Digital Twin: Energy Optimization at Chicago Skyscraper

This case study reflects an ongoing energy analytics initiative I lead in partnership with an ambitious Class A commercial office building in downtown Chicago.

Executive Summary

I led an end-to-end initiative to build a real-time energy digital twin for a Class A commercial office building in downtown Chicago, transforming fragmented building data into a system for continuous energy optimization and decision-making.

The core challenge was not basic efficiency (the building was already highly optimized) but rather unlocking the next layer of performance through visibility, modeling, and behavioral influence.

I defined and delivered a system that integrates submetering, occupancy data, weather modeling, and equipment telemetry into a unified analytics layer. This enabled:

  • Tenant-level energy visibility

  • Equipment performance benchmarking

  • Real-time anomaly detection

  • ROI analysis for capital upgrades

The result is a live decision-support system used by engineering and property teams to monitor performance, detect inefficiencies, and guide both operational and strategic energy decisions.

Introduction

This initiative was developed in partnership with a long-time Cohesion customer - a large Class A office building with strong ambitions to lead in sustainability and energy performance.

Despite prior investments in efficiency, the building lacked a unified way to:

  • Understand energy usage at a granular level

  • Connect energy consumption to real-world drivers (occupancy, weather, behavior)

  • Quantify the impact of operational or capital decisions

Through strategic planning sessions I led with building leadership and internal stakeholders, we aligned on a vision: Create a live digital twin of the building’s energy system that allows teams to see, model, and act on performance in real time.

This system would unify previously siloed data sources into a single layer of insight spanning:

  • Submetered energy usage

  • Occupancy and tenant activity

  • Weather conditions

  • Equipment telemetry

Problem Statement

How do you make meaningful energy improvements when your systems are already optimized?

That’s the question this building team posed but well, it’s a trick question. Their systems aren’t completely optimized, they just didn’t have the visibility and insight available to optimize beyond the traditional approaches.

Traditional approaches focused on:

  • Static reporting

  • Aggregate energy usage

  • One-time efficiency projects

So we planned beyond that:

  • Monitor energy consumption at the tenant and floor level

  • Isolate equipment-level inefficiencies

  • Model expected energy usage based on real-world drivers

  • Quantify ROI of upgrades with defensible analytics

  • Influence tenant behavior through visibility and feedback

Solution Architecture: Building the Digital Twin

We broke down the project into several parts, including these currently active solutions:

  1. Tenant-Level Energy Intelligence

  2. Equipment Performance & ROI Modeling

  3. Real-Time Energy Performance Baseline (“Health Gauge”)

Tenant-Level Energy Insights

We coordinated closely with building engineering to install tenant- and floor-level submeters across multiple phases. This effort aligned with a large tenant move-in and included the integration of CPower metering infrastructure. My job was to integrate with CPower and ensure the incoming submeter data flowed into Cohesion’s platform, mapped accurately to our space hierarchy (building > floor > space > equipment, etc).

The execution of this plan went as follows:

  • CPower data ingestion

    • Partnered with CPower to ingest submeter data via API

    • Led internal engineering coordination to build reliable ingestion pipelines

    • Ensured continuous, accurate data flow into Cohesion’s platform

  • Data modeling and alignment

    • Mapped submeter data to Cohesion’s building hierarchy:

      • Building → Floor → Tenant → Space → Equipment

    • Integrated HVAC equipment relationships to enable system-level analysis

  • Reporting

    • Created tools and live dashboards with visibility into each tenant suite:

      • Occupancy

      • Indoor Air Quality

      • Maintenance Requests

      • Associated Equipment

      • Power / Energy Performance

This mapping unlocks per-suite energy analysis and enables real-time feedback loops for tenant engagement.

Equipment Telemetry & Chiller Upgrade Analysis

The building provided a large data dump of an aging chiller: power draw, current, voltage, phase angle and other data across multiple chillers. Shortly after, they replaced the chillers with new unit in a phased approach.

I led the development of an analysis framework to isolate true efficiency gains.

We modeled energy usage as a function of:

  • Weather (degree days via NOAA data)

  • Occupancy (access control + elevator data)

  • Equipment runtime characteristics

By controlling for these variables, we could:

  • Normalize energy usage across conditions

  • Compare old vs. new equipment performance

  • Quantify actual efficiency gains

Outcome:

  • Identified 12–15% reduction in normalized energy use

  • Estimated <3 year payback period

This provided a defensible business case for future capital investments.

Real-Time Energy Performance (“Health Gauge”)

Working with engineering and analytics teams, I helped define a new concept: a real-time “energy performance gauge”. This visualization answers: “is the building performing as expected right now?”

Model Inputs:

  • Outside air temperature

  • Occupancy

  • Real-time power consumption (via submeter integration)

  • Day-of-week and seasonal patterns

Method:

  • Applied regression modeling to establish expected energy usage

  • Introduced feature transformations (e.g., cooling degree days × ln(occupancy))

  • Validated model accuracy (~5% average deviation on historical data)

Output:

A live performance signal showing:

  • Within expected range

  • Above baseline (inefficiency)

  • Below baseline (optimized performance)

Impact:

  • Enables real-time anomaly detection

  • Supports load-shedding and operational decisions

  • Surfaces both equipment issues and behavioral anomalies

This transformed energy monitoring from passive reporting into an active control system.

Data & Product Layer

Beyond individual features, this initiative required building a unified data and decision layer.

I led:

  • Definition of data models connecting energy, occupancy, and equipment

  • Development of pipelines integrating IoT, APIs, and external datasets

  • Creation of analytics tools used by engineering and property teams

This aligns with Cohesion’s broader platform strategy of unifying building systems into a single digital layer for operational intelligence .

Results (So Far)

  • Full mapping of tenant-level submeter data to building model

  • Operational dashboards for real-time energy and tenant insights

  • Regression-based baseline model combining weather and occupancy

  • Live energy performance gauge actively used by site teams

  • Chiller upgrade analysis framework supporting capital decisions

Early results show:

  • 12–15% reduction in normalized energy usage

  • <3 year estimated payback on equipment upgrades

Takeaways

This project highlights a shift in how advanced buildings operate:

  • Once systems are optimized, visibility and behavior become the next frontier

  • Energy optimization is no longer a one-time effort, it is a continuous system

  • The value of a digital twin is not visualization, but decision-making capability