Development and Demonstration of a Fuel-Flexible Burner for Fired Heaters

Paper from the AFRC 2014 conference titled Development and Demonstration of a Fuel-Flexible Burner for Fired Heaters by J. Jamaluddin.

The US Department of Energy and the project team members are co-funding the development of a fuel-flexible combustion system for refineries and chemical plants. This technology will enable operation of fired heaters on fuels ranging from conventional gases to bio-gases and synthetic gases. The burner's unique flame stabilization mechanism will allow operations to continue safely if and when the heating value of the fuel goes through wide and rapid swings. A previous presentation (1) reported that the fuel-flexible burner was initially developed by modifying Zeeco's Freejet-style, ultra-low-NOx burner. Since then, a new tile (hereafter referred to as the "Trapped Vortex", or TV, tile) has been developed based on Computational Fluid Dynamics modeling. Testing of this burner in a test furnace at Zeeco demonstrated superior performance, both in terms of fuel-flexibility and flame stability. This paper will present data on the performance of the fuel-flexible burner employing the TV tile, along with a demonstration of flame stability as the heating value of the fuel is rapidly switched from high to low and back to high. Fuel blends corresponding to natural gas, refinery fuel gas, and a variety of bio- and waste-gases were tested. The heating value (LHV) of the fuels ranged from about 1100 Btu/scf to 88 Btu/scf. To accommodate this broad range of fuel characteristics, the burner was equipped with multiple fuel manifolds. Late last year, a set of the fuel-flexible burners were installed in one of the fired heaters in a Shell manufacturing plant. These burners have one fuel manifold, and are meant to combust fuels with heating values in the range of 440 Btu/scf to 1100 Btu/scf. The burners are operating well. Based on extensive tests performed in both horizontally and vertically fired test furnaces at Zeeco Test Facility, we are confident that stable flames will be maintained even when the heating value fluctuates rapidly and unexpectedly. No commercially-available burner can respond to such changes as effectively; and as such, the fuel-flexible burner should find wide acceptance by the refining and chemical community once its effectiveness is demonstrated at the Shell operating site over the next few months.

Development and Demonstration of a Fuel-Flexible Burner for Fired Heaters Jamal Jamaluddin1 Charles Benson2, Roberto Pellizzari2 Seth Marty3, Jonathon Barnes3, Rex Isaacs3 Joseph Renk4 AFRC Annual Meeting September 8 - 10, 2013 Hyatt Regency Hotel Houston, Texas 1Shell Global Solutions (US), Inc. 2etaPartners 3Zeeco 4US Department of Energy Abstract The US Department of Energy and the project team members co-funded the development of a fuel-flexible combustion system for refineries and chemical plants. This technology will enable operation of fired heaters on fuels ranging from conventional gases to bio-gases and synthetic gases. The burner's unique flame stabilization mechanism will allow operations to continue safely if and when the heating value of the fuel goes through wide and rapid swings. A previous presentation [Ref. 1] reported that the fuel-flexible burner was initially developed by modifying Zeeco's Freejet-style, ultra-low-NOx burner. Since then, a new tile has been developed based on Computational Fluid Dynamics modeling. Testing of this burner (hereafter referred to as the "Trapped Vortex Burner", or TVB) in test furnaces at Zeeco demonstrated excellent performance, both in terms of fuel-flexibility and flame stability. This paper presents data on the performance of the fuel-flexible Trapped Vortex Burner, along with a demonstration of flame stability as the heating value of the fuel is rapidly switched from high to low, and back to high. Fuel blends corresponding to natural gas, refinery fuel gas, and a variety of simulated bio- and waste-gases were tested. The heating value (LHV) of the fuels ranged from about 1100 to 88 Btu/scf. To accommodate this broad range of fuel characteristics, the burner was equipped with multiple fuel manifolds. A set of fuel-flexible burners were installed in a fired heater in one of Shell's manufacturing sites in the US. Each burner is equipped with a single fuel manifold, and can fire fuels with heating values in the range 440 - 1100 Btu/scf. The burners performed well over a 6-month demonstration period, and remain in commercial service at the Shell plant. We believe that the fuel-flexible burner will find wide acceptance by the refining and chemical community. 