March 29, 2024

Big Data: Enabler of an Intelligent, Sustainable Grid

by Steve Collier

Big Data: Enabler of an Intelligent, Sustainable Grid
Beginning with Thomas Edison’s Pearl Street Station in the late 1800s, persistent economies of scale, declining prices, exponential growth in demand, and steady improvement in technology supported the development an electric utility grid architecture and industry business model that continued largely unchanged for almost a century. The OPEC oil embargo in 1973 marked an inflection point after which both grid adequacy and the traditional business model began to erode. A new, profoundly different grid model is emerging, involving new technologies, topologies, techniques, traders, and, transactions. This modern, intelligent grid brings with it massive amounts of data (i.e., ‘big data’). Managing and exploiting this data poses great challenges but offers even greater benefits for utilities, non-utility providers, and consumers.

The legacy U.S. grid is a patchwork of several loosely connected synchronous AC grids. Some 10,000 power plants containing about 20,000 generating units supply power and energy. Energy flows one way from these resources over high voltage transmission lines to remote load centers where it is distributed and delivered to electric consumers through some 150 million meters. Almost all of the generation, transmission, and distribution facilities, the marketing and sales, and the metering and billing have been owned and operated by regulated electric utilities who operated as monopolies within defined service areas. The post OPEC embargo industry has seen economies of scale and declining prices eroded by risk, rising costs, stalled growth (even reduction in demand), and concerns about environmental and economic sustainability.

Electric utilities began struggling with a relatively modest introduction to big data as they moved from monthly ex post facto meter reading and billing to more frequent meter reading via ‘smart meters.’ The goal was to facilitate customer demand response programs that would relieve some of the pressure on the electric utility grid and business model. The usual 12 monthly meter readings for each customer multiplied to 730 hourly readings or more each month. In addition, other data points were being created including power on/off status, voltage, power quality, meter base temperature, meter tampering, etc.

So, the industry went from a handful of data points per customer per year to ten thousand or more. Over less than a decade nationwide the industry went from less than 2 billion data points per year to more than a trillion. And this was just the beginning.

In essentially every technology based industry, disruptive enabling technologies change the landscape. This happens even in thriving industries, and the struggling electric utility industry is even more susceptible. New ways of producing and using products and services emerge. New providers bring them to market. Existing infrastructure and industry incumbents are challenged. A new electric grid model is resulting from just this phenomenon. There is a profusion of distributed energy resources (DER) occasioned by exponential improvement in technologies along with innovative, entrepreneurial business models. These include renewable and conventional energy production, electrical / thermal / mechanical energy storage, energy monitoring / management systems, smart end use devices, electric vehicles, microgrids, smart buildings, even smart cities. A plethora of non-utility companies are entering the market as providers of DER and related products and services.

As the penetration of DER increases, utilities, customers, and non-utility providers require more monitoring and analysis of more things. Suppose for example that every retail customer has on-site generation, storage, energy management, and independently monitored end uses (e.g., smart thermostat, smart end use devices, smart home / building, EV, et. al.). The volume of data would grow from a monthly meter reading and bill to thousands of new data points for monitoring, analysis, and control. The potential number of data points nationwide could grow to trillions annually! Now add retail transactive energy markets which already exist to some extent in nearly twenty states. With their physical real-time and forward derivative transactions, additional orders of magnitude of complexity and data volume are added to the mix. Now we are truly talking big data. Acquiring, analyzing, applying that data is and will be ever more crucial to ensuring a steady, safe, secure, and sustainable electric energy economy.

There is another strong motivation for big data and analytics. It is the need to plan, operate and manage the grid in the presence of adverse circumstances, all outgrowths of the foundational erosion and new industry developments discussed above.

The legacy grid is aging like most of the public infrastructure in the US. The American Society of Civil Engineers gave it a grade of D+ in its 2013 Report Card for America’s Infrastructure saying:

America relies on an aging electrical grid and pipeline distribution systems, some of which originated in the 1880s. Investment in power transmission has increased since 2005, but ongoing permitting issues, weather events, and limited maintenance have contributed to an increasing number of failures and power interruptions.

New capital investment by utilities declines as demand growth stalls, non-utility supplies proliferate, costs and risks increase, revenues decrease and sustainability concerns prevail. The grid is wearing out. As a result, major outages are on the increase, practically doubling every five years for the past several decades. As outages increase, the ability to detect, locate, analyze, and timely remedy them becomes ever more important. This requires more data, better data analytics, and improved operations and management applications.

As if an aging infrastructure were not challenge enough, the weather is getting worse. Severe weather events are growing in frequency, duration, and severity. And they cause more outages. They would do so even if the grid itself were not in decline. Approximately half of the growth in major grid outages over the past several years has resulted from severe weather events. Again, as these events increase as do their adverse effects on the electric grid, the ability to anticipate, monitor, analyze, and respond to them becomes ever more important. This is all about big data analytics.

