Austin Energy, 1993 to 2015.
Manual to ADMS, without leaving the room.
Grid modernization is a thing most people in this industry read about, sell, or advise on. I operated it. Every platform generation that arrived at Austin Energy between 1993 and 2015 arrived on a system I was responsible for, on a shift I was working, and I was the operations subject matter expert on the implementations. When they went wrong, they went wrong on my watch. When they went right, I knew exactly why.
That is the difference between knowing what a platform does and knowing what it does at two in the morning during an ice storm. I started when it was one hundred percent manual and I finished with ADMS, distribution automation, and distributed generation on the same system.
The system I was responsible for
Austin Energy is municipal, so none of this is confidential. In my years there the utility served a 437 square mile territory across Austin and parts of Travis and Williamson counties, just under 500,000 meters and close to a million people.
620 miles of transmission at 69 kV, 138 kV, and 345 kV. 11,400 miles of distribution at 12.5 kV primary with a 34.5 kV downtown network. 74 substations. Transmission voltage control across 12 auto transformers, 21 capacitor banks, four 345/138 kV reactors, and a 138 kV STATCOM. More than 3,000 MW of generation across Decker, Sand Hill, Fayette, the South Texas Project, wind, solar, and biomass, with Decker as the contracted blackstart resource. Interconnections to LCRA at 138 kV and 345 kV with shared stations, and to CenterPoint at 345 kV with a shared station. All time peaks of 2,714 MW in August 2011 and 2,195 MW in February 2011.
The utility publishes the current figures, now 633.5 miles of transmission, 12,337 miles of distribution, 81 substations, and 575,087 customers. The system grew around the work described on this page.
Before any of it was automated.
I joined in 1993 as an electrical system controller aide and came up through the desk to supervisor of system operations. The tools were a Harris SCADA system for breaker and device control, paper logs, a wall mapboard, radio dispatch, and a calculator. Outage position came from customer calls and from what a controller could hold in his head about circuit configuration.
Every switching order was written by hand, which meant performing the demand and loading calculations on conductors and devices before directing a crew to open or close anything. I maintained substation and system one-line diagrams for Austin Energy and for large customers, because there was no model in a system to consult. Storm restoration meant prioritizing trouble calls on paper, working out which crew could safely get to which device, and tracking every truck by radio.
Every platform I have implemented, sold, or advised on since is an attempt to automate something I used to do by hand. I know what each function is actually replacing, which is why I can tell when a vendor demonstration is showing me a capability and when it is showing me a screen. It is also why I can say plainly which parts of the job should stay with the operator.
We built the first one ourselves.
Before Austin Energy bought an outage management system, we designed one. I served as the operations subject matter expert on the design of the in-house trouble call and outage analysis system, which took customer call data and turned it into a prioritized restoration picture the control room could actually work from.
I also trained more than 200 employees at the Austin Energy call center, because an outage system is only as good as what gets entered at the front of it. That training covered the life of a trouble call, what the control center needs from each one, and why the quality of the intake determines the quality of the restoration.
Transmission got real analytics.
In 2003 the Harris system gave way to ABB Ranger for SCADA and energy management. I served as control center operations subject matter expert and implementation advisor: operational configuration of monitoring and control, integration between the EMS and existing SCADA telemetry, validation of control room workflows for transmission monitoring and switching, and testing and operational readiness through to go live.
What changed was analysis. Contingency analysis, power flow across the transmission network, voltage stability monitoring, and scenario analysis during disturbances replaced calculations we had been doing in our heads and on paper. I then operated that system on shift for the next 12 years, which is a longer and harsher acceptance test than any factory acceptance test produces.
Outage prediction arrived, and had to be taught to be right.
I was the operations subject matter expert across the full lifecycle of the GE PowerOn outage management system: research and selection, factory and site acceptance testing, operator training, go live, and post go live support. My work concentrated on outage detection logic, where the system fuses customer calls, SCADA telemetry alarms, feeder topology, and meter events into a predicted outage location and customer count.
