OCR Text |
Show single-stage mode without FGR while making the measurements presented here. The fuel and air flows were measured using laminar flow elements. Each of the eight sides of each furnace spool has an opening (406 mm wide x 267 mm high) for a refractory-lined or bare stainless steel panel or a frame holding a fused silica window. For the cold-wall measurements, all of these openings contained water-cooled bare metal panels, as shown in Figure 1, except when optical access was required for laser Doppler velocimetry (LDV). For the original hot-wall measurements, the panels were refractory lined. Where probes are inserted, special panels are installed that have 51 or 127 mm wide vertical slots fitted with covers having 32 mm wide openings for the probes. These openings are filled with refractory blanket to minimize leakage of gas into or out from the furnace. The heat extracted through the furnace wall is detennined from cooling water flowrates and temperature changes. In-Dame Measurements The follow-on measurements duplicated the 15% excess air (3.0% exit 02' dry) cold-wall baseline flame as measured by Sayre et al (1994). A summary of typical operating conditions is given in Table 1. Table 2 contains a typical fuel analysis obtained during the tests. The data was collected using the six-spool BERL configuration instead of the five-spool configuration to limit facility modifications. The change in BERL configuration and the location of the original and follow-on measurement traverses are shown in Figure 3. Combustion modeling cases completed prior to the start of the tests indicated that this change had a negligible effect on the data. The measurements reported here were made using a suction pyrometer with type S thermocouple, an extractive gas sampling probe, and an ellipsoidal radiometer, all from the International Flame Research Foundation. Gas velocities, both mean and RMS fluctuating components, were measured by LDV using an argon ion laser and Aerometrics Doppler Signal Analyzer, with the combustion air seeded using 0.05 micrometer alumina particles. The probes and LDV system were mounted on a motor-driven optical table to traverse the flow. The entire furnace is moved up and down in relation to the table in order to probe the flame gases at different heights. Gas composition was determined using nondispersive infrared (CO and CO2), flame ionization (total hydrocarbons), paramagnetic (02)' and chemiluminescence (NO and NOx) detectors and a micro-gas chromatograph (H2, °2, CO, CO2, CH4, and N2). When making the temperature and gas composition measurements using the continuous monitors, the output from each instrument was observed for approximately one minute and the highest, lowest, and typical readings were all recorded. Table 3 provides a summary of all measured data for the three test conditions discussed below. The tests began with a repeat of two of the original traverses at 27mm and 343mm to ensure replication of the baseline flame characteristics. These two traverses displayed excellent repeatability with the original data set for all measured quantities. A comparison of several main variables from both the original and follow-on measurements is shown in Figure 4. This figure shows that only quantities such as carbon monoxide and NOx, which are sensitive to small changes in fuel, show any significant variations. With successful repetition of the original baseline flame, data was then taken at eleven new traverse locations as shown in Figure 3: 0+,16,65,131,150,208,390,510,695,815, and 1156mm downstream of the quarl exit. Data collected at these traverses included axial and tangential velocity components, gas temperature, and major species as described above. The exceptions are the quarl exit traverse (x=O+), where axial, tangential, and radial velocity components were measured, the 16mm traverse, where only gas temperature and species measurements were made, and the 815mm and 1156mm traverse, where only velocity and gas temperature were measured. For a majority of the intrusive measurements, both minimum and maximum values were recorded once the probe reached steady conditions. An example of this is shown in Figure 5 for gas temperature. This min-max range for key variables provides a valuable piece of information for model evaluation, since it provides an envelope of possible values. As can be seen from Figure 5, this envelope can be significant in some areas were turbulent fluctuations are large. Limited data were also taken at seven traverses for a low excess air condition (2.5% excess air, 0.6% exit 4 |