Creations, Circulations, Tensions, Transitions (19th–21st C.)
Edited By Alain Beltran, Léonard Laborie, Pierre Lanthier and Stéphanie Le Gallic
What interpretation(s) do today’s historians make of electrification? Electrification is a process which began almost a hundred and fifty years ago but which more than one billion men and women still do not have access to. This book displays the social diversity of the electric worlds and of the approaches to their history. It updates the historical knowledge and shows the renewal of the historiography in both its themes and its approaches. Four questions about the passage to the electrical age are raised: which innovations or combination of innovations made this passage a reality? According to which networks and appropriation? Evolving thanks to which tensions and alliances? And resulting in which transition and accumulation?
Quel(s) regard(s) les historiens d’aujourd’hui portent-ils sur l’électrification, processus engagé il y a près de cent cinquante ans mais auquel plus d’un milliard d’hommes et de femmes restent encore étrangers ? Le présent volume rend compte de la diversité des mondes sociaux électriques et des manières d’enquêter sur leur histoire. Il actualise les connaissances et témoigne du renouvellement de l’historiographie, dans ses objets et ses approches. Quatre points d’interrogation sur le basculement des sociétés dans l’âge électrique jalonnent le volume : moyennant quelles créations ou combinaisons créatrices ? En vertu de quelles circulations et appropriations ? Selon quelles tensions et alliances ? Et produisant quelles transitions et accumulations ?
Bias in Electric Power Systems. A Technological Fine Point at the Intersection of Commodity and Service
Is electricity a commodity or a service? This question lies at the heart of the project of electrification. In 1883, Thomas Edison’s central station service offered electricity as a commodity in competition with gas-lighting systems. Within only two decades, public debate suggested that electricity should be a service available to all potential customers in North America. The very nature of America’s power system has kept this tension alive for more than a century, through extension of power lines to cities, industries, and the countryside; through war- and peace-time; through different approaches to regulation and deregulation; and through the process of building the world’s largest interconnected machine, the grid. Historically, power companies also sought energy efficiency, which had the effect of conserving natural resources, while promoting consumption, which had the effect of increasing resource use. This defined a second tension in America’s power system, between conservation and consumption.
Sometimes a microscope offers the best perspective on very large problems. In the mid-twentieth century, engineers and power system operators engaged in heated debate over a very tiny matter – a fraction of a percent, in fact – for the setting on a piece of apparatus that controlled the flow of power on the grid. Called the “bias setting,” this number determined the extent to which one power company aided a neighbor when an interconnection experienced trouble on the line. The type of “trouble” at stake was caused when a change in generation or demand led to a major frequency change on the transmission networks. Each individual power station on an interconnected network had to make adjustments to bring the frequency back to normal. By the 1950s, ninety percent of the industry used apparatus that provided automatic control, but each company selected its own settings. The debate unfolded around whether the setting should return a station to its own economically ideal generation as soon as possible, or allow the station to continue aiding the entire system ← 271 | 272 → until normal frequency returned. In other words, should the industry standard reflect the economic goals of a commodity market or the stability goal of a service network?
This paper will present the bias-setting debate as a lens through which industry ambivalence about the economic nature of electricity can be examined. Drawing upon professional publications, manufacturer records, and engineers’ private business papers, the study explores the significance of the bias setting problem for building the North American grid, while interrogating the relationship between this very small technical matter and larger questions about equity, reliability, and energy efficiency in electric power systems.
Keywords: Interconnection, Energy Efficiency, Power Control, Utilities, Reliability
The word “bias” suggests leanings in one direction or another, and an inherent tension in between. The “bias” that is the subject of this paper – a fractional setting on a piece of apparatus – reflects a tension at the heart of power systems: is electricity a commodity or a service? If electricity is a commodity, the selected bias setting might favor energy efficiency, lower production cost, and higher profit. If electricity is a service, the setting might instead assure greater reliability. The tension between commodity and service characterized the development of electric power systems in the United States through the industry’s entire history. The bias setting on the control apparatus in question determined how much a power company assisted a neighbor to maintain the stability of an interconnected system. Among the diverse entities sharing power on a network, each company determined its own bias setting, which then reflected whether the company favored aiding the network to ensure the delivery of an essential service, or protecting profitability in the sale of a commodity. Several unique characteristics of the US power industry framed the bias setting controversy: its variety of public and private power companies, its unusual organizational structure, and its reliance on a fraternity of technical experts to mediate socio-technical challenges. In this case, the importance of electricity as a service trumped both the economic opportunity presented by electricity as a commodity, and the imperative to improve energy efficiency at individual plants.
