Cable selection algorithms

Checking and selecting a cable, by voltage class

Four algorithms, because the checks are genuinely different. A low-voltage circuit is decided by ampacity, voltage drop, device coordination, k²S² and loop impedance. An MV circuit adds the system earth-fault category that sets the rated voltage, and checks the screen separately from the conductor. An HV cable system has to settle its bonding before it can be rated at all. A d.c. circuit has its own voltage and current criteria that come from the array, not from a load.

Each algorithm is a flowchart plus the same content as a table · every decision node cited to a clause

Pick the algorithm by the voltage of the circuit you are checking. All four end the same way: the largest cross-section that any single criterion demands is the one that gets installed, and the criterion that forced it is worth writing on the cable schedule — it tells the next engineer what the circuit is limited by.

The four algorithms

What all four have in common

Every one of them is a loop, not a list. You take a trial cross-section, test it against each criterion in turn, and go back to the trial size whenever a test fails. That structure is why a cable schedule cannot be produced one column at a time: raising a size to satisfy voltage drop changes the grouping factor, and changing the installation method changes the ampacity that the short-circuit check was resting on.

Three habits do most of the work across all four:

  • Rate the worst section, not the typical one. IEC 60364-5-52, 523.8 says this explicitly for LV, and the same physics applies at every voltage: a route that runs through a duct bank for 15 m is a duct bank cable for its whole length. Only a wall crossing of 0,35 m or less is disregarded.
  • Use the conductor temperature, not the ambient. Resistivity at operating temperature is 1,25 × the value at 20 °C in the LV voltage-drop formula of Annex G; d.c. cables next to PV modules run some 40 K above ambient. Both are in the standards, and both are routinely skipped.
  • Record which criterion governed. It costs one column and it is the only thing that lets a reviewer check the schedule without redoing it.