By Arnie de Castro, Product Manager, SAS
On February 17-20, 2019, IEEE will hold its tenth Conference on Innovative Smart Grid Technologies. One of the outstanding features of this conference are the sessions and papers dedicated to analytics. There are panel sessions on data analytics for resilience, and artificial intelligence applications in power operations and planning. There is a panel session on advanced analytics applications in the smart grid. These sessions show how analytics is helping the industry manage the challenges posed by the rapid changes in the grid, including the increasingly uncertain and variable character of the load, the proliferation of distributed energy resources and the explosion of data in volume and velocity and variability from synchrophasors, smart meters and sensors. There are also numerous papers on the applications of analytics in the forecasting and optimization of grid operations.
Changes in the grid
The last two decades have seen a lot of changes in the utility industry. Government incentives and advances in technology have resulted in a proliferation of distributed energy resources, particularly solar and wind generation. Storage devices have also been successful in test beds, providing responsive support to minimize the impact of sudden fluctuations in grid operations.
Many of the renewable energy sources have been behind-the-meter rooftop solar generators. These, together with the ever-increasing number of plug-in electric vehicles, and the increased efficiencies of electrical equipment, appliances and devices, have resulted in an electrical load that has an entirely different character than what the industry has known before.
Challenges and Impacts
These changes have resulted in more variability and uncertainty. For the last decade utilities in California have been talking about the duck curve, where thousands of megawatts of generation must be ramped up as the load peaks and the sun sets. These renewable resources also change by the minute as the wind gusts and as clouds pass by. And as the industry struggles to produce the next efficient large-scale storage devices to help mitigate this variability, we realize that up to this point, 99% of electricity storage is still provided by pumped-storage hydro.
Another result of the distributed grid is the bidirectional flow of energy and information. While the transmission grid is generally designed with protective schemes that allow for these, the distribution systems must be retooled to ensure that they are properly protected. At the same time, while before bidirectional information flow was largely between the control center and the remote terminal units (RTUs) and controlled equipment, it now occurs at the scale of millions of advanced metering infrastructure (AMI) meters and phasor measurement units (PMUs) and even sensor-equipped devices in what is called the internet of things (IoT).
The amount of information transmitted has also grown exponentially. While electric meters were once read monthly, we now have possibly 2880 reads per month if done on a 15-minute basis. While supervisory control and data acquisition (SCADA) systems usually read measurements every 2 seconds, PMUs do so 30 times per second for even more measured variables.
Understanding the Grid
To engineers, analytics applications in the smart grid tend to focus on keeping it resilient and reliable. This involves forecasting load at the different delivery points, monitoring the performance of equipment and evaluating their effects on the operation of the whole grid. Analytics is used to optimize the planning, maintenance and operations of the electric utility. And advanced analytics provide the utility with insights that provide better understanding of the grid.
Beyond traditional analytics, visual analytics technology is allowing us to help avoid outages by detecting wildlife incursion, predicting vegetation growth and identifying areas in need of wildlife removal and tree trimming. Transformer and transmission line photo and infrared images help in identifying transformers with hot spots and allow line heat detection and sag identification. Biometrics and facial recognition software can readily identify persons at monitored sites and detect intruders.
At the enterprise level, analytics is used in managing records, enabling the workforce, and improving security. Analytics helps utilities improve their understanding of customers and allow them to offer more appropriate and more personalized services. Analytics also helps utilities understand the risk to which they are exposed, enabling them to manage their customer portfolio more effectively.
Artificial Intelligence and Machine Learning – Advanced Analytics Applications in the Smart Grid
At the heart of all this is the ability to manage the volumes of data, many times analyzing the data as it streams. Technologies like Hadoop, faster and cheaper computers and event stream processing have made it possible to store and process more data of different types. In-memory analytics then allow the more advanced artificial intelligence and machine learning technologies to build models, predict results and optimize performance quickly and allow decisions to be made on time.
The last decade has seen the growth of the grid’s complexity in generation, load and in the amount of data being processed. A major tool that utilities are using to describe, predict and optimize the operation of the grid is advanced analytics. We hope you will attend these panel sessions and participate in the discussion on the use of advanced analytics to help utilities provide a more resilient, reliable and responsive grid.