Calibrating a Seven-Building Campus Behind a Single Utility Meter
Energy Modeling
The seven-building campus modeled in IES-VE, reconstructed from architectural and mechanical drawings spanning original 1970 construction and two rounds of renovation.
- 7-Building School Campus
- Washington, USA
- Existing-Building Calibrated Energy Model
Case Study Summary
Service | Energy Modeling — existing-building calibrated baseline model |
Building Type | K-12 school campus, 7 buildings, ≈59,600 ft² combined — built 1970, renovated 2009 & 2016 |
Location | Washington, USA |
Software | IES-VE |
Metering | Single utility meter for the entire 7-building campus — one electricity meter, one gas meter, no per-building submetering |
Calibration Standard | ASHRAE Guideline 14 (MBE ±5%, CVRMSE ≤15%, monthly) |
Deliverable | Calibrated baseline energy model (native IES-VE), handed off for the client’s own EEM analysis and compliance reporting |
Client | Confidential (U.S. energy services engineering firm) |
The Challenge
A 7-building, ≈59,600 ft² school campus built in 1970 with two rounds of renovation since (2009, 2016) needed a calibrated existing-building energy model as the foundation for a compliance audit. The complication: all seven buildings share a single utility meter for electricity and a single meter for gas — there’s no per-building billing data to check against. Every building’s HVAC system, schedule, and occupancy pattern had to be modeled accurately enough on its own terms that the sum of all seven, added together, would land within a few percent of the actual campus-wide bills — with no way to catch an error in any single building against its own separate number, because there wasn’t one.
Our Approach
We reconstructed each of the seven buildings’ existing HVAC systems individually in IES-VE: central non-condensing boilers serving most of the campus’s heating demand, a central chiller for cooling, and a mix of central air handling units, air-to-air heat pumps, natural gas furnaces, electric unit heaters, and natural gas make-up air units for specialized spaces — with no ERV or HRV recovery anywhere in the existing systems, consistent with the campus’s age. More than 45 individually configured HVAC zones were built across the campus’s five-plus distinct system types, each reflecting real equipment and thermostat groupings rather than a simplified campus-average system.
With one meter behind seven buildings, there was no shortcut available: get any single building’s schedule, occupancy, or system performance meaningfully wrong, and it doesn’t show up as an error against that building’s own bill — it just gets absorbed into, or masked by, the campus total. The only way to trust the aggregate calibration was to get each of the seven buildings right individually first, before ever comparing the combined total to a utility bill.
The completed model was calibrated against 12 months of historical utility data for both fuels, evaluated using the industry-standard ASHRAE Guideline 14 metrics: Mean Bias Error (MBE) and the Coefficient of Variation of the Root Mean Square Error (CVRMSE).
Results
Metric | Electricity | Natural Gas | Guideline 14 Threshold |
MBE | -0.8% | -4.5% | ±5% |
CVRMSE | 1.3% | 0.9% | ≤15% |
Both fuels calibrated well within ASHRAE Guideline 14’s acceptance criteria for monthly calibration — electricity closed within 0.8% of actual billed consumption against a ±5% allowance, with a CVRMSE of just 1.3%; natural gas closed within 4.5% against the same allowance, with a CVRMSE of 0.9%. That’s a tight match for a single-meter, seven-building campus, where a genuinely well-calibrated model and a lucky cancellation of errors can look identical from the outside — the only real assurance is in how carefully each building was modeled going in.
Why It Matters for Your Project
Our scope stopped at the calibrated model — the client’s team took it from there, applying their own proposed energy conservation measures and producing the final compliance submission themselves. That’s by design. The most time-consuming and error-prone part of an existing-building study is the front end: tracking down old drawings, reconstructing systems that have been renovated twice since 1970, and getting a model to actually match reality — not running EEM scenarios once a trustworthy baseline already exists. Handing over a model that’s already cleared Guideline 14 calibration means a client’s team can go straight to evaluating retrofit options, without spending their own time re-verifying whether the foundation underneath is solid.