This paper begins with a very broad overview of the process of electrification in the United States, including a snapshot of the industry in the 1950s and a discussion of the importance of interconnections. The section that follows delineates two challenges that faced the industry through the mid-twentieth century: load control and frequency control. The third section describes the operating challenges that led to the bias setting controversy, how the controversy played out, and its resolution. Finally, the paper offers some analysis of this process and the implications for the history of electrification. ← 272 | 273 → Examining this debate about the proper bias setting sheds light on the ongoing negotiation over the true nature of electricity in modern energy economies.
Electrification in North America and the Tension Between Commodity and Service
Over the course of decades, thousands of companies – public and private – electrified North America. The origins of today’s power networks date back to the late nineteenth century. In 1882, Thomas Edison’s Pearl Street Station introduced economically viable central station service for electric lighting. Two years later, Nicola Tesla invented a practical alternating current induction motor. At the Chicago World’s Fair in 1893, George Westinghouse demonstrated an alternating current electric system that simultaneously illuminated lights, ran a transit system, and powered a variety of motors.1 Power companies adopted these innovations to build markets in urban centers across the continent.2 For a variety of reasons, Americans readily embraced electrification and the industry grew quickly.3 By 1902, more than 800 public and 2,800 private systems produced and sold 2,300 million kilowatt-hours of electricity to consumers.4 Twenty years later, power companies produced more than 23,000 million kilowatt-hours of electricity, and by 1950, that number had again increased more than ten-fold.5 ← 273 | 274 →
Diversity has always characterized electrification in North America. In the late 1800s, both government- and investor-owned licensees of various lighting and power systems brought electricity to cities and towns. Companies also used a variety of energy resources, initially relying on coal, falling water, natural gas, petroleum, and a few other combustibles to produce electricity. Nuclear joined the mix in 1957. Additionally, the US regulatory system varied regionally and by level of government. For example, by 1914, 43 states regulated power companies, but Texas waited until 1975. Figure 1 aggregates a variety of data to provide a snapshot of the industry in the mid-1950s, the period of concern in this paper. Investor-owned utilities dominated more than three quarters of the market and power companies relied heavily on fossil fuels, especially coal, for primary energy resources.
In the United States, companies produced and delivered electricity at times as a commodity and at others times as a service. When Edison first offered practical electric lighting in the 1880s, he sold a commodity. His customers were wealthy homeowners, department stores, downtown boosters, and the like. By the early twentieth century, however, power companies promoted electricity as a service. As one engineer noted, “Every step upward in the overall efficiency means a chance for a more economical supply and a larger market […] it should be possible to make electrical supply a necessity and not, as it now is in many instances, rather a luxury.”6 Further, politicians and activists pushed for greater equity in access to electricity, lobbying to bring power especially to rural customers. The shift from commodity to service is evident in figures 2 and 3. As the graphs illustrate, the cost of electricity dropped rapidly in the early 1900s; and during those same years, the number of dwellings using electricity rose dramatically. ← 274 | 275 →
Once consumers began to depend on electricity for lighting, motive power, and transportation, reliable service grew in importance. In fact, the power companies recognized as early as the late nineteenth century that the assurance of power on demand was essential to luring and keeping customers. With the promise of electricity as a service came the responsibility and expense of reliability. Power companies installed storage batteries and extra generators to create “spinning reserve,” that is, surplus electricity available on short notice. They also linked with neighbors to access backup power for planned and emergency outages.
Government regulation further underscored the notion of electricity as a service. While many municipalities and some cooperatives built power plants, investor-owned utilities quickly and permanently dominated the US power market. In the early 1900s, when states began to regulate the private companies, state commissions established monopoly service areas, set rates, and oversaw planned expansions. In return, the companies promised equitable and reliable service for a fair financial return.7
Despite their status as regulated monopolies, investor-owned power companies did not entirely abandon the concept of electricity as a commodity. Indeed, they lobbied hard for rates that guaranteed a steady profit and they organized financial structures that further increased the returns for savvy investors.8 They also adopted technologies and operating practices that lowered their costs, they promoted consumption, and thus they realized profits. Although some states, like Nebraska, prohibited investor-owned power companies altogether, for the most part utilities continued to weigh their responsibilities as service-providers against their objectives as commodity vendors.9
In the United States, both government and investor-owned power companies operated without a central authority to plan and control the transmission networks. They began to interconnect with neighbors in the ← 276 | 277 →
late 1890s in order to improve system reliability, economically balance the use of different energy resources, and reach wider markets. At the outset, interconnection appeared to be a simple matter of connecting the lines with interties (also called tie-lines) and scheduling power trades. But companies quickly recognized the challenge of moving alternating current across multiple power lines with multiple owners. Engineers designed solutions that often caused new problems, and each network expansion amplified the complexity of power control. By the mid-twentieth century, interconnections among autonomous power systems became a salient feature of the North American electric grid.
