The typical economic applications at that time were the classical long-line applications, where the cost of conversion equipment was offset by savings in line costs by using HVDC. For overhead lines, this typically occurred with lines greater than 300 mi in length. For underground or especially submarine cables, the economic crossover point occurred at the 30–40 mi point. In North America, well-known examples included the Pacific Intertie as well as the submarine cable connecting Vancouver Island in British Columbia.

The unexploited but intriguing characteristic of HVDC was the possibility of controlling power flows within the otherwise generally free-flowing ac grid. The fact that a HVDC connection contributed very little short-circuit current was noted as well; in fact, a back-to-back (BTB) dc connection functionally appears to be an open circuit from an ac impedance point of view.

Detroit Edison is located in southeastern Michigan and is bordered on the eastern side with Canada, separated from Canada by rivers and lakes. As a result, Edison's extra-high voltage (EHV) transmission grid was configured like a "C," where the C was comprised of overhead 345-kV double circuit towers and where the bulk of Edison's generation was located at the upper and lower ends adjacent to the water. The Edison grid was also connected to the Ontario Hydro 230-kV system at the upper and lower ends of the C.

The city of Detroit resides along the Detroit River across from Windsor, Canada. Detroit was entirely supplied from the EHV system via underground cables operating at 120 and 345 kV that were, in turn, connected to the ends of the C.

With underground impedances around 40% of overhead impedances, and the cable system effectively bridging the C in parallel with the overhead EHV system, the system was very sensitive to the balance of generation between its northern and southern sections. In fact, generation planning had to be carried out in a fashion that maintained this balance to avoid overloading the underground circuits and setting off an endless series of cable additions, the effect of which was to increase fault current levels that ultimately exceeded the interrupting capabilities of circuit breakers.

An opportunity to consider exploiting the unique characteristics of HVDC at Detroit Edison arose with a project intended to address the generation balancing problem when it was decided to install the bulk of future generation at the north end of the system. This project called for converting existing overhead and underground ac transmission to dc. This enabled an increase in the capacity of the system without simultaneously increasing short-circuit currents. The overall capacity of the dc circuits was still far less than the capacity of parallel 345-kV circuits, but the effective capacity increase attributable to the controllability of flow on the dc circuit was remarkable, equivalent to several ac cables. Adding the necessary ac cables increased fault currents to unacceptable levels and also made the parallel underground system even more sensitive to generation imbalance.

The need for this robust solution was mitigated when generation plans were scaled back to a moderate level. However, the control aspect noted with the HVDC project was later exploited with the installation of a phase-shifting transformer in a key 345-kV underground cable, and its operation continues today.

My next exposure to the potential of HVDC came while working at Florida Power and Light in the late 1980s. A developer proposed constructing a 1,300-MW coal power plant on Grand Bahama Island, where a deep water port gave the project economical access to low sulfur coal throughout the world. The output of the plant was to have been delivered to South Florida via six submarine HVDC cables to ensure reliability.

The effect of this project on the geographically stressed transmission grid in Florida was remarkable. It injected 1,300 MW of coal-based generation right into the South Florida load centers, reduced ac power losses by nearly 80 MW, enhanced transient stability, and displaced generation otherwise fueled with oil.

Another project during the same time period called for Florida to purchase abandoned nuclear plants in the southeastern United States. The plants would be resurrected and completed, and the output delivered to South Florida via HVDC transmission—in this case, overhead transmission. The impact of effectively injecting generation into South Florida was very positive, and similar to that observed with the Grand Bahamas project described earlier. In this case, it was nuclear generation that was displacing oil, with the corresponding economic benefits to Florida customers.

Both of these very effective projects were victims of declining oil and gas prices as well as the ascent of combined-cycle technology projects that competed more effectively in the short term than the more ambitious larger-scale projects. However, the great positive impact of functionally "teleporting" economic generation into load centers where construction of such generation was otherwise infeasible was noted.

I returned to Michigan in 2000 and am now with ITCTransmission in Novi, Michigan, an independent transmission company formed by acquiring the transmission assets of DTE Energy. With the acquisition of the Michigan Electric Transmission Company, ITC Holdings is now responsible for most of the Michigan transmission grid.

