In This Issue
Summer Bridge on Issues at the Technology/Policy Interface
July 1, 2016 Volume 46 Issue 2

Electric Power and DC’s Renaissance

Tuesday, July 5, 2016

Author: Lionel O. Barthold and Dennis A. Woodford

In 1882, 82 customers with 400 of Thomas Edison’s new electric lamps signed on to America’s first central electrical supply system. Direct current (DC), at 110 volts, fed those loads through copper cables that emanated from dynamos at Edison’s Pearl Street generating station in Manhattan (figure 1). Edison, quite aware of Ohm’s law, must have lost sleep worrying about how he could keep up with the soaring demand for current using copper wires at such a low voltage.

 Figure 1

Meanwhile George Westinghouse, also aware of Ohm’s law, accepted the help of Nikola Tesla—an advocate of alternating current (AC) who, in frustration, had abandoned Edison—in adapting a new device called a transformer. It was modified to convert the output voltage of AC generators to a higher voltage, distribute it at that higher voltage, then step it down again to 110 volts for consumption. For the same level of power, higher distribution voltage requires less current and therefore smaller wires. That general system architecture flourished and prevails to this day.

As demand for electricity grew exponentially, transformers allowed distribution networks to be overlaid with successive layers of higher and higher voltage until America’s highest voltage lines, rated for 500 or 765 kilovolts (kV), now span several states and carry thousands of megawatts (MW); power is then progressively stepped down through those networks for delivery to their 110 or 220 voltage loads. At each voltage level, systems were made sufficiently redundant to lose one or two major lines without any interruption of power to users.

Early Power Generation/Supply Structure

Coal-burning power plants, a dominant source of early power generation, grew in rating to individual generators of up to 1,000 MW. The plants were built close to coal and cooling water resources and sent their output to load areas via high-voltage AC transmission lines on rights of way, which, with the help of eminent domain laws, took the shortest route to electrical loads and to one another. The breadth of interconnections grew to the point that generators in one area could back up those in another when one had to be shut down.

Power was supplied by vertically integrated, interconnected regional power companies. That pattern of growth ended in the 1960s when

  • concern for the environment ended an era of easy transmission line permitting;
  • standard AC transmission voltage levels ended their long progression of increases;
  • economies of scale in generator size reversed as gas turbines made big inroads into the generation mix;
  • nuclear accidents dimmed the bright promise of nuclear energy;
  • industry structure went from vertical to horizontal and from regional monopolies to regulated competition; and
  • transmission grid management became very sophisticated and software-dependent; “smart grids” progressed from concept to necessity.

In speculating where things may go from here, it will be instructive to track the alleged demise of DC and its subsequent resurgence in the power generation, storage, delivery, and consumption matrix.

Resurgence of Demand for DC Power

Demand for DC, overwhelmed by AC at the turn of the 20th century, did not exactly die. Though AC dominated early electricity usage, elevator systems and subways still needed variable-speed motors, which DC alone could then support. Hence DC stayed on as a footnote to the electricity supply menu in New York City until as late as 2007 and is still available from other metropolitan energy suppliers (Fairley 2012). Furthermore, with each passing decade new DC-only loads emerged, primarily in industry (e.g., for aluminum reduction, electrolysis, and metal finishing). That demand was usually met by AC power purchased from the grid and converted back to DC by the user.

But DC’s resurgence in demand was inadvertently started by Edison himself! While tinkering with incandescent light bulbs in 1883, he noticed that he could get DC current to flow through a vacuum from a hot filament to a plate, but not in the reverse direction. In 1904 Sir John Ambrose Fleming, a British scientist, used that idea to patent a glass-enclosed “diode,” a one-way valve for electricity. A decade later Lee de Forest put a metallic “grid” between the filament and the plate and, with a very small amount of current, controlled a much larger DC current flow through the tube, thus introducing the “triode” and with it the ability to progressively strengthen and control a very weak signal. That launched the electronic age, first with vacuum tube–based transmitters and radios, then with solid state valves whose expanded usefulness gave birth to a revolution in communications, computation, and control that has literally changed civilization. Every such device first needed conversion of AC to DC for its operation.

