SPINNING OUT AND SPINNING UP! Utopus Insights, a bright new energy analytics start-up
Interview with Chandu Visweswariah of Utopus Insights
Conducted by: Mark Johnson, Executive Director, Kim Gaddy, Vice President, and Vanessa Barbarisi, Director at Utility Analytics Institute
Kim Gaddy: As the CEO of Utopus Insights, please describe the vision of this new organization.
Chandu Visweswariah: Certainly. We have just launched Utopus Insights. We are in day three of our operations in our new office space in Valhalla, New York. It is a very exciting time.
We see a grid and energy space that is disrupting right before our very eyes. The first disruptive trend is the availability of digital information like smart meter data, PMU data, SCADA data, historical data (e.g., demand and production data), and relevant exogenous data (e.g., weather, satellite, soil data). The second big disruption is the growth of renewables driven primarily by economic forces as the cost of renewables declines. This is changing the way that the grid operates, the direction in which power flows, intermittency, power quality, etc. Thirdly, new devices are coming on-line like heat pumps, energy storage, and electric vehicles that we expect to really take off towards the end of this decade. This will change the pattern of energy consumption in a profound way.
We believe a new intelligence is required to plan and orchestrate the grid of the future. We would like to be the leading provider of these types of solutions.
Kim Gaddy: What is its relationship with IBM and VELCO?
Chandu Visweswariah: The starting employees of Utopus Insights, including its leadership, were all members of IBM Research until recently. We have been working in energy analytics for many years. IBM decided that we would be more successful and could unlock the value more effectively as a nimble, focused, and small start-up company. So, IBM decided to spin us out into an independent company. Our intellectual property, source code, servers, people, and patents have all been given to us by IBM in exchange for an equity stake in the new company. We will partner with IBM in specific areas. For example, we will be working with IBM over the next two years on hyper-local specialized weather forecasting. While we have IBM DNA in our blood, we are an independent company and intend to move quickly, be nimble and laser focused on the energy and analytics space in a way that would be difficult as part of a large company.
For the last three years or so, we’ve been in a Joint Development Agreement (JDA) with Vermont Electric Company (VELCO) and its 18 Vermont distribution utility owners. We’ve studied solar forecasting, wind forecasting, congestion management, demand forecasting, renewables integration, Distributed Energy Resource (DER) management, etc. Together, we have co-created the most innovative solutions in these areas. VELCO decided to step-up and become an investor in Utopus and will have three distinct roles: 1) VELCO will be an investor and equity holder in Utopus, 2) VELCO subject matter experts will support the JDA and collaborate with our deep data science and stochastic optimization experts to develop unique energy analytics solutions, and 3) VELCO will be our strategic partner, helping us from within the industry to better understand what we need to do to provide the most effective solutions to utilities.
Boston Consulting Group (BCG) helped us complete a business analysis and a business model to prove to investors that we have a viable business. In providing these consulting services, BCG became excited about the concept, decided to invest in Utopus, and are a minority stakeholder. The rest of the equity will be owned by Utopus management and employees, board members, and a pool of options will be used to attract future employees.
Kim Gaddy: What is your current staffing level and staffing plan?
Chandu Visweswariah: The number of Utopus employees is currently in the low twenties. We will also have an office in India, and some IBMers working with us in IBM’s Dublin Research Lab by means of an arrangement. We will also have some contract employees in Budapest, Hungary. In addition, we are hiring and plan to round out our talent with the appropriate finance, marketing, and sales resources. I expect by late Summer or by early Fall, we will be at 45 to 50 employees.
Kim Gaddy: We understand that Utopus Insights will be bringing a full-featured analytics platform to the utility market. Why should C-level utility executives care about data and analytics?
Chandu Visweswariah: We believe going forward that it is going to be “DER by DER”. I use that catch phrase to mean that Distributed Energy Resources will be managed and orchestrated by a Digital Energy Revolution. Digital algorithms will control our grid. We’ve all imagined a day when cars get charged when the wind is blowing at night, houses are pre-heated or pre-cooled when the sun is shining, batteries get charged and discharged depending on when energy is available or not available. We envision a keen appreciation of what the weather is going to do, how that will change demand minute-by-minute with a lead time of 72 hours, how solar and wind production will change minute-by-minute over the coming 72 hours, and what actions will be required to keep the power flowing and to maintain power quality.
If you buy into such a vision or such an intelligence that will help us run our grid in the future, then you must ask, what is the architecture that we need to embrace to make this a reality? We’ve been thinking about this for years – and concluded that a specific type of platform will be needed for the grid of the future. That platform would include:
- A toolbox of algorithms, including machine learning to support future forecasting needs.
