Fiber-Coupled Laser Diagnostics for Temperature, Species, and Flow Velocity Measurements in Practical Combustors

We present the recent advances made in fiber-coupled laser-diagnostic methods that are being developed for practical combustion applications. In particular, we describe fiber-coupled, picosecond, coherent anti-Stokes Raman scattering (ps-CARS) spectroscopy, fiber-coupled, nanosecond, ultraviolet planar-laser-induced-fluorescence (ns-UVPLIF) spectroscopy, and fiber-coupled, particle-image velocimetry (PIV) in details that employ long-length, multimode step-index fibers for temperature, species-concentration, and flow velocity measurements in harsh combustion environments. Gas-phase single-laser-shot thermometry using fiber-based ps-CARS spectroscopy of H2 is demonstrated in atmospheric-pressure, near-adiabatic laboratory flames. Also demonstrated is a fiber-coupled, 10-kHz UV-PLIF/PIV system for simultaneous detection of OH concentration and velocity in a realistic 4-MW combustion rig. The effects of delivering intense laser beams at visible and UV wavelengths through long optical fibers are investigated. The fundamental physical limits on the transmission of intense laser pulses through silica-based stepindex fibers for CARS, UV-PLIF, and PIV are studied. The potential system improvements are discussed. Development of such fiber-coupled diagnostic systems constitutes a major step forward in transitioning laser-diagnostic tools from research laboratories to practical combustion facilities such as gas-turbine test rigs. Corresponding author: [email protected] ILASS Americas, 25 Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 2013 Non-intrusive laser-based diagnostics methods have proven to be versatile and important tools for making measurements of temperature, chemical composition, and fluid-dynamic parameters in complex combustion flows. In particular, coherent anti-Stokes Raman scattering (CARS) provides temporally and spatially resolved “point” measurements of gas temperature and major-species concentrations under hightemperature and -pressure conditions [1]. Planarlaser-induced fluorescence (PLIF) [2] and particleimage velocimetry (PIV) [3], on the other hand, are optical diagnostic techniques for measuring twodimensional fields of species concentration (or temperature) and velocity, respectively. PLIF and PIV techniques have been widely used for characterizing reacting and non-reacting flow fields. Especially the simultaneous application of PLIF and PIV imaging enables new insight into the interaction between turbulent flows and chemical reactions. Recently, the CARS, PLIF and PIV techniques have been used for spray diagnostics such as temperature measurements in sooting flames [4], fuel concentration measurements in isothermal Diesel sprays [5], and air/fuelphase velocity field measurements in a dense fuel spray [6]. However, the harsh environment associated with combustors and engine test facilities is often challenging to access, even by means of the simple optical method of PLIF and PIV. This difficulty can be overcome by transmitting the required laser energy through optical fibers to the probe volume for CARS, PLIF, and PIV measurements. In addition, the use of an optical-fiber delivery system allows reduction of the required free-standing optics in the test-cell environment, isolation of the high-power laser system from harsh environments, and provision of safe, guided, and confined laser-beam delivery. Silica-based multimode fibers are best suited for long-length delivery of high-power laser pulses because of the minimization of bending and transmission losses [7,8]. However, realization of fibercoupled CARS, PLIF and PIV techniques in hightemperature gases is challenging because of the intrinsic optical-damage threshold of the fiber, which limits the maximum deliverable energy [7-11]. Additional difficulties arise when the laser wavelength is below ~300 nm, where solarization effects become prominent and cause poor fiber transmission [7]. Solarization results from UV absorption by fusedsilica material and the subsequent change in transmission properties due to the formation of color defect centers, which can further reduce the transmission. Moreover, the increase in input pulse-repetition rate (PRR) further exacerbates the difficulties associated with high-power fiber delivery because the cumulative thermal effects caused by high-PRR lasers can reduce the optical-damage threshold of the fiber [9,10]. As a result of the technical challenges described above, to the best of our knowledge, no fibercoupled, high-speed (~kHz) UV-PLIF and PIV systems and no fiber-coupled