21 Years of Real-World Low NOx Injection ("LNI")

Paper from the AFRC 2013 conference titled 21 Years of Real-World Low NOx Injection ("LNI") by John Newby.

Originating from a license for the late 1980's-developed "Fuel Direct Injection" (FDI) technology from Tokyo Gas, the author's company developed and widely applied several commercial burner products and systems. Known at Fives North American Combustion as "LNI" (Low NOx Injection), the technique has been used in the aluminum, steel, and glass industries and demonstrated outstanding low-NOx performance. Originally envisaged for gas firing only, the technique was evaluated and commercialized on both light and heavy oils, again with good NOx emissions performance. The technique has been commercially applied to cold air, recuperated air, regenerated air, and oxygen burners. An LNI burner operates as a conventional nozzle-mixing burner below 1450F furnace temperature. Above 1450F, fuel is switched to one or more strategically placed nozzles adjacent to the burner port. Fuel from the nozzle(s) and oxidant from the burner port mix intimately with furnace gases, becoming extremely dilute before combining in a combustion reaction. Local oxygen concentrations can be reduced to below 5%. The dilute streams autoignite and achieve complete combustion within the furnace environment. In the flame envelope, entrained gases limit the maximum in-flame reaction temperatures that generate high NOx emissions. Original laboratory work investigated the parameters affecting the rate of NOx formation and/or the combustion zone mixing and heat release patterns for LNI combustion, covering four burner-dependent aspects: orientation of fuel and oxidant injectors, oxidant exit velocity, fuel velocity, number of injectors; and four application aspects: furnace O2 concentration, furnace temperature, oxidant temperature and furnace geometry. The paper will present test results covering these parameters for conventional forward flame regenerative burners on gaseous and liquid fuels including heavy oil and ambient combustion air and oxy-fuel burners. More recently the ability of LNI systems to reheat metals in box furnaces with close temperature uniformity for high quality forming work has been studied Thermal and emissions performance in the real world of the production environment will be reviewed with examples of regenerative burner and recuperated air reheat furnaces from the ~300 commercial installations in operation worldwide.

21 YEARS OF REAL-WORLD LOW NOx INJECTION ("LNI") John N. Newby Fives North American Combustion, Inc. 4455 East 71st Street Cleveland, OH 44105 USA [e-mail] AFRC 2013 Industrial Combustion Symposium - Kauai, September 22-25, 2013 ABSTRACT Originating from a license for the late 1980's-developed "Fuel Direct Injection" (FDI) technology from Tokyo Gas, the author's company developed and widely applied several commercial burner products and systems. Known at Fives North American Combustion as "LNI" (Low NOx Injection), the technique has been used in the aluminum, steel, and glass industries and demonstrated outstanding low-NOx performance. Originally envisaged for gas firing only, the technique was evaluated and commercialized on both light and heavy oils, again with good NOx emissions performance. The technique has been commercially applied to cold air, recuperated air, regenerated air, and oxygen burners. An LNI burner operates as a conventional nozzle-mixing burner below 1450F furnace temperature. Above 1450F, fuel is switched to one or more strategically placed nozzles adjacent to the burner port. Fuel from the nozzle(s) and oxidant from the burner port mix intimately with furnace gases, becoming extremely dilute before combining in a combustion reaction. Local oxygen concentrations can be reduced to below 5%. The dilute streams autoignite and achieve complete combustion within the furnace environment. In the flame envelope, entrained gases limit the maximum in-flame reaction temperatures that generate high NOx emissions. Original laboratory work investigated the parameters affecting the rate of NOx formation and/or the combustion zone mixing and heat release patterns for LNI combustion, covering four burner-dependent aspects: orientation of fuel and oxidant injectors, oxidant exit velocity, fuel velocity, number of injectors; and four application aspects: furnace O2 concentration, furnace temperature, oxidant temperature and furnace geometry. The paper will present test results covering these parameters for conventional forward flame regenerative burners on gaseous and liquid fuels including heavy oil and ambient combustion air and oxy-fuel burners. More recently the ability of LNI systems to reheat metals in box furnaces with close temperature uniformity for high quality forming work has been studied Thermal and emissions performance in the real world of the production environment will be reviewed with examples of regenerative burner and recuperated air reheat furnaces from the ~300 commercial installations in operation worldwide. 