Application of Laser-based Diagnostics to a Prototype Gas Turbine Burner at Selected Pressures

Abstract

The matured laser-diagnostic techniques of planar laser-induced fluorescence (PLIF) and particle image velocimetry (PIV) were applied to a prototype gas turbine burner operating on various fuels. The work was performed to provide verification of computational fluid dynamic (CFD) models of the combustion of atypical fuels in a gas turbine combustor. The burner was operated using methane and three synthesized fuels of interest- one with hydrogen as the principle component and two with a low heating value (15 MJ/m3). Experiments were performed at pressures from 1 to 9 bar, with the fuel/air mixture at both ambient (~ 300 K) and elevated temperature.

The burner, which was supplied by Siemens Industrial Turbomachinery, is a down-scaled prototype of that used in the SGT-750 gas turbine. It is composed of three individual sectors that are arranged concentrically, a centermost pilot sector, intermediate sector and main sector. Each sector contributes a premixed fuel/air flow, while swirl elements in each sector promote flame stabilization and recirculation in the combustion region. There are dedicated fuel feeds allowing for localized setting of fuel/air mixture at each of the sectors. The central pilot sector of the burner was separable from the full burner assembly and was examined in detail.

Information was generated regarding the use of syngas to fuel the burner. This information is intended to be used for the validation of CFD models of the experiments, including optimization of reduced chemical kinetic mechanisms for the specific fuels. Laminar flame speed measurements were made for several syngas fuel candidates from which the high-hydrogen syngas fuel was selected. Burner performance at the lean stability limit was examined using the fuels of interest. It was found that increasing the fuel/air ratio in the central pilot sector improved the lean limit onset of flame extinction up to the point that the central pilot extinguished. Optimization of the burner nitrogen oxides (NOx) emission by fuel partitioning among the three sectors was performed. The response in emission level with fuel/air ratio was not universal among the fuels tested.

The largest portion of work in this thesis is the visualization of the burner combustion field by laser diagnostic methods. The flame shape was imaged by the PLIF of the OH radical distribution. PLIF imaging of the central pilot sector was recorded for atmospheric and elevated pressure for iterations of inlet air temperature, fuel type and equivalence ratio. When comparing the OH-LIF distribution for various fuels and pressures it was found that equivalence ratio had the greatest effect on the distribution of OH signal from the exit of the central pilot sector. Lean equivalence ratios showed a diffuse signal typical of the post combustion region. Near stoichiometric equivalence ratios yielded a distribution having a clearly defined inner edge indicating combustion occurring outside of the pilot sector. At rich equivalence ratios the OH signal was lifted away from the pilot burner exit. Comparison of OH-PLIF and chemiluminescence signal for methane combustion supported the characterization that the pilot sector efflux varied from post combustion to attached and then lifted flame in conjunction with the increase in equivalence ratio from lean to rich. OH-PLIF imaging was collected for staging of fuel to all three sectors of the burner at atmospheric pressure. The flow field in the combustion region produced by the full burner was visualized using PIV for each of fuels of interest, illustrating the recirculation zone. Finally the OH-LIF distribution was imaged for the combustion region of the entire burner at elevated pressure during operation at a single equivalence ratio with various dilutions of natural gas. There was little discernible change in flame shape as the pressure was changed from 3, 4.5 and 6 bar and energy content was changed from 30, 40 and 45 MJ/m^3 Wobbe index.

Abstract (Swedish)

Popular Abstract in English

There is a natural tendency to take advantage of the most convenient energy source available. In this usage, convenient may mean most abundant, easiest to harvest or simplest to handle. Convenient will also correspond to cheapest in many instances. However, some consider it a responsibility to think of the long term costs of the energy we use. No form of energy is free from environmental impact. The question is, are we getting enough value from the energy we use and are we properly evaluating the secondary penalties we pay for that energy?



Combustion is universally linked to human cultural development. For the last one hundred and fifty years coal, natural gas and liquid petroleum products- the big three fossil fuels, have been the power behind progress. These fuels are burned for heating, transportation and to generate electricity; they function quite well in the intended role. But not all energy sources are equal. For instance, coal which is cheap and abundant, is considerably dirtier when burnt than natural gas.



