Finding energy solutions for a changing global climate and advocating for women to choose careers in science and technology drive Sakshi Mishra’s mission as an energy researcher and power systems engineer.
Since joining the National Renewable Energy Laboratory (NREL) in 2018, Mishra has worked in the Integrated Applications Center researching artificial intelligence (AI) approaches for accelerating the transition to clean and resilient energy infrastructure through smart energy systems design. She also leads the Denver chapter of Women in AI (WAI), a nonprofit think tank that promotes women in AI fields.
Mishra spent some time talking to us about her research and mission at NREL.
How did your passion for addressing the challenges of climate change lead you to NREL?
Throughout my education and early career at American Electric Power, I focused on grid planning engineering to research the integration of higher levels of renewable energy generation into the grid to solve climate change problems from extreme events. Then, I discovered the research nexus of renewable energy and AI, and it made sense for me to join NREL. Working here allows me to blend my power systems knowledge with AI to develop forward-looking solutions to issues surrounding high penetration of intermittent energy generation into the grid and be part of the solution for the climate change problem.
How do you define AI, and how does it impact your research?
AI is an umbrella term for any computer technology that adopts an algorithmic set of rules, which appear to emulate human reasoning, like a building operations software is designed with a rule-based “schedule” to dispatch various generation technologies or dispatchable loads. Machine learning is a more complex subset of AI, in which an environment not only follows rules but can be trained to “discover” the rules, like fault-detection software using a training set of tagged events to figure out if the air conditioning system has a fault. Deep learning is a type of machine learning where the system not only looks at results from different algorithms but also uses a predictive model to train itself. It mimics the human brain by constantly testing a hypothesis against results and adjusting the hypothesis according to these results. My current research leverages machine learning and deep learning algorithms for predictive analytics applications that enable preemptive management of building energy usage by providing accurate forecasts of energy consumption.
Tell us about your current projects in the Integrated Applications Center.
As part of my predictive analytics work for NREL’s Intelligent Campus project, I have built a load forecasting framework that uses a neural network capability, or deep learning, to predict the energy consumption of various campus buildings and achieve better performance compared to traditional approaches.
I am also the development lead for REopt Lite API, an optimization tool aimed at studying the techno-economic feasibility of behind-the-meter distributed energy resources. REopt Lite API provides energy resource planners with concurrent, multiple-technology integration and optimization used to improve cost savings while achieving ambitious energy performance goals. The tool is being used by industry practitioners, consultants, renewable energy developers, research labs, as well as academia.
I recently wrapped up another project where we collaborated with industry partners to develop a machine learning–based tool for automatically tagging building management system metadata using Project Haystack schema. Having an automated way to tag building controls data lays a strong foundation for deploying advanced energy analytics applications where access to properly labeled data is a necessary requirement. Operational cost savings and decreased carbon footprints are the benefits associated with employing efficient energy management systems in buildings.
What do you see for the future of the “Energy + AI” field as a whole?
I believe that solving current global energy problems (including climate change) requires a blended approach consisting of classical knowledge and AI. AI is an ally for research that can augment existing knowledge to speed up our transition to clean energy.
You’re a strong advocate of helping women pursue careers in science, technology, engineering, and math (STEM). Where did your sense of mission originate?
Growing up in a developing country in the 1990s, I experienced a fair amount of resistance from elders in the family when I wanted to become an engineer. My grandfather had the mentality that investing in a girl’s education is not a good use of the constrained resources they had. Fortunately, my parents (who both had master’s degrees in nonengineering majors) were very supportive of my continued education, so I made the tough decision to choose an electrical engineering major and go to college, and it worked out well for me.
I strongly believe choosing to pursue a STEM career does not have to be a “swim against the tide” decision for young girls and women in today’s world. We need the huge, untapped potential of female talent whose contributions can help build a sustainable future for the planet. However, in the United States, women comprise only 28% of the STEM workforce overall and only 16% of the engineering workforce. Science and engineering fields cannot afford to lose talented individuals because of their gender.
What are you doing to help young girls and women get into science and engineering? How can other STEM professionals get involved?
When young girls are contemplating their career choices or young women are planning to pivot into the STEM field, they need affirmation from parents, mentors, and colleagues to help instill their confidence. My work is aimed at raising this awareness in society as a whole, like leading the Denver chapter of WAI and getting the message out through outreach, like in the Society of Women Engineers podcast in July.
There are multiple ways for STEM professionals to help encourage young women to pursue STEM careers: (1) Join WAI as a volunteer to help spread its message of increasing female representation in the field of AI, (2) encourage little girls and young women in their immediate families/circle of acquaintances to pursue STEM education, and (3) engage with communities like Soapbox Science or the Society of Women Engineers to promote role modeling for young girls.