BIM Electrical Circuits Can Report Zero Strictly Connected Edges

Seventy-eight attempts to connect one equipment connector directly to another returned the same interface exception, seventy-eight times: electrical device or equipment cannot connect to electrical device or equipment directly [MODELED]. That is one refusal, not seventy-eight reasons. The workaround that replaced them built a system object from each exact origin connector and added the target device to the circuit. It produced 78 systems with complete edge and port metadata, each storing its own edge identifier, and reported no blocked edge on the generator side. It also reads back zero direct connections and zero edges with both ends resolved on one system [MODELED].

The Platform Refused Every Direct Equipment-to-Equipment Join

A single cause across every attempt is the cleanest result a postmortem can get, and the record quotes the interface in its own words rather than paraphrasing it. What that cleanliness hides is that nobody reproduced it independently. On the current record the count comes from the generator's own attempt log. The platform behaviour was not checked against documentation anywhere in the read set. A refusal that is consistent is not thereby a refusal that is understood. The record does not claim otherwise. It states the prohibition, quotes the exception and moves to the workaround in three sentences.

This is a property of the platform version used in a concept-stage run. It is not a proof that equipment cannot be connected in a later design stage. The record says the limitation does not block concept drawing. What it blocks is a claim.

Origin Connectors Are Exact; Destination Connectors Are Not

The workaround is asymmetric, and the record says where. At the origin the connector is exact. At the target the interface adds a device to the circuit, so there is no way to show that the connector it selected is the one the edge table named. Four consequences are then listed together, and they do not agree with each other:

• the systems, the origin connectors, the ownership of the devices at both ends and the port metadata are all established;
• the direct-connection read-back returns zero, and the strict test of both connectors resolving to one system also returns zero;
• the 78 edges may not be described as 78 completed physical port-to-port connections;
• the shared-source relationship between the two incoming feeders can be expressed through the system graph with controlled edge metadata, but cannot yet be derived from connector references alone.

That third line is the sharpest thing in the document. It is written as a constraint on what may be said. Comments on the system objects are author metadata. They can be complete while the software-selected destination connector is the wrong one. Independent utility dual service remains false on this project. That statement is unchanged. It is not a deduction from connector references.

Uniqueness Passed; Coverage of the Audited Register Did Not

Against the audited register of 573 objects and 183 connections, the pilot touches 96 objects and 78 connections — 16.8 and 42.6 per cent [MODELED]. Both figures move with the snapshot. The audit is 14D on 29 August: 573 objects, 183 unique connections, 1,537 registered ports, 366 used, 1,171 unused [MODELED]. The frozen Rev1 connectivity register one day later is 196 unique edges. A citation of "the connection count" without naming the snapshot is invalid. Uniqueness passed. Coverage did not. Those are two different gates.

Of those 1,171 unused ports, the largest families are access, control, lift, low-voltage, path and primary liquid [MODELED]. Most are legitimate conditional branches, maintenance envelopes, replacement paths, routes or interfaces not yet developed. Forcing every port to connect would be as wrong as ignoring a real miss. The two zeros on the native run are the same class of result: an unconnected connector-to-connector reference is not, by itself, a missing object.

Summary Counts Look Complete While Both Strict Tests Are Zero

Every headline number this run offers is positive. Objects built, connectors generated, systems created, metadata resolvable, zero blocked edges on the generator side. The two zeros are the only measures pointing the other way, and they are the ones a summary drops. Reporting a run that is simultaneously complete and empty is the hardest case there is. The record simply prints both halves in one list.

The remedy was already written, before this run and not in response to it. The register audit requires that the next review report four numbers rather than two: how many ports must be connected and at what pass rate, how many are conditional, how many are legitimately unconnected and how many are not [TARGET]. It adds that an object count, a connection count and a zero-bad-reference result is not sufficient to prove a model complete. A zero-blocked-edge result on the generator side is exactly the kind of measure that requirement was written against. It reports what the generator managed to attempt, not what the model can demonstrate. That report does not exist. Neither does the equipment-family connector work, the distribution-system settings or the engineer review the record names as the route to closing the limitation. The configuration was restored to disabled and the process closed at the end of the run.

Limits and open items

Confirmed: AllRefs direct connections are zero; the strict same-system test on both connectors is zero; the 78 system objects may not be described as 78 completed physical port-to-port connections; uniqueness of the audited register is not coverage. [FACT]

Modelled and unfrozen: 78 ConnectTo attempts, each returning one native exception; 78 systems with metadata; 96 of 573 objects and 78 of 183 connections; 16.8 and 42.6 per cent; 1,537 registered ports with 1,171 unused; 366 used; the later frozen snapshot of 196 edges. The 183 and 196 figures are different snapshots. They are not two counts of one plant. [MODELED]

The four numbers the next review must report remain a requirement on a future pass, not a result [TARGET].

Open: an independent reproduction of the interface refusal; a read-back that both named connectors sit on one system; equipment-family connector work and distribution-system settings; the four-number coverage report; an Engineer of Record review of the limitation. Independent dual utility remains false; that statement is unchanged and is not derived from this run's AllRefs. The file is concept-pilot, partial, Engineer of Record false, IFC false. Nothing here has been sealed. Nothing is a connected plant. [HOLD]

Ask Whether the Strict Read-Back Still Prints Zero

The useful test on the next native run is not whether 78 system objects appear in the file. It is whether the strict read-back still prints zero. If the summary leads with the 78 and drops the two zeros, the run has been reported as a success it did not earn.

© 2026 K&K Data Service Inc. All rights reserved. Reproduction or republication is permitted only with clear attribution to K&K Data Service Inc. and a working hyperlink to the canonical URL of this article. Excerpts must preserve the technical context, maturity labels, assumptions, and limitations. No excerpt may imply project approval, field validation, certification, or endorsement that the original article does not state.

Source: K&K Data Service Inc., “BIM Electrical Circuits Can Report Zero Strictly Connected Edges,” https://www.kkdatasvc.com/lab/bim-and-navisworks/78-native-circuits-zero-strictly-connected-edges/.