Your browser is not supported. Please update it.

6 December 2019

Verdigris wins ABB deal to target smart buildings with ML algorithms

ABB has signed up startup Verdigris to augment its smart building offering, using Verdigris’ machine learning expertise to create a system that will predict unplanned energy consumption spikes, and then let ABB act accordingly. This could be a big win for the Silicon Valley firm, as ABB’s industrial reach is extensive.

Verdigris came to the fore through ABB’s Open Innovation program, a strategic plan to engage with accelerator programs and incubators. ABB uses the program to find potential partners that it could incorporate into its portfolio, as do many of its rivals and other large enterprises. For ABB in particular, it is looking to expand its ‘digital energy’ offering, and so is looking for anyone interested in the smart grid and smart building space. Thankfully, for the startup at least, Verdigris could be of use here.

Two new features, based on the Verdigris algorithms, have been added to ABB’s Electrical Distribution Control System (ECDS), to manage forecasting and alerts. The former aims to lower customer bills by helping them reduce their peak demand usage, essentially showing them how they could save money by better scheduling energy-intensive processes.

The latter is a predictive maintenance function, alerting customers to potential underlying problems that could be proactively fixed.  It appears to have a core focus of reducing false alerts and the consequent ‘alert overload’ that some systems can inflict on their operators.

Using the ECDS, a customer should be able to reduce their energy bills and unplanned downtime. Key to this are the Verdigris ML algorithms, which are used to process the data collected by the ECDS from all of the connected devices that the customer wants to connect. This will help study usage patterns, determine what are the expected parameters, and then issue alerts based on deviations from those points.

ABB says that it can then combine these with weather data and the historical data it holds, to accurately predict the energy usage for the next day, with the forecast being recalculated every 15 minutes. With this data, ABB will be able (in the near future) to become a very valuable partner to the energy suppliers themselves, as it would hold the key to much of the Demand Response applications that could leverage the building or facility via the ECDS.

“Our use of AI to help customers make better energy management decisions demonstrates ABB’s commitment to innovation in our products and quality in our services,” said Andrea Temporiti, digital leader for ABB’s electrification business. “With the new energy forecasting and smart alerts apps, AI drills down into the facility’s power data to pinpoint actionable opportunities for productivity improvements and energy cost savings. This innovative digital service makes it easy to take the necessary corrective actions to minimize any peak demand charges. The precision of the forecasting reduces hedging positions, narrows variability and produces meaningful energy cost savings for commercial and industrial buildings.”

Verdigris itself was founded in 2011, and is headquartered in the NASA Ames Research Center, in Silicon Valley. It has raised around $21.6mn in funding to date, with Jabil and Oyster Ventures both major investors. Public customers include the Grand Hyatt hotel in San Francisco, Marina Mechanical, the Orchard Hotel Group, W Hotels, Vention Medical, and in a different vein, Jabil and Arm. It has a pretty extensive list of testimonial use cases here.

The key component of the Verdigris offering is its IoT Energy Meter, a modular bit of kit that essentially clamps onto the electrical supply into the building, which then monitors the frequency of the cable to identify the devices being used. The device can be connected to a WiFi network, or backhaul via an LTE connection, and Verdigris claims that it takes a licensed electrician between half and two hours to install – with power for the device drawn from the nearby circuit panel, from which the monitored cable emerges.

If you want to measure the usage for a single building, then you would only need one such device. If you wanted a more granular view, you would want to install more of the devices, perhaps dividing by floor or process. These devices then feed data to Verdigris’ cloud-based analytics platform, where the real magic happens.

Verdigris has just launched its new Adaptive Automation offering, a tool for building management systems (BMS) that lets them react to changing conditions in the building without human intervention. The selling point is the recency of the data, with Verdigris arguing that conventional alternatives are using models that were last refreshed quarters or even years ago.

Thomas Chung, Head of Product Strategy at Verdigris said in the ABB announcement “Verdigris AI is 10 times more effective than traditional energy management methods. Our partnership with ABB enables our AI capabilities to reach a significantly larger ecosystem of ABB users. These energy and asset management tools will cut through the noise to deliver actionable insights, identify real energy savings and make resource allocation more effective.”