The distribution grid isn’t what it used to be. Traditionally, the lower-voltage grid carried electricity in one direction from large power plants to businesses and residents who needed it. Though distribution grids have always been inherently complex, the one-way flow of electrons from power plants and the transmission grid to customers made it relatively straightforward to plan, upgrade, and operate them.
The distribution grid that has emerged over the past decade is significantly more complex, characterized by a massive and growing influx of distributed energy resources (DERs), such as rooftop solar, battery storage, and electric vehicle (EV) chargers. Not only can these and other DERs consume electricity from the distribution grid, but they can also produce and export it in the other direction, creating two-way power flows that can both provide support during times of peak demand but also negatively impact grid voltage, frequency, and reliability.
While significantly more complex, the distribution grid is also arguably more important than ever before. With electricity demand rapidly increasing due to large loads such as data centers, domestic manufacturing, and the electrification of transportation, utilities and businesses are seeking capacity that can be accessed quickly and at scale. States like New Jersey, Virginia, and Illinois have enacted legislation and regulations to accelerate the adoption of virtual power plants (VPPs), which are collections of DERs and other resources managed by software to operate like a large power plant.
This evolution of the distribution grid has heightened the importance of grid models, which are mathematical representations of the physical power system that capture the electrical characteristics of power lines, transformers, switches, and other equipment that make up the distribution system. Utilities use these models to inform important decisions about future grid investments, to configure protection systems to guard against faults, and to manage daily grid operations.
While grid models have long been a standard tool for planning and operating the transmission system, they are a relative novelty for distribution grids. “We have not needed them because we have not managed the distribution grid actively,” said Sean Crimmins, a program manager in EPRI’s Information and Communication Technology program, which provides frameworks and methodologies for effective grid model data management. “To make the most of the grid we’ve already built, you need a good model of the distribution grid, and historically we have not had one.”
Why Data Management Matters
Getting the most out of the distribution grid today and in the future requires both robust models and a better approach to building and maintaining them. An EPRI supplemental project, Applied Grid Model Data Management (GMDM) for Distribution, aims to improve the management of grid model data. The project aims not to provide the grid models themselves, but to improve the availability and accuracy of the data that feeds the models, and to enable centralized management and sharing of the models for planning, operations, transmission, and neighboring utilities. Addressing the data that goes into grid models improves grid planning and operation by enhancing the simulations they run, making superior distribution grid management possible.
EPRI’s Applied GMDM project builds on earlier research that brought utilities and vendors together to develop a common architecture for enterprise-wide grid model data management. The underlying mission of the Applied GMDM project is to help utilities apply that architecture, improve quality and consistency, reduce data-prep time, and unlock the full value of grid models.
It’s a challenging task, and one that varies from utility to utility because each distribution grid is unique. But the most consequential transformation that needs to take place is streamlining, standardizing, and centralizing how grid data within a utility is collected, stored, and managed. In most cases, distribution grid data today is fragmented, stored in silos by the teams that use it, and transferred manually when other utility departments need to use it.
This balkanized approach to data management and sharing has many negative consequences. One is simply the time and effort staff must devote to assembling data. “When planning engineers work, they spend somewhere between 60 percent and 80 percent of their time gathering and preparing data. Then they get to do the engineering,” Crimmins said. “If a utility is also using a model in operations, in an ADMS (Advanced Distribution Management System), engineers are spending far too much time creating another model, from different sources, and then comparing models to figure out why one model is different from another and why their results are so different.”
Instead of fragmented, siloed systems, the long-term vision is to create a centralized grid model manager that automatically shares and updates data in near real-time. The model manager would gather data from utility planning, design, construction, materials, and asset management teams, then distribute validated models to those who need them. “The North Star is this central capability that’s doing all the gathering and all the sharing with people that need data, sometimes at different granularity, often for different time frames,” Crimmins said. For example, a planning engineer needs a model that reflects planned future projects and operations to understand the as-built and as-operated pictures of today’s grid.
