Tiza Group

Home / Guides / Studies and design

Common Errors in Protection Coordination Studies

The most common errors in a protection coordination study, how to catch them in a review and a checklist with worked numerical examples. We do the study.

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

34.5 kV34.5 / 13.2 kV50/51input3213.2 kV50/51feeder1three points of the study where errors creep in1 Units320 A in a relay that asks for secondaryvalues: pickup at 320 × 80 = 25,600 Acorrect: 320 / 80 = 4.0 A2 Voltage base2,000 A at 34.5 kV is 2,000 ×34.5 / 13.2 = 5,227 A at 13.2 kVplotting the point at 2,000 A underestimates it 2.6 times3 Crossing curvesat 700 A: 1.33 s and 4.53 s. At 2,400 A:0.511 s and 0.498 s: the incomer trips firstwith TMS 0.25 at the incomer: 1.245 sChainDataUnitsCurvesReview© tizagrp.com34.5 kV34.5 / 13.2 kV50/51input3213.2 kV50/51feeder1three points of the study where errors creep in1 Units320 A in a relay that asks for secondaryvalues: pickup at 320 × 80 = 25,600 Acorrect: 320 / 80 = 4.0 A2 Voltage base2,000 A at 34.5 kV is 2,000 ×34.5 / 13.2 = 5,227 A at 13.2 kVplotting the point at 2,000 A underestimates it 2.6 times3 Crossing curvesat 700 A: 1.33 s and 4.53 s. At 2,400 A:0.511 s and 0.498 s: the incomer trips firstwith TMS 0.25 at the incomer: 1.245 sChainDataUnitsCurvesReview© tizagrp.com
Three typical coordination-study errors with their figures: pickup units, voltage base and crossing curves. Location in the system and operating principle. Illustrative: real values come from the nameplate and the manufacturer's manual.
Short answer

Most errors in a coordination study are not in the theory but in the data and units: a wrongly applied CT ratio, a mixed voltage base, a curve that crosses another at maximum fault, or an assumed value that nobody confirmed. They are caught with a systematic review of data, units, curves and faults.

Short answer

The most frequent errors in a coordination study are data and unit errors, not theory: a wrongly applied CT ratio, a current on the wrong voltage base, a curve that crosses another at maximum fault, a transformer with its connection modeled wrongly, or an assumed value that was never confirmed. They are caught with an orderly review on five fronts: data, units, curves, faults and documentation. The list follows the spirit of the network protection audits CIGRE describes: apply faults, compare operation with the margin and point out whatever does not respond as expected.

Front 1: input data

  • Impedances and ratings of the transformer, the network and the cables: if they come from an old drawing, they do not match what is installed.
  • Real CT and VT ratios, not what the design says.
  • Current network operator settings: they must be the current ones, not those from two years ago.
  • Maximum and minimum faults with the network in the right conditions. See maximum and minimum short-circuit current.
  • Assumed data: it must be marked, with someone responsible for confirming it. See the data for the study.

Front 2: units and bases

Primary or secondary, amperes or per unit, one voltage base or another. Every current in a time-current diagram must be on the same base, and the diagram must say which.

Front 3: curves

Curves that cross, devices that do not coordinate at maximum fault, instantaneous elements that pass above the downstream device, damage curves that are not respected, pickups below load. To read them, see the time-current diagram and the damage curves.

Front 4: faults and cases

Which faults were run (three-phase, phase-to-phase, ground), at which points, with which network conditions and with which device out of service. See validating the study with fault simulation.

Front 5: documentation

If there is no report and no settings sheet per relay, no review is possible. See protection settings report.

Worked example: three typical errors

Example data (assumed): feeder relay with a desired pickup of 320 A primary, CT 400/5 (ratio 80), TMS 0.15; incomer relay with pickup 600 A, TMS 0.10; both with the IEC 60255-151 normal inverse curve, t = TMS × 0.14 / ((I/Is)^0.02 − 1); maximum fault on the feeder 2,400 A; a 2,000 A fault measured on the 34.5 kV side of a 34.5 / 13.2 kV transformer.

Error 1, units. The correct secondary pickup is 320 / 80 = 4.0 A. If 320 is loaded into a relay that expects secondary values, the relay would pick up at 320 × 80 = 25,600 A primary: it would not see any feeder fault. The control is simple: every value on the sheet must carry its unit and the ratio used.

Error 2, voltage base. A 2,000 A fault on the 34.5 kV side is equivalent, on the 13.2 kV side, to 2,000 × 34.5 / 13.2 = 5,227 A. If the point is drawn at 2,000 A on the 13.2 kV diagram, the fault is underestimated by 2.6 times and the trip times look much longer than they will be. (The conversion holds for phase-to-phase faults. For ground faults with a delta-star transformer the ratio changes; see star, delta and zigzag windings.)

Error 3, crossing curves. With the example settings, at 700 A the feeder relay takes 1.33 s and the incomer 4.53 s: plenty of margin. But at 2,400 A the feeder takes 0.511 s and the incomer 0.498 s. The incomer trips first: selectivity is lost. A single check at 700 A would have passed the study. With TMS 0.25 on the incomer relay, the time at 2,400 A rises to 1.245 s and the margin is 0.734 s.

Signs that there is an error

Some symptoms are visible in the diagram itself: a curve with a pickup almost equal to the load current, two curves nearly touching in the high-fault stretch, an instantaneous element that cuts right at the maximum fault, or a device whose trip time is much shorter than that of any other in the circuit. Others show up when comparing documents: a settings sheet that does not match the report, or settings programmed in the field that are not those of the study. And the most telling: a real trip on site whose order does not match what the study predicted. In that case, the oscillogram is the best starting point (see how to analyze a fault from the oscillogram).

How to run the review

A person other than the author reviews with the checklist and the original data. At least three points per relay are recalculated by hand. Each value on the settings sheet is checked against the report, and the report against the spreadsheet. Findings are noted with their correction, and the study is reissued with the new revision. If the study will be presented to the network operator, see the EACP guidelines.

Review checklist

  • Impedances, ratings and CT and VT ratios match what is installed.
  • Network operator settings are current.
  • Assumed data marked and with an owner.
  • Maximum and minimum faults under the right conditions.
  • All currents on one voltage base, noted.
  • Every sheet value with its unit and the ratio used.
  • Curves reviewed over the whole range, with no crossings.
  • Instantaneous elements above the downstream fault and above inrush.
  • Damage curves of cables, transformers and motors respected.
  • Faults simulated with weak grid, strong grid and a device out of service.
  • Three points per relay recalculated by hand.
  • Report and settings sheet per relay, with revision and author.

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 examples are our own.
  • Calculations and data in this guide: our own, with assumed data for the example.

Need this study?

We review coordination studies, yours or a third party’s, with this checklist and recalculate the critical points, as part of our engineering design and studies service. We do it for projects in any country. Write to us.

Frequently asked questions

What is the most frequent error in a coordination study?

Data and unit errors: a wrongly applied CT ratio, primary values loaded as secondary, currents on different voltage bases, or impedances that do not match the installed equipment. They are simple errors, which is why a review with a checklist is worthwhile.

How do you detect that two curves cross?

You calculate each relay's time at several currents, including the maximum fault of the zone, and compare. At one current there may be margin and at another there may not. A curve program shows it, but it can also be checked by hand with the curve formula.

Who should review a study?

Someone other than the author, with the checklist and the original data at hand. An independent review finds errors the author no longer sees, because the author repeats his or her own assumptions.

WhatsApp