Open orso82 opened 1 year ago
@bclyons12 , about this approach to account for Ip changing as a function of time: https://github.com/ProjectTorreyPines/FUSE.jl/blob/2bc33052badd980718ffea6749ffd6c333541375/src/actors/current/qed_actor.jl#L58-L70
@jmcclena was asking if this stop-n-go method perhaps breaks the ordering of the implicit time evolution in QED?
Should we then modify QED to accept Ip as a function (and the same for Vedge)?
Ip
Vedge
In fact, should we allow for all inputs to QED to vary in time? https://github.com/ProjectTorreyPines/FUSE.jl/blob/2bc33052badd980718ffea6749ffd6c333541375/src/actors/current/qed_actor.jl#L124
For example, I could foresee that we could solve the equilibrium at time t0 and then project the evolution of gm1, gm9, etc... using automatic differentiation?
t0
gm1
gm9
What do you think?
@bclyons12 says not an issue. Should use mid-bin value for Ip
@bclyons12 , about this approach to account for Ip changing as a function of time: https://github.com/ProjectTorreyPines/FUSE.jl/blob/2bc33052badd980718ffea6749ffd6c333541375/src/actors/current/qed_actor.jl#L58-L70
@jmcclena was asking if this stop-n-go method perhaps breaks the ordering of the implicit time evolution in QED?
Should we then modify QED to accept
Ip
as a function (and the same forVedge
)?In fact, should we allow for all inputs to QED to vary in time? https://github.com/ProjectTorreyPines/FUSE.jl/blob/2bc33052badd980718ffea6749ffd6c333541375/src/actors/current/qed_actor.jl#L124
For example, I could foresee that we could solve the equilibrium at time
t0
and then project the evolution ofgm1
,gm9
, etc... using automatic differentiation?What do you think?