Different solvers in FreeKiteSim

[x.g.ge...@gmail.com]

Dear Uwe,

When going through the code I found that there exists the option for selecting one of two different solvers of Assimulo:

- RADAU5

- IDA

I found that the IDA of Assimulo is actually a binding to Sundials' IDA.

Can you tell me something about the difference in speed between the two solver algorithms?

Additionally: is it true that the RADAU5 algorithm determines the Jacobian numerically?

Best,

Xander Gerrmann

[Uwe Fechner]

Dear Xander,

I am using Assimulo because the solvers I tried before, mainly from Simulink did not work well

with the very stiff differential equation system that we have if we model a real kite power system.

I found that RADAU5 is the best choice for complex simulations, but in more simple situations

(no wind turbulence, not so many particles) IDA can be faster.

Both RADAU and IDA determine the Jacobian numerically.

If I would have to rewrite this code I would use Julia and the package DifferentialEquations.jl:

(See: http://docs.juliadiffeq.org/latest/ )

Much more powerful due to automatic differentiation, very fast and convenient to use.

But 2014 Assimulo was the best choice. Still a good choice for learning.

Uwe

--
You received this message because you are subscribed to the Google Groups "Free KiteSim" group.
To unsubscribe from this group and stop receiving emails from it, send an email to free-kitesim...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.