Tools › Building Systems › Lighting and Electrical › Protective Device Coordination › Method
Protective Device Coordination: Method
Open the Protective Device Coordination tool
Protective Device Coordination: Method, Curve Equations, Assumptions and Limits
What the tool does
Overcurrent devices in series are coordinated (selective) when, for a fault anywhere, the device nearest the fault on its source side opens first and the devices above it stay closed long enough for that to happen. The check is made on a time-current chart: every device has a curve of operating time against current, and each curve must lie above the curve of the device below it, by a time interval, over every current the two devices can carry together. The tool draws that chart for up to eight devices listed from source to load, computes the operating time of each at the entered fault currents, the interval between each adjacent pair, and the ranges of current where an interval is lost. It is a quick check for a small radial study. It does not replace the manufacturers' published curves or a full coordination study.
One voltage base
Devices on the two sides of a transformer carry different amperes for the same fault. As in TM 5-811-14, paragraph 4-2c(1), every current is converted to one voltage base. With the reference voltage Vref and a device at Vdev, device amperes = reference amperes × Vref / Vdev. Fault currents are entered on the reference base. Settings are entered in the device's own amperes. The ratio is the ratio of line-to-line voltages, right for three-phase faults; a line-to-line or line-to-ground fault behind a delta-wye transformer divides differently among the primary phases, which the tool does not model.
Device models
Inverse-time relay. The operating time is
t = TD (A / (Mp − 1) + B), M = I / Ipickup > 1
and the relay does not operate at or below pickup. This is the trip equation of IEEE Std C37.112-1996 (its Table 1 gives the constants) and the dependent-time law of IEC 60255-151:2009, which writes it t = TMS [k / ((G/GS)α − 1) + c] and lists six curves in its Annex A. The defaults of this tool were checked against both documents: IEEE moderately inverse A = 0.0515, B = 0.1140, p = 0.02; very inverse A = 19.61, B = 0.491, p = 2; extremely inverse A = 28.2, B = 0.1217, p = 2 (A and B in seconds; these are also curves D, E and F of the IEC standard); IEC curve A, "inverse", often called standard or normal inverse, k = 0.14, α = 0.02; curve B, very inverse, k = 13.5, α = 1; curve C, extremely inverse, k = 80, α = 2, each with c = 0. The long-time inverse curve (k = 120, α = 1) is not in IEC 60255-151:2009. It is a curve many relays offer under that name; its constants here were typed from memory and must be confirmed against the relay manual. All constants are inputs (Advanced mode), and a Custom curve takes any A, B and p. A relay may also carry a definite-time stage (a second pickup with a fixed delay) and an instantaneous stage; the operating time is the shortest of the stages that are picked up.
Time dial. TD (the IEC time multiplier setting, TMS) multiplies the whole curve, so TD = 1 is the tabulated curve. The IEC standard takes TMS = 1 as the reference setting. The IEEE standard writes its equation without a dial and says that A and B vary in proportion to the time dial, with the tabulated constants describing the curve near the middle of the dial range; it does not fix the numbers on the dial. A dial marked 0.5 to 15 or 1 to 10 on a particular relay is therefore not the TD of this equation unless the manual says so.
A manufacturer's curve family with a familiar name is not always the standard equation. Some relays use other constants under the names "U.S. inverse" or "very inverse", and some scale the time dial differently (a dial of 1 to 10 on an electromechanical relay is not the TD of this equation). Electromechanical relays follow the equation only approximately. Enter the constants and dial that the relay manual gives.
Definite-time relay. No operation below the pickup, the entered time at and above it.
Trip unit. A simple long-time / short-time / instantaneous model of a low-voltage breaker trip unit: above the long-time pickup the time follows a constant I²t line through the entered delay at six times the pickup, t = t6 (6 / M)²; at and above the short-time pickup the time is the short-time delay (flat, with no I²t ramp); at and above the instantaneous pickup it is the instantaneous time. The model is this tool's simplification. A real trip unit has a band between its minimum and maximum clearing curves, and the published curves govern.
Point table. Fuses and thermal-magnetic molded-case breakers have no public equation: their curves are the manufacturers' property. Read points from the published curve and enter them. Times between points are interpolated along straight lines on log-log scales, t = t1 (I / I1)s with s = ln(t2/t1) / ln(I2/I1). Below the first point the device does not operate. Above the last point the last time is held and the result is marked. A point may carry two times, a lower curve (a fuse's minimum melting time) and an upper curve (its total clearing time). The upper curve is used when the device is the downstream one of a pair and the lower curve when it is the upstream one, which is the unfavorable choice in both positions. TM 5-811-14, paragraph 4-2d(5), calls for the total clearing time of a downstream fuse.
What the standards say about range, tolerance and reset
Range of the curve. Neither standard promises the equation at every current. IEEE Std C37.112-1996 defines conformance from 1.5 to 20 multiples of pickup. IEC 60255-151:2009 gives the effective range of the inverse part as 2 times the setting up to a limit the manufacturer declares, which may not be below 20 times the setting, and requires a definite minimum operating time: above that limit the time no longer falls with current but stays at its value at the limit (or a separate definite-time delay takes over). The tool follows this. In Simple mode every inverse-time relay is held at its 20-multiple time for larger currents, and such a result is marked ‡; in Advanced mode the limit is an input, and 0 lets the equation run on without limit, which gives shorter times than a relay with a limit. At 30 multiples the difference is small for the very and extremely inverse curves in absolute terms but it can decide a tight interval, so take the limit from the relay manual. At the low end the IEC standard only guarantees operation somewhere between 1.0 and 1.3 times the setting, and both standards leave the curve between pickup and 1.5 to 2 multiples outside the range they specify: the tool draws the equation down to pickup, where real relays are slow and scattered. Do not rely on intervals read below about twice the pickup.
