Data Analytics that Do More with What You Already Have

The power industry is sitting on a goldmine of data.

Data can unlock valuable insights on how to further improve grid operations, be it through the more efficient delivery of energy from generation to the end customer, better planning of capital investment, or even near real-time guidance of how to triage and navigate restoration activities in severe weather events. Right now, however, as an industry we are leveraging a small fraction of operational data – it’s generated for a specific use, and then relegated to a box somewhere in a remote office, rarely to be seen again. It’s not that the data is unusable, but it requires very specific expertise that spans both domain and software to unlock its potential and drive actionable insights.

Enter GE’s grid analytics portfolio – released in general availability this week at our Americas User Conference and engineered to make powerful analytics accessible to all. Leveraging our analytics on top of a utility’s current operational software helps move the needle on hyper-critical goals for electric utilities, moving from a reactive state to a proactive state through advanced analytics. Outcomes will generally include minimization of customer downtime via outage prevention and reduction of time to restore when outages do occur. This improves customer satisfaction and ensures greater control over operating costs for our utility customers.

What’s especially exciting to me is the convergence between the agility of the analytics available and the additive nature of each incremental analytic. Rather than asking our customers to make a huge investment in a data platform, we designed the analytics with speed to delivery of value in mind. At the same time, we built all analytics with the same production pipelines and infrastructural strategy to ensure that a customer could build out their data platform incrementally as they added new analytics. 

As a result of this laser focus in the design phase, the analytics can be deployed over the course of three months as opposed to an entire year, at a price level within budget for most Tier 1 utilities. We also have ensured that there will not be a separate workstream to create an analytics-ready environment; it simply comes with the analytic!

Best of all, this same schema is currently being used across our entire grid software suite. Unifying data on a secure, scalable and user-friendly platform drives efficiencies beyond the initial investment and across a utility’s entire operational system. These analytics and their supporting infrastructure will easily snap to a customer’s current GE grid systems but also has the flexibility to align to other vendors’ products. GE’s Grid Analytics enable data to be leveraged by many solutions across the energy value chain, from generation to consumption. This improves accuracy, reduces costs, preserves assets, and keeps field crews safe on a utility’s operational side, while driving efficiency within its IT environment.

Now without further ado, here are the analytics that we are excited to release:

Storm Readiness

Talk with any electric utility and they’ll tell you: Weather is the number one cause of outages. Our customers today want to monitor incoming weather activity to ensure that they can act quickly (or at times proactively) in preparation for major storm events. GE’s Storm Readiness analytic does just that.

The analytic combines high-resolution weather forecasts, outage history, crew response, and geographic information system (GIS) data to accurately forecast storm impact and prepare response crews and equipment ahead of time. We use a recurrent neural net model to learn from outage events and weather activity to predict where and when future outages will occur given incoming weather forecasts. This will result in the ability to accurately predict the number of device failures on an hourly basis, based on a 72-hour weather forecast.

In other words, we are helping utilities move from reactive operations to predictive operations. The benefits they are experiencing have tangible impact on the utility and the homes and businesses they serve, including more accurate storm crew staging decisions resulting in shorter outage duration as well as reduced costs of hiring contingency crews and other operational expenses.

Network Connectivity

One of the interesting paradoxes of this new digital world is that the volume of data created by and for operational systems is exponentially increasing, yet the data quality isn’t always reliable. And when data quality isn’t reliable, the foundation for all operations is greatly destabilized. This is certainly true for our electric utilities. GIS data errors can cause significant operations and maintenance waste and prevent electric utilities from realizing the benefits of advanced distribution management solutions and automation programs. 

With GE’s Network Connectivity analytic, algorithms specifically tailored to grid operations automatically detect association and phase errors in the grid. The analytic provides recommended corrections for increased GIS and CIS data integrity and is then configured to push these recommendations back into the operational systems themselves. These automated corrections enable grid operators to stop worrying about data quality and to focus on value-added activities in their day-to-day roles.  

Effective Inertia

The distribution network isn’t the only one that benefits from predictive analytics – transmission systems have a lot to gain as well. The operation of the grid is continuing to grow in complexity, in large part due to growing volumes of variable renewable generation. This has led to a massive displacement of “system inertia,” or the resiliency of power generation, given spikes in customer demand or reduced supply, due to unforeseen decreases in wind or sunlight. As you would expect, ineffective management of a transmission system could result in system-wide blackouts and major financial and reputational penalties. 

Given this massive influx of renewables, solutions requiring enhanced system visibility and understanding to deliver fast-acting response services are critical. However, traditional transmission operations are not built to accommodate this level of transparency, in this capacity. Analytics now facilitate the measurement and forecasting of effective system inertia to help with these new needs. We accomplish this by applying machine learning relating the notion of system inertia to known and predictable values: conventional rotating power generation, load, solar power, and wind power. Accurate inertia forecasting gives the transmission system operator confidence in a secure level of renewable penetration and the appropriate amount of reserve services, saving money by eliminating excess reserves and enabling a more stable grid.

Where We’re Headed

We are on the precipice of a massive transformation in our industry. A resource once considered to be a byproduct of operations and of use for a very limited amount of time before being cast aside has become the very fuel that will power the engines of efficiency and optimization for our customers. Like fuel however, data must be treated, stored and eventually delivered to the appropriate engines to unlock its potential. And even then, if the engines – advanced analytics in this case – are not appropriate for their users, they will never be plugged back into critical daily operations. This is the full cycle of digital transformation and is the closed loop that we seek when we build and release analytics like those that I’ve described today.

No two utilities are at the same stage in their digital transformation. However, all are seeking new ways to harness the continuous insights their assets and their network create. For utilities just getting started on their digital transformation, I challenge you to consider the untapped potential in the data your operations are creating every day. It’s real, it’s substantial, and it’s close – and we here at GE are excited to partner with utilities to help transform it into exceedingly positive outcomes.