1. Introduction Shell, Zeeco, and etaPartners have developed a fuel-flexible combustion system for refineries and chemical plants. This technology will enable operation of fired heaters on fuels ranging from conventional gases to bio-gases and syngases. Deployment of this system could enable the reduction of greenhouse gas emissions associated with plant operations. The goal of this project was to develop and demonstrate a full-scale combustion system which will allow a broad range of gaseous fuels to be utilized safely, cost-effectively and efficiently while generating minimal emissions of criteria pollutants. The project was conducted in three phases, as follows.  The Concept Definition phase was completed in 2011. A design specification was prepared by the team. Then, the preliminary designs of two preferred burner configurations were completed. Next, a series of Computational Fluid Dynamics (CFD) simulations was conducted to refine the burner designs.  The Technology Development phase was completed in 2013. Activities focused on testing and optimization of two prototype combustion systems at Zeeco's test facility in Broken Arrow, OK. The first of these, a modification to Zeeco's Free-Jet burner, was described in a prior AFRC presentation [Ref. 1]. The current paper focuses on the development and demonstration of a fuel-flexible burner that incorporated a trapped vortex flame zone for enhanced stability.  The Technology Demonstration phase was completed in April of 2014. A 6-month field demonstration of the fuel-flexible burners, retrofitted to a fired heater, was conducted at a Shell operating site. Zeeco is now prepared to manufacture and market the project's burner technology. Flexicoker Off-gas* Refinery Fuel Gas* Natural Gas* CH4 2.0% 43.7% 95.4% C2 0.5% 14.6% 2.1% C3 7.7% 0.5% C4 3.3% C5 1.3% CO 22.5% H2 12.5% 27.5% O2 0.1% CO2 6.5% 0.4% 1.3% N2 56.0% 1.4% 0.8% HHV (Btu/scf) 142 1148 1018 SG 0.88 0.65 0.58 Wo 151 1426 1332 *Typical, gas compositions vary. Table 1. Conventional Fuels Conventional fuels of interest are shown in Table 1. The composition of refinery fuel gas (RFG) varies considerably, depending upon the refinery configuration and operating condition. This variability has led to the development of combustion systems for process heaters which can accommodate a moderately wide range of fuel gases. Natural gas typically has a much narrower range of compositions. Often natural gas is blended with refinery-generated gases to supply the balance of a plant's energy requirement. Natural gas can also serve as a dedicated fuel for a unit or an entire plant. Flexicoker off-gas is an example of a low- Btu gas which is currently used in certain refineries. Since this gas is significantly different from the range of RFG's and natural gas, it is fired in specially-designed burners. The Wobbe number is a parameter used in evaluating the interchangeability of gaseous fuels in combustion applications. It is defined as the fuel's higher heating value divided by the square root of its specific gravity. For incompressible flow through a fixed fuel orifice with constant fuel supply pressure, the energy flow rate (i.e., firing rate) is proportional to the Wobbe number. Included in Table 1 are the Wobbe numbers for the various fuel gas compositions. Alternative gaseous fuels of interest include biogas from organic matter digesters, including animal and agricultural wastes, waste water plants, and landfills; as well as syngases from gasification of biomass, municipal solid wastes, construction wastes, or refinery residuals such as tar, pitch and petroleum coke. Table 2 shows the range of compositions of potential alternative fuels from gasification or anaerobic digestion of various feed-stocks. Bio-gas* Landfill Gas* Biomass Syngas* Wood Syngas* Charcoal Syngas* CH4 56.0% 52.0% 3.0% 1.0% CO 20.0% 20.0% 28.0% H2 18.0% 18.0% 4.0% H2O 9.0% 0.0% CO2 36.0% 47.0% 8.0% 9.0% 2.0% N2 8.0% 1.0% 45.0% 50.0% 65.0% HHV (Btu/scf) 568 528 128 153 113 SG 0.93 1.01 0.82 0.84 0.94 Wo 588 525 141 167 117 *Typical, gas compositions vary. Table 2. Alternative Fuels A key consideration in designing the fuel-flexible burner is the broad range of Wobbe numbers that these fuels span: from 120-150 for syngases, 500-600 for biogases, 1300-1400 for natural gases and 1100-1500 for RFG's. A combustion system which could accommodate all of these fuels would be working over an order of magnitude variation in Wobbe number. 