OCR Text |
Show ( schematically in Figure 2. For most tests, the unit was fired at a rate of 1.5 million Btu/hr, which gives a residence time and temperature history in the firebox representative of full-scale boilers in the radiative and early convective sections, as shown in Figure 3. The wall-fired burners were variable-swirl research burners equipped with a single fuel injector tube. Combustion air was preheated to 400°F upstream of the burner by an electric heater. Steam atomization was used for oil and waste firing. The convective section consists of approximately 20 tube bundle heat exchange drawers that are cooled to about 300°F with Dowtherm. For oil firing, a 2.2 percent sulfur residual oil with a heating value of 18,300 Btu/1b was used for the three burners fired on conventional fuel. The oil was heated to about 220°F by drum heaters and heat traced recirculation loops. For coal firing, an Illinois No.6 coal with a 3.6 percent sulfur content was pneumatically conveyed to the burners. The synthetic waste fired through the remaining burner was a blended mixture of 80 percent by weight No.2 distillate oil and 5 percent each carbon tetrachloride, trich10roethy1ene, monoch10robenzene, and trich10robenzene. The mixture was recirculated continually with a pump to retard stratification and fed to the steam-atomized burner at room temperature. For baseline testing in the absence of waste, all four burners were fired on conventional fuel. Thermocouples, sampling ports, and viewports were positioned along the furnace and convective section. Temperatures, f10wrates, and emission measurements were recorded and processed on a microcomputer or data logger. Continuous monitoring was done for 02, CO (two scales), C02, hydrocarbons, S02, and opacity. Volatile waste constituents and PICs were sampled in the flue gas with the VOST. Two methods of using the VOST were employed. For periodic comprehensive quantification, the fu1l-VOST sampling protocol was used with gas chromatography/mass spectroscopy (GC/MS) analysis following thermal desorption. To reduce the cost and turnaround time for multiple samples, a reduced protocol, termed "mini-VOST,· was used. The sampling was done on a single VOST trap for 10 minutes, and then the trap was taken to a nearby gas chromatograph and desorbed to a parallel flame ionization detector and a Hall detector to quantitate hydrocarbons and chlorinated species, respectively. This procedure can provide near-real-time feedback on volatile organic waste concentrations. Semi volatile wastes and PICs were sampled by a Modified Method 5 train fitted with an organics module containing 65 grams of XAD resin. The train was typically run for approximately 4 hours to a total sample volume of 4.3 dscm (150 dscf). The expended XAD was Soxh1et extracted, along with other catches from the Method 5, and analyzed with the GC/MS. RESULTS To address the test objectives, 30 test series were run over 28 test days during early 1987. The resultant sampling and analysis for these runs included 151 fu1l-VOST or mini-VOST runs for volatile organic speciation and 5 |