Meanwhile the incandescent light bulb, which spawned the centralized electric supply industry, gave some ground to fluorescent and neon lighting (both AC-dependent) for a few decades. Then electroluminescence—the tendency of certain crystals to emit light of their own characteristic color when subject to small DC currents—became the focus of research the world over and eventually led to light-emitting diodes (LEDs), which now enable visual interface with watches, smartphones, television screens, and advertising signs, and most recently have become the basis for area lighting. LEDs operate on DC from batteries in handheld devices, but from AC-to-DC converters in computers, TV sets, and even individual LED lamp bulbs.

In a modern household, air conditioners, fans, and portable power tools need variable-speed motors, a feature that DC motors can best manage. Where such motors are used, AC is converted to DC within each appliance. Other household loads such as resistive heating can inherently operate with either DC or AC.

Thus the majority of both household and commercial electric consumption may soon be DC, not AC. The advent of electric or hybrid automobiles and their DC charging load will tip the balance to DC’s favor in many households.

DC’s Emergence as a Form of Power Generation

Coal, gas-fired, and nuclear power plants still dominate electric energy production in developed countries, and are not easily adapted to produce DC. But new generation additions are another matter. In 2015 more than 65 percent of America’s new generating capacity came from solar power and wind (Alhart 2015)!


Solar’s share of world electrical generation, inherently DC, is growing by about 55 percent per year (about 40 percent per year in the United States), compared to worldwide growth in electricity demand of just 3 percent per year (SEIA 2013). While America’s solar growth has been driven in part by subsidies, today’s cost per watt is one-third of the cost 10 years ago (figure 2), and approaches parity with grid-delivered power in some regions. Furthermore, as rooftops demonstrate, solar generation can be local. As the effects of global warming are increasingly felt and energy storage becomes a more effective “levelizer” of irregular solar and wind energy production, solar’s rate of market penetration will likely continue at a very high rate.

Figure 2


Growth in wind generation too is dramatically outpacing overall electricity supply growth, expanding by roughly 20 percent per year over the past 15 years—six times the rate of worldwide electricity consumption (GWEC 2015). In Europe, where population density puts high pressure on air quality, roughly 30 percent of all new generation addition since 2000 has been through wind technologies (EWEA 2015).

Wind turbines generate AC, but offshore wind farms, taking advantage of steadier winds and lower aesthetic impact, usually do so at a frequency dependent on wind speed. There being no easy way to convert variable-frequency AC to standard-frequency AC, AC current must first be converted to DC and then back to standard-frequency AC. Thus at its initial first step, this form of wind generation too can be considered a DC source, leading to the question: With recently developed DC-to-DC transformers, why not take the “intermediate” DC voltage and transform it to a higher DC voltage for collection, then to a still higher DC voltage for cable transmission to shore (Barthold et al. 2015)? That approach simplifies electrical architecture, takes advantage of lower-cost DC cable, and allows that cable to supply reactive power to the shore system rather than consume it. One such offshore wind development, with the potential to link East Coast states from New Jersey to Virginia with up to 6,000 MW, is being proposed by the Atlantic Grid Development, LLC (figure 3; Taiarol et al. 2014).

Figure 3

Energy Storage, Too, Is Going DC

Ever since the classical (AC) system model gained sway, the incentive for storing electrical energy during periods of low demand and retrieving it during periods of high demand was apparent. Doing so eliminated the need for extra generators just for peaks in demand. As early as the 1890s, Swiss engineers arranged to pump water into a high reservoir at night, then discharge it through hydroelectric turbines to a low reservoir producing electricity during daylight hours. “Pumped hydro” remains the dominant and lowest-cost source of large bulk energy storage.

