Flare monitoring by infrared camera

[ThijsM]

Hello,

Due to constant problems with the addition of steam to the flare, we are looking for possibilities to imporve the current situation.
At the moment, we have a normal camera in the control room aimed to the flare. When we have a release of off gas, the operators open the steam further by adjusting the steam to the flare from the control room. This has got not the highest prio of the operators in case of emergencies (understandable, but the flare is our most visible process equipment for our neighbours) and results sometimes in inadequate steam supply or overdosing of steam.
There is a flow meter in the line available that can be coupled to the steam and put in a certain ratio. This ratio however is a problem because we have 2 plant on this flare with total different compositions of off gas.
A good idea (at least, I think it is a good idea, hope to hear from you) is to use a IR camera, aimed at the flame of the flare and that camera does give input to the steam valves to open or close if necessary.

My question to you is if you are having such a system in use and what experience you have with this. How is the behaviour in different weather conditions like fog or heavy rain? What is your vendor and specifications of the system and what are the installation costs of the package. Any other input is more then welcome!

 

[EngAbdelkader]

Hello ThijsM,

I'm not an expert in the matter but as the process is based on combustion and as in all the petrochemical plants we carry continuous monitoring of the flare emissions, why don't you think to make a steam flow control regulated by the limit values of the combustion product (NOx and CO emissions) and take a look at this page:

http://www.holanengineering.com/sitebuildercontent/sitebuilderfiles/pm-sm.pdf

The expert of flares here is "Flareman" so if you don't see him here take a look at his personal page:

http://www.geocities.com/flareman_xs/

you'll find there a way to contact him directly if needed (email)

Good luck!
kader

 

[ThijsM]

Kader,

Thanks for your reply, the site of Holan Engineering was one of my sources to ask the questions about your experience with respect to IR steam control of the flare.
But I would like to hear the opinion of people who have worked with it in pratice, instead of a "fantastic, great!" selling story. Some information on possible startup problems, weather influences and so on would be interesting.

Thanks anyway!

Thijs

 

[Flareman]

ThijsM
Been away a while but just saw your question.
I'll try not to give you a fantastic selling story.

There are several manufacturers who provide I/R based flare monitoring technology. As far as I am aware, all the flare suppliers sell someone else's instrumentation. I do have a preference but I'm not advertizing for them so you'll have to drop me an email on that.

The general experience is so-so to good. The basis is that a smoky flame puts out more radiant heat than a comparable clean flame, a bigger flame more than a smaller flame so if the smokiness or the flowrate increase the heat input at the receptor increases and calls for more steam.

The biggest problem is operator (and instrument engineer) misunderstanding of the operation. The system has to work as a direct feedback loop responding with more steam as the heat goes up. This is different from normal control which uses an offset signal from a setpoint to control the valve. There is NO SETPOINT for smokiness. The "tweak" is on the signal gain between the input and the output signals so that the ratio stays in the correct range. A linear trim valve usually works best. You set it up by manual trial and error in the first couple of weeks of operation by comparing the input and output signals (manually set for the correct appearance) and draw a line through the results which then shows you where the best operating conditions will be.
Sometimes the signal ratio line does not pass through the graph origin and you need a slight permanent offset to bottom load the valve at a minimum flow. Systems set up like PID offset control work well enough day-to-day when conditions are roughly the same but need constant resetting or tweaking of this offset when the flows or gases change.

A mild drawback with the feedback loop method (technically called "scan and bias" is that there is always a response time and you need to make sure that the instuments are properly set up with the correct damping and reset times (site set up)to ensure that you don't create a nasty positive feedback. I once saw a system with a thermal response time which was exactly half of the steam valve response time which caused the flare to puff black smoke and white steam like smoke signals on a 5 second cycle. Funny - but environmentally disastrous!!
Depending on the location and sizing of the steam valve and line to the flare, you may choose to incorporate a differential monitor on the input signal to advance the valve according to the rate of change of the input rather than the absolute value.

So, assuming that you get that part of it right, there are "blind" versions which mount directly on the stack (basically thermocouples with a fancy top) and optical versions which need a clear shot at the flame. The "blind" system receptors have to be mounted on the flare which makes them inaccessible except during shutdown (the instruments are in the control room). The optical systems are more accessible but have to be sighted correctly so that they never pick up the sun in the lens (pretty obviously) and the signal strength needs to be within the capability of the thermopile, which is a limiting factor on how far away you can be. Also, the visual field of the optical device is another factor because any part of the view which is not filled with a flame doesn't register a signal and that affects the sensitivity of the instrument. On the other hand, if you set a visual picture which doesn't represent most of the operation, you will blunt the signal by missing part of the flame when it gets larger or blows in the other wind direction. Also, forget a reliable visual picture if you have a plant which is always foggy or has cooling tower discharge drifting into the line of view.
The blind systems don't have those problems.

My personal experience has been that (in the main) people who use these systems tend to like them because of the forget-it principle (even some of those who set them up incorrectly). One user I know added "blind" systems to all of his flares. They may still have to adjust the setting a little, a few minutes into the relief, but it doesn't require primary attention. Unfortunately a lot of operators set it for a transparent flame condition to avoid having to readjust too frequently and that leads to periodic flame oputages. But you clearly already have that problem with no instrumentation.

So that's it. Not a sales pitch in sight!!
I hope I gave you enough to think about.

David