Tiza Group

Home / Guides / Codes and grid connection: US, UK, CA, AU

IEEE 1584-2018 Calculation: Inputs, Steps, Checks

IEEE 1584-2018 calculation explained: the inputs the model needs, the order of the steps and the checks that catch a wrong arc flash result before labels.

By Shadai García, CEO and protection engineer at Tiza Group.

480 V bus, bolted 18.0 kAfeeder breakerCT50/51feeder relayarcing 14.5 kAreduced 12.3 kA0.20 s at 14.5 kA0.32 s at 12.3 kAPowerMeasure / tripbolted 18.0, arcing 14.5, reduced 12.3 kA0.20 s at 14.5 kA, 0.32 s at 12.3 kA14.5 x 0.20 = 2.90; 12.3 x 0.32 = 3.94ratio 1.36: case 2 has 36% more energy5.0 cal/cm2 in case 1 gives about 6.8t (s)I (kA)1015200.10.20.5feeder relay curvecase 1: 14.5 kAcase 2: 12.3 kAcase 12.90 kA.scase 23.94 kA.sfirst-order check, not the full calculationCalculationBoltedArcingClearingEnergyWorst case© tizagrp.com480 V bus, bolted 18.0 kAfeeder breakerCT50/51feeder relayarcing 14.5 kAreduced 12.3 kA0.20 s at 14.5 kA0.32 s at 12.3 kAPowerMeasure / tripbolted 18.0, arcing 14.5, reduced 12.3 kA0.20 s at 14.5 kA, 0.32 s at 12.3 kA14.5 x 0.20 = 2.90; 12.3 x 0.32 = 3.94ratio 1.36: case 2 has 36% more energy5.0 cal/cm2 in case 1 gives about 6.8t (s)I (kA)1015200.10.20.5feeder relay curvecase 1: 14.5 kAcase 2: 12.3 kAcase 12.90 kA.scase 23.94 kA.sfirst-order check, not the full calculationCalculationBoltedArcingClearingEnergyWorst case© tizagrp.com
IEEE 1584-2018 plausibility check: a 480 V switchboard with an 18.0 kA bolted fault, the feeder relay with its CT and trip, and the clearing times of 0.20 s at 14.5 kA arcing current and 0.32 s at 12.3 kA reduced arcing current, with 2.90 and 3.94 kA seconds and a ratio of 1.36. Location in the system and operating principle. Illustrative: real values come from the nameplate and the manufacturer's manual.
Short answer

The IEEE 1584-2018 calculation turns a bolted fault current and a clearing time into incident energy at a working distance. The model needs the system voltage, the equipment type, its electrode configuration, the gap, the enclosure size and the working distance. It evaluates an arcing current and a reduced arcing current, takes the clearing time of each from the protective device, and reports the worse result. Most errors come from the data and the clearing time, not from the arithmetic.

The IEEE 1584-2018 calculation turns a short-circuit result into incident energy. Below are what the model needs, the order in which the steps run and the checks a reviewer can make without running the software.

The method described is the one of IEEE 1584-2018. IEEE 1584.1-2022 is a separate guide on specifying and performing the assessment.

The wider framework is in the US electrical codes and grid interconnection guide. For how the result feeds the workplace rules, see NFPA 70E 2024: what engineers need to know.

Short answer

The IEEE 1584-2018 calculation takes a bolted fault current, a system voltage and the geometry of the equipment, estimates the arcing current, finds how long the protection takes to clear that current and gives the incident energy at the working distance. It repeats the run with a reduced arcing current and keeps the worse case. It covers three-phase systems from 208 V to 15 kV. The two inputs that move the answer most are the clearing time and the electrode configuration.

What the model needs

Think of the inputs in four groups:

  • System: nominal voltage, bolted three-phase fault current at the equipment bus and, for the clearing time, the protective device that sees the fault.
  • Equipment: the electrode configuration, the gap between conductors and the enclosure height, width and depth. The configuration records whether conductors are vertical or horizontal, in a box or in open air, with or without a barrier.
  • Person: the working distance from the arc to the worker’s face and chest.
  • Time: the clearing time of the device that opens for the arcing current, plus any intentional delay.

The five configurations are named and compared in our guide on arc flash studies. The enclosure matters because the 2018 model corrects for box size: the same fault gives different energy in a shallow cabinet and in a deep switchgear cell. Entering the nameplate dimensions, not a generic size, is part of the job. The next guide in the series lists how to gather them: arc flash study data collection checklist.

