It’s big, complex, mission-critical, constantly in flux, and experiencing a convergence of information technologies (IT) and operations technologies (OT). The distribution grid is the link of the electricity supply chain from the substation to the residential or commercial meter. It has different characteristics and therefore different data analytics challenges in contrast to the transmission grid issues discussed in my previous blog. But like its prominent role in the transmission grid, data analytics will be killer apps in the distribution grid.
The term dynamic is often used to describe the distribution grid, especially as it transforms into a Smart Grid. First and foremost, there is a massive increase in the number of assets and devices that are monitored and/or remotely controlled. Then the distribution grid is impacted by new plug loads such as electric vehicles (EVs) that could introduce new consumption patterns to existing grid models used in operations planning. EVs and distributed generation and energy storage assets enable the distribution grid to accommodate bi-directional electricity flows – one of the key benefits of a fully deployed Smart Grid, but certain to add to operations management challenges. Aging legacy equipment needs to maintain uptime. And finally, smart meters offer the possibilities to collect data more often than the traditional and limited monthly meter read. The grid must deliver power reliably, safely, and cost-effectively – and now it requires similar performance in the communications networks that monitor, collect, process and control remote equipment and devices on both networks. Managing and making sense of all the data from all these devices on two different networks is a significant challenge for utilities – how to correlate, integrate, and analyze data to manage network + grid performance.
The Smart Grid buildout has surpassed the capabilities of traditional utility data management solutions for distribution grids with little to no communications capabilities. But from a data analytics perspective, the distribution grid shares common characteristics with another sector – specifically wireless service providers. Wireless providers have networks that are even more complex than power grids, with large numbers of devices consuming voice, data, and video in different and sometimes very elaborate subscription plans. While all of us typically get the same electric service and consume different quantities of kWh with minimal rate variability (although that will change over time), we have a range of choices in the subscription plans we buy from wireless providers.
Therefore, it makes sense for utilities to turn to data analytics solution providers that have extensive big data experience with wireless providers for guidance to solve the toughest Smart Grid operational and business challenges. One of these companies, PreClarity, shared several best practices to help utilities manage the data deluges hitting distribution grid operations centers. According to Bob Becklund, principal and co-founder of PreClarity, a mission-critical objective for any service provider is to maintain the reliability of services – whether these are delivering electricity or bandwidth. For example, a missing meter read could be the result of a meter transmitter failure, a power grid failure, or a wireless network failure. With smart meters, data analysis must integrate data from the utility’s operational systems, communications network and the distribution grid, correlate very diverse data points, and present it as actionable information. Fault location and identification analysis expedites realtime response resolution. While that’s a great application for data analytics – the best practice is to use predictive data analytics to identify patterns and trends and proactively respond with corrective actions (predictive maintenance) before those faults occur and thus avoid critical power outages and customer service issues. Another best practice is the use of harmonized data models that accept and integrate data from a variety of devices and software applications for function-specific dashboard views and data visualizations. For example, an operations center can use data visualizations to integrate and overlay high-resolution aerial or satellite maps with device status information and workforce applications and thus optimize deployment of field repair resources. It’s a clear demonstration of the “picture worth a thousand words” principle, and one that makes sense in operations centers working to keep the lights on in homes and businesses.
A common concern for wireless and utility operators is how to future proof against anticipated growth rates in devices and data. When the first wireless networks were built, designers did not anticipate smart phones and tablets, streaming video or social networks. Wireless networks experienced exponential data growth from 100Gbs of data per day to TBs of data per day, stressing some analytics applications while other systems were able to handle network and growth issues with marginal additions to their core platforms. The ones that gracefully scaled up in devices and data volumes exercised best practices in data architectural design – creating flexible and scalable data models that consider Event Driven Architectures (EDA) or Service-Oriented Architecture (SOA) decisions; real time event processing vs. historical analysis needs; data schemas to deal with diverse network elements; and the advantages of centralized vs. distributed architectures.
So while there is a learning curve for utilities to deal with all the data that the Smart Grid delivers, the climb doesn’t have to be as steep or as high as it is sometimes made out to be. There are other business sectors and solution providers like PreClarity that bring thought leadership and experience to share in the Smart Grid convergence of IT and OT and managing data deluges. And there’s no doubt that Smart Grid data analytics – with their potentials to improve operational performance and reduce service downtime – will be killer apps in the distribution grid.