Final Activity Report Summary - ESPHI (A European smooth particle hydrodynamics initiative)
ESPHI is a Marie Curie project that was initiated within a short circle of active scientific teams that shared a common vision of the potential that exist with Smooth particle hydrodynamics for fluid mechanics issues in the field of naval, turbomachinery, hydraulic engineering and aerospace applications. Various limited and isolated initiatives existed and ESPHI was the first European move to start networking the SPH developments to share knowledge on one hand and to confirm the potential of SPH for complex high-dynamics fluid mechanics problems. SPH represents indeed a paradigm in the field of 'Computational fluid dynamics'.
The work achieved during ESPHI has been very successful. With active ESPHI fellowships, achievements are visible.
* In modelling developments, ESPHI acted as an exchange trigger between the partners, taking profit of the competences within the Consortium and even inducing contacts form outside, to foster and refine understandings we had up to now in the field of multi-fluid, incompressible and turbulence modelling. Numerical basics have also been consolidated.
* For processing, data storage standards have been established and HDF5 IO for SPH codes is being step by step implemented in the various codes of the Consortium. ESPHI is playing a strong role of exchange platform in consolidat8ng the PARAVIEW software to post-process SPH results as no such existed yet.
* Parallelisation made a jump on the front line with the upcoming of GPUs and this technology and has been addressed. This work, disseminated within the SPH community, is the first consolidated evaluation of GPUs applied to SPH modelling. The immense value of GPUs for SPH is confirmed and opens the doors for massive high speed simulations that can be attained with difficulty by other simulation methods.
* Industrial test cases results can now be shown which was not possible after the first year. The limitations of computing power, raising the importance of parallel performances, emphasised the value of GPUs. Atop modelling test cases that enriched the potentials of SPH, some large scale simulations were addressed. EDF presents for instance a hydraulic structure case with 1 505 674 particles running on 1024 processors on a Blue Gene. VA TECH launched also multi-million simulations on a real two jets full Pelton turbine, representing hence a world "premiere".
* Dissemination has also been very active and ESPHI is seen as a reference within the SPH community. A number of publications and presentations have been made, especially by means of the ERCOFTAC SPHERIC workshops that have seen their attendance drastically increasing during ESPHI.
SPH was a rather confidential CFD technology when ESPHI started. It is now confirmed that SPH represents a breakthrough in numerical simulation. This is shown by the number of large scale National and European initiatives that now hold the keyword SPH.
The work achieved during ESPHI has been very successful. With active ESPHI fellowships, achievements are visible.
* In modelling developments, ESPHI acted as an exchange trigger between the partners, taking profit of the competences within the Consortium and even inducing contacts form outside, to foster and refine understandings we had up to now in the field of multi-fluid, incompressible and turbulence modelling. Numerical basics have also been consolidated.
* For processing, data storage standards have been established and HDF5 IO for SPH codes is being step by step implemented in the various codes of the Consortium. ESPHI is playing a strong role of exchange platform in consolidat8ng the PARAVIEW software to post-process SPH results as no such existed yet.
* Parallelisation made a jump on the front line with the upcoming of GPUs and this technology and has been addressed. This work, disseminated within the SPH community, is the first consolidated evaluation of GPUs applied to SPH modelling. The immense value of GPUs for SPH is confirmed and opens the doors for massive high speed simulations that can be attained with difficulty by other simulation methods.
* Industrial test cases results can now be shown which was not possible after the first year. The limitations of computing power, raising the importance of parallel performances, emphasised the value of GPUs. Atop modelling test cases that enriched the potentials of SPH, some large scale simulations were addressed. EDF presents for instance a hydraulic structure case with 1 505 674 particles running on 1024 processors on a Blue Gene. VA TECH launched also multi-million simulations on a real two jets full Pelton turbine, representing hence a world "premiere".
* Dissemination has also been very active and ESPHI is seen as a reference within the SPH community. A number of publications and presentations have been made, especially by means of the ERCOFTAC SPHERIC workshops that have seen their attendance drastically increasing during ESPHI.
SPH was a rather confidential CFD technology when ESPHI started. It is now confirmed that SPH represents a breakthrough in numerical simulation. This is shown by the number of large scale National and European initiatives that now hold the keyword SPH.