The order of the steps

  1. Bolted fault current at the bus, from the short-circuit study. See the ANSI short-circuit four-step method. Use the maximum credible source for the highest arcing current, and think about the minimum source too, because the lowest current can be the slowest to clear.
  2. Arcing current. The model estimates it from the bolted current, voltage, gap and configuration. It is lower than the bolted current. The guide on arc resistance and arcing fault current explains why.
  3. Reduced arcing current. A second value that allows for the variation in the arc. Both go to the next step.
  4. Clearing time for each of the two currents. Read it from the time-current curve of the device that opens, with real settings. The curve is not the same at the two currents.
  5. Incident energy at the working distance, and the arc flash boundary, the distance at which the energy falls to 1.2 cal/cm² (5.0 J/cm²), the onset of a second-degree burn.
  6. Report the worse case for each piece of equipment, with the inputs listed.

The model equations are in the calculation clauses of the standard. If your software implements them, the check below tells you whether its output is plausible.

Where results go wrong

The calculation is rarely the weak point. The inputs are:

  • A fault current taken from an old study, before a transformer or a utility source changed.
  • A clearing time read from the wrong curve, or from settings that are not the ones loaded in the relay.
  • A configuration chosen by habit. A horizontal-conductor box in a drawing marked vertical changes the result.
  • An enclosure size left at the software default.
  • A working distance that does not match how the worker really stands.

For reducing the result once you trust it, see how to reduce arc flash incident energy.

Worked example: a plausibility check on the clearing time

This example uses illustrative data, not from any project. It is a first-order check, not a replacement for the calculation. A 480 V switchboard has the following study values:

  • Bolted fault current: 18.0 kA.
  • Arcing current from the model: 14.5 kA.
  • Reduced arcing current from the model: 12.3 kA.

The feeder relay curve, read at the real setting, gives these clearing times:

  • At 14.5 kA, the relay clears in 0.20 s.
  • At 12.3 kA, it clears in 0.32 s.

For the same voltage and geometry, arc energy grows with the arcing current multiplied by the time. As a rough indicator:

  1. Case 1: 14.5 kA × 0.20 s = 2.90 kA·s.
  2. Case 2: 12.3 kA × 0.32 s = 3.94 kA·s.
  3. Ratio: 3.94 / 2.90 = 1.36.

Case 2 has about 36% more arc energy even though the current is lower, because the relay is slower. If the full model returns 5.0 cal/cm² for case 1, then case 2 should land near 6.8 cal/cm² (5.0 × 1.36). A result for case 2 far below that, such as 4.0, deserves a look at the curve reading. A result far above it may mean a different electrode configuration or a typo in the time.

The check also shows why the study reports the worse case, and why a setting change at the relay changes the label. For the arithmetic of the relay curve itself, see protection coordination step by step.

Checklist for the engineer

  • Confirm the standard edition used: IEEE 1584-2018, not the earlier edition.
  • Confirm the equipment voltage is within 208 V to 15 kV, or that the report states the method used outside it.
  • Verify the bolted fault current is from the current system model, with the sources that can run in parallel.
  • Verify the electrode configuration and enclosure dimensions against nameplates or photos, not defaults.
  • Verify the working distance for each task.
  • Verify the clearing time at both arcing currents, using the settings loaded in the device.
  • Check that the report lists the inputs next to every result.
  • Run a first-order check, as in the example, on at least the equipment with the highest energy.
  • Keep the software version and the model data with the report.

Who signs and seals

The study is engineering work, and in the United States it is normally reviewed and stamped by a professional engineer licensed in the state of the project. Tiza delivers the study ready for your engineer of record to review and stamp. Tiza does not sign, seal or certify work in the United States.

How Tiza can help

If you need the incident energy calculated for a plant, Tiza runs power-system studies remotely and delivers the results in English, ready for your engineer of record to review and stamp. Contact us with the single-line diagram and the relay settings, and we will scope it with you.

References

  • IEEE 1584-2018, IEEE Guide for Performing Arc-Flash Hazard Calculations.
  • IEEE 1584.1-2022, guide for the specification and performance of arc flash hazard assessments.
  • NFPA 70E, 2024 edition, section 130.5.

Figure: the calculation chain from bolted current to incident energy, with the arcing current and the reduced arcing current taking separate paths and the worse result kept.

Frequently asked questions

What does an IEEE 1584-2018 calculation need as input?

System voltage, bolted three-phase fault current at the equipment, the electrode configuration, conductor gap, enclosure dimensions, working distance and the clearing time of the protective device. The clearing time is read from the device's time-current curve at the arcing current, which is lower than the bolted current.

Why does the study use two arcing currents?

Arc current varies from event to event. The standard asks for a second, reduced arcing current to account for that. Protective devices often act more slowly at the lower current, so the reduced case can give the higher incident energy. The study reports the worse of the two.

Does IEEE 1584-2018 apply above 15 kV?

No. The model covers three-phase systems from 208 V to 15 kV. For higher voltages, the report must state which other method it used and why. Do not extend the model beyond its range without saying so.

WhatsApp