Establishing an authoritative, centralized model manager is a challenging, years-long effort. It’s a journey that also depends on a utility’s starting point and priorities. What systems are they using today? What are their goals? What are the drivers for new power flow analytics? EPRI can help at each stage of the journey by clarifying where a utility is, where it wants to go, and the necessary steps to get there.
Finding the Gaps, Identifying a Pathway Forward
The project begins with an in-person workshop that brings together utility teams from planning, operations, protection, and IT with EPRI power system engineers and data and enterprise architects. The initial goal is to better understand and document how grid model data moves through the utility. The exercise involves detective work to clarify which applications produce data, what protocols are in place, and where problems arise. Sometimes, the bottlenecks are straightforward, like manual data transfers, spreadsheet handoffs, and the need for multiple follow-up phone calls. Other times, the problems run deeper, rooted in workflows that were never clearly defined or communicated, or in teams that lack a clear picture of how their data connects to everyone else’s.
These lessons are all documented and summarized to explain the current state of grid model data management, its level of maturity relative to industry benchmarks, and short-term steps that can be taken to improve it—such as eliminating the need to exchange spreadsheets or make numerous phone calls by automating data exchanges. That clarity about the current state of a utility’s grid model data management practices serves as the foundation for the next step: collectively determining a long-term vision.
Lessons From Puget Sound Energy (PSE)
EPRI has applied this methodology at several utilities, including Ameren, AEP, and Hydro-Québec. More recently, Puget Sound Energy (PSE), a utility serving about 1.2 million electric customers in Washington State, worked with EPRI to assess its grid model data management practices and chart a pathway forward aligned with the utility’s priorities. Drivers of the initiative include PSE’s implementation of an ADMS, a sophisticated platform for real-time grid monitoring and control that relies on accurate, continuously updated models to be effective.
PSE has also seen widespread adoption of electric vehicles (EVs), battery storage, and other DERs across its service territory. The utility wanted to explore opportunities to improve its approach to hosting capacity analysis, enabling it to connect resources to the grid faster and in locations that provide the greatest benefits. To achieve these and other objectives, PSE reimagined its approach to grid modeling, decentralizing it with tools managed by individual groups. “This has resulted in data inconsistencies, inefficiencies, and a complex web of point-to-point interfaces with data sources and consuming applications,” said Rebeca Rogers, an engineer who works in PSE’s Grid Modernization Strategy & Enablement group. “Many of these integrations are a mix of automated and manual processes.”
On a practical level, the decentralization requires PSE staff in planning, operations, and engineering to build, maintain, and clean multiple versions of grid models.
An initial workshop led by EPRI included nearly 40 utility staff from planning, operations, IT, and GIS, reflecting the organizational importance of grid model data management. “PSE’s collaboration with EPRI played a key role in shaping the strategic direction for our grid model management efforts,” Rogers said. “EPRI helped define a clear vision for GMDM.”
One outcome from that initial engagement: PSE built a business case for establishing and funding a dedicated data governance working group. Another was to translate findings about current data flows into a living document that will update as workflows and systems change. PSE is now developing requirements for a grid model manager, software that will centralize how the utility collects, validates, and distributes grid model data across departments. “We aim to improve data consistency and traceability across systems, strengthen data governance practices, and establish a scalable foundation that can support the integration of future modeling and analytics tools,” Rogers said.
With EPRI’s grid model manager functional requirements report and experience from the Applied GMDM, PSE has a clear picture of what that software needs to do and how it fits within its organization. This will make the procurement process considerably more precise. “They have a good idea of where that fits in their organization because of the work this project did,” Crimmins said. “And we have generic requirements for what the software should be able to do to support model managers and consuming applications.”
Though pioneering, the work at PSE reflects broader changes underway across distribution utilities as the grid grows more complex and faces greater demands. “Having an accurate picture of the grid you’re operating and changing in the future is really critical,” Crimmins said. “And having an efficient way of doing that accurately is really important.”
EPRI Technical Expert:
Sean Crimmins
For more information, contact techexpert@eprijournal.com.
Banner and supporting images created by EPRI using Microsoft Copilot.