2 1 INTRODUCTION A dilute gas combustion scheme named "Fuel Direct Injection" (FDI) was developed by Tokyo Gas Company [Ref. 1, 2, and 3] in the late 1980's. As the exclusive licensee of this technology, Fives North American Combustion developed commercial products and extended their application to a range of industries as "Low NOx Injection" or LNI. 2 LNI COMBUSTION SYSTEMS An LNI burner operates as a nozzle mixing burner when the furnace temperature is below 1450F. Above 1450F, fuel is switched to one or more strategically positioned nozzles adjacent to the burner tile port. These injectors are displaced from the burner tile exit inside the furnace [Figure 1]. Fuel and oxidant streams mix thoroughly with furnace gases, becoming extremely dilute before combining in front of the burner tile. Oxygen concentrations can be reduced to below 5% in the oxidant stream. The dilute gas streams autoignite and achieve complete combustion within the furnace environment. In the flame envelope, entrained gases limit maximum in-flame combustion temperatures that generate high NOx emissions. All combustion takes place within the furnace, not inside the tile port, providing short high temperature residence times that further inhibit NOx production. After combustion, the gases lose their heat through radiation and convective heat transfer to the work. These cooled gases travel throughout the furnace and are again entrained by the burner oxidant and fuel jets, sustaining the NOx inhibiting process. Separated high velocity oxidant and gas jets can reduce the combustion system NOx emissions by as much as 90%. Figure 1 Schematic of an LNI burner 3 LNI BURNER DEVELOPMENT Developing LNI for commercial applications required selecting parameters of burner geometry that would provide the lowest possible NOx and CO emissions, maintain or enhance the capability of the burner(s) to heat the furnace and load, provide safe and 3 reliable operation of the furnace, all with minimum impact on capital and operating costs. The parameters which affect the rate of NOx formation and/or the combustion zone mixing and heat release patterns for LNI combustion are: The orientation of fuel injector to oxidant injector (offset distance and angle) Oxidant velocity leaving the oxidant injector (typically the burner tile) Fuel stream velocity leaving the fuel injector The number of injectors Oxygen level in combustion products in furnace Furnace temperature Air preheat temperature Furnace geometry The first four of the above parameters are burner dependent, the last four parameters controlled by the application. The original development effort focused on optimized combinations of the first four, and quantification of the effects of the last four parameters on emissions. As the parameters interact, application parameters can influence the selection of particular combinations of burner parameters. An LNI burner is required to perform two functions. It must operate as a conventional burner, providing a flame holder and mixing of fuel and oxidant to raise the furnace temperature above autoignition to sustain LNI combustion. It generally serves as the oxidant nozzle for LNI combustion. High velocity burners are ideally suited to both functions. In conventional firing mode, high velocity burners are inherently low NOx producers as they entrain significant amounts of "cool" furnace gases into the flame envelope prior to completing combustion of the fuel. The reduced port tile which is typical of high velocity burners provides an excellent oxidant nozzle for LNI combustion by entrainment of cool furnace gases with the oxidant stream prior to mixing with the fuel. The late 1990's development programs were devoted to commercial LNI products in the following Forward Flame Burner categories: Regeneratively preheated combustion air, natural gas, #2 and #6 fuel oil (air preheat temperatures typically from 500 to 2050F). Combustion air preheated by recuperator, natural gas fuel (air preheat temperatures typically from 180 to 1200F). Ambient combustion air, natural gas fuel. Oxygen enriched and pure oxygen combustion, natural gas fuel. 3.1 Regenerative Burners Fives North American Combustion's TwinBed™ II regenerative burners are characterized by air preheat temperatures approaching to within 200F of furnace temperature. A typical pre-LNI regenerative installation firing natural gas into furnace temperatures in the 2200 to 2300F range could result in NOx emissions levels of well over 1000 ppm (corrected to 3% O2 basis) as a result of the high air preheat level. The high potential for reductions in NOx emissions due to these high emissions levels, combined with high furnace and combustion air temperatures, made regenerative burners a natural first choice for implementation of the LNI technique. Application of 4 LNI to the TwinBed™ II burner results in NOx emission levels of less than 80 ppm at the same conditions. (Where ppm is used for NOx emissions in this paper, it is a volume concentration corrected to 3% O2 in the products of combustion unless stated otherwise.) The effect of the orientation of the fuel and air nozzles on NOx emissions was investigated in extensive laboratory testing. Two parameters were determined to be significant. One is the angle between the axis of the fuel and oxidant injectors; the other is the displacement (or offset) between the centerlines of the air and fuel nozzles in the plane of the burner wall. For the purpose of investigating the effect of varying the displacement of the fuel nozzle from the air nozzle, a dimensionless injector offset parameter l is defined as the ratio of the distance between fuel and air injector centers to the sum of the air and fuel injector radii:       2 D d l X Where: X = Distance between the air nozzle and fuel nozzle centers D = Diameter of the air nozzle d = Diameter of the fuel nozzles Results of some those tests are illustrated by Figures 2 and 3 below. 4343 TwinBed II LNI NOx emissions 0 0.05 0.1 0.15 0.2 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. angle = 15 degrees angle = 8 degrees Figure 3 4343 TwinBed II LNI NOx emissions 0 0.1 0.2 0.3 0.4 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. L = 1.7 L = 3.75 Figure 4 Further regenerative burner development focused on the application of the LNI technique to #2 and #6 fuel oils. A steam or compressed atomizer is used as a remotely located fuel injector for introduction of a stream of atomized fuel into the furnace. Figure 4 shows the effectiveness of the LNI technique in reduction of NOx emissions for oil firing, and the favorable comparison with the high reductions obtained on gas firing. 5 4343 TwinBed II LNI NOx emissions 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. No. 2 oil conventional f iring No. 6 oil conventional f iring Nat. gas conventional f iring No. 6 oil LNI No. 2 oil LNI Nat. gas LNI Figure 4 3.2 Recuperated Combustion Air With the effectiveness of the LNI technique in reducing NOx emissions from regenerative burners proven in both laboratory tests and commercial installations, the next step was to apply it to burners using recuperatively preheated combustion air. The development platform for low-to-moderate air preheat burners was a high velocity version of a proven 1200F air capable burner with capacity range up to 40 MM Btu/h. A similar philosophy to that used for regenerative burners was followed in the investigation process, with the addition of investigation of the effects of varying the air preheat on the LNI performance - Figures 5 & 6 refer. 4821-10-A NOx emissions vs. temperature 0 0.02 0.04 0.06 0.08 0.1 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. 130 F preheat 400 F preheat 900 F preheat 900 F preheat Figure 5 Figure 6 4821-10-A NOx emissions vs. temperature 0 0.02 0.04 0.06 0.08 0.1 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. L = 2 L = 4 L = 5.3 4 fuel jets, 900 F air preheat, angle = 18 deg. 6 As can be seen from Figure 7, LNI results in a dramatic reduction in furnace NOx emissions, from 290 ppm for a conventional burner firing into a 2300F furnace with 900F preheat temperature, to 66 ppm for the LNI burner firing at the same conditions. 