CARS had been realized for reacting-flow applications. In this paper, we discuss our recent advances in fiber-coupled picosecond (ps)-CARS (10 Hz) and nanosecond (ns)-10-kHz-UV-PLIF/PIV systems that employ long-length, multimode step-index fibers for temperature and species-concentration measurements in reacting flows. We have demonstrated gas-phase single-laser-shot thermometry using fiber-coupled psCARS spectroscopy of H2 in atmospheric-pressure, near-adiabatic laboratory flames. We have also developed a fiber-coupled, high-speed UV-PLIF/PIV system for measuring hydroxyl radical (OH) concentration and velocity in a realistic 4-MW combustion rig. Simultaneous OH-PLIF and PIV imaging at a data-acquisition rate of 10 kHz is demonstrated in turbulent premixed flames behind a bluff body. The fundamental physical limits on the transmission of intense laser pulses through silica-based step-index fibers for CARS and high-repetition-rate UVPLIF/PIV have been studied. Fiber-coupled ps-CARS thermometry in reacting flows The ps laser is ideal for fiber-coupled CARS because 1) the peak intensity of ps-laser pulses at the fiberdamage threshold is nearly a factor of twenty higher than that of ns-laser pulses, which is particularly advantageous since the CARS signal scales as the product of the intensities of the input laser beams; 2) sufficient ps-laser energy can be reliably delivered through a MSIF for gas-phase CARS without damaging the fiber [8]. A ps, laser-based, fiber-coupled CARS system of H2 was designed and demonstrated using a collinear geometry. A schematic diagram of the system is shown in Fig. 1. The nearly transformlimited, frequency-doubled, 135-ps output of a Nd:YAG regenerative amplifier at 532 nm was used as the pump beam, and the 115-ps output of a broadband modeless dye laser (FWHM ~5 nm) at 607 nm was used as the Stokes beam. The pump and Stokes beams were then each coupled into a 365-m-core low-OH fiber (OFS, CFO1493-52) using an F=+125mm spherical lens. The input end of the fiber was positioned behind the focal plane such that the beam expanded to fill 80% of the fiber core, thereby avoiding self-focusing-induced damage inside the fiber core. Temperature measurements were performed in an atmospheric-pressure, nearly adiabatic H2–air flame stabilized over a Hencken burner. The burner was operated at high flow rates (combined fuel and air flow rate ~50–70 slpm) to minimize heat losses to the burner and produce nearly adiabatic temperature flames. Shown in Fig. 2 are the single-shot spectra and associated temperature PDFs (probability density functions) derived from 500 spectra for equivalence ratios () of 4.1 with fiber lengths of 1 and 3 m. The energies used for the pump and the Stokes pulses were ~600 μJ/pulse and ~900 μJ/pulse, respectively. The Q-branch transitions of the v = 0 v = 1    vibrational band of H2 are shown in Fig. 2(a) and (b). The integrated intensities of the Q-branch transitions were employed to extract temperatures using Boltzmann plots [12]. The mean temperature and precision of the single-shot thermometry for the 1-m-long fiber case were found to be within ~1% and ~2% of the set value, respectively. These values were ~2% and ~5% for the 3-m-long-fiber case. The precision of the single-shot temperature measurement decreased with an increase in fiber length because of the decrease in signal-to-noise ratio (SNR) and the loss of spectral resolution. The longest fiber demonstrated for single-shot, fiber-based, H2-CARS thermometry was 3 m in an ~1500 K flame with an SNR of ~10:1; with the use of 6-m-long fibers, the SNR is insufficient for single-shot flame thermometry. Sufficient SNR for higher flame-temperature measurements could be achieved with 3-m or longer fibers by averaging the CARS spectra over multiple laser shots. With an increase of fiber length, a significant reduction in the CARS signal was observed, the most probable causes being (1) scrambling of the polarization state of the laser beam during propagation through the longlength, multimode fiber, (2) increasing M and degrading beam-focusing ability, and (3) increasing nonlinearity and, hence, broadening of the bandwidth of the fiber-delivered beam [11]. Application of this system to higher temperatures by improving the fiber-delivery capability through the use of end-capped fibers, thick clad fiber, or hollow-core, photonic bandgap fibers is the subject of a future investigation. Fiber-coupled 10-kHz OH-PLIF/PIV in a realistic 4MW combustion rig The experimental apparatus for the fiber-coupled, high-speed UV-PLIF/PIV system and the combustion Figure 1. Schematic diagram of the fiber-based psCARS