Independent Physics Engine

Roovie runs its own physics engine.

This is not a wrapper around EnergyPlus or a call to a third-party simulation service. Roovie maintains an independent thermal and energy simulation engine that takes a structured building model and produces a full year of hourly results across thermal, energy, financial, and emissions categories.

That distinction matters because it gives Roovie control over how buildings are simulated, how results are structured, and how fast the feedback loop is between design changes and performance outcomes.

In One Line

One building model in. 8,760 hours of thermal, energy, cost, and emissions results out.

What The Engine Computes

A single simulation run produces results across four categories:

Thermal

• Building net heat flux in watts and per unit area
• Heat flux component breakdown: conduction, convection, radiation, and solar
• Peak heating and cooling loads with hourly breakdown by source
• R-value and U-value calculations in both metric and imperial
• Thermal comfort metrics including hours above and below comfort, compliance percentage

Energy

• Total energy usage with electricity and natural gas separation
• End-use breakdown across 15+ categories: heating, cooling, ventilation, lighting, equipment, refrigeration, hot water, fans, pumps, heat rejection, and more
• HVAC subsystem analysis: compressor, indoor fan, outdoor fan, ventilation fan
• System efficiencies: heating COP, cooling COP, overall HVAC efficiency
• Energy Use Intensity in kBtu/sqft, kWh/sqm, MJ/sqm, and GJ/sqm
• Source EUI and site EUI with ENERGY STAR score equivalents

Costs

• Total energy cost with electricity and gas breakdown
• Cost per area metrics
• Cost by end-use category: heating, cooling, lighting, equipment, ventilation, hot water
• Peak demand charges
• Lifecycle costs including maintenance and replacement
• Full utility rate context showing which tariff rules were applied and why

Emissions

• Total CO2e emissions with Scope 1 and Scope 2 separation
• Emissions by source: electricity, natural gas, district heating
• Carbon intensity per area
• Net carbon balance
• Time-based emissions with peak hourly identification

All four categories are available at monthly, weekly, daily, and hourly resolution from a single simulation run.

How The Engine Works

The engine receives a structured building model containing geometry, zones, assemblies, materials, HVAC systems, schedules, and weather data. It returns a complete performance profile.

Roovie building model
  + weather file (8,760 hourly records)
  + simulation configuration
  → physics engine
  → 8,760 hourly results
  → monthly, weekly, daily aggregations
  → performance summary with all four metric categories

The engine runs as a dedicated service with a 10-minute computation timeout. Results can be saved to the portfolio database for comparison, calibration, and reporting workflows downstream.

Physics Modes

The engine exposes three physics configuration axes that control how thermal calculations are performed:

Surface Film Mode

Controls how convective heat transfer coefficients are calculated at building surfaces.

• Dynamic: Recalculates surface convection film coefficients at each timestep based on current wind and temperature conditions
• Static: Uses fixed per-surface film values for faster computation
• Auto: Selects per surface based on configuration and exposure

Solar Irradiance Model

Controls how sky diffuse radiation is decomposed for building surfaces.

• Perez: Anisotropic sky diffuse model for clear and partly-cloudy conditions. Highest fidelity for real-world solar analysis.
• Isotropic: Uniform sky diffuse assumption. Simpler and useful for benchmarking.
• ASHRAE Blend: Diffuse-fraction blend designed for ASHRAE 140 alignment and code compliance testing.

Exterior Convection Method

Controls how wind-driven convection is calculated on exterior surfaces.

• DOE-2: Roughness-dependent forced convection following the EnergyPlus reference standard
• TARP: Combined forced and natural convection model
• MoWiTT: Empirical wind-driven correlation from field measurements
• Adaptive Mixed: Blended forced and natural convection with height-dependent natural convection

These modes can be mixed and matched depending on whether the simulation is targeting ASHRAE compliance testing, real-world design analysis, or comparative scenario modeling.

Variable Selection

The engine supports granular control over which output variables are returned and stored. This matters for performance and storage efficiency at portfolio scale.