4821-10-A, NOx emissions vs. temperature 0.000 0.100 0.200 0.300 0.400 0.500 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. Conventional f iring LNI f iring, 2 injectors 900 F air preheat, l = 4 Figure 7 3.3 Ambient Combustion Air The success of LNI in meeting the requirements for preheated air applications suggested that it may be effective on ambient air burners. A proven conventional high velocity low NOx burner with capacities up to 18 MM Btu/h, the 4575 HiRam burner was an obvious choice as the development platform for an LNI ambient air burner. Since the NOx emissions for this burner were already fairly low (50 ppm in a 2000F furnace), an extensive development program was defined to achieve the maximum possible reductions in NOx emissions. The variables of injector offset, angle between air and fuel injectors, air velocity, fuel velocity, and number of fuel injectors were evaluated with the following objectives: Determine a configuration of burner and injector parameters to achieve the lowest NOx emissions with suitable combustion characteristics. Quantify the effect of the application-related parameters on NOx emissions. Quantify the effect of deviations from the optimum configuration to facilitate flexibility in application, particularly physically constrained retrofits. Air and fuel injector orientation. The effects of offset distance and relative angle between fuel and air injectors are interdependent. Intuitively, either decreasing the angle between injector axes (i.e. approaching parallel) or increasing the injector offset distance will have the same effect: greater distance between the injector nozzles and the point that fuel and air streams begin to mix, providing increased dilution and hence lower NOx emissions. The results from testing of LNI with ambient combustion air support the intuitive conclusion to a degree. With respect to the offset parameter l defined previously, offsets corresponding to l = 4 combined with angles of 5 produced unacceptable combustion results (stingers from furnace openings, swirling clouds of luminous combustion gases, and generally poor mixing) as did larger offsets with angles of 20 between air and fuel injectors. Comparison of NOx vs. temperature curves for various combinations of l and injector angle are shown in Figures 8 & 9. 7 Figure 8 Figure 9 Air velocity. As Fig. 10 shows, it was found that air velocity had a significant effect on NOx emissions at velocities from 40 to 130 ft/s. Increasing from 130 to 280 ft/s resulted in only a slight improvement in NOx emissions. HiRam air/LNI NOx emissions vs. air velocity; 10 MMBtu/hr firing rate; l =. 4; 2% O2 in POC's 0 0.01 0.02 0.03 0.04 0.05 0.06 0 50 100 150 200 250 300 Air velocity, f t/sec NOx, lb / MMBtu. 2300 degrees F 2200 degrees F 2100 degrees F 2000 degrees F 1900 degrees F Figure 10 Fuel velocity. Fuel velocities were varied from 226 to 750 ft/s. There appeared to be a small improvement in NOx emissions as velocity increased. Velocities below 300 ft/s often resulted in poor flame appearance and furnace pressure pulsation, and occasionally in higher CO emissions from the furnace. Fig. 11 shows a typical comparison of NOx emissions for varying fuel velocity. HiRam air/LNI NOx emissions; 3 injectors, injector angle = 5, l = 4 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. velocity = 226 fps velocity = 750 fps Figure 11 HiRam air/LNI NOx emissions; 3 injectors, l = 3 0 0.01 0.02 0.03 0.04 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. injector angle = 5 injector angle = 20 HiRam air/LNI NOx emissons; 3 injectors, angle = 20 0 0.01 0.02 0.03 0.04 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. L = 3 L = 4 8 Number of injectors. Early work regarding the optimum number of fuel injectors was inconclusive. Some configurations performed as well or better with one injector as with three. Others appeared to be slightly better with three injectors. Results. Results of the laboratory testing of LNI with the HiRam burner conclusively demonstrated that NOx emissions can be dramatically reduced by application of LNI method to burners fired with natural gas and ambient air. Fig. 12 shows a comparison of NOx emissions for an LNI burner compared with a conventional low NOx burner firing into the same furnace. The NOx emissions at 2000F were 14 ppm for the LNI burner, compared with 50 ppm for the same burner fired in conventional mode. The LNI technique reduced NOx emissions from the HiRam burner (which already met requirements for a "Low NOx Burner" in many areas) by a factor of three. HiRam air/LNI NOx emissions; Conventional vs. LNI firing 0.00 0.02 0.04 0.06 0.08 1600 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. Conventional f iring LNI f iring Figure 12 3.4 Oxy-fuel burners The HiRam LNI burner resulting from