Digital Mirror™ Proof of concept

Measured data and a calibrated model, side by side

Wireless sensors record how the building operates. A physics-based energy model is calibrated to those measurements and to the utility history. Comparing the two shows where operation has changed and estimates what each improvement would save. The kit installs in one day and does not need access to the building network or BAS.

The model as a reference

How the comparison works

The model is built from the building's geometry, systems and schedules, then adjusted until it matches measured energy use. At that point it reproduces how the building runs today, faults included.

After calibration, measured use that departs from the model points to something that changed, such as a new schedule, an override or a failed component. The same model estimates what a proposed improvement would save before anyone pays for it.

Measured performance against the calibrated model During calibration, measured and modeled energy use track each other closely. Later, measured use rises above the model, flagging a change or fault to investigate. A lower dashed line shows the modeled effect of an improvement. Calibration Energy use Time Measured use departs from the model Calibrated model Modeled effect of an improvement Measured and modeled agree
After calibration, the model is the reference for expected performance. Departures from it are findings, and the model estimates the effect of improvements.

Measures

What the model can evaluate

Each measure is run in the model of this building, so interactions between measures are included. A lighting retrofit changes the cooling load, and a new control sequence changes how the equipment runs. A spreadsheet calculation treats each measure on its own.

Envelope and loads

Windows, insulation, roofing and air sealing, lighting, and plug and process loads.

HVAC equipment

Equipment replacement, heat pump electrification, economizers and ventilation changes.

Controls and operations

Schedules, setpoints, resets, demand-controlled ventilation and optimized start. Detailed sequences are modeled with custom controls and treated as estimates.

Solar, storage and microgrids

The calibrated hourly load profile is the starting point for sizing solar, battery storage and microgrids, including battery dispatch against the utility tariff and resilience scenarios.

Project development

From findings to a project

The model is also the working tool for project development. Measures are grouped into bundles and run together, so a bundle's savings reflect how its measures interact.

  1. Select measures from the findings and the owner's goals.
  2. Bundle them and run the bundle in the calibrated model.
  3. Build the business case: cash flow, payback, incentives and the effect on utility costs.
  4. Lay out a decarbonization roadmap showing emissions and electrification steps by year.
  5. Hand off an implementation plan with scope, sequence, and the data needed to verify results.

Reports when they're needed

The same model and data produce static reports on request: audit-style energy reports, retro-commissioning findings, Building Performance Standard documentation and decarbonization plans. Updating a report means rerunning it with current data.

How it works

Each deployment starts with written success criteria. At the end, the owner decides whether to integrate, expand or remove the kit.

01 · Capture

Geometry

Where drawings are missing or unreliable, a 360° walkthrough scan becomes a measured floor plan and then the energy model's geometry. Otherwise the model starts from drawings and a site walk.

02 · Measure

Wireless sensors

Clamp-on current sensors on panels and key circuits, plus occupancy counters and CO₂, temperature and humidity sensors. They report over LoRaWAN to an on-site gateway with its own cellular connection.

03 · Calibrate

Energy model

An EnergyPlus model is adjusted until it matches the utility data and the circuit-level measurements. Calibration passes or fails on ASHRAE Guideline 14 statistics.

04 · Compare

Findings

Analytics compare measured and modeled performance as data arrives. Schedules, runtimes and loads that don't match become findings, and proposed measures are evaluated in the model before any money is spent.

Building network and BAS

Deployment without the building network

The kit includes its own sensors, gateway and cellular connection. It does not connect to the building network and does not read or write BAS points, so the first results don't depend on network projects or controls contractor schedules.

Data is exported in open formats and built for Project Haystack tagging. The gateway supports BACnet/IP, Modbus TCP and MQTT, so later integration with building systems can use the same installation.

When to integrate

Integration starts once IT/OT governance is in place (a named system owner, network decisions, and read-only access first), and when one of these applies:

  • A finding needs control action, such as a schedule, setpoint or sequence change, to capture the savings
  • The owner wants alarms and fault detection inside their existing operations workflow
  • Verifying a completed project needs data that only the BAS has
How the Digital Twin builds on this

Security and data

How the kit is secured

  • Nothing connects to the building network or the BAS.
  • The cellular connection has no public internet address, and no inbound ports are exposed to the internet.
  • Remote access runs over outbound-only encrypted tunnels. Internal services are firewall-restricted to the kit's private network.
  • Data is backed up nightly, with off-site copies transferred over an encrypted connection.
  • IT receives a device inventory and connection summary before install. If IT prefers, the gateway can run on the owner's network instead.
  • When integration happens, it starts read-only and runs under the owner's IT/OT governance.

Your data stays yours

The building's data and energy model belong to the owner. They're delivered in open formats during the engagement and handed over or deleted at the end, the owner's choice. Stride doesn't sell or share building data.

From the proof of concept

First deployment

  • 2.5electrician-hours for 16 current sensors across three panels, with no shutdown
  • 34sensors installed in one day, with no wall or ceiling penetrations
  • 99data points recorded at 10-minute intervals
  • 0connections to the building network or BAS

Before the model was calibrated, the install and the sensor data turned up a rooftop unit with no economizer, an unevenly loaded electrical service, and a night base load on the plug and process panels. Calibration is next. Read the Field Notes.

What you get

  • Circuit-level current with derived energy use, equipment runtime, occupancy, and CO₂, temperature and humidity by zone
  • As-built floor plan and model geometry, from a scan where warranted
  • Calibrated EnergyPlus model, checked against ASHRAE Guideline 14 criteria
  • Live dashboards comparing measured and modeled performance
  • Operational findings and prioritized efficiency opportunities
  • Measure bundles run in the model, with cash flow, payback and emissions results
  • Decarbonization roadmap and implementation plan
  • Audit-style, retro-commissioning and compliance reports from the same data when needed

The sensor kit belongs to Stride and is removed at the end. Data exports and models are delivered in open formats, and permanent metering installed in later phases belongs to the owner.

How AI is used

AI tools help with planning, configuration, data-pipeline code, data checks and documentation. Each step is checked against the hardware and the data.

The energy model is an EnergyPlus simulation, and calibration passes or fails on ASHRAE Guideline 14 statistics.

Partners

For ESCOs, consultants and energy managers

The Digital Mirror can be part of your service. The firm with the client relationship decides how much Stride does.

Deploy and set up

Stride installs the kit and builds the calibrated model. Your team takes the data, model and findings into its own deliverables.

Ongoing support

Stride stays on for analytics, model updates and reporting while your firm leads the client relationship.

Under your name

The Digital Mirror is available as a white-label service, delivered under your firm's brand.

In every arrangement, the building owner owns the data and the model. Talk to us about partnering.

What comes next

When the results justify it, the Digital Mirror's data is connected to the building's controls and maintenance systems, with fault detection on the same data. That is the Digital Twin phase.

See the full approach