Where Short-Circuit Current Flows: Fault Paths and Protection Coordination

A short circuit is the most violent fault an electrical system experiences. Understanding it comes down to one question: which path does the current take, how large is it, and who opens first.

Fault Types and Magnitudes

Low-voltage side distribution in data centers is almost always TN-S or TN-C-S, so single line-to-ground fault currents can be enormous. That is precisely the scenario protection coordination has to handle well.

The Current Path: One Long Impedance Chain

Short-circuit current is determined by the source EMF divided by the total loop impedance. The chain runs from the grid (treated as an infinite bus in engineering terms) through the transformer (with an impedance voltage, for example 6%), onto the low-voltage bus, through cable with resistance and reactance at every meter, to the fault point.

Two intuitions carry most of the insight:

• The closer to the transformer, the larger the fault current. At the terminals of a large transformer, three-phase fault current reaches tens of kiloamps; at the rack PDU far downstream, accumulated cable impedance has cut it dramatically.
• Transformer impedance is a natural current limiter. A higher impedance percentage means lower fault current but larger operating voltage drop. It is a trade-off, not a parameter to maximize.

The DC Side: A Nastier Scenario

Battery bank short circuits are chronically underestimated, for three reasons:

• No natural current limiter exists. The battery supplies energy directly from electrochemistry with very small internal resistance, so a terminal short can drive thousands to tens of thousands of amperes.
• DC has no current zero crossing. AC fault current crosses zero a hundred times per second at 50 Hz, giving the arc natural chances to extinguish. A DC arc, once struck, can only be interrupted by the forced arc-quenching capability of the switching device. Trying to clear a DC short with an ordinary AC breaker can fail to extinguish the arc and simply destroy the breaker.
• A short across the whole bank is the most severe event; a fault inside a single cell is milder but depends on the BMS and inspections to be caught in time.

For this reason, breakers or fuses in battery DC circuits must be DC-rated devices with verified interrupting capacity, and the connecting busbars need solid insulation protection. A metal tool dropped into a battery cabinet is an incident waiting for its moment.

Three Principles of Protection Coordination

• Dependability: the protection must operate when required. Sensitivity verification ensures that even the minimum fault current under the weakest operating condition trips the protection.
• Selectivity: only the stage nearest the fault opens. Upstream and downstream devices achieve this through current-setting margins and time-grading margins, typically 0.2 to 0.3 seconds per stage. In a data center with poor selectivity, one rack-level fault trips the entire UPS, a textbook design defect.
• Speed: the faster the better. Clearing time determines the fault energy, the integral of current squared over time, which directly affects cable withstand, equipment damage, and personnel safety through arc flash.

Data centers add one special consideration: when a short occurs at the UPS output, the UPS itself limits current, typically to 2 to 3 times rated current, holding for tens to hundreds of milliseconds before it blocks output. Downstream protection downstream of a UPS must therefore be verified against the UPS current-limiting curve, not against the grid short-circuit capacity. A large share of end-level breaker selection errors trace back to exactly this point.

In short: Short-circuit current is source drive divided by total path impedance, largest at the transformer terminals and decreasing downstream. The DC side has neither a current limiter nor a zero crossing, so it demands DC-rated protective devices. Coordination is a contest of three qualities: what must trip does, what must not does not, and what trips does so quickly.