OCR Text |
Show oil or gas or a combination of oil and gas. It is a natural circulation reheat type boiler with an approximate capacity of 2,400,000 Ibs./hr. of superheated steam at 2460 psig and 10050 F. The front wall contained 18 burners arranged in three horizontal rows. The burners were designed by Clark Chapman Ltd., International Combustion Division and are described as axialflow, tip shut off, low excess air burners. The burner vendors specifications indicate that maximum combustion efficiency is achieved when the burners are operated at 0.54% O2, The exhaust gases are split as they leave the boi 1 er. The gases from Burners 1, 2, 3, 7, 8, 9, 13, 14, and 15, in theory, exit the boiler through the East ductwork, and the gases from Burners 4, 5, 6, 1 0 , 11, 1 2, 1 6, 17 , and 1 8 , in theory, exit through the West ductwork. Our tests were conducted on the last set of burners and the West ductwork. Figure 8 is a graphic representation of the boiler that was tested. The test probes were located in an area immediately preceding the economizer section. The gases were flowing in the downward direction and were about 200 ft. from the burner locations. Figure 9 provides a cross-sectional view of the duct work and displays probe locations relative to burner center lines. The procedure followed to collect data went as follows: The boiler was allowed to operate at a constant 200 or 270 MW output. During this period, steam pressure and oxygen trim control were put in the manual mode and baseline data was collected - neither steam pressure nor oxygen varied significantly during the data collection operation. Then, one by one, each burner was perturbated. Fuel to a row of burners came from a common header. A val ve on the supply side of the burner allowed it to be shut off for physical removal from service and a val ve on the return side of the burner could be closed down to force more oil through the burner nozzle. An opacity monitor was located on the stack. When it indicated a smoking condition, data collection began. It was not always possible to create a smoking condition. In those instances, data collection began when a noticeable change in carbon monoxide readings were noted. These test results were usually the ones that were least conclusive. Data was collected manually using portable analyzers to measure ppm combustibles and percent oxygen. Each test, which included baseline data and data after each perturbation, took about an hour to run although the actual time for collecting the data took only 10 to 15 minutes per condition. Often, the last few data points were collected hurriedly. The increase in fuel would cause steam pressure to increase to its alarm point at which time the tests were aborted. Prolong periods of excessive smoking conditions violated local and state environmental regulations which also caused some tests to be 232 abbreviated. Data was collected for each point but could not always be rechecked because of the need to reduce boiler steam pressure or to eliminate a smoking condition. No tests were run twice; however, some of the data was confirmed by retaking it a second time during the same perturbation when time and conditions allowed. This was done, in particular, when there were large differences noted in our measurements from one side to the other of the ductwork, such as those that occurred when we tested burner 18 at 270 MW. All nine burners were individually perturbated at a 270 MW load and Burners 4, 5, and 6 were individually perturbated at a 200 MW load. Following each test run, the boiler was allowed to settle out, and another set of baseline data was collected in preparation for perturbating another burner. A portion of the data collected at the utility site is displayed in Figure 10 in a three dimensional form. The top displays are 6PPM CO and the bottom ones are 6 PPM O2, The large rectangular box in the "X-Z" plane represents the duct dimensions (36'6" W x 15'4" D) and the smaller rectangular box inside the larger one represents the outline of the area in which data was collected. It was noted during our test program that more mixing occurs between gases from burners above and/or below the burner that was perturbated as opposed to mixing with gases originating from burners located horizontally or diagonally to the perturbated burner. This was very evident with the results on burner 18. This test was not run again, but the O2 and CO readings were confirmed by retaking them again before the test was concluded. The objective of the EPRI sponsored program was to determine if flue gas stratification existed on a large multi-burner utility boiler and if that effect could be used to detect and adjust an individual maladjusted burner. The test results demonstrate that flue gas stratification is real and measurable. Baseline measurements show a relatively even distribution of carbon monoxide and oxygen across the exhaust duct. Following a burner perturbation, this distribution changed dramatically. In the majority of the tests, the effect was seen prominently in the location where it was predicted to have been found in the duct location. In most of the inconclusive tests, there was some question as to how effectively we were able to perturbate the burner being tested. A number of things become evident as the test program progressed. The first was under conditions near stochiometric, changes in carbon monoxide readings are a better indicator of a faulty burner condition. This is due to the increase sensitivity inherent to carbon monoxide analyzers. It also has to do with the characteristic fUnction of carbon monoxide in the classic relationship of excess air and carbon monoxide in any combustion process, wherein, the |