But DC-directed shifts in electrical generation and use are changing the storage game dramatically, in many cases shifting it to individual homes or businesses. Output of both solar and wind generators, being highly variable on an hour-to-hour basis, add a new point-of-use incentive for electricity storage, one benefitting both the homeowner and the electricity supplier but demanding downward scalable technology (Gilpin 2015).

While flywheels, compressed air, “supercapacitors,” and other downward scalable means share a small niche in that localized market, the big winner appears to be advanced lithium-ion (li-ion) batteries, the cost of which has dropped by about 65 percent in just five years. Their applications, forecast to increase at roughly 45 percent per year (IRENA 2015), now include home-size storage systems at prices approaching those of gas-powered generators. With that option homes and businesses will remain connected to the grid to purchase (and sell) electricity, but remain independent for critical loads if the grid fails. Electric vehicles, inherently DC, will boost domestic DC demand and be a source of domestic energy storage (Tweed 2015).

Battery storage is making inroads at the power system level as well. At the Imperial Irrigation District in Southern California, 2016 is the target operational date of a 30 MW li-ion unit.

Considering the prospect of eventual DC dominance in home energy use and the potential for both generation and storage of electricity at the consumer level, it’s entirely possible that the wires linking homes to the grid may one day include a DC supply along with, or perhaps eventually supplanting, AC . . . which would, at a massively larger scale and through a dramatically more sophisticated delivery system, take the industry back to where it came from.

DC’s Growth in Energy Delivery

As with energy supply, consumption, and storage, the same shift (albeit more slowly) from AC to DC can be seen in large-scale regional exchange of electric power. AC’s early dominance in power transmission, driven by the availability of transformers to boost voltage, was first challenged in the 1930s when engineers at General Electric’s Schenectady plant asked the following: Why not (1) step low-voltage AC up to high-voltage AC with a transformer, (2) convert the high-voltage AC to high-voltage DC for transmission to another location, (3) convert the high-voltage DC back to high-voltage AC at that location, and, finally, (4) step high-voltage AC back down to use-level voltage at the receiving end of the line?

To demonstrate the idea, General Electric built a 15 kV, 23-mile DC transmission line to bring 150 kW of DC power from a hydroelectric plant in Mechanicville, New York, to Schenectady. Conversion of electricity from AC to DC and back again was achieved with a complex array (“bridge”) of electrical “valves,” each able to conduct electricity in one direction only and which, by being “fired” in the right sequence, can either convert an AC voltage, which varies in both polarity and magnitude, to a constant and unvarying DC voltage or do just the opposite. Those valves were then composed of evacuated tubes containing mercury. It worked!

That prompted Swedish engineers to build the first commercial DC project in 1954: a 98-kilometer undersea cable link to supply 20 MW, at a voltage of 100 kV, from the Swedish mainland to the island of Gotland. That succeeded and was followed by a number of other projects, some using overhead DC transmission lines and inspiring a large US DC project, commissioned in 1970, that eventually sent over 3,000 MW of inexpensive hydroelectric power from the Pacific Northwest to Los Angeles at 500 kV DC.

Why go to the trouble of converting AC to DC at one end of a line, then back to AC at the other?

  1. DC is a more efficient way to transmit power. Because AC voltage varies in magnitude its effectiveness is roughly 30 percent less than a constant (DC) voltage.
  2. Power flow on an AC line cannot be directly controlled, for somewhat the same reason the tension on one of many load-bearing strings of a hammock cannot. In both cases individual line loads are determined by the system in which they serve. Thus large wires on an AC line can’t always be used to full advantage. DC power can be controlled, with the full current-carrying capability of wires used constantly.
  3. AC power lines have a characteristic (an electrical cousin of elasticity) that limits the practical distance power can be transmitted at a given voltage. There is no such limit, other than tolerance for losses, to the distance that DC can transmit power.
  4. Underground or underwater cables constantly need a certain amount of current per mile for “charging.” Charging power is called reactive power in the trade. If a cable is very long the charging current itself will use all the cable’s current rating. DC cables need no charging current.
  5. DC interconnections can help prevent the electrical breakup of systems that have only AC ties.