2. Development of the Trapped Vortex Burner 2.1 Computational Fluid Dynamics Modeling The general configuration of the Trapped Vortex Burner is illustrated conceptually in Figure 1. A new feature in this design is the flame stabilization zone that is created by diverting small portions of the combustion air and fuel gas into an annular chamber within the tile. Hot combustion products and reactive species generated by this zone are transported to the top of the tile where they mix and react with additional fuel and air in the secondary stabilization zone. These two zones work in tandem to enhance ignition and stabilization of the main flame. The initial prototype Trapped Vortex Burner was developed using Computational Fluid Dynamics modeling. Simulations were run with natural gas and syngas having Wobbe numbers of 1346 and 117, respectively. The TVB utilizes a trapped, annular vortex of reacting fuel and air to stabilize the flame (Figure 2). The geometry and location of the annular chamber that houses the vortex is engineered s0 that Figure 1. Illustration of Trapped Vortex Burner the circulating flammable gases in the vortex have sufficient residence time to ignite and react. The fuel and air streams entering this annular chamber, as well as the combustion product streams leaving the chamber, are configured to generate the desired circulating flow field. To provide a source of ignition for the main fuel stream as it mixes with the burner's core air, the vortex's combustion products are vented to the top of the burner tile through a series of slots. A portion of the hot gases also exits the vortex at its open side to provide a source of ignition at the tile's outer rim. As with its predecessor, the Free-Jet burner, the TVB accommodates a very broad a range of Wobbe number fuels by utilizing three rings of nested fuel injection manifolds to progressively increase the fuel injection flow area as the fuel's Wobbe number decreases. The inner manifold is active for all fuel types. Its injection ports are angled radially inwards so that a portion of this fuel, as well as the flue gas that it entrains, is directed into and drives the trapped vortex. The porting geometry supplying air from the windbox to the vortex chamber is fixed, so the air flow rate into the vortex is relatively constant regardless of fuel type. However, as the Wobbe number decreases, proportionally less fuel is fired from the inner injection manifold. Consequently, the equivalence ratio of the gas mixture within the vortex becomes leaner with decreasing Wobbe number. The strategy for achieving low NOx emissions becomes one of Figure 2. TVB - Velocity vectors colored by velocity magnitude on plane through fuel injector creating a rich, yet flammable mixture in the vortex when firing high Wobbe number fuels. This mixture shifts to a lean flammable mixture when firing low Wobbe number fuels. In addition, the burner's overall NOx emissions are minimized by burning only a small quantity of fuel in the vortex region. Figures 3 and 4 show the predicted temperature fields when firing high and low Wobbe number fuels, respectively. Since the initial CFD simulations indicated that the TVB design configuration would likely perform as desired, the decision was made to continue the development efforts experimentally. Figure 3. TVB - Contours of Temperature Figure 4. TVB - Contours of Temperature (F) for 1346 Wobbe Number Fuel (F) for 117 Wobbe Number Fuel 2.2 Burner Design and Fabrication Due to many design similarities, the TVB design was developed from a Zeeco Free-Jet platform. The design parameters were set by the operating conditions for the furnace selected by Shell for the field demonstration. These are provided in Table 3. Design Heat Release (MMBtu/hr) 5.00 Available Combustion Air Pressure Drop (in WC) 0.30 Combustion Air Temperature (°F) 100 Furnace Temperature (°F) 1450 Required Turndown 5:1 Table 3. Trapped Vortex Burner Design Basis In order to meet these design criteria, a GLSF-12 Free-Jet plenum and manifold was adapted to incorporate the Trapped Vortex tile. For the development testing, an air door was also fabricated to allow control of the amount of air passing through the Trapped Vortex holes. In addition, the tile was constructed from two pieces, thereby allowing different tile geometries to be evaluated efficiently. Also, in order to handle the wide variety of fuel gases, two additional staged fuel manifolds were utilized outside the primary GLSF-12 Free-Jet manifold. 