In 1859, a coronal mass ejection erupted from the sun and, just 17 hours later, hit the earth’s magnetosphere, causing the aurora to be seen as far south as the Caribbean and allowing some telegraphs to transmit messages even after being unplugged. If a geomagnetic disturbance (GMD) of this magnitude occurred today, the US alone would face up to $3 trillion in damages, with recovery taking up to two years due to the time required to rebuild critical power generating equipment. In 2012, a similarly sized coronal mass ejection just missed the Earth.
Today, the US power grid can be decomposed into two abstract portions: The generation side, which handles the production of electricity and includes nuclear power plants, coal-burning plants, and solar farms; and the distribution side, which transmits the generated power to your home. Most of the related Internet of Things (IoT) press has described the latest efforts to instrument the distribution side of the US power grid with smart meters that collect data for the proverbial “last mile.” On the generation side, a similar sensor rollout has been underway but with less fanfare.
Over 2,000 phasor measurement units (PMUs) have been installed to date throughout the US, capturing time series describing system voltages and currents at rates of up to 120 samples per second. All told, over two billion sensor values per day are captured that describe the daily operation of the grid. For the first time, we have the data and ability to explore the operation of power grids, and to identify and explain anomalous behavior in order to protect our power grids from the tremendous potential of damage.
PingThings is a well-funded data science startup that has been bringing big data techniques, agile methodologies, web philosophies, and open source technologies such as Python and Spark to an industry bound to physics-based modeling and simulation. We have already shown that, for a few examples, PMU data reveals a clear correlation between a GMD event and the anomalous behavior of reactive power in utility substation equipment. We are currently working to determine how early a signature becomes evident during a large GMD event, and how the subtle effects of smaller scale GMD events add up to impact the grid. Both of these could enable the use of PMU-based alerts and early warnings of growing GMD events to reduce GMD-related risks through proactive asset and grid management. Our large-scale vision is to unearth and identify all the patterns in this new source of data, and bring a new level of situational awareness to the power grid.
This talk is the story of bringing a completely new philosophy and approach to a 100-year old industry. We will discuss:
and answer the following questions:
To our knowledge, this is one of the first applications of ML techniques to big data in the power industry.
Sean Patrick Murphy, with degrees in math, electrical engineering, and biomedical engineering and an MBA from Oxford, has served as a senior scientist at Johns Hopkins University for over a decade, advises several startups, and is now the chief data scientist for PingThings. Previously, he served as the chief data scientist at a series-A funded healthcare analytics firm, and as director of research at a boutique graduate educational company. He has also cofounded a big data startup, and Data Community DC, a 5,000-member organization of data professionals.
Jerry has been involved in the enterprise software industry as a serial entrepreneur and consultant for more than 30 years.
In 1993 he founded Flashpoint Systems, a leader in internet and Intranet application development. Following Flashpoint System’s 1995 merger with a leading systems integrator, he served as CTO, directing product research and development. He has been a pioneer in the development of internet-based enterprise business applications across multiple industries.
In 1998, Jerry and Simon Arkell founded VERSIFI, Inc. The company’s initial product was considered the world’s most visionary technology in the area of enterprise content management by Gartner Group. They sold the company in 2000 to a European content management company funded by The Carlyle Group. In 2003, VERSIFI Technologies was re-established and acquired AdaptiveInfo, Inc., the leader in predictive content delivery to wireless devices using machine learning algorithms. VERSIFI Technologies was acquired in 2005.
In 2008, he co-founded Perssonas, Inc. Perssonas is a social network-based hyper-aggregator with a focus on delivering branded, content-rich experiences for the entertainment industry and major brands worldwide.
In 2014, he co-founded PingThings, Inc with an industrial internet incubator, Frost Data Capital. PingThings brings PredictiveGrid™ capabilities for the first time to the bulk power system and utilities worldwide. PingThings seed investors include General Electric, Inc.
Contact Debbie Holmgren at firstname.lastname@example.org
For information on how to participate, visit our Call for Startups.
For information on trade opportunities with O'Reilly conferences, email email@example.com
View a complete list of Solid contacts
©2015, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • firstname.lastname@example.org