I have read that in Mexico there is a project to build a network of distribution lines (over 1000 km), generators and substations using direct current rather than alternating current. These are rated to provide 3000 MW.
As far as I understand this would require too many substations to compensate for losses (due Joule effect) and the physical size of the lines would be too large.
Isn't this highly inefficient and too expensive?
While I cannot speak for the specific project you are talking about (my Spanish isn't so great), DC transmission lines are most definitely feasible. High-voltage DC (HVDC) has several advantages over AC transmission, one of the more significant ones being that it allows power to flow between AC grids with different frequencies or phase angles.
The List of HVDC projects should be enough to show you that it is definitely feasible under certain conditions, as there are quite a lot of HVDC transmission lines in use.
Additional advantages include lower losses in underwater/underground transmission lines (with AC underwater lines, the capacitive losses can be quite high), and even on regular air lines because HVDC lines do not transmit reactive power. Also, if an AC transmission line is built for a given AC voltage, the insulation must withstand the peak voltage, while RMS voltage and therefore power is ~70% of that. An HVDC system utilizing the same transmission line can operate at the peak rated voltage, achieving ~40% higher power throughput.
the physical size of the lines would be too large.
This seems like you are referring to the issue with low voltage transmission, which, irrespective of AC or DC, will incur I2R losses. But this is not something inherent to DC transmission (high voltage is used instead), rather this is inherent to low voltage transmission.
Skin effect can cause problems with AC transmission: -
For an aluminium conductor at 50 Hz the skin depth is about 10 mm so if the power carrying conductors are large (as would be expected with 3000 MW) the centre of the conductor will be passing very little current and naturally the resistance of the cable would be much higher compared to a DC carrying conductor: -
To mitigate this, AC cables are designed as multiple conductors with spacings like so: -
But you also get proximity effect with two conductors spaced a little apart to minimize skin effect problems: -
Just like skin effect, proximity effect will also cause a reduction in the amount of copper (or aluminium) being used in the electron transportation so it's a balancing act to design a decent conductor for AC power transmission.
DC power transmission doesn't suffer from skin or proximity effect and can use significantly smaller conductors for transmission of current.
The down side to DC voltage transmission is you can't use regular AC transformers. Having said that a 2000 MW cross-channel DC link was used to bridge between the UK and France in 1986 so these problems are no-doubt mainly overcome. The down conversion from 270 kV DC was done by: -
The system was built with solid-state semiconductor thyristor valves from the outset. Initially these were air-cooled and used analogue control systems but in 2011 and 2012 respectively, the thyristor valves of Bipole 1 and Bipole 2 were replaced by more modern water-cooled thyristor valves and digital control systems supplied by Alstom
The domino effect with AC grid systems and dynamic loading with regions running lower and higher in frequency makes HVDC transmission the solution, not the problem.
HVDC with a flexible HVAC distribution is far more favorable to increase transmission efficiency ( albeit greater insulation and converter cost) better utilization of networks, and balance capacity with 4 new HVDC feeders in Mexico.
The cost of construction, maintenance and ownership has proven that this mix results in the lowest cost of ownership from what I recall.
I recall working in new HVDC station almost 50 yrs ago. I was a summer student with a portapac climbing scffolding 6 stories high to clean the ceilings just after construction of the Dorsey HVDC station in WInnipeg. It was designed by English Electric and is a major source of cheap hydropower. On another personal note my former colleague spent $85 in fuel driving his Mitsibishi car there. But with the growth of e-cars and overnight charging, capacity must increase dramatically if it is ever going to replace carbon-based fuels.