The physical security of the grid is a growing concern even apart from weather issues. One of the greatest weaknesses of a centralized synchronous AC grid is this. If you disturb part of it enough (e.g., take a major power plant or substation or transmission line out of service) the entire grid is affected. In the limit removing critical generation or transmission facilities from operation for significant periods of time can mean the entire grid or substantial portions of it can be out of service. It goes without saying the adverse impacts of a prolonged system wide power outage as we recently learned from Hurricane Sandy and the derechos. As the possibility of physical attacks by terrorists increase, the ability to anticipate and deter or detect and respond becomes ever more important. This is all about big data analytics.

Rapid growth in renewable energy sources, both utility scale and distributed, also stresses the electric grid and business model. Wind and solar are not dispatchable and they can’t automatically follow load. Grid generation must be dispatched to accommodate their output, or load must be controlled to match their output, or energy storage must be utilized to time shift their output. This requires new kinds of data and applications. This is all about big data analytics.

Advancements in information and communications technologies have created opportunites for more data to be acquired, analyzed, and applied to improve the reliability, efficiency, security, and safety of the grid. New monitoring devices such as power measurement units (PMUs, aka synchrophasors) can sample the AC waveform hundreds, even thousands of times per second. With on board GPS and atomic clock, they can provide a time and location synched picture of the performance of the grid locally and system wide. Advanced computer algorithms can detect incipient problems with the grid from voltage and current variations. Or they can help monitor and control the grid more rapidly, accurately, and flexibly in the face of increasing complexity and uncertainty. A special case of this approach known as distribution fault anticipation (DFA) is being pioneered at Texas A&M University. Incipient faults can be detected and prevented. The traditional ‘run to fail’ mode for transmission and distribution can now be anticipated and avoid preventative maintenance to avoid failure and consequent interruptions of service. This can mean a dramatic improvement in the reliability, safety, security, and efficiency of the grid. So, think about moving from metering a customer once a month to metering 730 times a month to sampling points on the grid thousands of times per cycle. That’s two trillion data points per device per year! Now we’re really talking big data analytics.

How can this tremendous new volume and complexity of data be handled with adequate speed, reliability, and security? How can it be acquired, analyzed and applied effectively? In an unusually serendipitous circumstance, disruptive enabling information and communications technology (ICT) is paving the way. The ICT infrastructure and industry is already going through the erosion, disruption and transformation that the electric grid is just beginning. Just think. What else in the technology world has grown from a multiple of the number of people in the world to tens of billions? The Internet of Things (IoT) has. The number of things connected to the Internet surpassed the number of people in the world in 2008. Oft quoted Cisco has suggested that by 2030 there will be 50 billion things connected to it. That might be a tremendous underestimate given the discussion above of just the smart grid by itself?

In the presence of big data, the electric grid will become a hybrid electric, information and telecommunications grid. It will become part of the IoT. It will be an Enernet as named by Bob Metcalfe, a convergence of the smart grid with the IoT.

What about cybersecurity? As the grid becomes an Enernet, cybersecurity becomes of paramount concern. A recent book by reporter Ted Koppel paints a cautionary picture of a grid that is part of the Internet of Things. Without being naive or overly simplistic, this problem is its own solution. The entire world, government / business / daily life, is becoming a part of the Internet of Things. Like electricity, the Internet has facilitated improvement in the quality of life, productivity of business and national economies unparalleled in the history of man. We would not abandon it any more than we would abandon fire or electricity. So, the entire world will endeavor to ensure its security and sustainability. It is already way more reliable than the electric grid, having not been out of service in any substantial way in decades. It is the very definition of an intelligent, resilient, self-healing, sustainable grid. Will the cybersecurity challenge ever be eliminated? That is unlikely. Will it be manageable? That is likely. As inscribed on the Guardianship statue at the National Archives in Washington, DC, “The price of liberty is eternal vigilance.”

Big data will get bigger. Big data analytics will get better. The energy business will improve. Customers will benefit.
 

About the Author

Steven E. Collier is Director of Smart Grid Strategies at Milsoft Utility Solutions. Operating from his office in Austin, Texas, he assists Milsoft with corporate business development and industry relations. Since starting his career at Houston Lighting & Power in the early 1970s, he has worked as a consultant or executive with energy, telecom, and technology companies in the United States and abroad. He has degrees in electrical engineering from the University of Houston and Purdue University, and is a designated IEEE smart grid expert. Besides blogging as SmartGridMan, he writes and speaks widely on new and emerging energy, telecom, and information technologies. For more information on Big Data and Smart Grid, please visit http://bigdata.ieee.org and http://smartgrid.ieee.org.