roovie

The Roovie Physics Engine

A physics engine built for scale.

A fully coupled heat balance, solved from first principles — independently verified against ASHRAE 140 and designed end-to-end for automation. Not a wrapper. Not a shortcut. A new foundation.

Fully Coupled Heat BalanceASHRAE 140 VerifiedCloud-NativeAI-AcceleratedClimate-Aware

Fully coupled heat balance

One building. Every influence. Solved at once.

At every timestep, the engine resolves the weather outside, the loads and schedules inside, the envelope between them, and the HVAC system that responds — as one simultaneous equation set. No sequential decoupling. No lagged zones. Every surface, every zone, every air node, every moment.

▤ Heat balancet = 14:00· step 4,200 / 8,760
OUTSIDE

Weather sets the boundary

Real EPW files push dry-bulb, humidity, solar, and wind into every envelope surface — every hour, every zone.

INSIDE

People, lights, equipment

Occupants, plug loads, lighting power, and setpoints each run on their own schedule — all contributing heat to the zone balance.

HVAC RESPONSE

The system closes the loop

Airflow, coil capacity, and plant response couple back into zone temperature, humidity, and surface temps — all at the same timestep.

Why we built our own

A data structure designed for automation, scale, and AI.

Wrapping someone else's solver — or simplifying the math — inherits an architecture that can't keep up with cloud orchestration, AI agents, or custom reporting. Owning the engine means every layer above it can be accelerated, extended, and automated.

The inherited path

Wrap · or simplify

  • ×Built around file-based I/O: one IDF in, one HTML report out. Fights the cloud.
  • ×Serial runs. Orchestrating 10+ in parallel is someone else's problem.
  • ×Reporting is fixed. Want a different rollup, a custom KPI, an API endpoint? Rebuild the post-processor.
  • ×Simplified math loses the physics that ECM ranking depends on.
  • ×No room to inject AI agents, live feedback, or model mutation mid-run.
The Roovie path

An engine, end-to-end controlled

  • Cloud-first data model. Every object addressable, every result streamable. No IDF round-trip.
  • Concurrent by default. 10+ simulations run side-by-side on demand; scale elastically.
  • Flexible result layer. Any KPI, any rollup, any benchmark — shaped to the question, not the template.
  • Full heat-balance fidelity. Same physics class as reference engines — verified against ASHRAE 140.
  • Agent-native. AI agents can spin up runs, read results, mutate models, and propose ECMs.

Built for the workflows that matter

Real physics → real deliverables.

Rigorous heat-balance physics is the foundation every serious compliance and reporting workflow requires. One simulation feeds your 90.1 Appendix G baselines, LEED EA credits, ENERGY STAR scores, code-compliance paths, utility rebate filings, and carbon disclosures — no model forks, no re-runs.

Rigorous foundation
Full Heat Balance physics
Every surface, every zone, every moment — solved simultaneously. The same physics class as reference programs like DOE-2, BLAST, and ESP, and regression-tested against ASHRAE 140 cases.
ASHRAE 90.1
Appendix G modeling
Performance Rating Method · baseline vs proposed
LEED v4
EAp2 / EAc1 energy credits
Minimum + optimize energy performance
ENERGY STAR
Portfolio Manager + Target Finder
Score benchmarking · Designed to Earn
IECC / Title 24
Performance-path compliance
Envelope trade-offs · whole-building path
Utility M&V
Incentive & rebate programs
Measured ECM savings · IPMVP-aligned
Carbon reporting
GHG accounting & disclosure
tCO₂e · SEC climate · investor-grade
One simulation, every deliverable. No parallel model forks for each program — the same calibrated building powers every workflow your team needs to ship.

From physics to outcomes

Heat balance is the means. Outcomes are the point.

The solver resolves temperatures, flows, and fluxes. What teams actually need — energy, dollars, carbon, comfort — flows straight out of that physics. No bolted-on assumptions.

