Direct Answer

Is the gateway not plug-and-play? Why does a project have to spend time drawing the point table and building an object model first? Because "connected and producing data" and "the data can be used" are two different things. The object model abstracts "transformer → distribution cabinet → feeder → terminal" into digital objects with hierarchy and attributes — it decides whether an alarm can be located to a terminal, whether a report can be summarised by workshop, and whether AI can have position awareness. Skipping the model and connecting tables directly means that after go-live, alarms have no location, reports do not match devices, and every expansion means reworking the whole table; building the model first and connecting afterwards is doing it right once, otherwise every expansion reworks everything. In the implementation of the digital electrical intelligent supervision system, the first step is always object modelling, not installing hardware.

A scene: a factory completed phase one with 200 measurement points. Rushing the schedule, it directly connected bare tables by "device + point". The first alarm after go-live arrived — "point 3 current over limit". The electrician asked, "which device's which wire is point 3?" They searched for a long time and could not find it, because the point table had no hierarchy and no device ownership. In phase two, adding 300 more points, all reports, alarm rules and dashboards had to be redone. With the same workload, if two days had been spent building the object model at the start, none of this would ever have happened.

1. The Three Elements of the Object Model: Device Objects, Measurement Points and Topological Relationships

The object model is not one table; it consists of three elements:

Element 1: device objects — transformers, distribution cabinets, feeders and terminals, each a first-class citizen with its own attributes (model, location, workshop ownership, commissioning date); Element 2: measurement points — data points attached under device objects (three-phase current, temperature, voltage); a measurement point must "have an owner" and belong to one and only one device object; Element 3: topological relationships — the connection relationships between devices (which distribution cabinet draws power from which transformer, which feeder supplies which device), with the parent-child hierarchy clear.

In one sentence: device objects manage "what exists", measurement points manage "what is measured", and topology manages "who connects to whom" — with the three elements complete, the data has a sense of position.

2. Why "Model First, Connect Afterwards" Is a Discipline

Skipping the model and connecting tables directly looks, in the short term, like "saving two days"; in the long term it is three debts:

Debt 1: alarms have no location — an alarm has only a point number and a value, not "which device, which level, in which workshop"; on-site troubleshooting has nowhere to start; Debt 2: reports do not match devices — reports must summarise by workshop and by device type; bare tables have no ownership relationship, so every summary is manual stitching; Debt 3: expansion means reworking the whole table — when phase two adds devices, if the model did not reserve hierarchy and codes from the start, all alarm rules, reports and dashboards must be redone; the rework cost is tens of times the modelling cost.

In one sentence: the "two days" of modelling buy "three years without rework" — however you calculate it, this account is worth it.

3. A "Residential Building" Analogy: Structure and Door Numbers First, Then Meters

Compare object modelling to accepting delivery of a residential building: first the building structure (how many towers, how many floors, how many units per floor, how the door numbers are coded), then the electricity and water meters — every meter is attached to a definite unit, and meter readings naturally belong to "which household used it". Conversely, if the meters are installed first and the door numbers are not coded, the readings are all a pile of "some meter on the 4th floor of Unit 3", and reconciliation, troubleshooting and billing all fall into chaos.

Device integration is the same: define the structure first (the object model), then attach measurement points (data integration) — only then can alarms, reports and AI work at the right positions. If the order is wrong, what gets reworked is the entire data system, not one table.

4. The 18-Level Scene Tree: Position Awareness from the Park to the Contact Point

Why can the object model support "position awareness"? Because behind it is an 18-level scene tree: from park → building → floor → power room → transformer → distribution cabinet → feeder → terminal → contact point, every level is a locatable, queryable node.

The value of the scene tree: an alarm says "A-phase terminal overheating on cabinet 1, power room, floor 2, building 3, some park" — every word maps to a node on the tree, and the on-site engineer can walk straight to the terminal with this alarm; for AI to calculate "the load rate of a workshop", it follows the tree's "workshop → device → measurement point" path to get the data in one pass, without scanning the whole table.

In one sentence: the 18-level scene tree gives "data a position, alarms an address, computation a path" — this is how the object model supports position awareness.

5. The Unified Coding Rules: One Code per Device, One Code per Point

The object model runs stably thanks to coding discipline:

One code per device — every device has a unique code, with hierarchy information embedded (park-building-level-sequence); when a device is replaced the code stays, so historical data does not break; One code per point — every measurement point has a unique code, and the code tells which device the point is attached to and what physical quantity it measures; The code is the address — seeing the code tells you where it is on the scene tree, no need to look up a table to know "where this point is".

In one sentence: the code is the "door number" of the object model — with the door numbers coded well, finding positions, matching reports and doing analysis are all along the way.