2.3 Burner Testing Burner testing to evaluate the TVB design was performed during the fall of 2012 and spring of 2013. The burner was tested in both vertically-fired and horizontally-fired arrangements. During testing, several designs for the top tile piece were evaluated. The best performance was achieved using a slotted tile-top design as illustrated in Figure 5. Figure 5a. The slotted "Trapped Vortex" tile. Figure 5b. Top view of the TV tile. Tip drillings were also varied to achieve optimum fuel-to-air ratios in the trapped vortex stabilization region over a wide variation in fuel compositions. For each tip drilling revision, the amount of air allowed into the trapped vortex region was varied using a push-pull air door, shown in Figure 6. Figure 6. TVB Test Configuration The first round of testing on the TVB was performed in a vertically-fired, vertically-mounted configuration in Zeeco's Test Furnace No. 11 (Figure 7). This is a single-pass furnace having dimensions of 7 feet long by 5 feet wide by 45 feet tall. The furnace has 6 single-pass, water-tubes running from the top to the bottom of the heater to remove heat. To achieve the target furnace temperatures, the tubes were covered with 1-inch-thick ceramic fiber insulation from the floor to 6 feet above the floor. The remaining portions of the tubes were left bare. Emission samples were extracted at the base of the furnace stack below the stack damper. The firebox temperature was measured with a velocity thermocouple located about 14 feet above the furnace floor. The floor temperature was measured through the furnace door with a velocity thermocouple located about 1.5 feet above the furnace floor. Furnace draft was measured at the floor of the furnace. After sufficient testing had been performed to ensure proper performance on the range of fuels to be utilized during the field demonstration at Shell, the burner was moved to Test Furnace 1 for further optimization for bio-derived fuels. Furnace No. 1 (Figure 8) is a single-pass cabin style furnace that is about 37 feet long by 12 feet tall by 6.8 feet wide (inside of the tubes to inside of the tubes). The furnace has two sets of tube banks along the walls of the furnace. The west set (closest to the burner) has 9 horizontal, single-pass water tubes (4 on the south side and 5 on the north side) running from about 2 to 25 feet from the burner end of the heater. The tubes were left bare to maximize heat transfer. The east banks consist of 24 horizontal, single-pass water tubes (12 on each side) running from about 26 to 36 feet from the west end of the heater. These tubes were also left bare to maximize heat transfer. Flue gas samples for emissions were extracted at the base of the furnace stack below the stack damper. Flue gas temperature was also measured at that location. Firebox temperature was measured with a velocity thermocouple located 16 feet from the burner. Furnace draft was measured at the wall where the burner was mounted. Figure 7. Test Furnace No. 11 Figure 8. Test Furnace No.1 2.4. Burner Performance Test results for the optimized TVB configuration operating on natural gas and on a simulated bio-derived fuel are provided in Table 4. This bio-fuel represents one of the more challenging compositions with respect to flame stabilization, since the primary reactive specie is methane and the level of dilution with carbon dioxide is high. Test Point 1 2 FUEL GAS Natural Gas Bio-derived Natural Gas % 100.0 52.0 Carbon Dioxide % - 48.0 Lower heating value, Btu/scf 910 473 Wobbe Index 1323 517 FUEL GAS DATA Heat Release MMBTU/HR. 5.000 5.000 Inner Manifold Pressure PSIG 4.6 2.6 Inner Manifold Temperature F. 39 37 Middle Manifold Pressure PSIG 0.1 4.4 Middle Manifold Temperature F. 52 50 COMBUSTION AIR Ambient Air Temperature F. 42 40 Relative Humidity % 88 89 Barometric Pressure IN. Hg. 30.29 30.30 Furnace Draft IN. W.C. 0.31 0.31 Air Door Setting 3.75 4.50 T.V. Air Door Setting (in open) F/O F/O EMISSIONS DATA Oxygen % (Dry Basis) 2.9 3.1 CO PPMV 0.0 0.0 NOx PPMV 19.8 9.9 Firebox Temperature F. 1593 1607 Floor Temperature F. 1470 1485 Visible Flame Length Ft. 8 - 9 8 - 9 Visible Flame Width Ft. 3 - 4 3 - 4 Table 4. Representative Test Results For operation with natural gas, essentially all of the fuel was injected through the inner ring of tips. Due to the lower Wobbe number of the bio-fuel, both the inner and middle rings of tips were utilized. For each fuel, flames were established within the trapped vortex as shown in Figures 9 and 10. Hot products exiting the trapped vortex, as evidenced by the brighter color within the slots, supported stabilization of the main flame on the tile's upper surface. Air pollutant emission levels when operating with about 15 percent excess air were low for each fuel. The carbon monoxide concentrations were below 1 ppm. The