High-voltage transmission

Why is electricity pushed to hundreds of thousands of volts only to be dropped again at the far end?

A power station generates at perhaps twenty thousand volts. A transformer immediately raises it to four hundred thousand. It travels a few hundred kilometres, and then a chain of substations patiently undoes the whole thing, back down to the two hundred and thirty volts at your wall. Two expensive conversions, added so that the electricity spends its journey at a voltage nothing in the network actually wants.

The stock answer is "to reduce losses", which is true and explains nothing. Why would voltage be the lever, when we could simply use thicker wire? And what stops us going higher still?

Ask first where the loss actually occurs. The conductors have resistance, and current through resistance dissipates heat at a rate of I squared times R. Notice what that expression does not contain: voltage. The wire has no idea what potential it sits at. It only feels the current passing through it.

So the loss is a current problem, not a voltage problem, and the whole scheme is indirect. But the delivered power is fixed by the customers — a city wants what a city wants. Power is voltage times current, so for a fixed delivery, raising the voltage lowers the current in exact proportion. Halve the current and the heat falls to a quarter, because of that square.

Now combine the two. Loss is proportional to the square of the current, and the current is inversely proportional to the voltage, so the loss falls as the square of the voltage. Doubling the transmission voltage cuts the wasted heat to a quarter for the same power delivered down the same wire. That is the entire trick, and it is worth noticing that voltage never touches the physics of the loss directly. It is a way of buying a smaller current.

Which raises the objection I would want a student to raise: why not lower R instead? Use fatter conductors. You can, and utilities do — but halving the resistance halves the loss and costs you twice the aluminium, twice the weight on every tower, and stronger towers to carry it. Loss falls linearly with metal spent; it falls quadratically with voltage. Against a squared return, a linear one loses badly.

So what does voltage cost, if not metal? Clearance and insulation. Air breaks down at a finite field strength, so higher voltage means longer insulator strings, taller towers, wider corridors of land, and more demanding switchgear — a cost climbing roughly in step with voltage while the benefit climbs with its square. Hence voltages rose steadily through the twentieth century, and hence a short suburban feeder does not justify a 400 kV tower line.

Is there a ceiling? Yes, and it is not insulation. Push the field at a conductor's surface too high and the surrounding air begins to ionise — corona discharge, which wastes power and makes the hissing you hear under wet transmission lines. Engineers fight it by making the conductor effectively fatter, splitting each phase into a bundle of spaced sub-conductors. That is the practical brake on going higher.

One thread ties the arrangement together. Changing voltage this cheaply requires a transformer, and a transformer works only on a changing current. That is the substantive reason alternating current won the nineteenth-century argument — not because AC travels better, but because AC can be transformed.

Three refinements are worth keeping.

The first corrects the commonest misreading: high voltage does not help the electricity "travel further" or "arrive faster". Nothing about distance is improved. What changes is the ratio of heat wasted to power delivered, and the choice of voltage follows from how much of that ratio you are willing to buy with steel, land and insulation.

The second is that resistive heating is only the dominant loss, not the only one. There is corona, there are transformer losses at every conversion, and on long alternating-current lines there is the charging current drawn by the line's own capacitance, which does no useful work but does occupy the conductor. That last is why very long links and almost all submarine cables are built as high-voltage direct current instead. The break-even distance is genuinely argued over, but is commonly put at several hundred kilometres overhead and a few tens for cable.

The third is that quoted loss figures are for whole systems, not lines. Transmission and distribution together typically run somewhere around five to eight per cent of generated energy in a well-maintained grid — but most of that is incurred in the low-voltage distribution network, precisely where the voltage has been dropped back down. The high-voltage middle of the journey is the efficient part.

• Why long submarine links are almost always direct current, and what the converter stations cost.
• How reactive power and line charging set the real limit on how much a long AC line can carry.

Every term the collection defines is gathered in the glossary.