Smart-Grid Analytics


Intelligent networks support better decision making.

Fortnightly Magazine - October 2008

Sophocles once said, “Quick decisions are unsafe decisions.” Apparently Sophocles did not work in the utility industry. Utilities must make quick decisions every day to maintain a safe and reliable grid. As they have learned, the key to a quick and safe decision is making a well-informed decision. Yet utilities face challenges in providing enough information for their employees and automated systems to make these types of decisions. To address this challenge, some utilities are looking at intelligent grid analytics to help them more quickly process large amounts of complex data and make better-informed decisions.

Defining Analytics

At their most basic level, analytics applications are fact-based decision-support systems that turn data into information that a company can use to make a decision or take action. According to IDC, analytics applications include: query, reporting and analysis software - ad hoc query and reporting tools, production reporting tools, multidimensional analysis tools, and software dashboards; and advanced analytics— software that uses advanced technologies to discover relationships in data and make predictions that are hidden, not apparent, or too complex to extract using query, reporting, and analysis software.

Intelligent grid analytics can use both query, reporting and analysis software, frequently in the form of dashboards; or advanced analytical software to help utilities make decisions or take action regarding their transmission and distribution networks. However, intelligent grid analytics have some characteristics that distinguish them from typical business analytics, including the need to process large volumes of data (500,000 sensors collecting 15-minute interval data produce approximately 1 terabyte of data per day) that is complex—real-time, interval and time-series data from multiple sources such as smart meters, line sensors, transformer monitors, SCADA systems, weather feeds and GPS receivers.

Although there currently aren’t a large number of commercially available advanced analytics applications that can handle complex intelligent grid decisions on a large scale, vendors and utilities are developing and piloting them right now. Some vendors have seen success with their technologies, but most are still being tested on a smaller scale (see Table 1).

As the intelligent grid market matures and more intelligent devices are deployed, vendors will make a wider variety of advanced intelligent grid analytics commercially available, making it easier for utilities to visualize and analyze the grid on a real-time basis and at a broader level. The first push for more advanced analytics likely will come from more fully utilizing smart meters, followed by the addition of other sensors and intelligent devices. However, just because truly advanced analytics solutions aren’t here yet, it doesn’t mean utilities aren’t using some form of intelligent grid analytics today. Utilities are taking advantage of numerous intelligent grid analytics options available for more basic decisions and longer-term planning decisions, including dashboards, smart metering analytics and more traditional business intelligence applications.

The Need for Knowledge

Table 1

As the North American economy depends more and more on utilities to “keep the lights on,” any outage can be detrimental to today’s economy and a utility’s reputation. Yet maintaining a reliable grid is a challenge for many reasons.

For example, as many grid assets reach the end of their useful life and face an increased risk of failure, utilities cannot afford to replace all of the assets, so they must better understand and maintain their existing assets.

Further complicating this problem is the reality that utilities will have to maintain this infrastructure with a rapidly aging workforce. Many utilities ultimately want to replace their retiring workforce with fewer, more productive employees. However, new employees do not have the same extensive knowledge about the grid as longtime employees.

With the need to understand many complicated factors, such as power flows and voltage profiles, analyzing the grid is already difficult. New additions to the grid, such as distributed and renewable energy sources, smart metering, and demand response programs, are only going to increase the grid’s complexity. Energy trading also introduces more complexity because utilities have to deal with multiple power sources and providers.

Table 2

The North American Electric Reliability Corporation (NERC) reliability standards cover a broad range of issues, from resource and demand balancing to personnel performance, training, and qualifications. With the enforcement of the NERC reliability standards, utilities can face fines of up to $1 million per event, per day.

Given these needs, utilities are finding that the intelligent grid is not just for making operating decisions about their transmission and distribution (T&D) assets. To fully understand their T&D systems, utilities are also looking at decisions they have to make about areas connected to the grid—generation (e.g., energy storage, distributed and renewable generation, and new central generation plants), customers (e.g., residential, commercial, industrial, and municipals), and staff (e.g., mobile workforce) (see Table 2).

