Moneybulb: Data analytics and the future of the smart grid
Automation Insight – September 2011
For decades, companies and organizations as diverse as automobile makers and baseball teams have assessed how their products and players perform in different environments. They figure out what works and what doesn’t work and use this information to make improvements. With the creation of the smart grid, the opportunity to analyze data and immediately deliver results that could benefit consumers is becoming available to all participants in the electric utility value chain. This is the field of data analytics. In this field, each entity has a different role to play and different information and knowledge to contribute.
The consumer’s smart electric meter is the primary tool that gathers data. It tracks the time electricity is consumed. Various sensors along different points on the electric grid effectively perform the same function. They track when and what flowed past these points, and in some cases, how fast. This information is compiled in a large database. Here, the data analytics engine takes over to gain a deeper understanding about how and where power is consumed, and how it can be generated more efficiently to minimize costs for everyone.
Electricity is an unusual commodity. It has to be generated at almost exactly the same time it is consumed. Therefore, its data has to be gathered and analyzed very frequently to be useful—in some cases, several times each second. The smart grid’s full operational benefits and reliability can be realized when the rate of the flow of information matches the flow of electricity.
As utilities evolve and create their own data analytics infrastructure, they will gain access to a valuable information pathway, which will expand their operational knowledge. This targeted, contextually rich knowledge offers utilities a strong competitive advantage: the ability to create new programs that meet the needs of their consumers more precisely.
Data analytics and optimizing operations and maintenance
Some techniques, such as regression analysis, can predict the likelihood that an individual grid component will fail by monitoring a known set of conditions that a component has experienced and comparing them to a baseline. Operations & maintenance (O&M) personnel can conduct a cost-benefit analysis of this scenario to assess whether it is better to replace or repair the component. They can also identify the optimum time to carry out the work, because they will have a complete picture of all the components on their network.
When an engineer visits the component, he/she can further update the data set, report about the repair/replacement experience, and identify whether or not the work was necessary. Documenting how data errors are resolved can enable future repair iterations to be performed with greater efficiency.
To some extent, electric utilities have been doing work like this for the past 50 years. The smart grid brings a paradigm shift through its capability to collect data. Real-time data analysis takes the conversation about a component’s failure to a whole new place. By creating a contextual mash-up between the utility’s geographical information system and detailed component metrics, the utility now has the ability to see how a situation develops earlier than before and in much greater detail. As a result, the utility can take immediate action to support its customers. This minimizes the length of outages and automatically restores power to consumers inseconds, rather than in hours when done manually.
The new data analytic infrastructure
Once the core components of a data analytics infrastructure have been created, a new service-oriented architecture will start to appear in the form of a self-organizing network (SON). This will be especially relevant to electric grid participants whose location is not fixed, such as electric vehicles.
SON is a fundamental concept in the world of mobile communications1. In these networks, all devices are equal and operate with a plug-and-play protocol. So, if devices are connected to the network, they get to participate, and if not, they have to wait until another identical device offers to connect with them. The self organizing element kicks in through several features:
- Self-configuring: Once a device is connected to a network, it automatically downloads any software upgrades from the central server.
- Self-optimizing: The network organizes and optimizes itself to manage different customer loads as they evolve.
- Self-healing: When a device drops off the network or fails, the remaining network devices flag it as a faulty element and reorganize themselves to continue providing the same service until that device comes back online.
Self-securing2: When a single device is attacked, be it maliciously or accidentally, the devices around it can observe, log, and support or alienate the attacked device.
These concepts, designed for cell towers and cell phones, may appear abstract. However, when they are applied to electric vehicles, which are the ultimate mobile storage devices3, the possibilities to predict and manage electric demand, improve road safety, manage traffic congestion, and more, become almost limitless. Now, apply the same concepts to mobile devices that require more granular storage like laptops or to appliances with embedded technologies such as refrigerators, which do not currently employ storage but have potential to do so. For these devices and appliances, the opportunity for data analytics to support demand management and demand transfer multiplies almost exponentially.
All of these opportunities will become available as electric utilities learn more precisely how their customer groups will respond to certain stimuli. Like the Oakland Athletics demonstrated in Moneyball4, the recent movie adaptation of the team’s stunning 2002 season, data analytics can transform the rules of business in every industry. The electric utility industry is no exception. The winners will ultimately be those who use the new rules to build a better game.
Sources:
1) S. Feng and E. Seidel, "Self-Organizing Networks (SON) in 3GPP Long Term Evolution," Nomor Research GmbH (Munich, Germany: May 20, 2008).
2) G. R. Ganger and D. F. Nagle, "Better Security via Smarter Devices," HotOS-VIII (Carnegie Mellon University, May 2001).
3) W. Kempton and S. Letendre, "Electric Vehicles as a New Source of Power for Electric Utilities," Transportation Research 2(3) (1997): 157-175.
4) M. Lewis, "Moneyball: The Art of Winning an Unfair Game" (New York: W.W. Norton, 2003).