search

BlockResearchGroup / compas_tna

Form finding of funicular networks in compression with vertical loads using Thrust Network Analysis
https://blockresearchgroup.github.io/compas_tna
MIT License
13 stars 9 forks source link

Shouldn't we store q final as attribute for edges after vertical equilibrium? #11

Closed ricardoavelino closed 4 years ago

ricardoavelino commented 5 years ago

In other contexts after the vertical equilibrium it may be useful to keep the qs that resulted in the required height.

attr['q'] = attr['f'] / form.edge_length(u,v) or attr['q'] = f[index, 0]

Is that any reason why we don't store the 'q's?

tkmmark commented 5 years ago

I totally agree! Also, I would also argue for storing the self-weight calculated by the LoadsUpdater (at the final step) as a vertex attribute. Ditto for residual forces.

By the way Ricardo... do you mean attr['q'] = q[index, 0]

ricardoavelino commented 5 years ago

This also @tkmmark

An enhancement could be actually to allow Loads3D(True/False) since for some problems the updates on the loads don't need to occur. And in any case I guess storing the information at the end of the calculation always will always be helpful...

tkmmark commented 5 years ago

Yes it is very useful... especially when we are doing post-TNA-vertical correction with FD, so that we could simply feed one form diagram output (from TNA) to the FD method.

tomvanmele commented 4 years ago

let me know if it is indeed fixed