Digital Tools, Data, and Emerging Technologies
This domain tests engineering judgement applied to digital systems: information management through a common data environment, data quality and volume, model validation, and security and governance of connected plant.
Core rules
- BIM information management (ISO 19650) runs through a common data environment with defined states: work in progress, shared, published, archived. Only published information is relied upon; check suitability codes and revision before issuing.
- Level of information need is defined per purpose, not as one monolithic model. Excess detail costs money and creates rework; insufficient detail blocks the downstream task.
- Data quality dimensions: accuracy, completeness, consistency, timeliness, validity, uniqueness. An uncalibrated sensor fails accuracy even when the dashboard looks healthy.
- Analytics ladder: descriptive, diagnostic, predictive, prescriptive - each rung depends on the one below it.
- Operational technology security (IEC 62443) uses zones and conduits, least privilege, segmentation between IT and OT, and patching planned around plant availability.
Worked example
You plan vibration-based predictive maintenance on 200 pumps, one triaxial sensor at 1 kHz, 2 bytes per sample. Data rate = 200 × 1000 × 2 = 400 000 B/s ≈ 400 kB/s; daily = 400 000 × 86 400 ≈ 34.6 GB/day, about 12.6 TB/year. Streaming all of that to the cloud is costly and mostly background noise. Edge processing is the engineering answer: compute RMS velocity and envelope-spectrum features at the gateway, transmit summaries each minute (megabytes per day), retain raw waveforms only on alarm. Document sensor calibration and the validation data set behind any model.
Traps
Treating a clash-free model as an as-built record; accepting vendor AI output as verification; skipping data provenance and timestamps; patching OT on an IT schedule; running pilots with no owner, acceptance criteria or handover.