Innovations in Flare Stack Design for Low Calorific Value Waste Gases

Paper from the AFRC 2018 conference titled Innovations in Flare Stack Design for Low Calorific Value Waste Gases

Oftentimes flare stacks serve as thermal oxidizers in refinery and petrochemical plants as a way to ensure that hydrocarbons and other volatile organic compounds do not enter the atmosphere. Use of flares, however, is more widespread serving other industries as well. Several upstream steps in the steel making process (for example) produce combustible, but not always economically useful, by-product gases (blast furnace gas, coke oven gas, basic oxygen furnace gas, etc.). Because these gases bear Carbon Monoxide, and sometimes variable amounts of hydrogen, it is usually necessary to oxidize them, often through a flare stack. Because these by-product gases contain fewer combustibles by volume, they typically fall outside of the EPA limits for calorific value (40 CFR §60.18). These gases have unique challenges for such types of flare systems. Not only are there common challenges of exposure to the elements, supply of adequate oxygen, etc., there are additional difficulties associated with reliable ignition, proper mixing, and flame stability.; Innovative designs address some of these issues, leading to better and more reliable performance. Following a brief technical discussion of these innovative designs, (such as improved ignition systems, and mixing strategies), this paper will explore a case study of a domestic installation and will take a more in-depth look at details of a retrofit on an existing hood with a modern ignition system and other state-of-the-art design features.

Innovations in Flare Stack Design for Low Calorific Value Waste Gases David Schalles - VP Technical Services [e-mail] Michael Cochran - Marketing Engineer [e-mail] Abstract Oftentimes flare stacks serve as thermal oxidizers in refinery and petrochemical plants as a way to ensure that hydrocarbons and other volatile organic compounds do not enter the atmosphere. Use of flares, however, is more widespread serving other industries as well. Several upstream steps in the steel making process (for example) produce combustible, but not always economically useful, by-product gases (blast furnace gas, coke oven gas, basic oxygen furnace gas, etc.). Because these gases bear Carbon Monoxide, and sometimes variable amounts of hydrogen, it is usually necessary to oxidize them, often through a flare stack. Because these by-product gases contain fewer combustibles by volume, they typically fall outside of the EPA limits for calorific value (40 CFR §60.18). These gases have unique challenges for such types of flare systems. Not only are there common challenges of exposure to the elements, supply of adequate oxygen, etc., there are additional difficulties associated with reliable ignition, proper mixing, and flame stability. Innovative designs address some of these issues, leading to better and more reliable performance. Following a brief technical discussion of these innovative designs, (such as improved ignition systems, and mixing strategies), this paper will explore a case study of a domestic installation and will take a more in-depth look at details of a retrofit on an existing hood with a modern ignition system and other stateof-the-art design features. Introduction The basic premise of a flare stack is to remove unwanted constituents out of a waste gas stream through thermal oxidation (burning). Many flare stacks are in use for the petrochemical industries where these undesirable elements are hydrocarbons or other volatile organic compounds (VOC). These constituents are readily combustible, and generally produce a more robust and stable flame. Flares are in use, though, in many industries. For example, some upstream steps of steel making process produce certain process gases bearing carbon monoxide (CO) and trace amounts of hydrogen (H2). Because CO is toxic, it is imperative that any CO in the waste gas be properly oxidized (formed into CO2) instead of escaping into the atmosphere. Background As mentioned above, any of several processes can produce waste gases requiring flaring. Two of the most common (in the steel making industry) are coke oven gas (COG) and blast furnace gas (BFG). In some integrated steel mills, it is possible to make use of these gases as a fuel for other processes. However, in the case of an excess of waste gas, or an interruption of the downstream process using the fuel, it is necessary to have the option to flare, often the flare is designed to be an element in the gas pressure control system.Tables 1-2 show the relative composition and other thermochemical of these 1 two common process gases. (Table 2 also shows the thermochemical properties of natural gas- assumed as pure methane-for comparative purposes.) Constituent CO H2 CO2 N2 CH4 BFG [% by volume] 20-25 3-5 15-20 50-62 0 "Clean" COG [% by volume] 5-10 45-55 0-5 5-10 25-35 Table 1 - Chemical Constituents for sample waste gases 3 Heating Value [kcal/Nm (Btu/scf - HHV)] Theoretical Flame Temperature [°C(°F)] Specific Gravity Air to fuel ratio Waste gas to fuel ratio Blast Furnace Gas 700-850 (75-90) Coke Oven Gas 4200-5100 (450-550) Natural Gas 8500-10,400 (900-1100) 1245 (2275) 1990 (3600) 1950 (3550) 1.0-1.05 ~0.6:1 ~1.5:1 0.35-0.45 ~5:1 ~5.5:1 0.55-0.65 ~10:1 ~11:1 Table 2 - Thermochemical properties of some sample fuels Unique Challenges Flare stacks for such situations have been common for several decades. There are, however, some unique aspects to the fuels burned that require special engineering consideration. Not only are there concerns about the rigors of the weather and other elements, the very nature of the flare stack means that the burner is located as much as 80m (250 ft) in the air where maintenance and repair are particularly tricky. Because the fuel itself has a large proportion of inert constituents, it can be difficult to burn, and special design considerations to ensure proper mixing of the air and fuel are necessary. Ignition