As these advantages were increasingly recognized, the number of high-voltage DC transmission applications grew steadily wherever transmission distance exceeded a “break-even” distance—the distance at which DC’s savings in transmission line cost plus the value of its operating advantage became greater than the cost of converting AC to DC at one end of a line and back to AC at the other.

The Emergence of Supergrids

Figure 4

By 2015 over 2,000 gigawatts of DC projects were in operation throughout the world. DC’s growth in gigawatts of conversion capacity (figure 4) promises to accelerate as system planners the world over foresee DC “supergrids” overlaying and potentially interconnecting AC systems in all of the world’s major electric power generation and consumption areas—a development that could more than double the number of existing DC installations (Gellings 2015). Supergrids can

  • limit the need for new generation additions by “pooling” generating capacity over a very wide region, each with peak demands offset in time;
  • allow heavier loading of local AC networks without risk of blackouts; and
  • equalize irregular patterns of wind and solar generation to make such “green” resources more widely available.

Figure 5 shows what’s foreseen as a possible North American DC supergrid.

Figure 5

Supergrid Plans and Designs

Early in the evolution of the supergrid concept, planners recognized that its realization would require an economic DC-to-DC transformer to (1) allow the grid’s connection to or integration of existing DC lines of differing voltages and technologies and (2) control flow on long DC lines, formerly individual projects but now in a networked configuration. Foreseeing that need, Friends of the Supergrid, one of many European planning groups, recognized in 2014 that DC-to-DC transformation was an essential element in supergrids (FOSG 2014), in response to which a number of DC-to-DC transformer designs have been proposed—all based on capacitive energy transfer rather than the magnetic transfer characterizing AC transformers. One proposed solution, capable of very high ratings and very low losses behaves, in a DC system, exactly as a magnetically based transformer does in an AC system (Barthold et al. 2015). That innovation is expected to further accelerate DC’s inroads into the power transmission market.

Technical Advances

The dramatic growth of DC transmission has been fueled in substantial part by advances in AC/DC conversion technology that have kept costs relatively stable while increasing DC’s value to the host AC system. Mercury arc valves, for example, gave way to water-cooled solid state (Thyristor) valves in the 1970s and then, in 1997, to integrated gate bidirectional transistors (IGBTs), which allowed a bridge architecture (the voltage source converter, VSC) that changed the game significantly. Until then AC/DC converter stations needed to be fed “reactive” power from the AC network just as AC cables need to draw charging power from such networks. VSCs allowed terminals to supply reactive power to the AC system rather than drawing it from that system. Research into rectifying materials other than silicon, the basis for IGBTs, may lead to bridges capable of substantially higher current ratings and lower cost per megawatt of capacity.

Microgrids and Other Changes in Power System Architecture

Changes in technology notwithstanding, the evolution described above would maintain the classical hierarchical electricity delivery system structure. A DC grid would allow generation in California to benefit users in the Midwest and failures in the Midwest to be rescued by East Coast grids. This hierarchical architecture, together with growing sophistication in the control of power systems and competition among power suppliers (within constraints of the transmission network), has minimized electricity costs to consumers.

Yet while a hierarchical system implies interdependence, there’s a growing countertrend toward a degree of local independence. Microgrids started simply as a convenient way of distributing AC and (through conversion) DC power in a factory or commercial building. But the advent of local energy storage and site-based generation is bringing with it a trend to expand microgrids both functionally and geographically to form a partially autonomous source of power after full or partial loss of supply from the primary grid (Backhaus et al. 2015). Interest in a degree of local independence is fueled in part by concern over the vulnerability of the very large, highly sophisticated, and computer-dependent “smart” interconnected system to cyberterrorism.

Microgrids connected to the primary grid could buy and sell electricity through the primary grid, the latter acting as an energy marketplace. A supergrid would make that market continentwide and accessible even to the homeowner.