the ambient air work previously described was augmented with the addition of oxygen capability to allow either mixed air/oxygen or 100% oxygen firing capability. Several configurations were developed: oxygen nozzle concentric with air nozzle and external fuel injectors, both oxygen injector(s) and fuel injector(s) external to the central air nozzle, fuel nozzle concentric with air nozzle and external oxygen injectors. The magnitude of NOx emissions reductions achieved with the various configurations is similar enough that the selection of one configuration over another is more likely to be based on application requirements than on NOx emissions. Results of parameter evaluations similar to those described for the previous investigations are shown for the first configuration in Figures 13 and 14, in which can be seen the effects of injector offset (l) and number of injectors. Significant differences in NOx emissions levels are attributable to variation in both the number of injectors and in the injector offset. 9 HiRam O2 LNI NOx emissions 0 0.01 0.02 0.03 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. 1 injector 2 injectors 3 injectors 6 MMBtu/h firing rate L= 3 HiRam O2 LNI NOx emissions 0 0.005 0.01 0.015 0.02 0.025 0.03 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu . 1 injector 2 injectors 3 injectors 6 MMBtu/h firing rate L = 4 Figure 13 Figure 14 Figure 15 shows the dramatic difference in NOx emissions for the HiRam Oxy-LNI burner firing in conventional mode vs. LNI mode, both with 100% of the oxidant from oxygen. HiRam O2 NOx emissions: Conventional vs. LNI firing 0.0001 0.001 0.01 0.1 1 1700 1800 1900 2000 2100 2200 2300 2400 Temperature, F NOx, lb / MMBtu. conventional f iring LNI f iring: 1 injector, L = 3 LNI f iring: 3 injectors, L = 4 6 MMBtu/hr firing rate Figure 15 4 Post-Y2K work The foregoing material covers the foundation work that provided the basis for a series of commercial products applied to more than 120 commercial installations worldwide with over 700 burners in total by the end of 2001. Subsequent focus has been on application development and refinement, product rationalization, system control and safety code-compliance optimization, while the total number of LNI system installations increased to around 300, with close to 1,400 burners, at mid-2013. As the types of furnaces and processes to which LNI was applied were broadening over the 21 years, the various safety codes of the jurisdictions where the systems were in use were changing to reflect growing awareness of the potential for risk mitigation, primarily lead by European and US standards. At the same time emissions monitoring and control standards were ratcheting down. The combined effects presented some interesting 10 challenges for a furnace-temperature-dependent dual-mode system when used on batch processes. Either the material charge, or both the charge and the furnace may be below autoignition temperature of the fuel, thus involving a mode switch under closely controlled furnace emission constraints. When combined with aggressive productivity requirements that limit the possibilities of such strategies as making mode transitions only at low fuel input rates, system effects and choice of synergistic control components become very significant. Physical location of components in the piping systems, the "capacitance" of the oxidant and fuel piping, and the various control system timing and tuning parameters directly affect ability to achieve a satisfactory performance result. Of course, the same consideration of low temperature conditions is required with continuous process equipment which is designed to operate above autoignition temperature - all such furnaces must start up from cold at some time. Many continuous processes heat product in a counterflow configuration, using load recuperation, such that some portion of the furnace chamber may be always, or perhaps intermittently, below autoignition temperature. Low fuel prices de-emphasized the market interest in the use of oxygen; ambient air LNI systems have been only infrequently of interest. Product rationalization and improvement mostly focused on injector aspects - their number, their location and discharge characteristics relative to particular process conditions of the bulk POC flow in the furnace chamber, and physical life optimization. Single injectors per oxidant port (burner) are the norm today. Understanding nozzle discharge position and orientation relative to the oxidant jet and POC