The details of coding in practice: coding is not made up on the spot; three things must be defined — first, the code format (which segment means the park, which the building, which the device type); a unified format makes it machine-parseable; second, the code mapping (codes and scene-tree nodes correspond one-to-one, one code cannot map to two positions); third, the code-change rules (how the code is inherited when a device is replaced, how historical data is reattached — define the rules in advance to avoid broken data chains). The finer the coding rules, the higher the automation afterwards — this is a key link in "doing it right once".

6. The Three Rework Consequences of Skipping the Model: Look at Each

Break the three debts apart, and each is real money:

Alarms without location — the alarm object is only "point 3 current over limit", with no device and no hierarchy. Handling: look up the point table manually → find the device → then find the location. Each troubleshooting starts at half an hour; 100 alarms means 50 hours; Reports not matching — "workshop 3 energy consumption" must be summarised by workshop, but the data has no workshop field, so it can only be exported and matched by hand in Excel. Repeated labour every monthly report, and the report is always late; Rework on expansion — adding 300 points in phase two means rewriting all the original alarm rules, report formulas and dashboard configurations by the new points. The rework volume equals rebuilding half the platform.

In one sentence: the common cause of the three consequences — all come from "data without structure". The object model is that "structure"; making up the lesson later is always more expensive than doing it right from the start.

When is modelling "wrong": modelling is not "the more complex the better" — first, do not model positions that do not exist (levels not actually present should not be built yet; add them on expansion; doing it right once does not mean building everything at once); second, do not make on-site staff memorise codes (codes are for the system; on site use short names and labels, and let the system map them internally); third, do not build the model after connecting (the data has already been running bare for a while; when the model is added, the historical data ownership is untraceable, and it is as good as never connected). The right time to model: before connecting, build as needed, codes for the system, names for people — with the four in place, the model is an asset, not a burden.

7. The Relationship Between the Object Model and Digital Twin

The object model is not the end; it is the foundation of the digital twin:

Digital twin = object model + real-time data + rules/AI — the object model provides the "skeleton" (devices, measurement points, topology), real-time data provides the "flesh" (current, temperature, voltage), and rules and AI provide the "behaviour" (diagnosis, warning, optimisation); Model first, twin comes naturally — without the object model, the twin is only a "3D shell" and the data does not know where to attach; with the object model, real-time data naturally falls into place, and every component of the twin has data moving in it.

In one sentence: the object model built today is the skeleton of tomorrow's digital twin — modelling first is both preventing rework and paving the way for the future.

8. The Four-Step Method of Object Modelling: Usable On-Site

Step 1: take inventory — count the devices, measurement points and connection relationships, and draw a physical topology sketch; Step 2: define the levels — hierarchise the sketch by "park → building → power room → device → terminal", with one object per level; Step 3: code it — code devices and points by the unified rules, one code per device and one code per point; Step 4: attach measurement points — attach points under the device objects and verify that every point can be located to "which device, which level, where".

In one sentence: take inventory, define levels, code it, attach points — after the four steps, the object model takes shape and integration makes sense, with data having ownership; two days before go-live buy three years without rework.

9. Back to FEXLINK: How Object Modelling Lands

To land this methodology on products: the digital electrical intelligent supervision system takes object modelling as the first implementation step — the implementation team first builds the object model by the park structure and codes it uniformly, then attaches the measurement points of all kinds of monitoring controllers (ESF series, ESC series, FL series, etc.) under the objects, so they fall into place upon connection; on the platform side, alarms, reports and dashboards all address through the object model. A well-built object model lets alarms be located to terminals, reports be summarised by workshop, and AI have position awareness — this is the platform-level embodiment of "model first, connect afterwards". During selection you can also ask "does your implementation model first or connect first" — the one answering "model first" is the one truly treating data as an asset; the one answering "we patch it after connecting" will almost certainly be making up lessons after go-live. Modelling is not complex; what is hard is the discipline of "doing it before go-live" — and that discipline is the starting point of the project's data system.

Scope and Limitations

  • The conclusions in this article are intended for LV distribution and electrical-safety sites. Criteria are compiled from public codes and FEXLINK product documentation, and product parameters follow the latest selection handbook and product knowledge base.
  • Quantitative criteria (protection zone, discharge and residual voltage, earth resistance, thresholds and grading) hold under the preconditions stated in this article. Where the site voltage level, earthing arrangement, environmental category or load profile differs, the figures must be recalculated.
  • Specific product models and parameters are subject to the latest selection handbook and product knowledge base; this article does not constitute selection or procurement advice.
  • Standard numbers cited here refer to the currently effective editions; when a standard is revised, the related quantitative conclusions must be re-checked.
  • Deployment and implementation for specific projects must be carried out by qualified engineers based on site survey and design drawings.

Frequently Asked Questions

Why You Must Build the Equipment Object Model Before

Why does a project have to spend time drawing the point table and building an object model first