The “Internet of Things” (IoT) is creating a buzz across industries. It describes the integration of an advanced, interconnected information backbone into the functioning of physical devices, systems, and infrastructures. It is the convergence of the virtual and the physical. At a basic level, IoT applies the Internet as we know it to wirelessly connect machines, devices, systems, and other “things.” IoT will be transformational, and Gartner predicts that, by 2020, 25 billion things will be connected in industries ranging from automotive to food and beverage services.
In the energy industry, the momentum of IoT is having a tremendous impact as the grid becomes sensored, connected, and smarter. The “Internet of Energy” applies the premise of IoT to the infra-structure of the grid and is driving the acceleration of the dynamic smart grid.
The Advent of the Smart Grid Nearly 40 years ago, a Harvard sociologist predicted the advent of the Information Age. Years before the invention of the Internet, minicomputer, personal computer, and smart phone, Daniel Bell authored “The Coming of Post-Industrial Society.” He predicted that information and communications technology would cause a fundamental change in the structure of the global economy— a change as significant as the Industrial Revolution.
The Information Age, as we know it today, describes the free, nearly instantaneous transfer and access of data. It predicated the preeminence of the “knowledge worker” and resulted in the emergence and continued growth of information technology (IT). It drives ubiquitous changes in the ways we communicate, with lasting effects on professional and leisure activities.
With the genesis of the smart grid, Bell’s predictions from four decades ago are having a direct impact on today’s energy systems. The National Academy of Engineers identified the electric grid as the most significant scientific achievement of the 20th century. The smart grid will be the largest and most complex machine ever conceived and will likely prove one of the most significant scientific achievements of the 21st century.
It is estimated that as much as $2 trillion is being invested this decade in upgrading the power infrastructure globally to add sensors to the devices throughout the grid, creating part of what has come to be known as the IoT. These newly sensored things or devices are the basis for the physical infrastructure of the smart grid and include smart meters; thermostats; home appliances; heating, ventilation, and air conditioning equipment; factory equipment and machinery; transformers; substations; distribution feeders; and power generation and control compo-nents. Once sensored, these devices become remotely machine addressable, meaning information can be sent and received across a computer network.
Taking smart meters as one example, it is clear that the smart grid is advancing apace. As of 2014, nearly 400 million smart meters have been installed globally, according to Navigant Research. That number will more than double in the coming decade. Representing a fraction of the sensors on the grid infrastructure, smart meter installations serve as a proxy for the penetration and growth rate of the smart grid.
The “smart” sensored devices in and of themselves provide little utility. They simply provide the capability to remotely sense and/or change a device’s state. For example, is the device operative or inoperative? If operative, at what temperature, voltage, or amperage? It might allow us to know the amount of energy that the device has consumed or recorded over some period of time or is consuming in real time.
Collectively, these devices generate massive amounts of information—an increase of six orders of magnitude from before the connected grid. As utilities adapt from managing a relatively small number of non-communicating devices to connecting hundreds of millions of diverse sensored devices, data volumes are expanding exponentially. To effectively aggregate and manage this inﬂux of data, utilities require next-generation technologies to integrate, process, apply analytics, and intuitively visualize the data and analytic results in a way that drives business outcomes through a common data- and intelligence-driven solution.
By applying technologies and techniques commonly used by Google, Amazon, Netﬂ ix, and Twitter in the consumer industry, utilities can collect and aggregate the sum of increasing volumes of data to correlate and scientifically analyze all of the information generated by the smart grid infrastructure in real time. Computer science techniques, including elastic cloud computing, machine learning, and social human-computer interaction models, are now being applied to challenges utilities have faced for years, such as managing the operational health of advanced metering infrastructure (AMI) assets and preventing revenue loss due to theft and meter malfunctions.
Utilities can apply the same technology concepts that Twitter uses to process 15 million tweets per second or Netﬂ ix uses to stream more than one billion hours of videos per month to integrate, aggregate, and analyze the massive amounts of incoming data. One such powerful computer science technique is machine learning, or the ability for computers to learn without being explicitly programmed. Machine learning simply uses mathematical equations known as algorithms that can learn from data. The algorithms are trained on historical data to make predictions. After predictions are made, actual confirmed results are fed back into a machine-learning algorithm to refine it. As a result, over time, it “learns” and evolves so that the analyses generated are increasingly accurate, reflecting real-world conditions specific to the utility.
