Page 31 - Energize June 2021

VIEWS AND OPINION

“Utilities sharing operating experiences, use cases, and just as importantly, their data across scenarios to play forward.
the community we’re building with our AI. EPRI initiatives will enable the acceleration of AI On the energy generation side, EPRI
technology deployment.” continues to explore machine learning
Feldman hosted the last panel discussion at the Reverse Pitch event, where speakers from models to reduce O&M costs. One project
Stanford University, Massachusetts Institute of Technology (MIT), Idaho National Lab (INL), SFL that has advanced rapidly is wind turbine
Scientific and EPRI discussed the future of AI for electric power. component maintenance. EPRI research
“The utility sector by nature is a risk-averse industry, but it’s time to think about how shows the current gearbox cumulative
to adapt their business models to embrace new AI technologies,” said Liang Min, managing failure rate during 20 years of operation
director of the Bits & Watts Initiative at Stanford University. “If utilities dedicate resources to is in the range of 30% (best case scenario)
identifying right use cases and conducting pilot programmes, I think they will see benefits, and to 70% (worst case scenario). When a
it will eventually lead to enterprise-wide adoption.” component like a gearbox prematurely
“Validating different AI applications will help end-users and regulators determine their fails, operation and maintenance (O&M)
effectiveness, without eroding safety and reliability,” said Idaho National Lab Nuclear National costs increase, and production revenue is
Technical Director, Craig Primer. “We need to overcome those barriers to drive adoption and lost. A full gearbox replacement may cost
reduce the manual approaches used today.” more than $350000.
In 2020, a large California investor-owned utility and EPRI member inspected 105000 EPRI is researching and testing a physics-
distribution and 20500 transmission structures. Conservative estimates gave the utility 750000 based machine-learning hybrid model that
images for staff to review and evaluate. That’s about 3500 person-hours and cost more than can identify gearbox damage in its early
$350000 at a standard utility staff rate for inspection review work. stages and extend its life. If a damaged
With the wider adoption of drone technology in the very near future, significantly more bearing within a gearbox is identified early,
images will be available than ever before. However, without augmented evaluation capabilities the repair may only cost around $45000, a
offered by AI, evaluation costs will correspondingly and exponentially increase. Inspections are savings of nearly 90%.
complex tasks which become more complicated by using drones. These projects all demonstrate real
EPRI is working with utilities and the AI community to build a foundation for machine solutions that are deployed and are
learning to facilitate models that can detect damaged T&D assets and assist staff in more showing real results and increases in
efficiently managing the volume of images. But just as critically, it’s also taking on the tasks of efficiencies. Many are set to be further
collecting, anonymising, labelling, and sharing imagery for model development. These data deployed to enable the global energy
sets, along with a utility consensus taxonomy and data labelling process, are needed to achieve systems transition.
desired improvements in efficiency, predictive modelling, damage identification and repair/ “AI is at a point where I believe the
replacement of equipment. technology has advanced to support
During the Reverse Pitch event, Boston-based SFL Scientific, an AI consulting company, scaling up adoption. Meanwhile we
highlighted the significant technical and operational challenges associated with development know that society depends on electric
of end-to-end AI applications, including validating machine and deep learning models, power continuously, to run everything
optimising their performance long-term, and integrating the output into workflows and from health care and emergency
production pipelines. resources to communications
“AI is hard, it’s not easy,” said Michael Segala, CEO of SFL Scientific. “Introducing AI is infrastructure and in today’s current
essentially breaking people’s workflow and injecting risk into their process, which can break situation, working from our homes,”
down adoption. This is maybe significantly more difficult for utilities based on the regulations said Neil Wilmshurst, Senior Vice
that are set and consequences of getting things wrong. But there’s a great ecosystem, like the President of EPRI’s Energy System
folks here (at the Reverse Pitch) that will help with the journey and be a part of that adoption, Resources. “Reliability and resilience
so utilities don’t fail and risks are reduced.” have never been more essential in a
time when we’re also making a critical
Now is the time to accelerate adoption of AI technologies energy system transition to meet global
There’s a new layer to consider: the increasing urgency to protect against threats to our climate goals and demand needs. AI
energy infrastructure, recently heightened following the May cyberattack on one of the must be a tool in the toolbox, and
US’s largest fuel pipelines. the time is now – not tomorrow – to
“As physical threats to energy grids increase, connecting measures to ensure grid accelerate those applications.”
readiness, energy security and resilience becomes critical,” said Myrna Bittner, Founder
and CEO of RUNWITHIT (RWI) Synthetics, an AI-based modelling company. “Add on the Jeremy Renshaw is Senior Programme
pressures of electrification, decentralisation, climate change and cyberattacks, and the Manager, Artificial Intelligence, at the
demand grows for even more adaptive scenario planning, mitigating technology and Electric Power Research Institute (EPRI).
education.”
Bittner presented RWI’s Single Synthetic Environment modelling approach at the EPRI Acknowledgement
Reverse Pitch event. These geospatial environments include hyper-localised models of the This article was first published by POWER
people and businesses, the infrastructure, technology and policies, then enable future Magazine, www.powermag.com

energize | June 2021 | 29