 |
|
|
|
|
|
At a ceremony in London on July 13th, Steve Pope is admitted to the Royal Society by Sir Martin Rees, President of the Royal Society.
|
|
|
|
Signing the Charter Book of the Royal Society |
Professor Steve Pope elected Fellow of the Royal Society
Along with 44 other distinguished scientists, professors and researchers from around the world, Steve Pope was admitted to the Royal Society at their annual Admisson Day Ceremony in July 2007, where he signed the Charter Book and the Obligation of the Fellows of the Royal Society. Election to the Royal Society is the highest honor a scientist may achieve in the United Kingdom and recognizes their contributions to science in fundamental research resulting in greater understanding.
"Professor Stephen Pope has made outstanding contributions to many aspects of turbulent flows, from physical models to efficient numerical algorithms. His "intrinsic low-dimensional manifolds" (with Maas) was a major advance in dimension-reduction methodologies for combustion chemistry; his "in-situ adaptive tabulation" algorithm produced huge improvements in computational efficiency; and his pioneering contributions to pdf methods produced a much followed approach to turbulent combustion." - from the Royal Society website
About his Research:
Turbulent combustion, the process of fuel burning in a turbulent air flow, accounts for the bulk of the global utilization of fossil fuels in power generation, transportation, heating and other applications. It also accounts for more than half of the greenhouse gas emissions. In view of the projections of increased use of fossil fuels in the coming decades, there is strong motivation to develop a better fundamental understanding of turbulent combustion to facilitate the design of improved combustion equipment, yielding improved efficiency, and the capability for carbon capture and sequestration.
There are many aspects of turbulent combustion that make it a formidable scientific challenge. The flow may be multi-phase (containing coal particles, liquid fuel spray, soot and particulates); it is three dimensional and unsteady with length and time scales varying over 10 orders of magnitude, and with large turbulent fluctuations; and 50 to 1,000 reactive gaseous species may be involved.
In Steve Pope’s group, several computational methodologies have been developed to facilitate the accurate modeling of turbulent combustion. Probability density function (PDF) methods have been developed to treat turbulence-combustion interactions. In PDF methods, a particle method is used to solve the modeled transport equation for relevant fluid properties (e.g., species, energy, velocity). The great advantage of this approach is that chemical reactions are treated exactly, even in the presence of large turbulent fluctuations in the properties. The ability of PDF methods to represent fully large turbulent fluctuations is illustrated in Figure 1. On the left is shown a scatter plot of species concentrations obtained in a turbulent flame with strong turbulence-chemistry interactions, and on the right is the corresponding scatter plot from the PDF calculations.
|
|
|
Fig.1) Scatter plot of the mass fraction of CO against mixture fraction in a piloted jet flame: left, experimental data; right, PDF calculation. |
A major challenge in modeling turbulent combustion is dealing with the complexity and non-linearity of hydrocarbon combustion. To alleviate these difficulties, several “dimension reduction” methodologies have been developed to allow the combustion to be accurately represented by many fewer variables than the number of chemical species involved. Recently (in collaboration with Guckenheimer and Vladimirsky in Mathematics), Zhuyin Ren and Steve Pope developed a powerful and general dimension-reduction method called ICE-PIC (Invariant Constrained-equilibrium Edge manifold/Pre-Image Curve). Figure 2 shows results from a test calculation comparing the errors incurred by different dimension reduction strategies. It may be seen that ICE-PIC with 7 degrees of freedom yields smaller errors that other methods even using 12 degrees of freedom.
|
Fig. 2) Dimension-reduction errors for an auto ignition test problem, comparing ICE-PIC with 7 degrees of freedom to other methods (RCCE with 7 degrees of freedom and QSSA with 12 degrees of freedom). |
|
A complimentary strategy for reducing the computational burden of combustion chemistry is a storage/retrieval algorithm. In a typical turbulent combustion calculation, the chemistry needs to be treated 1010 times (e.g., for each of 106 grid nodes or particles on each of 104 time steps). To treat the chemistry once, termed a “direct evaluation” may take just 104 microseconds (µs), but then doing so1010 times takes over 3 years. Steve Pope developed the method of in situ adaptive tabulation (ISAT) in which a table of chemistry calculations is built up, as needed, during a turbulent combustion computation. Each time the result of the chemistry calculation is needed (termed a “query”), the ISAT table is queried to see if it already contains the needed information. If so, it is “retrieved”, at very low computational cost (e.g., 10µs). Otherwise the direct evaluation is performed and the result “added” to the table. Figure 3 shows the “adds” and “retrieves” for a test calculation, and the average time for different events. The average query time is less than 10µs (a factor of 1,000 less than the direct evaluation time), thus reducing the total time needed from 3 years to less than 1 day.
|
Fig. 3) ISAT performance for a test case using a 16-species chemical mechanism for methane combustion: left, numbers of “adds” and “retrieves” against number of “queries”; right, average CPU time required for different events. |
|
Steve Pope has worked with Fluent/Ansys to incorporate PDF methods and ISAT in the market-leader Fluent CFD code. This code is used extensively in industry, and also in Cornell courses (e.g., MAE 423 Intermediate Fluid Dynamics; MAE 643 Computational Combustion).
|
|
|
