped with pressure/vacuum safety valves. Setpoints are unknown at this time. Gas pressure inlet to pad valve is 5 psig. The discharge of the depad valve goes to a suction scrubber for the vapor recovery compressor. The compressor runs mostly unloaded with a -10"wc to -13"wc (vacuum) suction pressure.

I am visiting the site Thursday, Sept 21 to verify actual conditions, setpoints, pipe sizes, etc. I know a picture is worth a thousand words, so once I verify actual conditions I can post a P&ID of the system, if that would help.

I will be calculating gas flow requirements per API 2000 and verifying the pad/depad valve sizing (more questions may follow). I will also be selecting new pressure/vacuum safety valves for Tank A.

Based on my research so far, this system seems to have a large potential for pressure control problems. I am most concerned about the common pad/depad and sensing lines running vertically to the top of the tanks. All references I have seen indicate the pad and depad valves should (or must) be located on top of the tank with minimal piping and should be piped separately.

Any input would be appreciated.

Thanks,

Tom

[djack77494]

Tom,
I think you have cause for concern. Like you, I would greatly prefer very short (and dedicated) sensing lines. Any bit of condensation that might form in the sensing lines would nullify your control and could lead to an unsafe condition. You might want to observe whether or not the tank conservation vents appear to be at all active during your site visit. Though it is common practice, I have a pretty strong aversion to the use of fuel gas for blanketting large atmospheric storage tanks for several reasons. The simplest to explain is the potential for admitting air into such a system as you have described. I'd welcome the comments of others on this topic.

Getting closer to the original problem, you can appreciate that even quite small amounts of condensate that could form in the sensing lines will contribute significantly to the sensed pressure. I would be very inclined to use 1/2" tubing to minimize this effect, though I've often seen 3/8" used. I had one experience measuring gas flowrates using an annubar where condensate wrecked havoc with the accuracy. you could face a similar situation, depending on the particulars.
Doug

[proinwv]

Tom, soryy I did not reply sooner.

First, and foremost, the blanket valves should never, never, be below the sensing point. As pointed out the smallest amount of liquid accumulation in the sensing line can cause the pressure at the sensing diaphragm of the blanket valve to sense a "false" pressure and not open in response to a drop in tank pressure below setpoint. So the tank blanket valves need be mounted on top of the tank, above the sensing connection. This may not be always popular, but for your valve to function properly, you need to do this. Knock out tanks and blowdown valves at the bottom of the vertical piping will not "prevent condensation accumulation in the sensing lines". What they might do is allow one to drain the accumulation if they are opened before they fill and put a "false pressure" on the valve.

Second, be aware that the valve servicing two tanks is not responding to the pressure in either tank but to some intermediate pressure.

Third, the sensing lines for the two different blanket valves should not be connected together for similar reasons.

Setpoints of +0.3 and -0.3 in. wc are too close together. There will likely be interaction between the pad and depad functions and the depad should be set at a pressure above the pad valve. Your statement is the opposite. Further a pad setting this low is an invitation to ingest air thru the vacuum vent.

Bluntly, this system needs a serious review and I do wish you the best.

PAUL

[TCA Consulting]

Many thanks for the input.

First off, I reversed the setpoints in my original post - the pad valve is set to open at -0.3" wc and the depad valve is set to open at +0.3" wc. Sorry for the confusion on that one.

Of course you are correct, Paul - the knockout tanks do not prevent the condensation - only give it a place to collect for a while without causing a problem. Poor choice of wording on my part.

The site visit last week was informative. It is a bigger can of worms that I originally thought. The incoming product is a mix of oil, water and gas. The gas is drawn off in a fourth vessel ahead of the storage tanks into the common equalizing line that is connected to all three storage tanks. The incoming gas quantity is enough that the pad valve does not open (confirmed by upstream turbine meter). This system is in a constant venting mode.

The pressure at the ground level sensing line does indeed vary between 0.2" wc and 1.0" wc. The surprising thing is how rapidly it varies - within seconds between the high and low of the range. This is happening while the depad valve constantly hunts for a satisfactory position. There is a constant vacuum of 10" wc to 18" wc downstream of the depad valve. This is the suction line for an unloaded compressor that regularly overheats due to lack of load. I realize that this manometer does not represent the actual tank vapor space pressure. Until revisions are made, I encouraged the operator to put a pressure transmitter on the top of the tank to get a feel for what is actually happening in the tank vapor space.

Tanks B and C are piped such that only oil enters Tank C. It is possible that water could be drawn off Tank B (requiring pad gas flow) while a thermal heating requires Tank C to be vented at the same time. With the common pad-depad line, it is not possible for the system to address this scenario. These two tanks should have separate blanketing valves. One of the conservation vents for Tank B seemed to be constantly venting gas.

I am sorting through the flow rates and process connections. I'm sure I will be recommending major revisions to this system. I expect I will post additional questions later. Thanks again for the input.

[proinwv]

Tom,

Well you sure have a project on your hands but it seems that you have a good understanding of the issues. I believe that you got my paper from my website and if it helped it pleases me greatly.

If you feel that you need additional help on this project, feel free to contact me thru my website.

Of course, your posts here are also most welcome, but it seems that this problem exceeds what can be accomplished on a forum posting.

Too often, blanketing and venting are not given the attention they require. In your case, it is fortunate that you have identified these problems and are in a position to influence their solutions.

Best wishes
PAUL

[TCA Consulting]

Paul,

I did indeed get your paper from your website and it did help me considerably, especially since this is my first tank blanketing project.

Yes, this project is a challenging one and has grown in complexity to the point that it is difficult to address in this forum. My intent is to simplify this system so it is easier to operate and more reliable. Among other things, I will be recommending the pad/depad valves be relocated and piped separately.

Thanks for your input.

[proinwv]

Tom,

Sounds good and let us know if you need anything else.

You might, after you complete the work, post your experiences on this forum for others to read.

Best wishes
PAUL