Need Help w/Fundamentals of Tank Vapor Recovery System

[KernOily]

This might be the wrong forum, but... here goes. Please re-post if you feel this better belongs elsewhere.

Guys I need a shove in the right direction on the fundamentals of tank vapor recovery systems. The system in question consists of fourteen API 650 atmospheric storage tanks, ranging in size from 1500 bbl to 7500 bbl, manifolded together into a header with the endpoint of the header going to the suction of a compressor. The gas is then compressed and sent off to be mixed with utility fuel gas and subsequently burned as fuel.

The job of the system is to (1) collect tank vapors that off-gas from the product in the tanks (14° API crude) since the tanks can't be vented to atmosphere due to permit conditions (can you guess where I live yet? ) (2) control the tanks’ pressure to prevent damage to tanks due to overpressure. This system has no makeup gas because the product off-gasses fast enough to prevent the pressure from dropping as the tank levels are dropped when pumping out. Gas composition is about 0.6 mole fraction CO2, 0.3 CH4, 0.14 H2S, with the remainder being C2+ (mostly C6+). The gas is saturated.

The system currently does not work well. There are frequently leaks from the tanks (aka “releases” - a Very Bad Thing – can you say: “Notice of Violation”?) from the P-V roof hatches, the compressor controls are not stable, i.e. it cycles back and forth from recycle to full load, flange leaks, etc.

The task before me is to redesign the system. Scope of the project is (1) Replace the entire CS piping system (laterals to each tank and the header) with SS due to corrosion – this is the easy part; (2) “fix” the controls so the system is stable, reliable, and prevents unwanted releases. There is the tough part.

Here is where I am stuck. I am not knowledgeable on how these things are controlled and properly set up. Is it better to use one large P-V control valve/regulator, somewhere in the header, with large diameter piping to reduce pressure drop, or should each tank have its own P-V valve/regulator? Wouldn't this make the piping system smaller at the expense of more valves to maintain? And wouldn't this method provide faster system response? Since I don't have makeup gas I can use a separate tank hatch on each tank for vacuum protection.

I think the compressor control issue is just that somebody has messed with the setpoints and the speed of the PID loops, but I need to check into that.

Any help/suggestions/pointed barbs are welcome. If you could point me to a fundamental reference, that would help a lot. I got the Fisher info off the web and that helped a little but I need something more fundamental.

Thanks!!!! Pete


Thanks!
Pete

 

[pmover]

i'm familiar with 2 different system types: 1 utilizes ejectors and the other recip compressors in recovering vapor gas. in both cases, the recovered gas is burned in equipment at facility. yes, the control system is essential and must be functional with how system is piped.

at the very least, continuously monitor tank pressure (pressure xmtr) as that is primary process variable and compression equipment will operate based on that input signal. for example, as tank pressure rises, control system responds to increase vapor gas recovery flow rates and vice versa. other monitoring should be processes downstream of compressor.

is existing control system pneumatic or analog/digital?

regarding the p-v valve/regulator system, i'm not familiar with such as system, but perhaps further explanation from you would be helpful.

fyi, a control system will function properly ONLY if sensor, controller, and control element are functioning properly. when one component fails or deviates from design, then the control system as a whole deviates/fails from design.

good luck!
-pmover

 

[Montemayor]

Pete:

I believe the starting point in your project is the mechanical design of your “atmospheric” tanks. The whole issue of how you control (and when) centers on the MAWP (Max. Allowable Working Pressure) and MAWV (Max. Allowable Working Vacuum) of your tanks. This is the issue that makes this a tougher-than-usual problem. The underpinnings of the issue is the most important of all underpinnings: Safety Hazards.

The primary and most important of the design features on this application is the stuff that makes for process engineering nightmares: the probability of sucking in atmospheric air into a combustible gas stream prior to compressing it. This always raises the specter of an explosive and hazardous mixture. No process engineer will allow this to proceed without redundant instrumentation to mitigate that event. Your mention of the “frequent leaks” in the tanks just makes it even a worse possibility. First and foremost on your Scope of Work should be the confirmation of the mechanical soundness and integrity of all the tanks in question. A thorough mechanical rating should be done on all the tanks before anything else. This includes actual existing physical condition. For example, are the tanks all anchored? Without this being done, I wouldn’t touch the project or go any further with it. Coupled with this mechanical analysis is the positive identification of all the MAWPs and MAWVs involved. If you have API 650 tanks as you say, there is the possibility that you may have up to 2.5 or possibly 15 psig MAWP available in your tanks. The higher the MAWP, the more secure and reliable the instrumentation will work due to wider dead bands between set points in the vapor capacity controls required of the blower or compressor employed. With this much combustible vapor involved, I would go for as high an MAWP and as low a MAWV as I could justify. In my opinion, the concern for under pressure (vacuum) should be just as critical as for over pressure. I just finished a crude oil recovery project last year and I had to go through the same, old-fashioned dilemma: crude oil is traditionally stored in minimal pressure API tanks vented to atmosphere and suddenly today, producers want to control the vapors in such tanks when they have never invested money in having the tanks designed for pressure (max. 15 psig -API 620).

The next point of design concern has to do with eliminating all potential sparking or ignition events – type of compressor or blower, static electricity, etc. Many of the vapor collection systems that I’ve seen or heard of blowing up have been handling much less powerful flammables than the one you have. And the causes for the explosions have been as simple as friction inside the rotating lobes of blowers.

From what little I know of your configuration and scope, I would not employ a control valve or any throttling device on the header vapor flow. I would allow it to float with the capacity of the blower or compressor being the control point. The moment you start to throttle a flow stream that is being induced by a machine, you are potentially creating a source of low pressure (and possibly a vacuum) and you certainly don’t want a vacuum. I would stay away from having to break a vacuum with air into such a mixture as you have – and I would use it only as a last resort to avoid a disastrous over pressure. I realize that you must be prepared to break a potential vacuum should it suddenly appear, but working with a wide dead band allows you to stay within the confines of a positive pressure and maintain positive control over vacuum. That is why I would use one central header collection with laterals – all sized for minimal pressure drop. This collection header would float in pressure with the set point for capacity control on the compressor or blower.

I assume that with so many tanks and the location being a tank farm, the site is probably remote and, as such, one that is not conducive for a reciprocating compressor – or for any compressor as a matter of fact. Unless you’ve a large flow pressure requirement for subsequent burning of the fuel gases, I don’t visualize a need for a compressor of any sort. This, in my opinion is an application for a centrifugal blower – primarily because of the need to have no metal contact (no heat or sparks), little or no wear or maintenance (remote location), relative small pressure head required (a few psigs) and quick, large capacity when needed.

I cannot overstress the importance of obtaining liberal and ample deadbands on your header pressure setting in order to have a dependable and controllable system. A lot of serious accidents have occurred in systems much more benign than the one you have and most of them have been centered on trying to control between a inches of water column pressure. I can’t comment further because you haven’t really furnished a complete scope of work. But I hope what little I’ve said is of some help in orienting you to the important and critical features to design around.

Best of Luck.


Art Montemayor
Spring, TX