NOx concentration for operation with natural gas was about 20 ppm. For the simulated bio-gas this level dropped to about 10 ppm. The visible flame envelopes for these fuels were similar. At the design firing rate of 5 MMBtu/hr the flame length was about 8.5 ft. and the diameter was about 3.5 ft. Figure 9. Natural Gas (Test Point 1) Figure 10. Simulated Bio-derived Gas (Test Point 2) 3. Technology Demonstration Three TVB's, each with a design firing capacity of 5 MMBtu/hr, were installed in a process heater in one of the Shell plants in October, 2013. The heater started up in early November, and has been in operation since. Figure 11 shows the burners in operation in the process heater. In spite of the tight spacing of the burners the flames have maintained their shape, with no visible interaction. Although the burners were tested with fuel gas heating values ranging an order of magnitude, the historical data at the plant show a range of between 400 and 1100 Btu/scf. Non-conventional gases are unlikely to be used in process heaters in the near future, but the sudden fluctuation in heating value has caused many flame-loss incidents within our industry. Such flame losses have the potential to cause heater explosions with severe consequences. The emphasis for this endeavor, from a plant perspective, was therefore on demonstrating that the burner can withstand rapid changes in fuel heating value. The tests at Zeeco demonstrated that the burner does maintain a stable flame when the fuel gas heating value is either decreased or increased by a factor of three instantaneously. The photos in Figure 12 and Figure 13 illustrate the ability of the burner to maintain stable flames as the fuel gas switches were performed. This Trapped Vortex Exit Slots demonstrated flame stability will contribute significantly to improving safety and reliability of burner operation in manufacturing sites. Figure 11 Trapped-Vortex Burners in operation in Shell process heater. Figure 12 Photos showing the effect of rapid fuel switch from natural gas to high-H2 fuel (before, during and after). Figure 13 Photos showing the effect of rapid fuel switch from high-H2 fuel to natural gas (before, during and after). The only known issue with the burner has been a crack in the top piece of the two-piece burner tile in one of the three burners, as shown in Figure 14. The likely cause was identified in discussions between the three parties, and as a remedial measure, the angled Zeeco pilots were replaced with the more conventional vertical pilot design. Figure 14 Crack in the top piece of one of the burner tiles. Figure 15 shows the cracked tile after 4 - 5 months of operation since the crack was detected. Its condition does not appear to have worsened, while the tile looks cooler since the pilot flame is not impinging on the tile where the crack had formed. Figure 15 Cracked tile with new up-fired pilot. 4. Conclusions The development tests and the field demonstration have shown that the trapped vortex version of the fuel-flexible burner is able to utilize fuels having an order-of-magnitude variation in Wobbe number while maintaining stable flames and generating very low levels of NOx and CO emissions. Rapid and wide changes in fuel heating value were accommodated without noticeable changes in flame stability. The burner passed all tests required for commercial application. The burners have been in operation at the designated Shell site for 9 months, and have been performing well. The fuel-flexible burner development, testing and field demonstration are facilitating the commercialization of a new generation of burners that will be suitable for the fuels of the future. Significant benefits are expected in the following aspects of heater operation:  Reduction of greenhouse gas emissions  Reduction in energy costs  Lower NOx emissions  Mitigation of vulnerability of the manufacturing industry to natural gas price increases  Reduction in vulnerability to fuel quality swings 5. Acknowledgements The authors wish to thank the US Department of Energy for initiating and supporting this project through DOE Project Number DE-EE0000069. Also, the valuable contributions of Zeeco's staff, including Chris Parker, Jonathon Barnes, Cody Little, and Michael Richardson, and Shell's plant staff, including Dottie Williams and Ted Queener, are appreciated. 6. References 1. Jamaluddin, J., Benson, C., Pellizzari, R., Marty, S., Young, T., Isaacs, R. and Renk, J., "Development of a Fuel-Flexible Burner for Process Plants", American Flame Research Committee Annual Meeting; Salt Lake City, Utah; September 5-7, 2012.