February 05, 2026

Powering NC State’s Future with a Self-Healing Grid

RMF Engineering is featured in the 2026 Q1 issue of District Energy Magazine, where RMF’s Chris Skillestad, PE and North Carolina State University’s Tate Boulware, PE detail the design, implementation, and testing of a modern, self-healing electrical distribution system developed to support N.C. State’s growing campus. The article explores how a multiyear, multiphase upgrade addressed aging and unreliable infrastructure to deliver a safer, more resilient power system capable of reliably serving everything from research labs and classrooms to student residences—while enabling future growth with minimal campus disruption.

A previous version of the article below appeared in the First Quarter 2026 issue of District Energy Magazine. To learn more about District Energy or read the full publication, please visit their website. Subscriptions are FREE.

Planning, Building and Delivering North Carolina State’s New Self-Healing Grid

By Chris Skillestad, PE and Tate Boulware, PE

As a leading public research university, North Carolina State University depends on a steady, reliable supply of power to every building, from cutting-edge research labs and classrooms to student residences. N.C. State’s electrical grid supports two major district energy plants that combined have over 16,000 tons of chilled water generation capacity, 300,000 PPH of steam generation capacity and a co-generation plant that produces 11 MW of electric power, which is used to base load electrical demand from the local utility company.

But because of age, N.C. State’s main campus medium voltage electrical system, multiple energy plants and service to some of the university’s most historic buildings had grown unreliable in recent years. Further, sections of underground electrical infrastructure had been in place for more than half a century, posing a risk of failure and safety issues during maintenance. The system ultimately needed to be replaced to continue supporting the mission of the university, located in Raleigh.

Thus a multiyear, multiphase power system replacement project was initiated. The electrical distribution upgrade aimed to strengthen reliability, enable future growth and connectivity, improve operational safety and simplify maintenance – all while transitioning to a modern, self-healing grid. RMF Engineering worked closely alongside N.C. State and Balfour Beatty, the construction manager, through years of planning, phasing and testing to facilitate the transition while ensuring human safety and system quality with minimal campus impact. Ultimately, the project delivered a system with 50% more electrical capacity than the previous one, equipping the university to support its next 40 years of growth.

Six Modern Circuit Loops Across a 260-Acre Project

RMF’s design replaced all electrical distribution infrastructure, working within the university’s coordinating master plan with a smart 15-kV power system. Incorporating new SCADA controls, the self-healing grid is comprised of six circuit loops (a simplified configuration from the previous) that are fed from two double-ended, metal-clad switchgear lineups and connected with automated padmounted switches designed to detect outages, isolate faults and reroute power when needed.

Prior to construction, fieldwork was completed on every electrical manhole, switch and transformer on campus to document the initial conditions and verify connectivity to develop the sequencing plan. Four total construction phases were packaged to ensure no major outages would be necessary as six miles of duct bank were replaced, eight new pad-mounted transformers were installed (decommissioning the previous vault-style transformers), 52 new selfhealing pad-mounted switches were connected and 100 building services were transitioned to the new grid.

By the end of the replacement project, every pad-mounted switch on N.C. State’s main campus had an automated-style switch with remote operation capability. This feature provides a few main benefits:

It allows for remote operation and verification of the medium-voltage distribution equipment, increasing operator safety.
It provides the electrical system with self-healing capabilities that, as a result, adapt and restore power to affected buildings during an unplanned outage within seconds, instead of the hours it would have taken previously to identify, troubleshoot and manually operate the switches. This approach also helps standardize maintenance procedures for the university, and the standardization of the switches themselves makes it easy to maintain spare parts in the event of a component failure.

Real-World Testing a System Still Under Construction

SEL-451 relays at each padmounted distribution switch provide overcurrent status on Ways 1 and 2 (the loop main lines), as well as fault indications on Ways 3 and 4 (building feeders). This information is used by the distribution automation controller running on an SEL-3555 RTAC, or realtime automatic controller, to diagnose fault locations and determine the appropriate corrective action. The loop’s distribution switches are not rated for fault interruption; therefore, the substation breakers will isolate any detected faults before the distribution switches are operated.

Given that testing the distribution automation system when fully installed would have had major impacts to the campus, RMF developed a validation methodology as segments were built iteratively – from in-factory demos to full-scale software modeling. This strategy helped establish confidence in the system while minimizing interference with campus activities.

At the switch level, the team confirmed that the control writing matched manufacturer drawings and verified that the motor operators and battery chargers were all functioning correctly. SEL distribution automation simulation software was used to build the electrical system as designed and to perform a multitude of tests. Currents and voltages were injected into smallscale, factory test sets of switches to imitate power system events and confirm that the physical system’s response matched how it reacted in the simulation software. In one simulation, 10 of the pad-mounted switches were staged in a parking lot and connected with simulated voltage and current input wiring. Each pad-mounted switch communicated back to a central SEL RTAC to aggregate the data. Multiple electrical test sets were used to inject voltages and currents at different locations in the system to simulate faults and confirm the system’s functionality. This ensured the modeled scenarios in the distribution automation controller software testing environment were realistic and reliable.

Once the switches were installed in the field, they underwent NETA testing and were connected to the communication fiber loop, which allowed the status readings to be verified at a central HMI, or human-machine interface. It was also confirmed that open and close commands from the SCADA system were received by each pad-mounted switch and that they were operating correctly.

Once all of the pad-mounted switches of an individual 15kV distribution circuit loop (of the six in total) were installed and cabling was complete, a final, integrated system test was performed. To prevent campus disruption, the motor operators on the loop ways of the pad-mounted switches were decoupled and the SCADA system feedback status was reconfigured temporarily to read the motor operator location, rather than the actual physical switch blade position.

This allowed for the team to inject current into the test switches on padmounted switches in various locations throughout the circuit loop to simulate electrical faults and observe how the system reacts without the switch physically opening and disrupting power to campus. Because the automation system was tested with the switch position feedback wiring in the parking lot model, as well as validated with the software simulation testing, the team had full confidence in the holistic functionality of the self-healing system – given that each of its components had already been verified independently and iteratively.