The Smart Grid Transformation
The primary enabler to a modern, data-driven utility is the adoption of smart grid technologies. Unlike other industries, which have for decades used sensors and online networks to digitize, communicate, and manage resource delivery and utilization information, utilities have only just begun to adopt automation technologies during the past few years. Historically, highly trained workers were the sole means by which to read utility meters, identify broken or stolen equipment, ensure delivery, and prevent waste.
Although many utilities still employ traditional systems for managing and delivering resources, the combination of new regulations, lower device and infrastructure costs, higher human resource costs, customer demand, industry consolidation, and increased competition from next-generation suppliers (e.g., wind, solar, biofuels) has sped adoption of smart technologies that rely on telemetry and present opportunities for the information-driven utility. According to a 2013 Oracle study, the average large utility expected to spend approximately $180 million each on smart grid and smart metering technologies during the next five years.
However, smart grid technologies produce large volumes of data that require additional processing, storage, and management capabilities. As previously seen in other industries, the explosion of data—in this case, from switches, reclosers, and other field devices—initially presented a challenge to early-stage architectures, which could warehouse or archive some data, but could not scale to accommodate the flood of unstructured data from multiple sources that the modern utility needs to efficiently and effectively manage day-to-day operations.
Predictive Optimization Beyond Forecasting
In addition to reliability and consistency in the delivery of services, utilities are shifting their focus to optimizing voltage in anticipation of demand. Service providers are using big data to prioritize control of both capital expenditure and existing resource management, as well as to prevent spiky increases in customer pricing. A data-driven approach to systems efficiency transcends the current goal of standardized average utilization and better meter reads. Instead, it targets constant and dynamic voltage balancing on a per-meter basis, resulting in an objective of 100% allocation/utilization parity at the premise level across the entire grid. This not only drives energy efficiency for the utility, but also lowers customer costs and smooths upgrade and service cycles.
Hadoop provides a platform on which predictive models can be built to improve energy planning forecasts, increase efficiency, and enhance operating conditions at both conventional and distributed generation plants. Machine learning algorithms draw insights and identify patterns based on massive historic data that greatly improves a system’s advanced analytical capabilities for accurate and dependable forecasting and the ability to determine causality and correlation between utility conditions and energy outcomes into the future. As a result, these systems must be enriched with much larger and more diverse data sets to isolate meaningful signals that can be further tested and, eventually, used to determine better processes and drive outcomes.
The Business Value of an Enterprise Data Hub
The greatest promise of utilities driven by big data resides in the business-relevant questions firms across the entire value chain have historically been unable or afraid to ask, whether because of a lack of coherency in their data or the prohibitively high cost of specialized tools. An enterprise data hub encourages more exploration and discovery by building capabilities on top of Apache Hadoop’s open-source platform. For the first time, utilities can overcome the obstacles that have prevented them from catching up with other organizations, to not only embrace smart grid and other technologies as modernized systems, but also to being using big data to drive planning, strategy, and operations.
 Oracle Utilities. Utilities and Big Data: Accelerating the Drive to Value. Oracle. 23 July 2013.