Definition
BSTDI measures divergence between digital twin predictions and actual building system performance.
The Measurement Gap
The Decision This KPI Enables:
After a maintenance event that changes the physical characteristics of a building system, does the digital twin reflect the change, and how far has the twin diverged from measured physical performance?
Why Legacy Metrics Miss This:
Traditional digital twin metrics focus on model availability and data feed integrity: is the twin running, are sensors feeding it, is the visualization current. These are necessary but insufficient. A digital twin can be fully operational with 100% sensor connectivity and still produce inaccurate predictions because its underlying model parameters no longer match the physical system.
The divergence pattern is driven by maintenance events that change system characteristics without triggering model recalibration. When a facilities team replaces a chilled water pump, the new pump has a different performance curve than the one the twin was calibrated against. When water treatment chemistry changes, heat exchanger effectiveness coefficients shift. When a BAS controls technician modifies a sequence of operations, the digital twin's control logic model no longer matches reality. Each change is documented in the CMMS/EAM as a completed work order, but the digital twin platform has no automated mechanism to consume that information and recalibrate.
Over months, these unreconciled changes accumulate. The twin shows the chiller plant producing 850 tons of cooling capacity; physical measurement shows 720 tons. The 15% gap means that maintenance windows, load shedding decisions, and capacity planning based on twin predictions are built on stale assumptions. BSTDI measures this accumulated divergence so organizations can identify when recalibration is needed and which maintenance events have the greatest impact on twin accuracy.
Formula
BSTDI = (|V_twin − V_measured| / V_measured) × 100
Where:
- V_twin = Digital twin predicted value for a key performance parameter (capacity, flow rate, efficiency, energy consumption)
- V_measured = Physically measured value for the same parameter under comparable operating conditions
Unit of Measure: Percentage divergence (0%+), where 0% indicates perfect alignment and higher values indicate greater divergence
Measurement Frequency: Monthly for critical building systems; quarterly for secondary systems; additionally after any maintenance event that changes system components, control sequences, or operating parameters
Data Requirements:
- Digital twin output data: predicted performance parameters under current operating conditions
- Physical measurement data: sensor readings, portable instrument measurements, or test results for the same parameters
- Maintenance event log: work orders that changed system components, control sequences, or operating parameters, with completion dates
- Operating condition normalization: outdoor air temperature, building occupancy, equipment staging, and other variables needed to compare twin predictions with measured data under equivalent conditions
Calculation Notes: BSTDI should be calculated for each major building system independently (chilled water, hot water/steam, air handling, electrical distribution) rather than as a single building-wide average. A building-wide average can conceal critical divergence in one system behind good alignment in others. Operating condition normalization is essential; comparing twin predictions at design conditions against measurements at part-load conditions produces misleading divergence numbers. Use matched-condition comparisons or normalize both values to a common reference condition.
Proposed formula, pending validation pilots.