OCR Text |
Show KBTU/HR with step increases to 860 KBTU/HR for two minutes every ten minutes, totaling an RDF load of 540 KBTU/HR. PCDD/PCDF emissions during the unsteady incinerator operation were significantly higher than during steady operation. Increased PCDD/PCDF emissions during unsteady operation are likely due to insufficient oxygen concentrations during periods of high load resulting in lower combustion rates and increased survival and entrainment of unburned RDF particles with PCDD/PCDF precursors. Low O2 concentrations were measured in the furnace exhaust « 5% dry) during the periods of high RDF feed. These tests show incinerator operating conditions can have a significant impact on PCDD/PCDF emissions. Emissions are minimized by steady operation of the combustion system at non-excessive fuel loads with sufficient levels of excess air. Influence of Natural Gas Co-Firing. Natural gas co-firing, as a means to improve RDF combustion and reduce PCDD/PCDF emissions, was studied. It was presumed that the injection of natural gas would reduce PCDD/PCDF emissions by providing higher temperatures and combustion radicals to increase combustion kinetics of PCDD/PCDF and their precursors. Tests were run to study the effect of natural gas co-firing on PCDD/PCDF emissions and to determine optimum co-firing strategies for maximum reduction of emissions. These tests coincided with the RDF load reduction studies. A reduction in RDF load was replaced with injection of natural gas such that full load heat input (540 KBTU/HR) to the system was maintained. Injection ports for the natural gas were located in the CTT at or downstream of the RDF feed point. The natural gas was introduced to the furnace with a water-cooled radial injector. The injector was aligned at the center of the circular combustion chamber. Injector design and transport gas set the injector jet velocities to barely impinge the flame on the furnace walls, providing good cross-sectional area coverage. The flame pattern was visually verified from an observation port at the top of the CTT. The flame acquired a green tint caused by the copper in the RDF making flame pattern identification quite simple. Air was used as the transport gas in the injector at a volumetric flow of 1 0-1 5 % of theoretical air required for combustion of the co-fire natural gas. The substoichiometric air/natural gas mixture produced a diffusion flame zone requiring oxygen from the flue gas for combustion. A diffusion flame was used rather than a pre-mixed flame to provide a higher influx of combustion radicals. Two series of tests were run. The first series examined variations in co-fire natural gas load. The furnace was operated with substitution of 15, 20 and 30% RDF load with co-fire natural gas. The natural gas was injected at the port furthest from the RDF feeder, approximately 0.7 seconds flue gas residence time downstream. This injection created a high temperature afterburner zone downstream from the main RDF combustion. Results from the first series of tests are presented in Figure 8. PCDD/PCDF emissions are reduced with the co-firing of natural gas. Minimum PCDD/PCDF emissions are seen with 20% natural gas co-firing. Higher emissions with 30% co-firing may be due to a trade-off between excess air and RDF combustion zone temperature. Reductions in RDF load increase the SR at the RDF feeder, increasing the combustion rate in the well-mixed zone at the feeder. This is reflected in lower PCDD/PCDF emissions with lower load. However, further decrease in RDF load may reduce heat input to the point where low temperatures reduce the kinetics of suspended RDF combustion. An increase in PCDD/PCDF emissions is observed as the natural gas input is increased from 20 to 30% of load. A comparison of this data with results from the reduced load studies, presented in Figure 9, shows a similar trend. Emission levels are close to those from the earlier 6 |