Parallel numerical methods for large-scale DAE systems
- Borchardt, Jürgen
2010 Mathematics Subject Classification
- 65Y05 80A30 65L05 65H10 65F50
- Systems of differential-algebraic equations, Block partitioned systems, Newton-type methods, Parallelization, Chemical process simulation, Dynamic simulation of distillation plants
For plantwide dynamic simulation in chemical process industry, parallel numerical methods using a divide and conquer strategy are considered. An approach for the numerical solution of initial value problems for large systems of differential algebraic equations (DAEs) arising from industrial applications and its realization on parallel computers with shared memory is discussed. The system is partitioned into blocks and then it is extended appropriately, such that block-structured Newton-type methods can be applied which enable the application of relaxation techniques. This approach has gained considerable speedup factors for the dynamic simulation of various large-scale distillation plants, covering systems with up to 60 000 equations.
- Computers and Chemical Engineering, vol. 25 (2001), pp. 951-961