availability have been studied in detail over many applications. Given that relocating injectors after initial installation is at the least "inconvenient" as most are installed close enough to the burner discharge port to dictate that they are in dense refractory, picking the optimal location at design time is desirable. "Non-standard" nozzle discharge ports with application-specific discharge velocities have been deployed to produce particular mixing characteristics, as have bifurcated nozzles. Air cooled injectors cover the infrequent application need for supplementary cooling to prevent fuel passage coking or integrity issues. In terms of application development, ensuring close temperature uniformity in batch reheating of steel and alloys to meet AMS 2750 requirements has been the aspect of LNI combustion most examined. Studies in the late 90's had indicated reduced peak temperatures and more even temperature distribution in the LNI combustion zone of a furnace chamber at design input rates than for conventional forward flame burners. Little study was devoted to the characteristics of the heat transfer in high turndown systems because of lack of call from industry. In the last few years, however, even with low fuel costs that limit the attractiveness of investment in heat recovery, far-sighted users have been requesting regenerative burner systems for their high temperature heating furnaces. Aerospace and oil/gas-field-use alloys require very close temperature uniformity tolerance certification. Accordingly we outfitted an 11.5 ft long, 8.25 ft wide, 7 ft high laboratory box furnace with a highly-instrumented and configurable, fully industrial practice, 2-burner regenerative system and carried out the typical certification testing with test racks that would be required of a production-certified AMS 2750 furnace. 11 Figure 16 Laboratory regenerative burner test furnace Figure 17 View inside empty regenerative burner test furnace LNI firing at temperature 12 Figure 18 Test stands for 2250F testing in laboratory test furnace Figure 19 Typical lab test furnace temperature uniformity profile at 1750F control set point 13 Through this process we acquired a thorough understanding of the system configuration, control and operational nuances of such requirements, secured patent protection (Ref 4) and have successfully executed against them in the field. Most recently we have expanded our R&D lab testing capability with the construction of a very large high temperature air heater (200,000 scfh to 2000F) and downstream flexible-configuration test chamber/exhaust system to allow detailed industrial scale study of high temperature preheated air burner configurations in a wide range of process environments. Figure 20 View of high temperature air heater (RHS) and burner test chamber (LHS) Initial study work with this facility has investigated the use of injected high momentum fuel in combination with lean premix oxidant jets in high temperature air streams. 14 5. EXPERIENCE WITH PRACTICAL APPLICATIONS OF THE LNI TECHNIQUE To date there are approximately 300 commercial installations in operation worldwide with around 1,400 burners in total. Regenerative burner-fired furnaces in aluminum melting and steel reheating applications dominate the LNI installed base. Reported here is the LNI burner conversion of a recuperative hot air fired rotary steel reheat furnace. Also covered are a regenerative LNI aluminum melting installation and a new high temperature uniformity box furnace designed for reheating steel for rolling to AMS 2750D with regenerative LNI burners. 5.1 Low Emissions Upgrade to a 72 ton/h Rotary Hearth Reheat Furnace for seamless pipe production This 81 ft OD, 43 ft ID, 5.5 ft high chamber "donut" rotary furnace reheats steel to 2175F for forming into seamless steel pipe. Customer desired a quick-to-install (2 week shutdown), NFPA86-compliant system, turnkey upgrade to achieve NOx <0.08 lb/MM Btu to ensure ability to meet a 0.1 lb/MM Btu limit. The 850F preheated air capability of the original combustion system was to be maintained. A 55 LNI burner solution with 20 burners on the inside wall, 35 on the outside, used the original burner box locations where possible for the new LNI burners. Figure 21 Burner locations on the 81 ft diameter furnace, showing unfired preheat zone 15 Large custom burner mounting plates and tiles containing the injectors were made to be installed in holes cut in the refractory walls at the original burner box locations. Figure 22 Cross