OCR Text |
Show urements at various duct locations are shown, as made with one mill out of service. The duct areas with the highest oxygen content can be clearly associated with the two burners in service. Note that excess air on the left side is appreciably higher than that of the right. It is this type of unbalance that can be corrected by observing the stratified analysis data. Results obtained from the experimental runs follow the expected pattern. Oil injected into each burner produced an observable reduction in oxygen at all sampled duct locations, with the magnitude of the change greater at the expected duct locations. Figure 4 diagrams the results of a test using the oil injection technique. Vertical lines represent the amount of oxygen reduction in each duct area resulting from the introduction of oil through each burner. Variations seen strongly support the existence of the postulated stratification effect. The differences are very definite in horizontal directions though less evident but still identifiable vertically. This result is not unexpected due to the long but narrow duct configuration. SECOND DOE INDUSTRIAL SITE DESCRIPTION, PROCEDURES, AND RESULTS - The information obtained from the first test site was reviewed by DOE and Bailey representati ves. A recommendation was made to obtain additional data from a second test location. Observations at another site would provide confirmation the effect exists on various designs of boilers and process heaters. Additional tests are especially significant if the boiler and burner configuration differs from the first one tested. It is also of interest to investigate different fUel composition, as the volume of combustion products produced may affect the stratification pattern. The boiler made available for our use was a Babcock & Wilcox industrial bOiler, constructed in 1965. It had a design capacity of 300,000 pounds per hour continuously of superheated steam at 1000 psig and 910or. The boiler was equipped with six B&W burners, arranged in pairs at three levels on the front wall. The products of combustion turn upward from the burners, pass a baffle area, and made a single 900 turn before leaving the boiler. The burners are B&W mul tiple fUel type designed to utilize dirty by-product gases while burning pulverized bituminous coal as the primary fuel. We followed the same basic procedure as had been developed at the first test site. Portable analyzers were then used to make several traverses of the exit duct under various operating conditions. Figure 5 is typical of the data recorded during such surveys. Several tests were done in order to develop procedures for momentarily varying individual burners and observing the effect on oxygen in the flue gases. A procedure was used similar to that 231 employed at the first test site. Use of natural gas and coke oven gas made it much easier to make reproducible changes. This was especially effecti ve since the operator could set val ves from the control room while observing the fUel flow rate on the panel. Procedures were established to make a series of measurements in a planned sequence. With the boiler under manual control, the operator added gas to a single burner at a predetermined rate and time. The procedure was repeated at set intervals, allowing change in sample location to be made between gas additions until all duct areas had been sampled. With val ving set to allow gas addition to another burner, the sequence was repeated. In this manner data was obtained relating burner disturbance to flue gas oxygen content through all burner to duct area combinations. A Bailey Type OL oxygen and combustibles analyzer, Figure 6, was used for the measurements in conjunction with the portable analyzer used to obtain the preliminary data. Bailey's Analytical Engineering group made an experimental multiple point sampl ing assembly available for use at the test site. As an accessory to the Type OL, this allowed a selection of sample points and greatly facilitated data acquisition. The assembly was left in operation in the flue after our tests were done, in order to evaluate its long term performance under actual operating conditions. Excellent operation obtained from the mechanism is significant, as multiple point sampling is necessary to make effective use of observed stratification to achieve improved performance. The accessory product will effect major cost saving over the use of many individual analyzers. The results obtained from a typical sequence of tests are shown in Figure 7. The number and length of lines in each duct area represent the reduction in oxygen content observed when the fUel input to the indicated burner was momentarily changed. Based on the boiler configuration, a change in burner 111 would produce a maximum v"ariation in oxygen content in the lower left section of the duct, if stratification of the gas streams exist. The data plot, Figure 7, does show this, although the area of maximum influence is displaced away from the corner. Similar plots for all burner-duct area combinations produced patterns confirming the stratification effect; many showing more distinct correlation than the example shown. EPRI UTILITY SITE DESCRIPTION, PROCEDURES, AND RESULTS - The boiler tested in the EPRI program was designed and built by the Foster Wheeler Corporation. It was designed to burn |