Challenges of Interconnection: Frequency Control and Load Control
Because electricity is dynamic, even the most straightforward systems require finesse in order to function properly. Among the many physical and technical challenges, frequency control and load control persistently plagued interconnected networks throughout the twentieth century. To assure that consumers had the right quality of electricity at the moment they turned on a switch, power system operators had to control the “frequency,” also called the “speed,” of the alternating current. In addition, to meet financial obligations and to achieve greater energy efficiency, operators had to control how much of consumers’ demand, called “load,” was assigned to each generator. Together, management of both frequency and load produced complicated control issues.
Generators and all other devices connected to the same network must operate at exactly the same frequency. Today, in the United States, the grid operates at 60 cycles per second, and this frequency is held steady across the system.10 Generators, on or off a network, speed up and slow down (in other words, change frequency) in response to changing load. For example, when demand increases, it has the effect of slowing a generator, just as extra weight might slow a horse pulling a cart. Speed governors, in use since the earliest days of electrification, automatically act to return generators to the correct frequency when a load changes, but this takes time. The elapsed time is called the “governing characteristic,” or “natural characteristic.” This holds true for a collection of linked generators as well. For many decades, when the automatic adjustments ← 277 | 278 →
made by governors were inadequate to system demands, operators made any additional frequency adjustments by hand.
A simplified example helps to explain this phenomenon. Power Plant A provides electricity to a department store and a factory. In the morning, the plant operator starts the generators to assure that the lights will go on when the store opens. The moment the store’s owner turns on the lights, the generators slow a little bit. The speed governors bring the generators back to the correct frequency. The elapsed time, or governing characteristic, is so brief that the store’s owner barely notices any change in the quality of the lights. Later in the morning, the factory owner starts up an assembly line, and this additional load also slows the generators. With the larger load, it takes the governors longer to restore system frequency. As a result, the lights in the department store may flicker noticeably.11 While a brief and barely perceptible flicker of lights may not matter in a department store, for other applications, like electric clocks or textile machines, this is unacceptable. When frequency changes, an electric clock loses or gains time. Similarly, with varying frequency, a textile machine produces uneven and flawed goods. The power plant operator is committed to delivering high quality and dependable electricity to customers. To hasten the return to system frequency, he or she must manually adjust the generators, rather than waiting for the governors to do the job.
By the 1920s, power plants provided electricity to many different types of customers, and also sometimes shared electricity with each other. This made it very complicated for system operators to make appropriate frequency adjustments on a large system. Companies began to experiment with additional automatic control devices to relieve operators from this burden and keep their systems more stable. The recorder depicted in figure 4 is an example of a precursor to one of the first automatic frequency control devices. ← 278 | 279 →
Between 1928 and 1932, utilities tested multiple automatic frequency control devices on their interconnected systems to compare methods and determine which technologies held the greatest promise. In the tests, the devices proved successful at maintaining a steady 60 cycles per second across the systems, but the apparatus caused new problems.12 Companies sharing power on an interconnection carefully scheduled the distribution of load in order to achieve greater energy efficiency and keep costs down. Automatic frequency control upset those planned schedules. This, in turn, undermined the distribution of the load between different companies and the related financial arrangements. ← 279 | 280 →
The hypothetical system in figures 5 and 6 illustrates how this might work. There are two power plants: Power Plant A and Power Plant B. There are two customers: the department store and the factory, and in this instance the factory has three assembly lines. Power Plant A sells electricity to the department store and Power Plant B sells electricity to the factory (figure 5.a). The plant operators have determined, however, that Power Plant A generates electricity more efficiently than Power Plant B, and also can provide electricity for both the department store and just two of the factory’s three assembly lines at this lower cost. The cost will rise, however, if Plant A also supplies electricity to the third assembly line. So, the operators decide to create an interconnection and they agree that Power Plant A will generate electricity for the store and the factory’s first two assembly lines. But when the factory starts up the third assembly line, Power Plant B will provide the electricity. Money will exchange hands accordingly (figure 5.b).
