A time-current overlay can look selective and still drop the wrong customers. The plot cannot confirm that the tank sits on the feeder segment the utility meant to isolate. Source-impedance values that disagree with conductor limits stay wrong after the printer finishes. Neighboring protection can still refuse the sequence the family appears to prove.
This article stays on distribution feeders. Device location belongs in the same package as the cases the model actually ran. Settings the utility can issue close that package. Reach here means the fault conditions detectable under studied system states. An untested switching state is not protected because a curve family was printed.
Recloser Engineering Services Begin With the Feeder Objective
Recloser engineering services put automatic interruption on a distribution feeder through a studied settings package. IEEE C37.104-2022 is the application guide we use on those feeders. The utility still owns protection philosophy, and settings approval stays with the owner. A tidy plot is not the product.
The usual failure is picking a convenient pole before anyone names the protection function, even though climbing access and a spare bracket are construction facts that cannot define which customers the utility intends to take out of service for a temporary fault. Distance from the source bus will not answer.
Freeze a protection objective that names the feeder segment and the fault types considered before detailed production starts, because a later settings file cannot invent an isolation goal the utility never accepted or a continuity target the feeder still has to meet after a successful reclose. We will not start curve production against a hallway sentence.
We build that objective from the dated one-line and the feeder topology the utility actually operates. Conductor data retain the study's grounding basis so a later reviewer can reconstruct ground-fault reach without sorting a mixed file. Source conditions bound the cases we will run. Existing devices bound them too. Load data and owner protection philosophy enter as owner input, with source date and revision on the same line.
Recloser Engineering Services Before Device Placement
IEEE C37.104-2022, Guide for Automatic Reclosing on AC Distribution and Transmission Lines, published 28 April 2022, documents accepted application and coordination practices. This article uses that guide on distribution feeders. Transmission-line reclosing is a different application of the same document. A feeder recloser study is not a station-bus study. Adoption and professional responsibility still decide how a particular feeder uses the guide.
The limitation of our preferred curve-first method is that we will still print a family that looks complete while one switching state never entered the model. That is our process defect, not a software defect. We now stop production until that state is named as studied or as outside the issued settings. Locate the isolation decision in the one-line first. The field pole comes later.
A protection engineer who did not build the model then picks one source impedance and walks it onto the named feeder segment. Recollection from the original author is the wrong test. If that walk dies in a hallway note, the settings package records the break. The setting waits. We would rather hold a pole pick than issue a file that pretends the isolation goal was already accepted.
Build a Feeder Model That Exposes Uncertainty
Protection studies are only as precise as their source data. A curve family printed from an unlabeled impedance case cannot become a feeder study merely because the final plot looks orderly, and source impedance plus conductor segments still need provenance a stranger to the folder can follow before transformer connections can support a settings decision. Minimum-fault and maximum-fault conditions are not one plot.
Model Cases Before Curve Tweaks
The working file is a study model with normal and relevant alternate configurations plus a visible assumption register. The utility one-line and GIS records must agree on identifiers a later reviewer can actually join. Manufacturer curves describe device behavior. They do not supply the feeder's available fault current. Completeness is not a cell count.
If the study runs one maximum-fault snapshot and treats it as the complete feeder envelope, coordination may look acceptable in the studied state while a lower-fault configuration or an alternate-source configuration remains untested. Name those cases on the device they affect. Do not bury them in a cover-sheet paragraph nobody will reread.
DER status changes available fault current and can change whether automatic reclosing engineering is even acceptable on that feeder, so if the utility's circuit now has generation the last study never saw, the objective is stale and the cases have to be rebuilt. Do not keep last year's pole because the bracket is already there.
The related guide on distribution construction release controls addresses a neighboring handoff without replacing this model. The utility supplies or approves system conditions, and we record applicability inside the settings package beside the decision rather than in a detached standards paragraph nobody will reopen. The assumption register matters most where the model is weakest.