In 2003, ITC Holdings invested in Conjunction's Empire Connection project. The Empire Connection, similar to the Florida projects described above, was intended to transport generation from upstate New York down to New York City via 2,000 MW of HVDC transmission capacity and was formulated as a merchant (as opposed to regulated) transmission project. The circuits were intended to be overhead, where feasible, and underground when overhead transmission was precluded.

This project would have transported gas-fired capacity and energy from transmission-constrained upper New York State, where construction costs were nominal, to New York City, where such construction was far more expensive. The costs of the line were basically offset by the differential in generation construction costs between the two areas. This project was abandoned when auctions for use of the line capacity failed to produce sufficient revenues to support financing of the project.

All of the above applications of HVDC, except for the control example, were ultimately hampered by the high initial cost of dc terminal equipment, as the high initial cost is only offset by future benefits. Should breakthroughs occur that would lower such costs, HVDC applications, with their unique benefits, could compete with cheaper but less effective technologies.

For example, the transmission interface between Michigan and Ontario consists of four transmission lines, each currently operated at 230 kV. A phase-shifting transformer (PAR) is installed in series with each interconnection for the purpose of minimizing the unscheduled power flows across these lines. Such flows have been (incorrectly) characterized over the years as "Lake Erie circulating power," and the path is part of the "Lake Erie loop." While the PARs control a portion of the parallel path flow over these circuits, they are large mechanical devices that react only slowly to effect changes in flow. Under transient conditions, the PARs assert essentially no control at all.

BTB dc terminal equipment can also control power flows. In fact, BTB installations appear to be an open circuit to the ac system, and do not respond at all to ac disturbances or events. The control systems of BTB equipment can be designed to fully emulate ac flow across such equipment, but any such flow must be proactively scheduled, unlike the free-flowing ac grid.

This suggests an important application of HVDC technology that would have had enormous economic impact during the 14 August 2003 blackout. The blackout occurred when events in Ohio spread to Michigan as transmission lines tripped out in a cascading fashion, leaving Michigan and Ohio load connected only via the Michigan-Ontario interface, which subsequently dragged down the systems of Ontario and New York.

Had BTB HVDC converters, rather than PARs, been employed to control flow on the Michigan-Ontario interface, the interface would have looked like an open circuit. There would have been no power swings across Ontario, New York, or Pennsylvania, and those systems would have remained completely intact. Much of the US$6-10 billion cost of the blackout would have been avoided.

In retrospect, even at current BTB prices, such an installation is not prohibitive. Assuming a nominal interface capacity of about 2,000 MW and the cost of BTB equipment at US$200/kW, conversion of the interface to HVDC control would be in the US$400 million range. Such conversion would maintain total interface capacity and transmission use at current levels, but events such as the 14 August 2003 event would be precluded, with attended reliability savings and without giving up the benefits of fully interconnected ac operations.

George Loehr and others have proposed similar, but far more ambitious, grid reconfiguration projects—such as the above for the U.S. grid as a whole—arguing that the interconnections are getting too big, with consequent risk of increasingly catastrophic failure. They argue for breaking the grid into presumably more manageable mini-grids along natural break points, bridging the break points with BTB HVDC to maintain current levels of commerce.

Other unique characteristics of HVDC include the potential ability to reliably get more power through existing lines and rights of way, either through conversion of existing circuits to dc or installation of dc transmission circuits within existing ac rights of way. Bipolar dc circuits, with neutral or ground return, appear more like two circuits than a conventional three-phase ac circuit. A dc line can still be operated at no less than half capacity for failure of a single pole or wire, while a fault on any phase of an ac circuit requires disconnecting all three phases.

It is likely true that continuing to overlay the existing fully interconnected ac grid with higher-voltage "superhighways" is the best overall long-term solution to transmission capacity shortfalls. A truly "self-healing" grid is one with sufficient reserve ac capacity to accommodate outages and disruptions. However, HVDC can enable a fallback option in cases where such ac EHV expansion is not feasible, and certain carefully targeted implementations can decouple overly sensitive parts of the ac grid without giving up the existing benefits of the ac grid. The single most enabling event to wider HVDC applications would be to see significant breakthroughs in costs of conversion equipment.