That logic is only as good as the model and the assumptions behind it, and the people who find out first are the operators working a storm. Sitting on both sides of that, configuring the prediction and then living with it on shift, is where I learned how much of outage management accuracy is data quality rather than software.
Then we put telemetry on the distribution system.
The outage management upgrade came paired with a distribution SCADA deployment, which is a bigger operational change than it sounds. Until then, distribution visibility came from customers calling. After it, the control room had real-time monitoring and control of distribution feeders and devices, and the outage system had telemetry to detect against rather than only phone calls.
I supported both halves as distribution operations subject matter expert and implementation advisor, with the integration between SCADA telemetry and outage detection as the part that mattered most: validating that what the field reported and what the outage system believed were the same thing, and redesigning control center workflow around the new information.
One platform instead of three.
The ADMS program unified outage management, distribution management, and SCADA into a single operational platform. I served as distribution operations subject matter expert and implementation lead, and the work that consumed the most time was the network model: modeling feeders and switching devices in the platform, validating feeder topology against GIS, defining switching configurations and protection devices, and making the model accurate enough that outage detection and restoration logic could be trusted.
Beyond the model, it was control room workflow design, outage detection and event visualization, switching operations management, alarm and operator notification handling, and then scenario testing and tuning before production. This is the same platform I would spend the following six years implementing and upgrading at other utilities, which is only possible because I had first taken delivery of it as a customer.
The hardest modeling problem on the system.
Downtown Austin is a secondary mesh network, not radial distribution, and the two behave nothing alike. I was distribution system modeling lead and ADMS subject matter expert for representing that network inside the platform: designing the modeling approach for the downtown secondary grid, representing network protectors and transformer banks, making load flow calculations simulate bidirectional flows correctly, validating feeder interactions across the mesh, and then testing and tuning until the model reflected what the network actually does.
I single this one out because mesh network modeling is substantially harder than radial feeder modeling, most ADMS deployments never attempt it, and it is the clearest example of the kind of problem that cannot be solved by someone who only knows the software or only knows the network. It needs both.
Letting the system restore itself, carefully.
Fault location, isolation, and service restoration schemes hand a piece of the operator's job to the network. I supported the design and deployment of those schemes as distribution operations subject matter expert, advising on the switching logic inside the automation sequences, coordinating integration between the automation devices and control center systems, and testing automated restoration scenarios against what actually happens during a real fault.
The operational question underneath automation is always the same: what is the scheme allowed to do without asking, and what must it hand back to a human. I was answering that question from the operator's chair, which is the chair that inherits the consequences.
Restoring a system from zero.
I built Austin Energy's blackstart restoration plan from nothing. That meant developing the procedures for restarting the grid from a zero power state, defining the restoration sequencing step by step, identifying the generation and transmission resources used to get there, establishing the coordination procedures with ERCOT, and documenting restoration priorities for critical infrastructure.
A plan on a shelf is worthless, so it was updated annually and exercised: blackstart planning sessions with ERCOT and regional utilities, multi-utility restoration training exercises, and grid restoration simulations, every year from 1993 to 2015. I also maintained the load shedding procedures and incident command activation alongside it, and led the NERC compliance audits performed by Texas Reliability Entity.
Almost nobody in the grid modernization market has written a restoration plan for a system that has gone completely dark, and then rehearsed it with the ISO and the neighboring utilities. It is the deepest form of understanding how a power system is actually put back together, and it is the reason resilience conversations with utility executives do not stay theoretical when I am in them.
DER, before anyone called it that.
Multiple independent power producers were wired into the system I operated. The University of Texas campus cogeneration complex tied in at 69 kV is a combined heat and power microgrid of roughly 50 MW that can run the campus islanded. Three waste-to-energy cogeneration plants connected on automatically switched overhead primary. Beyond the interconnected producers, Austin Energy contracted wind in South and West Texas, the 30 MW Webberville solar project, and biomass.
Coordinating switching, clearances, and restoration around generation embedded in the distribution and transmission system was ordinary control room work for me in the 1990s and 2000s. When the industry started using the term DER, and later built DERMS products to manage it, I was not learning a new concept. I was watching a name get attached to something I had been operating for years.