section of LNI hot air Tempest high velocity burner with injector and tile Figure 23 Interior view of furnace chamber outer wall during installation of burners, prior to ramming around the new burners to seal them into the furnace wall refractory 16 New gas piping and controls were installed to complement the new burners and provide the hardware configuration to meet NFPA86 compliance. Figure 24A New zone gas headers were installed concurrently with the mechanical and refractory work to install the burners Figure 24B Zone gas safety trains and flow metering were installed in the center core of the furnace 17 Figure 24C The old Figure 24D The new Pre-Retrofit Post-Retrofit Charge Rate 72.3 ton/h 72.5 ton/h Steel Surface Temperature 2,175 °F 2,180 °F Preheat Zone Fuel Rate 49.7 MMBtu/h 51.2 MMBtu/h Heat Zone 1 Fuel Rate 50.2 MMBtu/h 46.1 MMBtu/h Heat Zone 2 Fuel Rate 44.0 MMBtu/h 24.1 MMBtu/h Soak Zone Fuel Rate 1.7 MMBtu/h 8.5 MMBtu/h Specific Fuel Consumption 1.95 MMBtu/ton 1.79 MMBtu/ton Figure 25 Before and After Conversion statistics and performance When the dust settled after commissioning and performance testing, we received an email from the customer's project engineer which included the following: "Find the third party stack tests below... The data shows that when we are in production the NOx levels are below the target of 0.08 lb/MMBTU and are typically around 0.05 lb/MMBTU..." 5.2 Low Emissions Upgrade to a Round Top-Charged Aluminum Melter With this furnace upgrade we experienced unexpected challenges in the application of LNI. The furnace is a 21 ft diameter round top-charged tilting melter of 45 tons capacity, to be retrofitted with a low emissions regenerative burner system for an increased melt rate of clean scrap from 15.75 to 17 tons per hour and a specific fuel consumption decrease of 28%. 18 Figure 26 Tilted furnace during tapping, showing three of the four LNI regenerative burners A 2 pair regenerative burner system of 40 MM Btu/h capacity replaced the original 750F air, 36 MM Btu/h conventional recuperated burners. With the 2280F-capable furnace melting and raising metal to 1380F, within a couple of weeks of commencement of commissioning trials the furnace was consistently meeting and exceeding the production and fuel consumption warranties. But emissions were extremely high - NOx at 2-3 times expected levels and CO "out of sight", though the melt capacity was easily reached with a 2160F furnace temperature. Probing the furnace chamber with an exhaust gas analyzer showed oxygen concentrations varying between zero and 10%+ during the burner firing cycle. All the usual culprits were investigated, resulting in the replacement of some valves and actuators that proved to be either slow to operate or to fail to close tightly, for a small improvement in emissions performance. Rotation of injectors to change air/fuel/POC mixing characteristics was to no avail, and changing overall XSA levels failed to provide directional indication. The furnace was modeled to attempt to understand potential in furnace mixing influences. Figure 27 shows a typical furnace load pattern after charging from the overhead clamshell "bucket". 19 Figure 27 Top view inside furnace after loading a fresh charge Figure 26A Path lines for air and fuel from two firing LNI regenerative burners 20 Figure 26B Path lines for air and fuel from two firing LNI regenerative burners As may be observed by following the red path lines in Figures 26A and B above, the fuel was being diverted from its intended meeting with the air jet by the in-furnace circulating POC, passing over the top of the air and flowing unmixed and unburned into the direct flue or the exhausting burners. Rotation of the injectors was modeled with the same null result as the field tests. We decided to evaluate the use of higher momentum injectors, despite the fact that the original momentum chosen was considered in the optimal range for a typical LNI burner. A revised momentum of double the original value was chosen as the maximum available from the furnace's fuel gas supply system. 