of the flare's pilot system can occur remotely (through a flame front generator) or with a direct spark ignitor at the top of the stack requires a considering and balancing ease of maintenance, with safety and reliability. Finally, flame monitoring, from such a remote location at the top of a tall flare stack, requires some careful planning. This paper will take up each of these design considerations in the following sections. (Defining terms: The terminology can become somewhat muddled in any discussion about these types of systems. Although some people refer to the entire system as the "flare stack", meaning to include the burner, the equipment and the actual "chimney", the term properly applies only to the actual chimney itself. The burner and associated equipment are the "flare tip", "flare hood" or "flare burner". These three terms are essentially interchangeable, and for this paper, flare hood will be the favored term.) Remember that the gases burned in these systems is sufficiently dilute with inert constituents as to fall outside of EPA jurisdiction for high calorific value flare hoods (40 CFR §60.18). Furthermore, the diluting 2 constituents (carbon dioxide, nitrogen, etc.) sufficiently lower flame temperature such that at ground level, thermal radiation is usually not so intense as to be harmful. A final caution about these process gases is that under some circumstances, it is possible for the fuel source to become supersaturated with water. In those cases, if the water condenses (due, for example, to a temperature change), it can cause mechanical and chemical attack on the flare stack hoods. Operating Principles The main purpose of any burner is to bring fuel, an oxidizer (in this case, atmospheric air) and an ignition source together in the right proportions and in the right location to ensure that combustion occurs. For this application, it is crucial to eliminate any existing carbon monoxide. The challenges mentioned above necessitate unique considerations for proper design of the flare hood system. Air supply is effected through two main mechanisms. First, the flow of waste gas out of the top of the stack will entrain ambient air into the waste gas due to the velocity disparity between the flowing gas and quiescent air. In addition, once the flare is burning, the temperature difference will create a natural draft that will draw additional air into the hood. These two influences will contribute sufficient air to maintain combustion. Concepts normally associated with industrial burner performance such as thermal efficiency, flame shape, or process temperature are not really of a concern in this application. The purpose of the flare stack burner is to provide an ignition and flame stabilizing platform only. It is impractical to provide the combustion air manually (i.e. with a blower). Instead, relying on a combination of natural draft and prevailing wind is a workable solution to ensure complete combustion. In order to enhance the mixing of air and fuel, the flare hood has vanes around the stack that agitate the fuel in order to enhance mixing. These vanes are fixed in position determined in the field based on prevailing winds and other site-specific factors. In Figure 1 the fixed vanes are clearly visible. Figure 2 shows a flare hood before and after installing such vanes. Obviously, the flame becomes more reliable with these stirring vanes, even in adverse wind conditions. In fact, for a hood with vanes, a wind load actually enhances the mixing. Without the vanes, an adverse wind load could cause the flame to deteriorate. 3 Figure 1 - Triple Headed Flare stack with fixed mixing vanes installed (a) (b) Figure 2 - Flare hood without (a) and with (b) mixing vanes for better combustion. (Note the stronger, more compact flame.) Because the flame temperature is sufficiently low, in many cases, it is sufficient to construct the flare hood from a stainless steel alloy. Depending on the specifics of the operation (such as fuel composition 4 and rate), though, the flare hood might require refractory lining to protect its structural integrity. Sometimes, a wind deflector shield (Figure xxx) might also be necessary to protect the flame from extremely strong winds. Turndown The basic system can easily operate through a 10:1 turndown (i.e. operate with only 10% of the intended fuel flow). If the fuel supply is sufficiently unreliable such that supply would fall below 10%, it is possible to design a multi headed system with two or more burners sitting atop the stack and connected to it with a manifold. A valve (or valves) can effectively isolate one burner, thus extending the turndown to approximately 20:1 (for two heads), or even further for more burner heads. Pilot System for Ignition of Main Flame Several factors make ignition of a flare stack particularly tricky. There is, of course, the fact that the combustion is occurring as much as 80 m (250ft) elevation while exposed to the elements and weather. Furthermore, the low calorific value of the flare gas can make it difficult to light. It is typical to supply the flare stack with continuous pilots, (generally on natural gas or another hydrocarbon) to ensure combustion even in the face of irregular BFG flow/pressure. Two types of pilots for a flare stack exist: inspirated and forced air. Inspirated pilots require approximately 70 kPa (10 PSIG) natural gas pressure in order to function properly. Forced air pilots are an option which can improve reliability over a wider range of atmospheric conditions. Furthermore, they may allow the use of lower cost fuels and reduce pilot fuel consumption. It's crucial to engineer the system properly to ensure that the blower for the combustion air is sufficient in volume, and that it can reach the proper elevation. These pilots often have a small protective covering to guard against wind impingement. While it is not generally a concern to maintain good flame characteristics on the main flame (the only purpose of the flare burner is to eliminate volatiles and/or carbon monoxide), it is