Microgrids, regardless of purpose, also support a shift away from AC-to-DC conversion within individual devices in favor of single-point, more efficient conversion to a common DC supply source. They will encourage an end to double conversion of rooftop DC generation (i.e., DC to AC for household supply, then back to DC for loads that demand DC), a transformation now achieved by individual AC-to-DC converters. Microgrids may encourage delivery of DC as well as AC or ultimately instead of AC directly to homes and businesses.

Where from Here?

AC power has served both suppliers and users of electricity extremely well, having the inherent advantages of simplicity and, to a large degree, self-regulation. In elementary terms, when a new electrical load is switched on in an AC system, incremental power flows to that load without instruction or control. Simultaneously, elsewhere on the same system, a governor responds to increase its power output by the same increment . . . not by a signal sent from the new load but from what that generator sees in system behavior at its location on the system. What could be simpler?

In contrast, a DC transmission line needs to be told, through its converter terminals, how much power to send at one end and how much to receive at the other. However, as in many facets of modern industry, the advantage of simplicity is increasingly threatened by today’s explosive growth in sophistication of control and communication. Thus AC systems may one day be rendered vulnerable to the continuing advances in DC technology. Further fueling that prospect is the fact that AC equipment and transmission technologies are very mature, while those supporting DC forms of generation, conversion, and use are still evolving rapidly.

Edison would smile.


Alhart T. 2015. Cloudy with a chance of electrons. GE Reports. Online at

Backhaus S, Swift GW, Chatzivasileiadis S, Tschudi W, Glover S, Starke M, Wang J, Yue M, Hammerstrom D. 2015. DC Microgrids Scoping Study—Estimate of Technical and Economic Benefits. LA-UR-15-22097. Los
Alamos National Laboratory, NM.

Barthold L, Woodford D, Salimi M. 2015. DC-to-DC capacitor-based power transformation: PS 1: Planning study and future requirements for DC system. Paper No. 14. Presented at HVDC and Power Electronics International Colloquium, International Council on Large Electric Systems (CIGRE), September 21–26, Agra, India.

EWEA [European Wind Energy Association]. 2015. Wind in Power 2014 European Statistics. Brussels. Online at statistics /EWEA-Annual-Statistics-2014.pdf.

Fairley P. 2012. San Francisco’s secret DC grid. IEEE Spectrum, November 15. Online at francis cos-secret-dc-grid.

FOSG [Friends of the Supergrid]. 2014. Supergrid preparatory phase: Review of existing studies and recommendations to move forwards. Online at REPORT-rev212.pdf.

Gellings C. 2015. Let’s build a global power grid. IEEE Spectrum, July 28. Online at build-a-global-power-grid.

Gilpin L. 2015. The energy storage market is about to boom. Forbes, September 9. Online at storage-market-is-about-to-boom/#158d071d76df.

GWEC [Global Wind Energy Council]. 2015. Global Wind Power Statistics 2015. Brussels. Online at

IRENA [International Renewable Energy Agency]. 2015. Battery Storage for Renewables: Market Status and Technical Outlook. Abu Dhabi. Online at storage_report_2015.pdf.

SEIA [Solar Energy Industries Association]. 2013. Solar industry data. Washington. Online at

Taiarol PVI, MacPhail GA, Pathirana VS, Mampaey B. 2014. The Atlantic Wind Connection—Building the foundation for offshore wind energy. International Conference on Innovation for Secure and Efficient Transmission Grids, International Council on Large Electric Systems (CIGRE), March 12–14, Brussels.

Tweed K. 2015. GE signs largest battery storage deal to date. Greentech Media, September 2. Online at battery-storage-deal-to-date.

About the Author:Lionel O. Barthold (NAE), a power transmission consultant, is a Life Fellow of IEEE and retired founder and CEO of Power Technologies, Inc. Dennis A. Woodford is former executive director of the Manitoba HVDC Research Centre and now president of Electranix Corporation of Winnipeg.