Baseline · one 8,760-hour run
The physics
Zone balance · every hour
Tzone · ṁair · Qflux
Solved simultaneously across every surface and zone, 8,760 hours per year.
Modeled monthly EUI
Calibrated to 12-month utility meter · CVRMSE 8.2% · NMBE 1.3%
71.4
kBtu/ft² · annual EUI
portfolio total $4.82M · 14,240 tCO₂e
Cooling36%
Heating17%
Fans17%
Lighting15%
Plug15%
Measured meter
With ECM bundle · HVAC + envelope + lighting
The physics
Same solver · perturbed inputs
COP · LPD · U-value · ACH
Every ECM changes boundary conditions — the coupled solve shows what actually happens.
ECM bundle impact
Delta between two coupled physics runs — no stock percentages.
Energy
−27%
18.3 GWh/yr avoided · coupled-run delta, not a stock %
Cost
−$1.30M
USD/yr · payback 2.4 yr · rolls into IRR / NPV per measure
Carbon
−4,812
tCO₂e/yr · SBTi-aligned · GRESB & city disclosure ready

Concurrent runs · autonomous agents

Ten simulations at once. Agents driving them.

The engine runs elastically in the cloud. A modeler can launch ten parallel simulations and pick the winner — or hand the wheel to an AI agent that runs the sim, reads the result, tunes the model, and proposes the next move.

▤ 10 concurrent runs · Dallas Tower A
live · 10 / 10 nodes
Run · scenarioProgress · annual hoursStatus
01BaselineTMY3
✓ done
02ECM · VFD on fans
✓ done
03ECM · DCV
◉ running
04ECM · LED retrofit
◉ running
05ECM · heat-recovery chiller
◉ running
06ECM · roof R-45
◉ running
07Climate · SSP2-4.5 · 2050
◉ running
08Climate · SSP5-8.5 · 2050
◉ running
09Setpoint sweep · 72–78°F
queued
10Agent-designed HVAC v2
queued
elapsed 03:42● 10 annual runs in < 7 min
Agent · autonomous
Simulation pilot
Launch baseline + 5 ECMs
Read results · baseline EUI 71.4 · cooling-heavy 14:00–17:00
Draft HVAC v2 · DOAS + VFD fans + DCV
Submit run #10 · rank on payback + tCO₂e

Climate change weather module

Not just today's weather. Tomorrow's.

Real EPW files for today. Generated future weather for any year, scenario, or warming pathway the modeler cares about. The same heat balance, re-solved against a shifted boundary — so resilience decisions are made on physics, not hunches.

Dallas · TX · climate zone 2A
Weather · today vs. 2050 projections
✓ CMIP6 · downscaled
● 2050 · SSP2-4.5 · peak dry-bulb +4.8°F · CDD +22% · resilience risk on peak demand
Generate future weather
Dial in the warming pathway, horizon, and heat wave intensity. Roovie synthesizes the EPW and re-runs the same heat balance.
Horizon2050
Mean warming+2.1°C
Peak heat wave+6.4°F / 9-day
Humidity shift+7%
Scenario
SSP1-2.6
SSP2-4.5
SSP5-8.5
9:41●●●●LTE
▤ RoovieLive
Dallas Tower A
on-site · 142,400 ft² · run 4a3f
EUI
71.4
Δ vs baseline
−27%
Savings
$1.3M
tCO₂e/yr
−4.8k
JFMAMJJAS
◉ Running on cloudstep 8,412 / 8,760
Insight · from agent
VFD on supply fans pays back in 1.9 yr — 41% of total portfolio carbon benefit.

Run from anywhere

In the client's lobby. Simulate live.

Because the engine runs in the cloud, a modeler doesn't need a workstation to simulate. Kick off a run from an iPad during a site walk. Watch it converge on a phone. Show the owner the savings before leaving the parking lot.

  • Hand off to the cloud. No solver on the device. No waiting for a laptop.
  • Results stream back live. EUI, load profile, ECM deltas — visible as the solver progresses.
  • Share any result. Any KPI, any rollup, any chart — flexed to the question, shaped for the client.