- A data curation engine able to take data from various sources including incomplete or inaccurate data and make sense of it all in an automated fashion.
- An uncertainly engine because the amount of uncertainty in the grid is only going to grow with more renewables penetration. We need to deal with the uncertainty mathematically – and not by simply worst-casing it all.
- Open APIs. We need a platform against which third parties can develop applications. Even we at Utopus know that we cannot be the only company that develops analytics solutions. We need an entire ecosystem of solutions.
- Specialized, hyper-local weather forecasting. We need to know solar radiance and shading impacts, wind interactions between multiple turbines, and wind at 80-meter hub height, not at the ground level. Integration of weather forecasting is vital.
- A common data model. The problem with a lot of utility software is that it is siloed. The EMS, OMS, and DMS each represent a pole, a transformer, or a substation in their own way. These systems cannot really talk to each other. We foresee an integrated series of applications that can deliver additional value by communicating with each other, enabled by a common data model.
- Support for hybrid on-premises, cloud (Software as a Service) and distributed Internet of Things (IoT) operation.
These are the pieces of the puzzle for the platform of the future. We are immediately bringing to market an energy analytics platform called MaestrOS that embodies all the above attributes.
Kim Gaddy: What “killer apps” do you envision?
Chandu Visweswariah: We are beginning with two “killer app families.” One is GridPulse. The traditional role of the utility is to maintain its assets and keep the grid as reliable as possible. GridPulse is a series of applications that help utilities take advantage of their data to fulfill that role in a much more intelligent way – e.g., replacing time-based maintenance with maintenance based on “electrical age” of assets, correlating smart meter data with SCADA data to make a better connectivity model to determine which transformer and which phase are connected to each meter. We have applications around managing assets better, increasing the reliability of the grid, quantifying network risk, and spending less money to achieve reliability goals.
The second family of applications is around distributed energy resources, starting with an application we call HyperCast. HyperCast takes advantage of hyper–local, specialized weather prediction to get the most accurate solar farm and wind farm energy forecast possible with a lead time of 72 hours. We are presently forecasting all farms across the state of Vermont and are achieving unprecedented accuracy. Our goal is to take energy production and energy consumption forecasts, both of which are weather dependent, and marry them with the uncertainty to understand power flows into the future. The goal is to enable utilities to make better decisions – pricing decisions, congestion decisions, relay decisions, curtailment decisions, etc. We have been building this system with VELCO for three years. In this family of products, we are bringing HyperCast to market immediately.
Kim Gaddy: Will Utopus Insights primarily be a product company?
Chandu Visweswariah: Utopus will be a product company, a service company, and above all an innovation company. We understand that some utilities have not fully grasped the power of cloud solutions and may have security concerns about cloud solutions. While we are familiar with and have addressed security requirements and needs, we architected our solutions from the ground up to support hybrid, on-premises and distributed configurations. The distributed configuration is an important distinction because we foresee a lot of intelligence being distributed in the grid rather than in the control room. Transactive energy management is a perfect example of a distributed solution. We will offer our products in the cloud, in software as a service (SaaS) form, as well as on-premises.
More than anything, we are an innovation company. The type of joint development we are doing with VELCO and Bonneville Power Administration (BPA) are prime examples of how we collaborate with clients so that two plus two becomes five. Our client’s expertise in running the grid, combined with our deep expertise in IT, data mining, business intelligence and analytics, produces a result that could not be achieved absent collaboration.
Kim Gaddy: Do you have any concrete results to date that you can share?
Chandu Visweswariah: There are many examples. We have been scrupulously benchmarking the accuracy of HyperCast. We achieved 94 percent accuracy on solar and 86 percent accuracy on wind, unprecedented accuracy when it comes to renewable forecasting. The connectivity model that I mentioned earlier is associated with a pilot that we conducted with a large utility in the Southern part of the United States, where we helped improve the accuracy of their connectivity model significantly and also predicted underground loop failure probabilities. We partnered with VELCO on demand forecasting and achieved 97.9 percent accuracy in our statewide demand forecast.
The main point that I want to make is that over the years while we were with IBM Research, we collaborated with many utilities on pilot projects and studies and achieved very, very interesting results. This ultimately led to IBM’s decision to spin us off as an independent venture based on an exciting business plan so that these innovations could see light of day as commercial products.
Utopus Insights Leadership
Chandu Visweswariah, PhD
President & Chief Executive Officer