Digital twin programs stall on data, not technology. The platforms exist - AVEVA, Bentley iTwin, Siemens Xcelerator, AWS IoT TwinMaker - and the sensors and connectivity infrastructure are deployable. The bottleneck is the structured equipment data that the twin needs as its foundation: the component records, the connection topology, the asset attributes that turn a 3D model or a process simulation into something that reflects an actual facility. For most industrial operators, that data exists in one place and one place only: the engineering drawings, in a form that machines cannot read.
The Legacy Drawing Gap
A facility built in 1985 has 40 years of engineering drawings accumulated across its operating life: the original design package, modification sheets, field redlines, as-built revisions, OEM submittals for equipment replacements. Some of these exist as original paper prints. Some were scanned to PDF in the 1990s. Some are in AutoCAD R12 DWG files on a network drive that nobody has opened since 2003. Almost none are in a format that a digital twin platform can consume directly.
The conventional approach to bridging this gap is to commission an engineering contractor to manually extract data from the drawings and enter it into the twin's data model. For a large facility, this takes 18–36 months and costs millions. Many operators conclude that the data foundation work costs more than the twin itself, and defer or cancel the program. The drawings stay on the shelf, and the digital twin initiative becomes a perpetual future project.
What You Need Before the Twin: Structured Component Data
A functional digital twin needs, at minimum, an equipment record for each physical asset - its type, specifications, location, and connections to adjacent equipment. For an electrical substation, that means a record for each breaker, transformer, relay, and protection device with its ratings, device numbers, and wiring connections. For a process facility, that means a record for each vessel, pump, instrument, and control valve with its tag, service, and process line connections.
That structured record set is exactly what engineering drawing digitization produces. A digitized relay panel schematic is not a cleaner picture of the panel - it is a database of component records extracted from the drawing, ready to load into the twin's equipment model. A digitized P&ID is an instrument index, a line list, and an equipment list in machine-readable form.
The Digitization-to-Digital-Twin Pipeline
The pipeline from legacy drawings to a populated digital twin runs in four stages:
Drawing triage and digitization: The archive is inventoried by drawing type (electrical, P&ID, layout) and condition. High-priority drawing types for the twin scope are processed first. Automated extraction produces structured component records with confidence scores. Low-confidence items are reviewed by domain engineers.
Data model mapping: The extracted attribute schema is mapped to the data model of the target twin platform. AVEVA E3D uses a different object hierarchy than Bentley iTwin; the mapping is a configuration exercise done once per platform. OpenDrawing's team provides standard mapping templates for major platforms.
Incremental loading: Validated records are loaded into the twin in batches, building the equipment layer incrementally as drawing sets are processed. A facility does not need 100% of its drawings digitized before the twin becomes useful - a partial model populated with the highest-criticality assets provides operational value while the archive digitization continues.
Ongoing synchronization: As the facility generates new or revised drawings, the digitization pipeline re-processes modified drawing sections and updates the twin's equipment records. This keeps the twin current with the as-built condition rather than creating a point-in-time snapshot that drifts.
Platform Integrations
AVEVA products (E3D, AssetWise, Unified Engineering) accept structured equipment data through their standard import APIs. Digitized electrical and P&ID records map to AVEVA's tag and equipment object models.
Bentley iTwin uses an open schema accessible via the iTwin API. Equipment records from drawing digitization load as assets in the iTwin's synchronization framework.
Siemens Xcelerator (formerly Comos) uses a tag-based plant model. Extracted instrument and equipment records from P&ID and electrical drawing digitization map to Comos object types during configuration.
AWS IoT TwinMaker accepts equipment data through workspace entity definitions. Digitized component records create entity instances with properties derived from drawing attributes.
Starting Point: High-Criticality, High-Value Scope First
No operator digitizes an entire legacy archive before building a twin. The productive approach is to identify the 10–15% of drawings that contain 80% of the highest-criticality asset data and start there. For an electric utility, that is protection relay panel drawings for transmission substations. For an oil and gas operator, that is P&IDs for primary process units. For a manufacturer, that is schematics for production-critical electrical systems.
A bounded initial scope delivers a working twin foundation in months rather than years, provides a validated accuracy baseline on the facility's specific drawing vocabulary, and generates measurable operational value - faster maintenance planning, better outage management, more reliable equipment records - before the full archive program is complete.
See also Engineering Diagram Digitization Software: The Complete Guide and Electrical Schematic Digitization for Asset Management.