HECToR-enabled Step Change in Turbulent Multiphase Combustion Simulation
Lead Research Organisation:
University of Southampton
Department Name: Faculty of Engineering & the Environment
Abstract
Over the past few years, we have concentrated efforts on a relatively new class of turbulent multiphase combustion, that is, turbulent combustion diluted by a liquid phase. Such diluted combustion is different from the conventional spray combustion (e.g. in diesel engines) and has applications in several low-emission high-efficiency energy systems and fire safety. Through a systematic approach and a series of jounral and conference publications, we have established a computational prototype of the phenomena, which is generic, phenomena-rich, scientifically interesting yet computationally amenable (on HEC!) and efficient. HECToR Phases 2a and 2b offer unprecedented opportunities for advanced simulations in turbulent multiphase combustion. The new Baker system, in particular, will have a peak speed of about 340 Tflops and 60 TB available, which is anticipated to consist of 44,544 cores (2x12 core chips and 32GB memory per node) giving a theoretical peak of around 340TF.We aim to conduct landmark simulations of turbulent multiphase diluted combustion on HECToR in order to make a step change in our understanding of key phenomena and potentially optimize applications of the technology. On the Baker system, we will conduct DNS of turbulent combustion diluted by evaporating droplets in order of increasing size and sophistication on up to 12,000 cores. Thereafter, the Lattice Boltzmann Method will be dynamically coupled with DNS to form a multi-scale simulation of fully resolved droplets in turbulent combustion. The results will be published in top international journals, promising to advance fundamental knowledge and impact on society and industry in the longer term.
Planned Impact
The project is a fundamental research that aims to advance the fundamental science of turbulent multiphase combustion by exploiting, in a timely way, the best capability computing service available in the UK. Generating and disseminating original, significant and fundamental findings are the main outcome during the project period. The best way to achieve research impact is to publish in top journals in the field. We aim to publish two or three papers in the top journals of the relevant fields: Journal of Fluid Mechanics, and Combustion and Flame. In the meantime, we believe our research outcome will be of sufficient fundamental interest to publish in Philosophical Transactions of the Royal Society and/or Science. In addition, two major international conferences during the project period have been selected for participation. Dissemination within the UK will benefit from the PI's roles in two on-going national HPC/HEC consortia: the Consortium on Computational Combustion for Engineering Applications (COCCFEA) and the UK Turbulence Consortium (UKTC). DNS databases will also be uploaded onto the COCCFEA website at http://www.coccfea.ac.uk/ for external access. Dissemination of results internationally will also be helped by our participation in the Spring School and Conference on Turbulent Combustion, 3-29 May, 2010, Stockholm (http://agenda.albanova.se/conferenceDisplay.py?confId=573). Over a longer term, exploitation of the research outcomes will be explored with our partners in the energy industry (E.ON Power Technology, Siemens and Rolls-Royce) and fire safety agencies (Building Research Establishment and Health and Safety Executives). The phenomena studied here are directly related to an important technology, turbulent combustion diluted by a liquid phase, which has applications in a range of zero and low-emission, high-efficiency power and energy systems as well as fire safety engineering. One application is water injection on an aeroengine to reduce NOx emissions (through reduced peak combustion temperatures), specific fuel consumption, and engine hot-section temperatures while maintaining constant thrust. Another application is in capturing the carbon emissions from combustion through burning fuels in pure oxygen instead of air - oxy-fuel combustion. A relatively new concept is the gas turbine (GT) - solid oxide fuel cell (SOFC) hybrid system, which can achieve an overall efficiency of over 70%, with a simultaneous reduction in NOx emissions. Such an hybrid system also involves turbulent combustion diluted by water. The development of these new generation low-emission high-efficiency power/energy systems requires extensive research and tests before being commercialized. Industry has increasingly exploited computational combustion tools for diagnostics and R & D. The ultimate goals are to conduct numerical experiments in place of laboratory tests and to harness the combustion processes for a wide range of applications in a predictable way in order to save the industry billions of pounds in the product developments. The results from the proposed research can inform designers directly or through other CFD tools which use the data from this study as benchmarks for tool development and validation. As an indication of potential economic and social benefits, the Aerospace America reported that between 1997 an 2007, the US airline industry improved fuel efficiency by 110%, saving 2.5 billion metric tons of green house gas, equivalent to removing 18.7 million cars from the road each year. Another indicator is that, according to Prof. Geoff Cox former Director of Fire Research Station, Building Research Establishment, the cost of fires worldwide is equivalent to 0.5 to 1.5 % of global GDP. Therefore, the fundamental research proposed could in the longer term make a big impact on the real world.