crucial that the pilots have reliable flame to ensure combustion in the main flame, and to ensure that there is a good signal from the pilot flare sensor for flame safety (see below). Note that for a flare stack with sufficiently low flow of flare gas, and with elevated concentration of molecular hydrogen, it is possible to ignite the main flame with direct spark ignition. Such cases are relatively rare, and require careful analysis to ensure safe performance and operation. Ignition System for the Pilot Burner After selecting the proper pilot system, it is still necessary to provide an ignition source for the pilot itself. (These pilot burners can range up to 0.5 Gcal/hr (2 MMBtu/hr). Direct spark ignition of the main flame is not generally a viable option.) There are two main ways of igniting the pilot. The older way, which is falling out of favor because technology improvements are making alternate ways more reliable, is to generate a small premix flame at ground elevation that is allowed to travel through a pipe up to the main pilot. This method is known as a flame front generator (colloquially as the "fireball" method) and requires extra precautions necessary for any premix combustion system. The increasingly popular way is to furnish the pilot with a high energy spark ignitor that sits at the top of the flare and will ignite the pilot directly. Equipment quality has improved sufficiently over the past few years to make this option viable, and it is even possible to retrofit this style of igniter in place of the fireball approach. For larger size flare stacks (more volume) it is suggested that there be two (2) or more pilot assemblies located around the periphery of the hood in order to guarantee that the BFG (or COG) will come in 5 contact with the pilot. Otherwise, the possibility exists that due to wind, or to other drafts, the flow of fuel could drift sufficiently far from the pilot as not to ignite. Flame Safety Flame monitoring is typically by use of thermocouples inserted through the pilot burner port (or ports in the case of multiple pilots on the larger systems). For BFG, there is no direct monitoring of the main flame, primarily because the flame shape and intensity are not sufficiently consistent to monitor. As long as the pilot flame is lit, it is sufficiently sized and properly located that the main flame will also be lit. Generally, there are two (2) thermocouples in each ignition burner in order to accommodate the vagaries of wind patterns, thus ensuring that the flame will always be in contact with at least one (1) thermocouple for positive flame signal. A newer approach which may be considered is via thermal imaging cameras. These cameras can be located at ground level for easier maintenance. By proper selection and filtering, they could be applied even for BFG flames, which are often invisible to the unaided eye. Case Studies Case Study I -Ohio In order to come into compliance with government mandated emissions specifications (particularly in regards to CO), a US customer required the best available technology. Design requirements were for a 60" (1500mm) diameter stack passing 123,000 ft3/min (3,300 Nm3/min) of BFG, wet saturated at 120°F (49°C). Because of the large physical size of the stack, the burner hood required four (4) pilot assemblies, each lit by a flame front generator. After a successful commissioning, which included , the flare hood system was easily able to meet the agency set limit of 50 ppm CO at a downwind distance of 100 ft (30m) and 150 ft (45 m) from the stack. Generally, CO monitoring, conducted in accordance with the agency consent agreement indicated CO emissions of 30 ppm. Case Study II -Indiana. Although BFG flares are probably the most common for integrated steel mills, any of several process offgases can require flaring. One such process produced a process gas rich in CO. Not only did the composition vary (in terms of CO concentration) but also the amount or flow varied as well. Because of these inconsistencies, and the elevated flame velocity with increasing CO concentraion, it was necessary to provide a flame arrestor in the stack itself to ensure that the flame did not travel back into the system. Otherwise, standard design of a single hood with an inspirated pilot and a high energy spark igniter was sufficiently flexible to work with this unusual fuel. Case Study III -Illinois Originally installed about fifty years ago, a flare stack at a US integrated steel mill for BFG used a "fireball" ignition system for lighting the pilot burner. In order to increase reliability, the customer chose to upgrade to a modern high energy spark igniter. The stack, (shown in Figure 3) is a dual hood design, with two (2) igniter pilots on each hood. As discussed above, these design features allow for a huge range of waste gas flows, while still ensuring proper combustion. 6 Figure 3 - Dual headed flare hood with upgraded pilot system The scope of this project was for a pre-piped and pre-wired pilot fuel train with all safety interlocks and devices, a stand-alone control panel, and a complete ignition system of two (2) high gas pressure inspirated pilot burners with local direct spark ignition on each of two (2) hoods on the stack (for a total of four (4) pilot burners. Although the original hoods were almost fifty (50) years old they were of sufficiently rugged design that the only changes required for the actual burner (aside from refractory replacement) was to install deflector shields around the pilots for better flame stability. Commissioning and start-up were uneventful, and the improved system has performed reliably in several different weather and BFG flow conditions. Conclusions With proper design of both the hood and ignition system, accounting for such factors as fuel composition and flow, and stack geometry, use of a flare hood is a safe and effective way to oxidize combustible components from process gases in the steel making industry. Recent advances in ignition technology have allowed for more reliable and safer ignition. 7