Tolerance. The curves are drawn as single lines. The IEEE standard's band of 0.85 to 1.15 times the curve is a test of whether a relay's shape may carry the standard name; it says itself that the band is not an accuracy or repeatability tolerance. Under the IEC standard the manufacturer declares a timing error at the upper limit of the curve, and the permitted error is 1.5 times that between 5 and 10 multiples of the setting and 2.5 times it between 2 and 5 multiples. Timing is thus least certain at low multiples, where the curves are also steepest. The required interval has to cover the error of both relays; the tool applies no band of its own.
Reset. The tool treats every fault as a single event starting from rest. A relay that has partly timed toward a trip and is then relieved needs time to return. An induction-disk relay turns back slowly, and both standards give a reset equation for relays that imitate it, t = TD ⋅ tr / (1 − M²) for currents below pickup, with a reset constant for each IEEE curve of about 5, 22 and 29 seconds at zero current and TD = 1. For the IEC curves A, B and C the standard leaves the reset behavior to the manufacturer, who may also provide an immediate reset or a fixed reset delay. This matters where a fault is interrupted and re-applied: with a recloser or a fuse-saving scheme downstream, with arcing or intermittent faults, and with a slow-resetting upstream relay behind a fast-resetting downstream one. The upstream relay can then accumulate travel over several shots and trip although the chart shows a comfortable interval. Both standards also describe the response to a changing current as an integration of 1/t(I) over time; the tool uses constant currents.
Intervals and flagged ranges
For an adjacent pair the interval at a current is the upstream operating time less the downstream operating time. It is printed at every entered fault current that both devices carry, against the required interval. TM 5-811-14, paragraph 4-2d(2), gives 0.3 to 0.4 second between inverse-time relays in series, made up of 0.08 second of breaker opening time (5 cycles), 0.10 second of relay overtravel and 0.22 second of safety factor; the default is the larger figure. The same paragraph (which the manual reprints from IEEE Std 242-1986) allows less with solid-state relays, which have no overtravel, with field-calibrated relays, between a relay and a downstream fuse, and between direct-acting trip units, whose curves need only not overlap.
The interval at the fault currents is not enough, because curves of different shapes can cross between them. Each pair is therefore examined over every current from just below the lower of the two pickups to the largest entered fault current that both devices carry. The range is cut at every current where either curve starts or steps (pickups, short-time and instantaneous pickups, table points), each piece is sampled at logarithmically spaced currents (about 600 over the range), and wherever the finding changes between two samples the boundary is located by bisection. A current is classed as:
crossing: both devices operate and the upstream one is as fast or faster; short interval: the interval is positive but below the required one; upstream alone: the upstream device operates at a current where the downstream device is not picked up (its pickup, on the common base, is the higher). The smallest interval in the range and the current at which it occurs are also printed. A crossing narrower than the spacing of the samples (about 1 % in current) that opens and closes between two samples can be missed.
Transformer point
In Advanced mode the tool plots the full-load current of a three-phase transformer, kVA / (√3 kV), and its magnetizing inrush point, and checks that the primary device does not operate there. TM 5-811-14, paragraph 4-6c: "8 to 12 times full-load current for 0.1 second are commonly used for coordination purposes". The default is 12 times. The transformer's through-fault (damage) curves of IEEE Std C57.109, which the primary device must lie below, are published as curves in a copyrighted standard and are not evaluated; nor are the code limits on the rating of the primary and secondary devices.
What it assumes
Every curve is a single line at the entered settings: no tolerance band, no relay overtravel, no breaker opening time, no current-transformer saturation. The required interval is the allowance for these. Devices are in series on a radial circuit and listed from source to load, and each pair of neighbors in the list is checked; devices that are not neighbors are not compared. A device carries the same fault current as the fault entry says, referred by the voltage ratio: there is no motor contribution or infeed between devices. The current is constant during the fault. Instantaneous elements operate in one entered time at any current above their pickup.
When it misleads
When the entered curve is not the device's curve: a relay whose manufacturer uses other constants or another dial scale, a fuse read coarsely from a small chart, a trip unit with an I²t short-time ramp or a wide band. When the fault is not three-phase and a delta-wye transformer lies between the devices, the primary device sees a different per-unit current than the secondary device (by circuit arithmetic, for a secondary line-to-line fault one primary phase carries 2/√3 ≈ 1.15 times the three-phase equivalent, and for a secondary line-to-ground fault 1/√3 ≈ 0.58 times), so curves that clear each other on this chart may not. When instantaneous elements respond to the asymmetrical first-cycle current, which TM 5-811-14 takes as 1.6 times the symmetrical value (its equation 4-3) and which is not modeled here, an upstream instantaneous element set just above the downstream symmetrical fault current can still operate. When current-limiting fuses or breakers operate in under half a cycle the time-current chart no longer describes them, and selectivity is judged from the manufacturers' selectivity tables or let-through energy. When the fault current decays (a generator) or when several sources feed the fault, a fixed-current chart is not enough. Cable short-circuit damage curves and ground-fault devices are not on the chart; the Voltage Drop and Short Circuit tool gives fault currents and the conductor withstand time.
Units
Currents are in amperes, times in seconds and voltages in kV in both unit systems; the unit-system selection changes nothing on this page.