Kevin D. Jones, Ph.D., Consulting Engineer – Electric Transmission, Dominion Energy
These days, it is hard to have a conversation about the future of any business without talking about data. This is certainly no different in the utilities sector. At Dominion Energy, we’ve known for a long time that data we can gather from our equipment is an asset. While hard assets such as transmission lines and transformers may have a book value, the value of data as a soft asset is no less in its enterprise value. Our “data-as-an-asset” mindset is reflected in our deployment strategy for synchrophasor technology. Phasor Measurement Units (PMUs) are sensors designed to measure and continuously streamhigh resolution, time-synchronized, phasor quantities from the grid. We havedeployed hundreds of these sensors across equipment on our electric transmission network providing observability of nearly all substations and transmission lines.
Our fleet of PMUs creates a deluge of data, approximately 275TB per year. Naturally, a great deal of resources is requiredto maintain and curate this flourishing data asset, and these resources must be justified. How can we quantify the return on investment (ROI) from the data in order to justify escalating support, maintenance, and investment? What are we going to do with the data?
First, we tap into the latent value of the literature from the research community. Through detailed in-depth evaluation, we identified nearly 5,000 publications addressing synchrophasor data analytics (a number that has nearly doubled in the last 5 years). Each paper is a possible candidate for a Dominion Energy use case. Second, we can treat each grid event or equipment failure as an opportunity to distill lessons learned into an algorithm to either track or predict future system behavior. Lastly, we mustactively experiment with analytics to test new hypotheses specific to our company. Each of these approaches have in common the need to quickly and effectively test and evaluate a particular analytic.
The answer to the question of use cases is not singular - there is no “killer app”. Value will be derived from a continually evolving ensemble of analytics that, not only vary over time, but also vary from one enterprise to another. This is good news indeed! A broad spectrum of opportunity for extracting value from different analytical approaches exists.Our challenge is to determine which approaches provide the most value to our company. Guiding us is a vision that prioritizes rapid, cost-effective hypothesis testing and analytic prototyping where the most promising analytics can evolve into production tools.
How can we realize such a vision? One of the mantras established by our safety and training program, and embodied by our employees, is to “always have the right tool for the job”. This isnot only true for hands-on jobs but also applies to massive archives of sensor data from the grid. When working with data, the right tools are the difference between rapid, cost-effective analytic development and failure to establish sustainable use cases for the business.
To assist us in our efforts, we are collaborating with a startup based out of nearby Washington, DC –PingThings. PingThings provides a platform-as-a-service called the PredictiveGrid which is designed to ingest, store, visualize, analyze, and learn from high-resolution, physics-based time-series data from the grid – in our case, synchrophasor data. We view this technology as a high performance analytics sandbox environment that can enable our subject matter experts to swiftly and successfully identify high-ROI use cases.Even more groundbreaking for Dominion Energy is that we have worked with PingThings to host our instance of the PredictiveGrid in AWS GovCloud. We invested the time needed to verify that we can be just as secure, if not more so, by leveraging the cloud so that we can benefit from its flexibility and scalability. Because of the success of this experience, we can now begin to bring the cloud into other conversations about the future of our tools and technologies.
Now that we are in possession of a high-resolution data set with such wide-reaching observability, and equipped with the right tools for working with this type of data, where will we first turn our focus and energy? For Dominion Energy, there are at least five areas that have the potential for significant ROI. First, generator model validation and parameter identification can help us determine the dynamic properties of the machines directly from the data much more frequently than traditional physical testing. This will not only improve our models but provide us with a means to detect emerging issues with the machine or its control systems. Second, we can better understand the behavior of loads on our system, particularly those impacted by greater renewable resource penetration, which feeds back into significant improvements in our long term planning capabilities. Third, we can observe the patterns and signatures of aging or failing equipment to facilitate proactive decommissioning and decrease safety risk. Fourth, post-event and post-mortem analysis will provide us with greater situational awareness and serve as a vehicle for socialization of synchrophasor data and our new data analytics platform. And finally, we can create a rich set of transmission level key performance indicators (KPIs) that will provide multi-dimensional insights not only into the present health of the grid but also the effectiveness of our investments in infrastructure and, potentially, guidance for future investments.
Getting to this point did not happen overnight. October will mark the 10 year anniversary of our synchrophasor program at Dominion Energy. In a sea of ever-evolving technology, the path has not always been clear or straight. However, recognizing that having the right tool for the job is critical for leveraging our data assets has served as another enabling force that has allowed Dominion Energy to embrace change.