Organisations
People |
ORCID iD |
Kai Luo (Principal Investigator) |
Publications
Bruchmüller J
(2011)
Modeling the thermochemical degradation of biomass inside a fast pyrolysis fluidized bed reactor
in AIChE Journal
Bruchmüller J
(2013)
Tar formation variations during fluidised bed pyrolytic biomass conversion
in Proceedings of the Combustion Institute
Lycett-Brown D
(2015)
Droplet Collision Simulation by a Multi-Speed Lattice Boltzmann Method
in Communications in Computational Physics
Meares S
(2015)
Simultaneous planar and volume cross-LIF imaging to identify out-of-plane motion
in Proceedings of the Combustion Institute
Prasad V
(2013)
Investigation of extinction and re-ignition in piloted turbulent non-premixed methane-air flames using LES and high-speed OH-LIF
in Combustion Theory and Modelling
Prasad V
(2013)
Investigation of auto-ignition in turbulent methanol spray flames using Large Eddy Simulation
in Combustion and Flame
Xia J
(2013)
Inert-droplet and combustion effects on turbulence in a diluted diffusion flame
in Combustion and Flame
Xing J
(2021)
Large eddy simulation of Cambridge bluff-body coal (CCB2) flames with a flamelet progress variable model
in Proceedings of the Combustion Institute
Zhou L
(2015)
Large eddy simulation of spray and combustion characteristics with realistic chemistry and high-order numerical scheme under diesel engine-like conditions
in Energy Conversion and Management
Description | The project has achieved advances in several fronts in the multiphase flow without and with chemical reactions: (1) direct numerical simulation [DNS] of turbulent combustion with evaporating droplets on an unprecedented number of computing cores; (2) lattice Boltzmann method [LBM] simulation of the dynamics of fully resolved droplets with physical boundaries and interfacial properties rather than point sources; (3) discrete element method [DEM] simulation of solid particles with internal distributions of properties (such as temperature) rather than point sources; and (4) coupled simulation of discrete phases in the Lagrangian frame and gas phases in the Eulerian frame. In the early part of the project, the PI was awarded 6-month distributed CSE (dCSE) support to optimize DSTAR specifically for multi-core architectures installed at HECToR. The code optimization was performed to incorporate new and efficient parallel algorithms including more flexible domain decomposition and a hybrid MPI/OpenMP algorithm. As a result, DNS of turbulent combustion laden with evaporating droplets was conducted on computational grids up to 15363 with linear scalability up to 18,432 cores, which was unprecedented. In the realm of LBM, access to HECTOR allowed grid-refinement and the use of high-order multi-speed lattices. This, in combination with improved treatment of the collision terms and careful selection of the equation of state, led to LBM simulation of droplets collisions with more realistic physical parameters than ever before (density ratio around 1000, Weber number around 100 and Reynolds number around 30,000). As a result, more scenarios of droplets collisions observed in experiments can be reproduced. DEM of particle-laden flow was an extension of the project but is consistent with the initial objective of multiscale coupling of discrete phases in the Lagrangian frame and gas phases in the Eulerian frame. The standard DEM methodology was further developed to account for internal distributions of physical properties within particles and chemical reactions associated with biomass pyrolysis. The improved DEM method was applied to simulate bio-fuel production in a fluidised bed, in which 600,711 silica sand particles were in the bed region and 37,000 biomass particles were injected every second. Multi-step, Arrhenius-type reaction mechanisms were employed for the pyrolysis of the three main components of biomass: cellulose, hemicellulose and lignin. In terms of the complexity of physical phenomena studied and the scale of computation, the project has achieved and in some aspects far exceeded the original objectives. The project also led to collaboration with research groups in the UK, Switzerland and USA, which was not planned in advance. |
Exploitation Route | 1. The research has created new knowldge in the fields of multiphase flow, turbulent combustion and bioenergy, which has been widely published and well cited. It will continue to inspire research work in these fields. 2. The project has further developed two computer codes (for DNS and DEM) and created a new one (for LBM). These codes will be exploited in follow-on research. 3. The fundametal knowledge and software created can be jointly exploited with industry. |
Sectors | Aerospace Defence and Marine Agriculture Food and Drink Chemicals Energy Environment Manufacturing including Industrial Biotechology Transport |
Description | Computational Science and Engineering: Software Flagship Project Call |
Amount | £513,863 (GBP) |
Funding ID | EP/P022243/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 05/2017 |
End | 05/2020 |
Description | EPSRC Standard Grant |
Amount | £67,872 (GBP) |
Funding ID | EP/J016381/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 09/2012 |
End | 09/2015 |
Description | EPSRC Standard Grant |
Amount | £580,960 (GBP) |
Funding ID | EP/J020184/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 01/2014 |
End | 09/2015 |
Description | EPSRC Standard Grant |
Amount | £49,517 (GBP) |
Funding ID | EP/J016381/2 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 01/2014 |
End | 03/2016 |
Description | EPSRC Standard Grant |
Amount | £325,834 (GGP) |
Funding ID | EP/I012605/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 04/2011 |
End | 05/2014 |
Description | EU 7th Framework Programme |
Amount | € 300,000 (EUR) |
Funding ID | No. 246772 |
Organisation | European Union |
Sector | Public |
Country | European Union (EU) |
Start | 06/2011 |
End | 06/2015 |
Description | Enhancement and Control of Turbulent Reactive Flows via Electrical Fields - A Mesoscopic Perspective |
Amount | £357,032 (GBP) |
Funding ID | EP/S012559/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 06/2019 |
End | 01/2022 |
Description | High-end Computing Consortia |
Amount | £397,424 (GBP) |
Funding ID | EP/L00030X/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 05/2013 |
End | 05/2018 |
Description | Newton International Fellowship |
Amount | £100,000 (GBP) |
Funding ID | NF110280 |
Organisation | The Royal Society |
Sector | Charity/Non Profit |
Country | United Kingdom |
Start | 02/2012 |
End | 04/2014 |
Description | UK Consortium on Mesoscale Engineering Sciences (UKCOMES) |
Amount | £331,316 (GBP) |
Funding ID | EP/R029598/1 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 05/2018 |
End | 05/2022 |