A coordination study is validated in two steps: a few curve points (pickup, a middle point and the maximum fault) are calculated by hand and compared with the times the simulation program gives. If the difference is small, the model is loaded correctly; if it is large, there is almost always a data error, such as a wrong CT ratio or dial.
Short answer
Validating a protection study means checking that what the simulation program says will happen matches what was calculated and what will be programmed into the relays. It is done in three parts. First, hand-calculated trip times are compared with the program’s at three points on each curve. Then faults are simulated on buses and along circuits and it is verified who trips, in what order and with what coordination margin. Finally, the extreme cases are reviewed: weak grid, strong grid and failure of a device.
What is validated
The model and the settings. With the model, you check that impedances, transformers and sources produce fault currents consistent with those calculated separately (see the short-circuit calculation procedure). With the settings, you check that each relay responds as its curve indicates and that the curves coordinate with each other, with the chosen margin (coordination time interval).
Step 1: control points by hand
For each relay, three curve points are calculated by hand: near pickup, an intermediate value and the maximum fault in its zone. With the IEC 60255-151 normal inverse curve: t = TMS × 0.14 / ((I/Is)^0.02 − 1). The program’s times are read for the same currents. It is worth recording the three chosen currents and the result of each calculation: they are the evidence of the validation and are reused if the study is redone. The comparison immediately exposes a wrong dial, curve family or CT ratio.
Step 2: simulated faults
The program applies faults at each bus and at intervals along lines and cables, and reports the trip time of each relay that sees the fault. You check that the nearest relay trips first, that the next one backs up with the right margin and that no unrelated relay trips in a shorter time. It is the same approach described by network protection performance audits (CIGRE TB 800, public summary): applying faults along each circuit, with and without fault resistance and with equipment out of service, and comparing operation against the coordination margin.
Step 3: extreme cases
- Weak and strong grid: fault current changes, and with it times and margins.
- Fault with a device out of service or failing to operate: backup is reviewed.
- Fault with resistance: current falls and times rise.
- Variable generation: with inverters, see their contribution and the setting.
Worked example
Example data (assumed): feeder relay A (downstream), pickup 320 A, TMS 0.15; incomer relay B (upstream), pickup 600 A, TMS 0.25; both with the normal inverse curve; fault on the feeder at 2,400 A; expected coordination interval at least 0.30 s.
1. Hand calculation at 2,400 A.
- Relay A: 0.15 × 0.14 / ((2,400/320)^0.02 − 1) = 0.021 / 0.04112 = 0.511 s.
- Relay B: 0.25 × 0.14 / ((2,400/600)^0.02 − 1) = 0.035 / 0.02812 = 1.245 s.
- Difference: 1.245 − 0.511 = 0.734 s, above 0.30 s. They coordinate.
2. Program result. The program reports 2,385 A for that fault and a time of 0.512 s for relay A. With 2,385 A the hand calculation gives 0.512 s: it matches, and the difference from the value at 2,400 A (0.511 s) is 0.3%. The model is loaded correctly.
3. A detected error. Suppose that in another run the program reports 0.64 s for relay A at 2,400 A. The difference from 0.511 s is 25%, too much to be explained by rounding. On review, the program has a 600/5 CT loaded instead of 400/5: the primary pickup is 4 A × 120 = 480 A instead of 320 A. With 480 A: 0.021 / ((2,400/480)^0.02 − 1) = 0.021 / 0.03271 = 0.642 s. The CT value is corrected and the run returns to 0.512 s.
What the simulation does not see
The program works in steady state, with phasors. It does not reproduce CT saturation with the DC component of the fault, transformer inrush or switching transients. Nor does it know about breaker opening times or relay reset time, which add to the curve time and count in the coordination margin. So simulation validates the study’s logic, and site tests validate the real relay. To see the fault as it happened, the oscillogram of an event remains the best comparison.
Record of the validation
The validation is documented in the study report: the control points, the calculated and simulated values, the difference, the extreme cases run and the corrections made. That way the reviewer sees what was done and, if the study is redone, a reference remains. See protection settings report.
Mistakes to avoid
- Accepting a result because the program «adds up» without comparing it with the hand calculation.
- Simulating only the maximum fault and not the minimum.
- Not checking the CT and VT ratios loaded in the model.
- Using the program’s curve library without checking that it is the real relay’s family.
- Forgetting the fault with a device out of service.
Checklist
- Model fault currents compared with an independent calculation.
- Three points of each curve calculated by hand and compared.
- Differences explained or corrected.
- Faults simulated on buses and along circuits.
- Weak grid, strong grid and device-out-of-service cases.
- Coordination margins verified over the whole range.
- CT and VT ratios in the model equal to those in the field.
- Validation recorded in the report.
Related guides
- Protection coordination step by step (IEEE 242)
- Time-current curve (TCC): how to read it and set pickup
- Coordination time interval (CTI)
- How a short circuit is calculated in an industrial plant
- How to analyze a fault from the oscillogram
- Lack of selectivity: several feeders trip
References
- IEC 60255-151:2009, Measuring relays and protection equipment, Part 151: Functional requirements for over/under current protection (normal inverse curve).
- CIGRE, Technical Brochure 800, network protection performance audits: the public summary was consulted; the wording and example are our own.
- Calculations and data in this guide: our own, with assumed data for the example.
Need this study?
Validating the study with fault simulation and documenting the results is part of our engineering design and studies service. We do it for projects in any country. Write to us.
Frequently asked questions
Why validate if the program already calculates the times?
Because the program calculates what it is given. A wrongly loaded dial, CT ratio or curve gives a plausible and wrong result. Comparing a few points with a hand calculation catches that kind of error before it reaches the relay.
What difference between calculation and simulation is acceptable?
A difference of tenths of a percent is explained by rounding of the fault current. Differences of several percent or more point to different data. The limit is set by the study's criterion and the tolerance declared by the relay manufacturer; there is no single value.
Which faults need to be simulated?
Three-phase, phase-to-phase and single-phase-to-ground faults on the buses and along the circuits, with the network at its highest and lowest contribution, and with one device failing to check backup. The set is defined by the scope of the study.