Parallel dynamic grid refinement for industrial applications

This paper covers a description of the algorithms and data-structures used 
\nin the parallel DLR TAU-Code Adaptation tool.
\nAttention is given to the basic algorithms of the parallel anisotropic 
\nrefinement and de-refinement, as well as the data-structures for 
\ndistributed grids and grid hierarchy. The parallelization issues for 
\ndistributed data are discussed along with the dynamic repartitioning and
\nthe load-balancing algorithm. 
\n
\nTo demonstrate the dynamic character of the TAU-Code adaptation, we choose a 
\nsimulation for a complex aircraft configuration and perform a parallel CFD 
\ncalculation using varying numbers of CPUs. The obtained results show good
\nparallel scalability with respect to memory and CPU-time consumption.