Together with the ideal combustion products which are water, CO2 and energy, combustion produces unintended pollutants. Some of these are resultant from a presence in the fuel, such as heavy metals or sulphur compounds in coal and oil. Other pollutants are formed by the process of high temperature combustion, specifically nitrogen oxides (NOx), or by incomplete combustion such as carbon monoxide and soot. Research into reduction of these pollutants has progressed for several decades. Attention is increasingly directed to the importance of the combustion product CO2, as there is concern that elevating CO2 levels in the atmosphere will affect the environment adversely. In response, there is a drive to find reduced impact and CO2 neutral alternatives to the energy sources that permit the current standard of living.



It would be impossible to completely replace combustion based energy conversion in the short term, and so efforts are underway to create cleaner, more efficient combustion systems. A good example is the area of gas turbine engines for electrical generation. Companies like Siemens, General Electric, Alstom and others have met the increasingly strict regulation of pollutive emissions. New design strategies that ensure better blending of fuel and air, and operation at lower combustion temperatures are developing. Simultaneously, there is interest in learning to operate these cleaner burning engines on fuels other than natural gas, e.g, carbon neutral fuels synthesized from biomass, coal gasification and low energy content gases.



How are laser diagnostics involved in this development cycle? New ideas in burner designs are being combined with alternative fuels that may not burn in the exactly the same way as natural gas. To understand how these interact it is useful to measure various aspects of the flame. Some laser-based measurements are aimed at recording a value, such as an amount of soot, the concentration of a certain chemical at a point in the flame, or a temperature. Laser-based techniques can measure these values without interfering with the local flame conditions as could a conventional probe. As a result, the measured value should be more representative of the true conditions. Laser-based techniques also enable measurements in harsh environments that are incompatible with direct sampling. One of the more useful applications of lasers in combustion research is for taking images.



Combustion in an engine is turbulent featuring a constant fluctuation of the flame edge. An image from a traditional camera will show the edge of the flame blurred by the amount of time that the camera was collecting emission, and the range of depths in the flame. If a two dimensional sheet of laser light of a selected wavelength passes through the flame, chemicals in the flame can give off photons, known as fluorescence. The image of this laser-induced fluorescence shows a slice through the flame frozen in time with no blurring. Depending on how these images are formed, certain qualities of the flame may be found: where the combustion reaction occurs, how quickly the flame edge fluctuates, how well the fuel mixes with the air before burning and so on.



The greater strength of laser-induced fluorescence imaging is realized when it is combined with computer modeling of the combustion process. By comparing the real images of the combustion region with the computer model predictions, the quality of the model can be verified and improved as needed. These more accurate models may then influence the design of next generation combustion machinery. It is for this reason that the thesis work was undertaken: to generate a set of data for a prototype burner operating at defined conditions while burning several fuels that typify renewable or reduced carbon deficit fuels. The burner was operated at several pressures from atmospheric pressure up to nine atmospheres. The result is a wealth of data describing the flame shape as characterized by laser-induced fluorescence, information regarding fuel combustion properties, burner temperatures and emissions data for documented operating conditions.

• author: Whiddon, Ronald
• supervisor: Marcus Aldén
• opponent: Grisch, Frédéric, INSA-Rouen, CORIA, Université de Rouen, France
• organization: Combustion Physics
• publishing date: 2014
• type: Thesis
• publication status: published
• subject: Atom and Molecular Physics and Optics
• keywords: OH, Laser Diagnostics, Elevated Pressure, Syngas, DLE, NOx, Combustion, Gas Turbine, PIV, PLIF, Pilot, SGT-750, Fysicumarkivet A:2014:Whiddon
• pages: 155 pages
• defense location: Lecture hall A, Department of Physics, Professorgatan 1, Lund University Faculty of Engineering
• defense date: 2014-05-07 10:15:00
• ISBN: 978-91-7473-883-4
• language: English
• LU publication?: yes
• id: c9d5b1cf-5463-4e65-b083-978dea60957c (old id 4392448)
• date added to LUP: 2016-04-04 12:12:46
• date last changed: 2018-11-21 21:09:40