Big Data to Optimize Energy Systems

By Amit Narayan, CEO, AutoGrid

Amit Narayan, CEO, AutoGrid

We are living through an era you could call the electricity dilemma. On one hand, electricity remains a driving force of economic development. On the other hand, electrical infrastructure's reliance on fossil fuel sources is one of the major contributors to carbon pollution threatening the very existence of life on earth. Over 1.3 billion people still aren't connected to the grid because, primarily, of the cost and complexity. Nevertheless, electric power generation is expected to grow by 93 percent by 2040 and the vast majority of that growth will occur in non-OECD nations. We simply can't meet this anticipated demand the old-fashioned way.

Fortunately, Big Data and predictive analytics provide a way forward. These technologies, combined with real-time actuation & control of the power system what I call the Energy Internet of Things-will dramatically alter the future of electricity by enabling us to utilize our energy infrastructure in a more effective way.

The needs of the electric power industry are vastly different most other industries. Utilities provide an essential service to everyone within a service territory with a vital impact on public safety and well-being. The Department of Energy estimates that blackouts and power quality issues currently cost American businesses more than $100 billion each year.

To achieve the demanding levels of performance, utilities have focused on integrating multiple levels of redundancy and control. Peak power plants constitute about $400 Billion dollars worth of investment in the US alone, and are used on an average for about 50 hours in a year (<1 percent of the time). Utilities need these because they are proven (if inefficient) tools for counteracting temporal spikes in demand. Some of the objections to renewable sources like solar and wind have been because of the variability that they can introduce. Utilities have compensated for this uncertainty through buffering, brute-force engineering and deliberately circumscribing options for the sake of control and consistency.

An Energy Internet of Things changes this paradigm by providing utilities with real-time feedback and insight for the first time. Simply put, utilities are finally able to know what their customers are doing and what they want, and are able to make better decisions to serve them.Blackouts become less frequent as predictability replaces uncertainty.

Today forecasting, a fundamental function that drives practically every operational and planning decision at the utility, is done at the system level. With the Energy Internet of Things, forecasts can be issued for millions of customers every few minutes to fine-tune predictions for power consumption across an entire region, in specific geographic areas or users along particular distribution branches. Decisions based on these micro-forecasts can be made to unobtrusively direct the flow of electrons to improve the quality of service and shave billions in operating expenses. Just to understand the magnitude of improvement, a mere 0.1 percent improvement in forecasting at a mid-size European utility with one million customers can reduce about $3 million/year in operating cost.

Software-based systems also improve as they absorb more data. Over time, the self-learning capability allows these systems to become more surgical in how they harvest power. Consumers and businesses won't know they are saving power until they get a pleasant surprise on their bill. New technologies such as solar, wind, EVs and storage will be integrated safely and more easily when supplemented by software and predictive analytics, and will give their owners a more rapid return on investment.

A virtuous cycle is easy to imagine. Big Data and software analytics will help spread the cost of capital investments in power plants and transmission infrastructure over a wider customer base and can incentivize adoption of more Internet connected devices that can participate in making the grid even more efficient. This will lower the cost of power, which in turn will make it more affordable and practical to bring the power of electrification to more people.

Implementing and integrating data systems will take time. Caution and security must still underpin any changes. Still, change is inevitable. We are at the dawn of a new energy revolution driven by software and predictive analytics, and the winners in this new world will be the enterprises who are able to harvest their data most effectively.

In this new Energy Internet world, data is power.

Advice to the Young Entrepreneurs

Having a desire to tackle really big problems, and a unique insight towards solving it is key to is the key solicit of support from people who share the same vision. While there is always some competition, there is also a very powerful ecosystem of supporters that will come to your help once you are able to articulate your vision. In my case, when I decided to tackle the energy issues, my unique insight was that the energy system, even though it looks very different than semiconductor systems that I had a lot of experience in, operated on fundamentally the same laws of physics. So the algorithms and techniques used to tackle electrical systems on the chips that we have perfected through three decades of keeping up with Moore's law could be applied to make our electricity grid run more efficiently. I got connected with the TomKat Center at Stanford, which gave me the opportunity to spend some time at Stanford and develop these ideas further. Even though I didn't know much about electricity grid at that time, I was able to partner with domain experts in the area to bring some of my ideas to market.
 

Current Issue

Featured Startups