Converter stations are very expensive and also take more space. You need one every time you tap in/out of the transmission line, great for point to point links say Offshore windfarm to major IC, but general transmission gets tapped into and out of much more frequently. Even if today you plan for this link to be P2P from city A > city B today what happens tomorrow when someone builds a generation plant, or a new town, datacentre campus on that route?
The "efficiency" gain is debatable, you do lose less on transmission but you now have this cost of getting from DC>AC AC>DC which costs roughly 1.1-1.6% - in the grand scheme of things for most schemes any overall efficiency gain is marginal to nil.
Overall the most flexible thing to do is build AC at the highest voltage your towers/interconnect points support (and consider increasing that).
In the UK we are building more lines and converting more substations from 275KV to 400KV.
And new towers with simply higher voltage on the old cables may be cheaper than these really expensive cables. And has the advantage that you can upgrade towers piecemeal as part of your maintenance cycle.
For good reason, you can’t do that.
Electrical distribution conductors are insulated by air (and distance). If you crank up the voltage, you could have line-to-line arc flashes if you don’t increase the conductor spacing. Increasing the conductor spacing requires new towers, so…
Capacitive losses. They're worse in water (dielectric), but they're there in air lines as well.
Compatibility and isolation of grids. If grid A is 35 degrees from grid B it's very hard to couple them. Also, cascading failures tend not cross DC boundaries.
HVDC only needs one copper conductor. HVDC has no reactive power loss. HVDC has less corona effect and self-inductance to overcome so conductors can be smaller. HVDC is not subject to grid phase problems.
In addition, since HVDC cabling can be cheaper at a given voltage, the voltage can be increased which minimizes resistive losses (Ptrans = IV while Ploss =I^2R--if you double V you halve I which reduces your losses by a factor or 4).
> Chojkiewicz says her team’s modeling neglected those alternatives because their goal was simply to lay out the “nationwide potential,” of reconductoring.
They only compared it to buying new land and putting in completely new lines.
They ignored simply increasing the voltage, switching to HVDC and any solution other than "putting in whole new lines".
In particular, the fact that they just ignored HVDC is problematic. HVDC gets you not just cheaper transmission but lower losses so makes better use of what you have even if you don't immediately boost capacity.