310 likes | 406 Vues
This research paper explores techniques for simulating physical systems efficiently using inlined implicit Runge-Kutta algorithms. It discusses the advantages and challenges of employing explicit and implicit algorithms, and introduces Inline-Integration as a method to merge integration algorithms with models. The paper also delves into Step-Size Control to optimize computational costs while maintaining accuracy. Various embedding methods like Alt. HW-SDIRK and Lobatto IIIC(6) are examined for their accuracy and stability. Numerical experiments are conducted to test the algorithms on benchmark ODEs. An application scenario involving real-time embedded control systems is presented, focusing on missile dynamics simulation for trajectory shaping. The study concludes by addressing potential complexities and suggesting inlining as a solution. For more details, read on.
E N D
Efficient Simulation of Physical System Models Using Inlined Implicit Runge-Kutta Algorithms Vicha Treeaporn Department of Electrical & Computer Engineering The University of Arizona Tucson, Arizona 85721 U.S.A
Topics • Introduction • Techniques for Simulation • Results • An Application
Introduction • Stiffness • Widely varying eigenvalues • Explicit algorithms • Straightforward to implement • Step size limited by numerical stability • Implicit algorithms • More difficult to implement • Additional computational load • Needed to simulate stiff systems • May use larger step sizes
Inline-Integration • Merges the integration algorithm with the model • Eliminates differential equations • Results in difference equations (∆Es) • Easily implement implicit algorithms • Circuit example inlining Rad3
Inlined with Rad3 Evaluate at Rad3 time instants Eliminate derivatives Integrator equations
Sorting • 10 equations immediately causalized • Need to perform tearing • Make assumptions about variables being ‘known’
Tearing Tearing variable Residual Eq.
Tearing Residual Eq. #2 Tearing variable #2
Tearing • Completely causalized equations • 2 iteration variables, vc and i1 • Could use this set of equations for simulation • Want step-size control
Step-Size Control • Want larger step sizes • Reduce the overall computational cost • Maintain desired accuracy • Compute error estimate • Embedding method • Shares computations with original method
Step-Size Control • Explicit RKs • Embedding methods have been found • Implicit RKs • Difficult problem • Algorithms are compact • Can find embedding methods using two steps • Linear polynomial approximation
HW-SDIRK Embedding • 3rd-order accurate • Behaves like an explicit method • May unnecessarily restrict step size for stiff systems • Search for an alternate embedding method
Alt. HW-SDIRK Embedding • 3rd-order accurate • Implicit method
Alt. HW-SDIRK Embedding Stability Domain Damping Plots
Lobatto IIIC(6) • No embedding method exists • Expensive to perform step size control • Can search for an embedding method
Lobatto IIIC(6) Embedding Method • 5th-order accurate • A-Stable • Large asymptotic region
Lobatto IIIC(6) Embedding Method Stability Domain Damping Plots
Numerical Experiments • Tested various algorithms with selected benchmark ODEs • Implemented in Dymola/Modelica
ODE Set B Inlined with HWSDIRK and alternate error method ode15s
ODE Set B Error estimate stays near 10-3 Step size grows and shrinks appropriately
ODE Set D Inlined with Lobatto IIIC(6) ode15s
An Application • Real-Time, Limited Resources • Embedded control systems • Model Predictive • Add additional system dynamics • Simulate missile dynamics in flight for trajectory shaping • First solution is faster computer • Model may still be too complex • Try inlining