Simulation language TKSL

The project is aimed at analysing direct use of the Taylor series method in finding numerical solutions of dynamic systems. The project also includes practical implementation and evaluation of the designed algorithms on different computing systems. This will enable an objective comparison of the proposed method with currently used methods. The cooperation with students and doctoral students also plays an important role. Particularly the doctoral students should find the project useful in focusing on a clearly defined professional field.

The first stage of the project consisted in designing and implementing a suitable experimental simulation system. The first version of the simulation language is already available and it can be used for solving some of the very difficult problems including stiff systems, systems with discontinuities and systems described by partial differential equations. The required accuracy is maintained by dynamically setting the order of the Taylor polynomial (more than 100) and at present an accuracy of the order of magnitude of 10^-20 is by no means exceptional.

The HPC team has created eight versions of the simulation language TKSL: TKSL/386, TKSL/ORCAD, TKSL/WINDOWS, TKSL/TRANSP, TKSL/XILINX, TKSL/STIFF, TKSL/PDE, TKSL/REAL. The accuracy and the number of equations (now the maximum number of equations is 150) is limited by the speed of the processor and therefore a gradual implementation of the simulation system on effective work stations, then on parallel systems with transputers and then on supercomputers would provide the final solution of the simulation of dynamic systems.

People

• Development of dynamic math models from basic principles
• Selection of numerical integration algorithms to secure accuracy and stability
• Special hardware design
• Special software design
• Graphical interface
• Stiff systems analyzes
• Design of control systems
• Simulation of electronic circuits
• Simulation of mechanical systems
• Partial differential equations solving
• Modelling of hydrophobic materials, solution of large problems involving contact flow. Usage of domain decomposition (FETI methods) for parallelization of numerical calculations 

Cooperation

Publikace

Applications

• Development of dynamic math models from basic principles
• Selection of numerical integration algorithms to secure accuracy and stabili ty
• Design of transputer systems
• Design of Xilinx systems