lec16 (1)

# lec16 (1) - Introduction to Simulation - Lecture 16 Methods...

This preview shows pages 1–12. Sign up to view the full content.

Introduction to Simulation - Lecture 16 Methods for Computing Periodic Steady-State - Part II Jacob White Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Outline Three Methods so far – Time integration until steady-state achieved – Finite difference methods – Shooting Methods Shooting Methods – State transition function – Sensitivity matrix – Matrix-Free Approach Spectral Methods – Galerkin and Collocation Methods
SMA-HPC ©2003 MIT Basic Definition Periodic Steady-State Basics • Suppose the system has a periodic input • Many Systems eventually respond periodically () N N input state dx t Fxt u t dt ⎛⎞ ⎜⎟ =+ ⎝⎠ ( ) 0 xt T xt f o rt += > > t T2 T 3 T

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
SMA-HPC ©2003 MIT Computing Steady State Periodic Steady-State Basics Time Integration Method • Time-Integrate Until Steady-State Achieved • Need many timepoints for lightly damped case! ( ) () () () 1 ˆˆ ˆ ( ) ll l dx t F xt ut x x tF x ul t dt =+ = + ∆+
SMA-HPC ©2003 MIT Basic Formulation Boundary-Value Problem Differential Equation Solution Periodicity Constraint () () () N Differential Equations: ii d xt Fxt dt = ( ) ( ) N Periodicity Cons traints: 0 xT x =

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
SMA-HPC ©2003 MIT Finite Difference Methods Boundary-Value Problem Nonlinear Problem ( ) () N [] () () periodicity constraint 0, input dx t F xt ut t T xT dt =+∈ = ±²³²´ Discretize with Backward-Euler ( ) ( ) 11 ˆˆ ˆ L xx t F x ut −∆ + ( ) ( ) 21 2 ˆ 2 xx t F x u t + ( ) ( ) 1 ˆ LL L t F x u L t + 1 2 ˆ ˆ ˆ FD L x x H x ⎛⎞ ⎡⎤ ⎜⎟ ⎢⎥ = ⎣⎦ ⎝⎠ # 0 = Solve Using Newton’s Method
SMA-HPC ©2003 MIT Shooting Method Boundary-Value Problem Basic Definitions ( ) () dx t F xt ut dt =+ Start with And assume x(t) is unique given x(0). D.E. defines a State-Transition Function ( ) ( ) 01 1 ,, yt t Φ≡ ( ) 0 where ( ) is the D.E. solution given y =

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
SMA-HPC ©2003 MIT Shooting Method Boundary-Value Problem Abstract Formulation Solve Use Newton’s method () ( ) ( ) ( ) 00 , 0 , 0 0 Hx x T x xT = Φ− = ±²²³ ² ²´ ( ) ,0, H Jx I x ∂Φ = ( ) ( ) 1 kk k k H Jx x x H x + −=