21 Images representing contours of temperature at a specific CH4 fraction are show below - Figure 27A has the original injector momentum, Figure 27B shows the high momentum version: Figure 27A Contours of temperature for a defined CH4 fraction Figure 27B Contours of temperature for a defined CH4 fraction 22 Figure 27B shows a very significant improvement in the fuel jet reaching its intended destination. Accordingly, high momentum injectors were installed on the furnace and emissions testing recommenced. The first day of testing showed that the CO emissions issues were all but eliminated. "Stingers" at the direct flue outlet were no longer visible. Only radiation from the refractory was evident, no "flickering of flames" reflected from inside the furnace chamber. There were no CO peaks at reduced input or at firing/exhaust switching events. Modifications of the system software for enhanced XSA control bought both CO and NOx within the warranted values in subsequent tests. 5.3 New Construction Close Temperature Uniformity Box Reheat Furnace Customer desired to increase his processing capacity for high density nickel-based alloys. The new furnace would preheat material to 1500F to 2200F over 8 to 12 hours for rolling, or homogenize an ingot at around 2200F for 2 to 4 days. Regenerative burners were desired to minimize fuel consumption, low emissions were a given, and uniformity of <+/-15F at 1650F and 1900F, <+/-25F at 2250F should be achieved in an empty furnace AMS 2750 rack test. To accommodate the typical load of around 35,000 lb, and provide the desired loading capability, a 3 door furnace of internal dimensions 4' high, 12' 8" wide and 16' deep was constructed, having 2 pairs of LNI regenerative burners, one pair on each side wall; Figure 28. Figure 28 23 Figure 29 During installation and instrumentation of temperature uniformity test rack After commissioning the furnace was rack-tested for temperature uniformity, Figures 29, 30 and 31. Figure 30 Burners firing before closing door for rack test of uniformity 24 Figure 31 Typical temperature uniformity profile at a 1900F control set point Emissions were tested over 12-hour cycles comprising 2 hours heating from 1330F to 2150F and a 10 hour load soak cycle at 2150F. Data was recorded at 1 second intervals throughout the test except for a 5 minute "analyzer rinse cycle" after each hour of operation. The higher of the emissions readings before and after each rinse cycle was chosen to represent the emissions in that period. The 12-hour cycle average NOx was determined at 0.084 lb/MM Btu, with a specific fuel consumption of 1.6 MM Btu/ton, both below the warranted values. Figure 32A Furnace at temperature Figure 32B Charging at side door25 6 SUMMARY Development work and application successes have demonstrated the power and versatility of the dilute gas combustion LNI techniques in reduction of NOx emissions in high performance combustion systems without sacrifice of efficiency or heating capability of the furnace. Optimization of LNI burner parameters has been aggressively followed in conjunction with development of a substantial understanding of the application parameters required to match LNI in-furnace combustion with the fired process. The LNI technique has been evaluated for several combinations of fuels and oxidant-supply conditions. It has shown to be a powerful NOx inhibitor with all commonly used practical industrial choices. For regenerative burners, where high air preheat capability offers low fuel consumption, LNI offers striking NOx emission reduction - to the point that such burners are most likely to offer the lowest mass emissions per unit of production of any combustion system type - ambient air, recuperative, or regenerative - in many high production processes. Application of the technique to liquid fuels significantly broadens the available high-performance system options for their users. Extension of LNI to oxygen and air/oxygen burners also provides high levels of emissions reduction. With around 1400 burners in operation in just under 300 furnaces in the 21 years of LNI work at the time of writing, we continue to believe that LNI dilute gas combustion is the most practical and powerful NOx-inhibiting technology available for high-temperature process furnaces. 10 REFERENCES 1. US Patent 4,945,841, 07/07/1990, Nakamachi et al., Apparatus or Method for Carrying Out Combustion in a Furnace. 2. Shigeta, E., Kanazawa, H., Koizumi, T., & Nagata, T., (1991) "Low-NOx Combustion Technique for High-Temperature Furnace", AFRC/JFRC International Conference on Environmental Control of Combustion Processes, Honolulu, Hawaii. 3. Saiki, N., & Koizumi, T., (1994) "Application of Low-NOx Combustion Technique for Regenerative System", AFRC/JFRC Pacific Rim International Conference on Environmental Control of Combustion Processes, Maui, Hawaii. 4. US Patent 8,083,517, 12/27/2011, Newby et al., Method of Operating a Furnace.