Delivering cost-effective, high-granularity, virtual-meter energy data

Thomas Chung
Verdigris Tech
Published in
4 min readAug 21, 2020


One year ago we launched our fifth generation of Verdigris hardware. Our proprietary hardware crucially samples current and voltage at an extremely high frequency (8 kHz, or 8 thousand times every second). Our algorithms use this information to create a virtual physical layer of data. One of these enhanced data streams is virtual equipment-level energy information. This process is commonly known as high-frequency energy disaggregation.

The big problem

Modern commercial buildings need consistent, reliable energy to operate. These operations range from space heating and cooling, to lighting, to conveyance, refrigeration, and machine operation. With each passing year these environments become increasingly complex and energy intensive. Our presence in these buildings correlates. According to the EPA, the average American spends about 90% of our lives indoors.

Yet, of the $4 trillion spent on energy in buildings in 2016 to support our activity, 50% of it was waste.

The industry problem

The number one strategic challenge for commercial building operators keeping CEOs up at night should be energy. In 1996, energy spend was less than 5% of the operating budget for Factories, Distribution Centers, Hospitals and Hotels. 20 years later, in 2016, nearly ¾ of their profit was spent on electricity alone. The rapid shift to renewables and changing demand tariffs in the next 5 years will force 1/3 of these company out of business.

It gets worse. Finding ways to become more energy efficient have not progressed in the last two decades.

The incumbent energy industry has invested primarily in control solutions. These solutions are expensive. With the added cost of setup and service contracts, these investments yield marginal returns and put their affordability outside the reach of all but the largest commercial entities. No wonder a meager 5% of buildings have these types of controls in them.

Energy Management Software technology has not kept pact with the changing needs of buildings. And many undifferentiated, snake-oil solutions have emerged, promising easy guaranteed returns and no real insights for customers.

Building audits have show some effective, but expensive and manual processes. Within a few operating quarters, building energy drift can reach of 20%, necessitating additional periodic audits and putting returns always just out of reach.

The solution — Change the cost structure of getting high quality data inside the building.

Energy disaggregation is considered the holy grail of energy efficiency because it solves this problem for the industry. From a single sensor on the electrical panel, Verdigris is able to track and separate individual appliances from the context of the circuit and panel.

What is Energy Disaggregation

Disaggregation is the process of taking an aggregate electrical signature and parsing out individual electrical signatures.
A live circuit where we have used machine learning to cluster individual waveforms into their underlying equipment signatures
This video clip illustrates an aggregate energy waveform decomposed into individual devices.

Virtual Device Data via API, now available

As part of our goal to solve this industry problem, we are proud to announce we’ve taken the further step to make time-series virtual device data available via API for our customers.

Here are some example results from running disaggregation on one phase of a Process Chiller circuit with one of our customers. These are the devices it finds — 7 devices plus the baseload.

Spark line plot of the devices and the real power of the circuit.

With a single physical sensor on the electrical panel we have effectively created 7 virtual meters on the devices on this circuit. This data is wirelessly accessible, available at minutely granularity, and cost orders of magnitude less than retrofitting wireless sensors on downstream equipment.

NIALM is one of the most promising ways to cost effectively capitalize on the energy efficiency opportunity in the commercial built environment. I’m enthusiastic about demonstrating the potential for driving both top and bottom line dollars for our partners and customers.

I can’t wait to see what our team releases in next iterations as we work to improve automated labels and equipment classification on disaggregated devices.

The enthusiasm from system integrators, enterprise-level customers, and ecosystem developers has been truly exciting. We look forward to working with these partners to develop solutions for the industry at scale.

Are you interested in working with us?

Click here to work full time on meaningful problems with exceptional people.

Click here to talk about partnering with us.



Thomas Chung
Verdigris Tech

Builder, Day Dreamer, Idealist, Persister, Reader, Runner, Learner, and Father.