Office review works backward from one proposed setting. A completeness percentage cannot tell the protection engineer whether a later reader could rebuild the join from the one-line into source cases and manufacturer curves without asking the original author. Expand the sample when the join problem looks systemic. A stale open-point on last year's diagram will place this year's tank on a feeder the utility no longer operates that way.
| Decision | Evidence form | Consequence if unresolved |
|---|---|---|
| Recloser Engineering Services Before Device Placement | One-line diagram joined to feeder topology. Owner protection philosophy remains labeled as owner input. | The installed device interrupts the wrong exposure or fails to improve sectionalizing as intended. |
| Model Cases Before Curve Tweaks | Utility one-line and asset data joined to source cases. Manufacturer curves remain labeled as device behavior. | Coordination looks acceptable in one studied state while another configuration remains untested. |
| Curves Explain Relationships, Not Territory | Fault results with time-current curves. Upstream settings and conductor limits sit on the same review. | A temporary fault becomes an unnecessary broad outage, or a persistent fault stays energized longer than intended. |
| Ratings and Controls Share One Design Basis | Calculated fault duties paired with the utility's control architecture. Sensor and power-source requirements remain on that sheet. | The device fits the voltage label yet fails a required interrupting or communications interface. |
| The Five-Part Protection Release | Final study report paired with settings-file identity. As-left settings return with owner acceptance. | The fielded control differs from the coordinated case even though each file was once approved. |
Coordinate Reclosing With Upstream and Downstream Protection
A recloser curve lives among fuses and relays, and the operating sequence remains unproven until sectionalizers and breakers have been tested against the same source conditions used for the recloser itself. Conductor limits can veto a curve that looked selective on paper. Transformer protection can do the same. Coordination is the relationship among those devices across selected fault and load conditions. It is not a single crossing point on a chart.
Curves Explain Relationships, Not Territory
IEEE C37.104-2022 discusses automatic reclosing application and coordination, yet the utility's protection philosophy still determines whether reclosing is used on a specific feeder and which fault types are allowed to reclose. Settings optimized for one device pair while ignoring another relationship can turn a temporary fault into a broad outage. They can also leave a persistent fault energized longer than intended. We will not release a sequence that only works for one pair.
Communications in the same corridor can change what operating states exist. See communications interfaces in power corridors when that interface is real. It does not replace feeder protection coordination. It can change which cases must be run. A protocol that dies when the feeder dies cannot support the isolation the overlay assumed.
We test selectivity against the utility's reclosing philosophy, and sensitivity plus equipment protection sit on that same review with fast and delayed operations where they are used. Fuse behavior belongs there. Load pickup after a successful reclose can veto a setting that looked perfect on a bolted-fault plot, and inrush on a long rural feeder can do the same before anyone treats the overlay as issued. Show those checks. Or hold the file.
Conductor damage limits are coordination parties. They sit on the same review as the upstream relay. Ignoring them to make a device pair look tidy is how a conductor becomes the fuse. Put the limit on the plot. If the recloser cannot live with it, say so in the settings package rather than hoping the field never sees that current.
If fuse-saving and fuse-blowing are both still hallway options: study them as distinct cases. A single overlay cannot hold both philosophies.
Fuse-saving and fuse-blowing are operating philosophies, not drawing layers. If the utility has not picked one, the overlay is pretending. Study them apart. Then issue the one the operators will actually live with. A single family that claims both will be argued in the first storm.
Specify Recloser Hardware With Its Control as One System
Interrupting equipment and its control cannot be procured as unrelated line items, because a tank that meets voltage class can still fail the calculated duty or the control interface the feeder study actually required. Sensors, control power and communications affect that same issued design. Mounting is part of the same system. Start from that constraint, not from a catalog page.
Ratings and Controls Share One Design Basis
IEEE Std C37.60 and IEC 62271-111, 2018 edition, address automatic circuit reclosers for alternating current systems up to and including 38 kV. That dual-logo document remains the published equipment standard we check against calculated feeder duty. Project specifications and manufacturer data still control the selected product. A catalog model chosen before duty and integration requirements are reconciled may fit the voltage label. It can still fail a required interrupting interface.
