POTT, Weak-Infeed Echo & Three-Terminal Lines

A distance relay at one end of a line cannot tell a fault at 95 % of the line from one just past the remote bus; CT/VT and line-impedance errors blur the boundary. So Zone 1 is set short (typically ≈ 80 %) and trips instantly, and Zone 2 over-reaches (typically ≈ 120 %) but waits, commonly 15–30 cycles (20 in the lab). That leaves the last 20 % of the line cleared slowly from one end.

Pilot schemes fix this by letting the two relays talk. Each tells the other what it sees, “fault is in front of me” or “fault is behind me”, and both can trip at high speed for any internal fault. The two classic philosophies are:

• POTT (Permissive Overreaching Transfer Trip): trip if my overreaching forward element sees it and the remote says it sees it too. Secure by default, because a missing signal means no pilot trip.
• DCB (Directional Comparison Blocking): trip if my forward element sees it and no block arrives from the remote within a coordination time. Dependable by default, because a missing signal means trip.

Both assume each terminal can see the fault. For phase faults an IBR plant terminal often cannot: the inverters' positive- and negative-sequence current is limited to roughly their rating, so 21 and 32Q/67Q forward elements may never pick up. That is the weak-infeed problem, and solving it is the topic of this module. Ground faults are different: as Module 4.1 shows, the grounded-wye winding of the MPT makes the plant end a strong, passive zero-sequence source, so 67G/32G elements there usually see ground faults well.

The lab opens on the classic failure: a 2-terminal POTT line, an internal fault at 85 % (beyond the strong end's Zone 1) and a weak IBR at B. Press play, then step the scheme through + Echo and + Echo + WEI. Then try the security cases (“Behind A”, “Behind B”, “Reversal”) and the three-terminal tap with outfeed.

Echo turns the weak terminal into a mirror. When it receives a permissive and sees nothing itself (no forward, no reverse), it sends the permissive straight back, so the strong end can trip. The preconditions are what keep it secure:

• No reverse element and no guard timer. If the fault is behind the weak terminal, echoing would trip the healthy line.
• Echo pickup delay (a few ms) rides through channel noise and brief permissive blips.
• Echo duration and block timer. The echo is a fixed pulse (3–4 cycles), then echo is blocked for a while so two echoing terminals cannot latch the channel in a loop.
• Open-breaker echo. If the local breaker is open, echo any permissive, so the remote can trip at high speed for an end-of-line fault.
• Weak-infeed trip = echo conditions AND local undervoltage (27) or zero-sequence overvoltage (59N) AND breaker closed. Voltage supervision proves a real nearby fault before the weak end trips itself.
• Multi-terminal lines. On a three-terminal line, echo and weak-infeed trip must require the permissive from all remotes, or be disabled. Otherwise one remote's permissive for an external fault, plus the voltage dip, trips the weak tap. Pair the rule with open-breaker keying at each remote, or an open remote strands the tap. The Topology section and the lab show this case.

On double-circuit or parallel paths, sequential clearing of an external fault on the other line can reverse the current in the healthy line within a cycle or two. Before the reversal, terminal A sees forward and keys a permissive; B sees reverse. After the near-B breaker on the faulted line opens, A sees reverse and B sees forward. But A's permissive is still in flight, delayed by A's element dropout plus the channel time. If B's forward element picks up while that stale permissive is still arriving, POTT trips the healthy line.

The current-reversal guard blocks keying and tripping for a set time. In common relay implementations the extension starts when the local reverse element drops out, and it must outlast the remote's stale permissive: the remote forward element's dropout plus the channel's reset time, plus margin. Nothing is subtracted for the local forward pickup, because the local forward element can pick up almost as soon as the reverse element resets. Vendors name this timer differently: SEL calls it the Zone 3 reverse-block delay (Z3RBD), and other relays call it current-reversal guard or transient blocking. Check the manual for when your relay starts timing.

Tapping a collector substation into an existing line creates a three-terminal line, and two new effects appear:

• Infeed: current from the tap raises the voltage seen by the outer terminals, so their distance elements under-reach. Zone 1 may not cover the faulted section at all, as at terminal C in the lab.
• Outfeed: if an external path connects a terminal's bus back into the network (for example a parallel line from C's bus to A's), an internal fault near A can draw current out of the line at C. C's relay declares reverse and withholds its permissive. A three-terminal POTT needs permissives from both remotes, so the pilot trip is lost everywhere. Outfeed needs that external path, so it is a feature of utility terminals; a radial IBR tap has no path back to the network and rarely sees outfeed for phase faults. The lab therefore places it at C. Treat the lab's outfeed toggle as a logic demonstration of a terminal declaring reverse for an internal fault: with strong sources at A and C and a weak, current-limited tap, C would normally see infeed for a fault on the A–tap leg. Real outfeed needs a source strong enough to push current through the line and out at C, which an outfeed study has to show.
• Weak-infeed echo at the tap: for a fault behind A, C keys its permissive and the weak tap sees only the voltage dip. If the tap echoes on any remote's permissive, its weak-infeed trip fires for an external fault. Require the permissive from all remotes before echoing (the lab's default), or disable echo and WEI on multi-terminal lines. The rule has a dependability cost: any remote that cannot key (an outfeeding terminal, or an open breaker) strands the tap. Pair it with open-breaker keying at each remote. The timeline lab does not model open-breaker keying (it appears only in the logic lab), so with outfeed on and the rule on, B never trips; with the rule off, B trips by WEI at 21 ms on the 8 ms channel.

Mitigations for outfeed include Zone-1-initiated direct transfer trip (DUTT) to all terminals, hybrid echo logic supervised by pilot-blocking zones, sequential-trip logic that re-evaluates once the first breaker opens, and line current differential where the channel allows. DUTT only helps when the fault is inside some terminal's Zone 1: in the lab it clears all three terminals at high speed for a fault at 60 % from A, but at 85 % no Zone 1 picks up: A waits for Zone 2 (344 ms), and B and C are left to time-delayed or sequential clearing (in the lab neither trips within the 450 ms window). Because a DUTT receiver trips with no local supervision, its channel needs security: a dual channel, a continuously monitored guard tone, or security counts.