The events I worked
Twenty-two years on a control desk in Central Texas means the weather eventually sends you everything it has. I worked all of it, and these are the ones with a public record behind them.
The ice storms of December 1998, late February 2003, and January 2007, the last of which laid down up to three quarters of an inch of ice and took out power to 35,000 customers in Austin. The ERCOT rotating outage events of April 17, 2006 and February 2, 2011, each of which has its own official investigation. In the February 2011 cold snap I was the supervisor overseeing the control center when ERCOT directed emergency load shed across the state. Load shed at that point is not a customer service decision, it is what stops the interconnection from collapsing, and the magnitude was past anything we were set up to calculate automatically. I figured the shed blocks by hand, on shift, and Austin Energy delivered the reduction it was directed to deliver. The August 2011 heat emergency, which is the same stretch that set the all time peak of 2,714 MW. The Halloween flood of October 2013 on Onion Creek, and both the Memorial Day and Halloween floods of 2015. Tornadoes: the record shows 21 tornadoes on 10 separate dates touching Travis County between November 1993 and November 2015, the strongest being the F4 in the Pedernales Valley near Lakeway on May 27, 1997, on the same day as the Jarrell outbreak. Windstorms and severe thunderstorm outbreaks throughout. Blackouts and load shed events. And ERCOT blackstart training and restoration exercises every year of it.
What I did that nobody asked me to do
None of the following was in my job description. I wrote them because the control room needed them and nobody else was going to.
A transmission circuit switching program that takes the line a controller needs cleared, pulls every associated piece of equipment from a database, and populates the full clearance order including special switching instructions and the other utilities to notify. A distribution load switching calculator. A system wide equipment rating database for quick retrieval during restoration and switching. Controller training scenarios drawn from real system events the control center had lived through. A NERC certification training program for new controllers. Interdepartmental notification and crew routing programs. Operating instruction documentation for the outage management system, and the call center training curriculum and pamphlet used to teach 200 plus employees how a trouble call becomes a restoration.
That habit did not stop when I left. It is the same instinct behind the offerings, frameworks, and platforms I have built since, and behind GridCORTEX now.
This is why Schneider came looking.
Schneider Electric did not recruit me for a sales record. They recruited me because their ADMS and DERMS platforms were being designed by people who had never run one, and sold to people who had to. I had taken delivery of that platform as a customer, modeled a secondary mesh network in it, and then operated it on shift. That is a short list of people.
So I took the control room into the product. I led the subject matter expert team across presales, sales, deployment, and services, and my operator experience fed enhancements straight back into the engineering roadmap, including the ADMS and DERMS hybrid architecture at Arizona Public Service and the wildfire mitigation capability built into ADMS for Pacific Gas and Electric, which moved wildfire risk out of planning and into real time control room workflow. I built the cloud based virtual server environment for ADMS demonstrations that became the global default, devised phased digital grid strategies, and set new sales records every year from 2016 to 2022. Then I deployed and upgraded that platform at 14 utilities across North America and Australia.
Operate the system. Take delivery of the platform. Go build the platform better. Sell it to operators who can tell in one question whether you have ever done the job. Then consult on it, integrate it, and now architect the AI layer above it. Each chair only made sense because of the one before it, and all of them trace back to the desk in Austin.
Why these 22 years are the foundation of everything after
Consultants, integrators, vendor subject matter experts, and executives in this market are mostly working from one vantage point. Mine started at the desk where the consequences land.
I did not arrive at grid modernization. I was in the room for all of it, on the same system, as each generation of technology landed: manual operations, then EMS and transmission analytics, then outage management, then distribution telemetry, then a unified ADMS, then automation that acts without asking, and distributed generation throughout. I was the operations subject matter expert on those implementations and then the person operating them afterward, which means I have seen what every one of those platforms promised and what each actually delivered on shift.
That is what I bring to a utility executive conversation now. Not an opinion about what modernization should look like, but a record of having done it, in order, under load, and with the audits and the storms attached.