system. HW, half-wave plate; P, polarizer; CCD, charged-coupled device; L1, fiber-input coupler; L2, fiber-output collimator; DM, dichroic mirror. Figure 2. Experimental single-shot, fiber-based, psCARS H2 spectra in a H2–air flame stabilized over a Hencken burner at = 4.1 for (a) 1-m-long fiber case and (b) 3-m-long fiber case. Corresponding temperature PDFs for 500 consecutive laser shots are also shown in Panels (c) and (d). rig is shown in Fig. 3. The 10-kHz, 8-ns, 283-nm laser pulses were generated by frequency doubling the output of a narrowband dye-laser system (Sirah, Credo) that was pumped by a Nd:YAG laser (EdgeWave, Innoslab HD Series). The frequencydoubled output of the dye laser was tuned to excite the Q1(9) transition in the (1,0) band of the OH A – XΠ system. The 10-kHz, 160-ns, 532-nm laser pulses were generated by a dual-head PIV laser (Quantronix, Dual-Hawk). The separation time between the two PIV laser pulses was 20 μs. The OHexcitation pulse was fired between the two PIV pulses to eliminate PIV-particle light scattering. The PIV laser beam passed through two 0.25° diffractive optical elements (DOEs) (HOLO/OR, RD-203-Q-Y-A and RPC Photonics, EDC-0.25) and was then combined with the 283-nm PLIF beam by a dichroic mirror. The two laser beams were coupled into a 600m-core, deep-UV-enhanced fiber (Polymicro, FDP600660710) with an f = 50 mm spherical lens. The FDP fiber has been experimentally identified as the most suitable commercial fiber for high-power UV beam delivery because of its superior ability to resist solarization effects [7]. Both fiber ends were terminated with specially designed, epoxy-free highpower connectors (Polymicro) for dissipating the thermal energy that could burn the surrounding material near the fiber core. The use of DOEs not only smoothes the input-beam profile but also increases the number of spatial modes existing in the beam. This setup minimizes the formation of hot spots that could potentially damage the entrance surface of the fiber. Furthermore, it also prevents the occurrence of self-focusing effects within the fiber. As a result of using DOEs, the laser-induced damage threshold (LIDT) of the fiber was increased by approximately a factor of three over the optimum achievable using conventional optics, enabling transmission of the required 532-nm pulse energy through the fiber for high-speed PIV [9,10]. No DOE was applied for the UV laser-to-fiber coupling, as the selected FDP fiber is capable of coupling the full-pulse-energy (~200 μJ/pulse) output from the PLIF laser. The output of the fiber was focused onto a probe volume using an f = 100 mm, 50.8-mm-square cylindrical lens. The resulting laser sheet was ~80 mm tall, with a thickness of ~2 mm at the probe volume. Fluorescence detection was performed using a CMOS camera (Photron, SA5), coupled to an external two-stage intensifier (Lavision, HS-IRO). Collection of scattered light from the PIV seeding particles (1-m Al2O3) was achieved using another Photron camera, coupled with an 85-mm f/1.8 lens. Measurements of 2D OH concentration and velocity field were made behind a bluff body (V-gutter) in a 4-MW atmospheric-pressure combustion rig using the fiber-coupled, high-speed UV-PLIF/PIV system. The crosssectional geometry of the experimental test section is 152 mm x 127 mm, and the length is ~1 m. Optical access for the cameras was available on the top window of the rig. The optical fiber was used to access the rig perpendicular to the cameras. The flame employed for OH-PLIF/PIV studies was a premixed propane–air flame with an equivalence ratio of = 1.05, an airflow rate of 0.7 lb/s, and an adiabatic flame temperature of ~2300 K. A strong acoustic flame instability is present under these conditions. The transmission characteristics of the fibercoupled, 10-kHz OH-PLIF/PIV system are shown in Fig. 4. For UV fiber transmission as shown in Fig 4(a), the signal was averaged over 10,000 laser shots at each input-energy step. As the input UV-laser energy level increased, a substantial decrease in coupling efficiency (i.e., ratio of output energy to input energy) occurred. Since no evidence of fiber damage (i.e., entrance surface damage) was observed during the measurements, the rapid decrease in transmission efficiency is thought to result from nonlinear absorption in the fiber. Because of the limited energy output from the 10-kHz, 283-nm laser, the UV LIDT of the fiber was not determined. For the fiber transmission of the 10-kHz, 532-nm laser pulses, the coupling efficiency was maintained at ~75%, and the visible Figure 3. Schematic diagram of the combustion rig and the fiber-coupled, high-speed OH-PLIF/PIV system.