Distribution conversion planning that changes feeder topology is covered in distribution conversion planning. A converted segment can invalidate a recloser location that was correct last year, which is why we rebuild the objective when the feeder changes instead of keeping the old pole by habit. Distribution recloser design follows the topology the utility will actually operate, not the topology on last year's drawing.
Procure the recloser as a protection and control system whose ratings trace back to the study. Do not procure a tank with accessories. Control functions, protocol and owner cybersecurity requirements belong on the same specification as interrupting duty. Control power that dies in the event the recloser is supposed to handle is not an accessory miss. It is a failed specification.
Calculated fault duties have to meet the selected unit, and basic impulse level plus sensor arrangement have to meet it before anyone treats a catalog page as a released specification. If any one of those is still open, hold procurement. Voltage class alone does not prove interrupting duty. Sensing is a separate check.
The in-house engineering model is how that study is staffed. We perform this engineering in-house. Power work covers distribution and make-ready, and substations are inside that offering even though this article is a feeder study rather than a station-bus walkthrough. Transmission stays outside it. A feeder recloser package may need a station-side change at the distribution substation that sources the circuit, and that station work is in-house engineering when the utility buys it under the same engagement.
Settings notes also have to survive outside the study meeting. A later reader should be able to tell a cited reclosing guide from a manufacturer curve. An owner criterion is not an engineer's project control. That split keeps IEEE C37.104-2022 from being treated as the utility's protection philosophy.
Choose Recloser Engineering by the Settings the Utility Can Release
Cooperative with one long rural feeder: require the study to state the isolation objective before a field location is chosen, because a curve family optimized for one device does not prove the feeder will sectionalize as intended when inrush and load pickup after a successful reclose were never modeled. Hold the pole pick until the objective is written.
Municipal feeder with dense laterals and fused taps: insist that fuse-saving and fuse-blowing philosophies appear as distinct studied choices, because a single overlay that pretends to hold both is not a settings package the operators can live with in the first storm. Pick the philosophy the utility will actually operate. Equipment selection follows the calculated duty and control architecture after that choice is written.
Utility already circulating a settings file: compare that file's identity with the coordinated study and the issued drawings before the switching window opens. If any one of those files points to an earlier case, stop the affected device. Accepted feeder work around it can remain intact. Do not energize a maybe.
Study completion is not permission to energize. Reviewed settings, construction drawings and approved equipment have to converge with the communications configuration, the commissioning plan and the utility's switching authority before anyone treats a calendar date as a release. A calendar date is not that convergence.
The Five-Part Protection Release
The five parts we will not skip are the feeder objective; the study model; coordination; the equipment specification; the field-release record. Draftech provides in-house distribution engineering within its approved scope. When construction is included, Draftech-managed subcontract crews work under Draftech QA/QC and safety oversight, while the utility still controls switching, settings approval and energization as owner acts.
A settings file whose identity is not on the issued drawing will be the file someone emails the night before switching. That is how as-left drifts from as-studied. We put the identity on the issued drawing, the commissioning record and the as-left return so a late email cannot impersonate the coordinated case. Performing-party identity for separately procured turnkey work sits in the managed crew framework.
Use the current settings package when reviewing our distribution feeder protection work. Email the same coordinated file identity to info@draftech.com. A convenient pole and a circulating file that no longer matches the study are the same class of problem. The field will install whatever those files currently say. Recloser settings review is the last chance to catch that drift before the window opens.
If the settings-file identity and the issued drawings disagree: stop commissioning and bring the mismatched revision to protection review. The feeder does not have one approved protection decision.
Protection settings come after the feeder route is issued. When that route is new, Draftech designs its first 20,000 linear feet at no charge through permit approval, and the utility keeps switching, settings approval and the drawings themselves. The free design offer is reviewed request by request before we commit.