Machine learning is used to classify assets at high risk of failure, segment customers for targeted marketing campaigns, identify non-technical loss (which includes measurement errors, recording errors, theft, and timing differences), and predict future load, among many other applications. For example, like a credit card company can use historical spending data to ﬂ ag potential fraud, utilities can use a variety of historical and real-time data to identify cases of energy theft. Baltimore Gas and Electric Company (BGE) proved this use case when it deployed the C3 Revenue Protection™ application across its full two-million-meter service territory. In six months, the solution identiﬁed more than 8,000 new cases of potential theft, higher than its original goal.
Leveraging not only machine learning but also a full stack of tools asso-ciated with the science of big data, analytics enabled by the smart grid provide efﬁciencies across the energy value chain. Examples far outreach revenue protection and include:
- real-time pricing signals to energy consumers;
- management of sophisticated energy efﬁciency and demand response programs;
- conservation of energy use;
- reduction of fuel necessary to power the grid;
- real-time reconﬁguration of the power network around points of failure;
- instantaneous recovery from power interruptions;
- accurate prediction of load;
- efficient management of distributed generation capacity;
- rapid recovery from damage inﬂicted by weather events and system failures;
- reduction of fuel needed to power the grid;
- and substantial reduction in adverse environmental impacts.
Data Analytics Solutions
Data analytics solutions for utilities must integrate massive amounts of disparate data, apply sophisticated multi-layered analytics, and provide highly usable portals that generate actionable real-time insights. Utilities need end-to-end system visibility across supply-side and demand-side smart grid operations.
Data analytics enable grid operators to realize dramatic advances in safety, reliability, cost efficiency, and environmental benefits by correlating and analyzing all of the dynamics and inter-actions associated with the end-to-end power infrastructure as a fully interconnected and sensored network, including current and predicted demand, consumption, electric vehicle load, distributed generation capacity, technical and non-technical losses, weather reports and forecasts, and generation capacity across the entire value chain. In another example from BGE, the utility used the C3 AMI Operations™ application to identify 3,600 meter health issues with 99-percent accuracy in order to stream-line critical maintenance on AMI assets.
The growth of the smart grid and the necessity for next-generation analytics solutions are not limited to the United States. The European market is seeing strong drivers for analytics, including the increased number of smart meter deployments in Italy, the United Kingdom, France, and Spain, and the European Union’s recommendation of 80-percent smart meter penetration in member countries by 2020. Enel, a leading integrated player in the world’s power and gas markets with the largest customer base (61 million) among its European peers, is deploying data analytics solutions to enable smart grid and smart city services. A smart grid pioneer, Enel was the first utility in the world to replace traditional electromechanical meters with digital smart meters, a major operation carried out among Enel’s entire Italian customer base. By 2006, Enel had installed 32 million smart meters across Italy; Enel has since deployed a total of approximately 40 million smart meters in Europe, representing more than 80 percent of the total smart meters on the continent.
To detect and prioritize reduction of energy theft, Enel has deployed C3 Revenue Protection™ across 8 million meters in Italy. Implemented in less than eight months, the AI-enabled solution ultimately more than doubled the energy recovered per inspection, relative to Enel's baseline performance that was a function of an expert driven process honed over a 20+ year timeframe.
For Enel, C3 Energy integrated, normalized, and aggregated more than 50 billion rows of data from 11 Enel sources, including the customer information system, billing system, work order system, outage management system, producer system, meter data management system, validated theft-case data, external weather data, and Google for address verification.
Leveraging investments in analytical algorithms, machine learning, data integration, and cloud-scale infrastructure, Enel deployed over 100 unique and sophisticated energy ﬂow analytics to identify anomalous meter activity. The company is leveraging these analytics to execute rule-based and machine-learning algorithms to unlock insights from both batch and streaming data, and to generate increasingly targeted and accurate results.
The initial deployment proved that the data-analytics solution could readily handle Enel’s smart grid data processing and aggregation needs. Based on the results from the one-million-meter demonstration, Enel and C3 Energy are working to expand the deployment of this solution more widely across the group’s distribution network. Enel also is installing additional applications to expand on what will be the largest deployment of software-as-a-service smart grid analytics in the world.
Leading utilities are driving innovation toward the Internet of Energy. They are at the cutting edge of technology advancements and are realizing significant returns by applying data-analytics solutions that combine the sciences of cloud-scale computing, advanced smart grid analytics, and machine learning to the benefit of their communities, consumers, stakeholders, and the environment.