Prime 3.0 can't solve BVP with units?

[ViktorKorobov]

Prime 3.0 can't solve BVP with units?

Without units - no problem:

With units - error:

Viktor

[Werner_E]

C:=Odesolve..., not C(x):=Odesolve...!

Unfortunately this does not solve the problem.

[ViktorKorobov]

Thank you Werner, but the problem remains

[Werner_E]

Yes, I already relaized that. It seems to be a bug in Prime and has to do with the units in combination with that kind of boundary conditions.

If I replace one of your conditions with a more "standard" one, Prime seems to be happy. Of course thats not a solution to your problem, though.

[ViktorKorobov]

Yes, I am also considering this option. However, the physical meaning of my task is distorted. This is the Cauchy problem, and not the boundary value problem ...

[HarveyHensley]

Viktor,

This problem is discussed at this link. It's a problem caused by sbval not attaching the units to the estimated initial condition. If you run sbval separately as Josue eventually did and then reattaching the unit to the new initial condition, it works.

Also, see here for other units problems with odesolve. It explains why odesolve sometimes works with units and sometime not. When it doesn't work, try using Rkadapt, line solver routine.

Harvey

[ViktorKorobov]

Thank you Harvey.

Viktor