1. Figure 1 shows the general arrangement of the combustion-control experiment. Details of the laser system and the operating conditions employed to control the lasers have been described previo~sly?~~ The system includes two independently operated distributed feedback (InGaAsP) semiconductor diode lasers tuned at a 2-kHz repetition rate over the desired transitions by ramp-modulating the individual injection currents to yield single-sweep spectrally-resolved absorption records every 500 p. The individual laser outputs were combined into a single path using appropriate single-mode fiber splitters and couplers. The multi-wavelength beam was directed through the flowfield using a gradient index of refraction (GRIN) matrix of 0.5-mm internal diameter, 16.7-mm long stainlesssteel needle tubes that carry fuel (CH4)from a reservoir fed by a circular annulus. At the top of the burner, the fuel exits the tubes and mixes with streams of co-flowingair to yield 4 6 0 diffusionflamelets. For the typical operatingconditions(CH, flowrate 4.5-5.3 Ilmin, CH, velocity 45-53 cm/s in the tubes; air flow rate4 0llmin,air velocity in the flowstraightener=120 cds), the flowwas laminar (ReDc 100)and thetips of the flamelets were typically less than 3 mm above the burner surface.
2. The system was used to record H,O absorption spectra throughthe product gasesof a CH,air diffusion flame generated by a Hencken burner. The burner, shown schematicallyinfigure 1,consistsof a 2-inch (on aside)squarearrayof diffusionflameletssustainedby a valve. Fluctuationsin gas temperaturewereinduced by modulatingthe fuel flow rate using two methods while maintainingaconstantairflowrate. In thefirstmethod, a speaker (4-inch dia, 8-R impedance, frequency response 50 - 3000 Hz), mounted at the base of the burner, modulated the fuel flow rate acoustically. The waveforms that were used to drive the speaker were synthesized by a computer program and converted to voltage signals using a 12-bitD/A card. In the second method,the fuelflowrate was varied by modulatingthe controlvoltagesuppliedtothe flowcontroller.6
3. of the combustor to the actuators. A sampledata trace illustrating the applied voltage amplitude scheduling is shown in figure 3. With the flow controller voltage maintained at 2.71 V ($ = 0.8),the amplitudeof a 50-Hz sinusoidal input to the speaker was increased incrementally. The resulting temperature change (recorded in 2-msec intervals) was measured from the ratio of peak H20 absorbance~.he transientdips in the mean temperature for relatively large driving voltages to the speaker (duringthe measured interval 9-10 sec) may be due to reductions in the extent of reactionatthemeasurementlocation.
4. transfer function. Temperature is recorded for various values of actuator input voltage at a particular driving frequency (50Hz,inthisfigure). Thesolidlinerepresents the measured temperatures. The dashed line represents the mean temperaturevalue. The inset clearly shows the measured 50-Hztemperatureoscillations. 60 - 60
5. rn wn~.~dl.h 60