Refinery flares have been used for many years to serve as safety devices at refineries. In case of over-pressurization of any process vessels, the flare acts as a pressure relief system. However, refineries also use flares to dispose of "excess" gas generated at various process units. It is this "routine" flaring that is the subject of the project.
Long ago, US EPA decided that smoking flares were bad and developed rules to restrict this smoke. In response, flare manufacturers developed "assisted flares" designed to reduce smoking. Smoke is a result of incomplete combustion, so the design of these assisted flares allows the introduction of more air into the combustion zone which improves combustion and reduces smoking. There are air-assisted flares and steam-assisted flares. Steam assisted flare are most common and have been the focus of this program.
In a steam-assisted flare, air-enriched steam is injected into the vent gas at the tip of the flare. This serves two purposes. First, the educted air improves combustion and second, the steam acts to create turbulence to mix the air and vent gas more thoroughly. This also improves combustion. The problem is that some flare operators seem to believe that if a little steam is good, more steam is better. Steam and air obviously dilute the vent gas and at some point if the gas become too dilute combustion becomes less efficient or stops altogether. This situation is known as "over steaming".
In addition to flare smoking, US EPA is also concerned about Volatile Organic Compound (VOC) and toxic emissions from flares. The point of flare combustion is to destroy these compounds so they are not emitted to the environment. It has long been assumed (based on some studies in the early 1980's) that flare combustion is 98% efficient. That is, 98% of the organic compounds in the vent gas are destroyed during combustion. However, with the advent of assisted flares and the problem of over steaming, US EPA and others began to question this 98% assumption. The concern is that in over steaming situations, combustion efficiency declines and more VOCs and air toxics are released to the environment
Now it's easy to tell if a flare is smoking -- just use your eyeballs. But how do you tell if a flare is not operating efficiently? You can measure the organic compounds in the vent gas going into a flare, but how do you know how much of this remains after combustion? These flare tips are typically elevated 100-200 feet above the ground and the flame is flopping around in the wind. Over the years, many attempts have been made to analyze the post-combustion flare gas. Small scale flares (an inch or two in diameter) were tested in a lab but the results don't really scale up to actual operating sizes (24-72 inches in diameter). Sampling probes were lowered by helicopter or raised by cranes into the flare plume to try to extract a sample for analysis but wind and other logistical issues made this very complicated and expensive.
Recently, remote sensing technologies have become available that potentially simplify the analysis of flare plumes. All hot gases give off an infrared signature that can be analyzed to determine the chemical compounds in the gas. Using this technique means that the flare plume can be measured from the ground. Extracting a physical sample is not required. If this technique can be shown to work with flare plumes, we would have a way to measure the combustion efficiency of flares and determine the effects of over steaming. One vendor of this remote sensing technology (IMACC from Round Rock, Texas) has developed a technology called Passive Fourier Transform Infrared Spectroscopy (PFTIR) that has potential to fill this role.
Beginning in 2009, Clean Air has worked with Marathon Petroleum, IMACC, US EPA, and others to develop the technology and procedures to measure flare combustion efficiency using PFTIR. While US EPA is interested in the end result -- flare combustion efficiency -- when we started down this road, we did not know whether PFTIR technology would actually work. So our research approach has been both to develop and validate the test method WHILE SIMULTANEOUSLY determining the effect of over steaming on flare combustion efficiency. Ideally it would have been nice to have a validated test method prior to actually using it to collect real data but life is not always so simple.
To date we have conducted flare tests at two Marathon (MPC) refineries - Texas City and Detroit and also at a Flint Hills Resources (FHR) chemical plant. We currently have two additional tests scheduled with Marathon -- one at Garyville, Louisiana and one at Cattletsburg, Kentucky. Right now, the Clean Air/IMACC team is the only group in the country (I guess in the world, actually) that can do this testing. So we have been receiving inquiries from other refiners and chemical plants regarding future testing.
We work very closely with US EPA on this project. To date, the work has been driven by Consent Decrees issued by the Department of Justice. However, EPA is using the data we generate to craft new flare regulations that will affect the operation of all flares across the country. They will also use the testing approach that have been developing to conduct this testing over the past two years. The test method cannot yet be considered "final" at this point but we are close. Clean Air is currently leading a group at ASTM to formalize this method. The group consists of refiners, flare manufactures, EPA and state regulators and ourselves. We expect to have a formal test method complete sometime in 2012.
So our path ahead at this point is to plan and execute the upcoming tests, continue to work with EPA and others to develop the test method, talk to other refiners and chemical plants about what we are doing, and look for ways to extend the PFTIR technology into other areas.
Quite a lot done, quite a lot to do.
Passive FTIR flare test