Knowledge about the grid can help answer these sorts of questions. Intelligent grid analytics build this knowledge by bringing together complex data and turning it into information that utilities can then use to make well-informed, effective decisions about the grid. Some utilities are finding that faster, more real-time analytics are important not only in preventing grid failures, but also to support other grid initiatives such as demand response, real-time pricing, and more distributed and renewable generation. Yet despite these needs, today’s utilities cannot effectively make these decisions because they have limited, and mostly historical information. In terms of information, utilities must deal with:

–  Few data points: Utilities generally have good data about transmission networks, but they collect little data about distribution networks beyond major substations. In particular, utilities are now collecting more metering information, but neglecting to monitor other critical assets between the substation and the meter, such as transformers. Furthermore, most utilities do not collect real-time data about their distribution networks, which is critical for making split-second decisions.

–  Limited access to data: Even with the data utilities do collect, often they have limited access to this data. Personnel might maintain data in spreadsheets on their own computers, or the utility may separate data into silos across the company. For example, a utility may store meter data—which is applicable for operations purposes—primarily in its retail business unit for billing purposes.

–  Lack of advanced analytics: Even if utilities do have their data in order, they might not be able to analyze all of the data in a real-time situation. Utilities do have some real-time applications in use today, but most are not ready to handle the complex needs of an intelligent grid. For example, control applications, such as SCADA, EMS, and DCS, have not advanced quickly enough to incorporate factors like consumer behaviors or weather information to provide useful real-time analytics to utilities.

Opportunities for Change

Some utilities are finding that the intelligent grid provides a way to overcome these challenges. As utilities gain more information—particularly about their distribution networks and end users—through intelligent grid initiatives, they now have more information to feed into their decisions. The key issue now is how utilities collect, sort, and analyze this information to make better decisions. Some utilities are taking advantage of new technologies to help them improve their data availability and analytics capabilities.

Utilities that are organizing their data have two main approaches for organizing data: putting the data in a central repository or building better connections between different data sources. Some utilities are working to bring their disparate data sources into a common format and database. For example, a large investor-owned utility (IOU) in the western United States had energy delivery data scattered across its offices, with many engineers maintaining their own personal databases. As the company took on more proactive maintenance practices, it realized that its employees needed to access data remotely and understand asset trends across the dispersed service territory. As a result the utility brought all of its operations data into one central database of time-series data.

Getting information to work together does not necessarily require a utility to pull all data into one database. Instead, utilities are developing other ways to begin connecting disparate systems, including an integration bus model, service-oriented architecture (SOA), cloud computing, or a mashup of technologies. For example, The California ISO uses an SOA solution to better display and analyze spatial information. These efforts include overlaying GIS data on satellite imagery and overlaying time-series data on GIS. The system can visualize the grid, substations, generators, nuclear power plants, and wind farms as well as mash up weather data and forecasted demand supply. Google Earth is used to visualize the real-time analytics.

Analyzing and Interpreting The Data

As utilities get their data in order, the question then becomes how to make use of the data. Some utilities are taking advantage of opportunities to make better use of the data.

One way they’re doing this is by applying more advanced dashboards. Green, yellow, and red aren’t just for traffic lights anymore. In the current intelligent grid analytics market, dashboards—geared toward everyone from the field crews to senior executives—provide reporting tools that calculate and consolidate metrics on a single screen to allow employees to easily monitor information. Metrics can include anything from transformer and substation status to SAIDI, CAIDI, and SAIFI indexes. For example, one U.S. utility is using software from Obvient Strategies to pull data from work management, asset management, outage management and other legacy systems and display it on dashboards tailored for each major business function such as field force management and operations.

As utilities put more intelligent sensors on the grid, such as smart meters, they need some way to collect and use the information from those devices. As a result, another stream of analytics applications comes from smart metering and intelligent grid device vendors wanting to help utilities better leverage information coming in from their devices. Thus, vendors like Itron, GridPoint, Current Communications Group, and Landis+Gyr are offering analytics applications to support their devices. Although most of these applications cannot make complex decisions in real time, they do enable utilities to make use of newly installed devices for operational decision making.

Finally, utilities are making better use of traditional business intelligence applications. Although not heavily marketed for intelligent grid applications yet, more traditional business intelligence applications, such as those from SAS, IBM (Cognos), Oracle (Hyperion), or SAP (Business Objects), can assist with some decisions. Utilities can collect information and track key performance indicators (KPIs) across business areas and sort through performance and budget factors. Vendors like SAP are looking at how they can leverage their capabilities in the smart metering/meter data management space, but are not yet aggressively pursuing broader intelligent grid applications.

Nevertheless, over time the industry will find more ways to use the increasing amounts of data available in smart grid systems. Advanced analytics will allow companies to turn this complex data into useful intelligence that will support operational and strategic decisions.