Variables are organized into preset groups:

• thermal: heating and cooling loads, comfort metrics, heat flux
• energy: consumption by fuel type and end use
• hvac: equipment-level performance, COP, fan energy
• weather: dry bulb, humidity, wind, solar radiation
• emissions: carbon by source and scope
• cost: energy cost by source, end use, and lifecycle
• solar: solar gains by surface and orientation
• assembly_summary: per-assembly performance summaries
• assembly_detailed: per-assembly timestep data
• assembly_layers: layer-by-layer thermal performance
• dashboard: curated set for overview displays

Users can combine presets, add specific variables, exclude others, and filter by scope (building, zone, assembly, or weather level). This keeps result payloads focused on what the analysis actually needs.

Batch Simulation

For portfolio-scale work, the engine supports batch execution across multiple buildings with centralized configuration.

A batch run includes:

• A named simulation profile with default parameters
• Per-building overrides where needed
• Execution controls: concurrent or sequential mode, max concurrency (up to 10), retry count and backoff
• Scheduling: run immediately, start at a specific time, or rate-limited windowed execution
• Naming templates for automatic result labeling

Batch status tracking covers the full lifecycle: draft, scheduled, queued, running, completed, completed with failures, and cancelled. Individual items within a batch track their own status independently.

Simulation Profiles

Reusable simulation profiles store standardized configurations for recurring analysis patterns. A profile includes:

• Default request parameters (date range, physics modes, output options)
• Weather source mode (use the building's assigned weather file or a fixed reference)
• Execution settings (concurrency, retries, scheduling)
• Naming template for generated results

This allows organizations to define a standard simulation methodology once and apply it consistently across buildings and over time.

What Feeds The Engine

The engine operates on the same building model that users create and edit inside Roovie:

• Zones: Thermal blocks with geometry, occupancy, and schedule assignments
• Assemblies: Walls, roofs, floors, windows, and doors with layer-by-layer construction
• Materials: Thermal properties including conductivity, specific heat, density, and surface characteristics
• HVAC Systems: Equipment type, capacity, efficiency, fuel source, and control strategies
• Schedules: Occupancy, lighting, equipment, thermostat setpoints, ventilation, and HVAC operation
• Weather: 8,760 hourly records with temperature, humidity, wind, solar radiation, and sky conditions

The building does not need to be exported or converted to a separate format. The engine reads the Roovie building model directly.

Ground Contact Analysis

When enabled, the engine produces detailed ground-contact and slab thermal analysis:

• Floor heat flux in watts and annual kWh
• Peak floor heat flux with timestamp
• Zone heat flux and mean temperature
• Hourly floor heat flux time series
• Surface temperature arrays for slab and soil boundary points

This is critical for buildings with significant ground-contact area where slab losses represent a meaningful portion of total energy load.

Calibration Integration

The engine connects directly to Roovie's calibration system. After a simulation completes, results can be compared month-by-month against actual utility data:

• Normalized Mean Bias Error (NMBE) for systematic over or under-prediction
• Coefficient of Variation of Root Mean Square Error (CV-RMSE) for prediction variability
• Per-fuel-type comparison: electricity and natural gas tracked independently
• Configurable thresholds aligned with ASHRAE Guideline 14

This means the physics engine does not operate in isolation. Its results are tested against real-world building performance and refined through calibration.

Why An Independent Engine Matters

Most building energy analysis tools depend on a third-party simulation kernel. That creates constraints:

• The tool cannot control the simulation's internal logic or output structure
• Result formats are dictated by the kernel, not the product
• Performance optimizations require working around the kernel's architecture
• Adding new output variables or physics modes requires waiting for upstream changes

Roovie's independent engine removes those constraints. The simulation logic, output structure, variable system, and performance characteristics are all under Roovie's direct control.

That makes it possible to:

• add new output variables without waiting for a third-party release
• optimize for specific building types or analysis workflows
• structure results in the exact format that downstream features need
• run at portfolio scale with batch execution and variable selection
• evolve the physics independently as building science and code requirements change

Bottom Line

Roovie's physics engine is the computational foundation that everything else in the platform builds on. Thermal visualization, calibration, compliance testing, investment planning, AI-powered design agents, and portfolio analysis all depend on the engine's ability to turn a building model into a complete, hourly performance profile.

It is purpose-built for the kind of work Roovie does: turning building geometry and systems into actionable energy, cost, and carbon outcomes at scale.