If the interconnection operates according to plan, both power plants will contribute electricity to the system once both the store is open and all three assembly lines are on at the factory (figure 6.a). After Power Plant A installs an automatic frequency controller, however, operations do not unfold as intended. As before, the operator of Power Plant A starts up the generators in time for the department store owner to turn on the lights, and for the factory owner to start up the first assembly line (figure 6.b). With an automatic frequency control device, the generators stay so close to the correct frequency that there is no noticeable difference in the quality of electricity delivered to the store and the factory. This holds true when the factory owner starts up the second assembly line, demonstrating that the automatic frequency controller works as intended (figure 6.c). Yet the ← 280 | 281 → same thing happens when the factory owner starts up the third assembly line (figure 6.d) and herein lies the problem. The automatic frequency controller works so well that there is no dip in frequency. The frequency change would otherwise serve as a signal to the operators to start bringing electricity from Power Plant B into the network. The operators would have to rely on telephone communications or some other mechanism to know that the time had come to share the load. Without additional control activity, Power Plant A provides all the electricity for the interconnection, and at a higher cost, and Power Plant B provides none, despite the plans made by the plant operators.
← 281 | 282 →
This example minimizes the complexity of actual operations. Plant operators addressed many additional factors in assuring that generators, transformers, transmission lines, and distribution networks all stayed in synchrony and delivered electricity as expected by customers. Some used the latest high-technology tools, such as telemeters, totalizers, and network analyzers to gather and analyze data; and nearly all relied on telephones to discuss system changes.13 Facing various incarnations of the problem caused by automatic frequency controllers, plant operators turned to automatic load control techniques as well. Managing load distribution while maintaining steady frequency was essential to successful interconnection. Combined automatic frequency and load control proved to be quite difficult and occupied engineers for the next several decades.
During the 1930s, companies experimented with numerous methods of controlling both frequency and load, some relying on a single plant to control for an entire network, others using distributed controls. By the late 1930s, many of the largest systems found that they achieved better frequency control with closer adherence to planned load distribution, without overburdening single plants, by using the approach called “tie-line bias control.”14 In this method, one station automatically controlled frequency for the entire network, and devices located at the tie lines automatically maintained the scheduled load distribution. When there were frequency changes anywhere on the system, the load controllers at the tie lines allowed the other stations to aid the frequency control station with adjustments to bring the system-wide frequency back to 60 cycles per second. But the devices allowed only minor and gradual adjustments, and only for a set period of time, regardless of whether or not this was sufficient to stabilize the frequency across the system. This resulted in brief diversions from the scheduled load distribution. The term “bias” referred to the amount the load controller allowed the load to vary before returning to the schedule. As one engineer explained, “The frequency controlling station had but to start to increase its generation and the load increase was automatically spread throughout the system so that no one ← 282 | 283 →
station had any appreciable increase in burden. This was all done without the use of telephone calls.”15
Operators set the bias to be a fractional percentage of the governing characteristic of the system as a whole. Each company individually determined how high or low to set the bias and there were two opposing views about this. One group believed that the tie line was there strictly for buying or selling power, and should be held to a specified load. This represented the “commodity” view. The other group believed that a tie-line was there to share the job of keeping the system stable, while allowing an average interchange between systems that met the average of the sharing agreements between power companies. This represented the “service” view. As one engineer explained, through the 1940s operators arrived at their bias setting through observation and “by some calculation and considerable arbitration, which often included a sizable factor of ignorance.”16 For many years, the effects of multiple bias settings on interconnected systems were barely noticed. In the early 1950s, however, operators began to care deeply about fractional differences in bias settings.
The Nub of the Problem
The heightened interest in bias settings came with the growth of interconnections in both size and complexity. By 1949, numerous interconnected systems served large regions of the country, and within the next decade, most merged into just seven giant power pools. The Federal Power Commission maps reproduced in figure 7 illustrate the consolidation that took place during the 1950s. Power production also grew tremendously during and after the Second World War, as shown in figure 8. ← 283 | 284 →
With the rapid expansion of interconnections and the greater quantity of electricity moving across the networks, systems experienced big load changes. This in turn caused more pronounced frequency changes. These disturbances magnified the effect of bias settings on operations. By 1955, the disagreements about bias settings left engineers and operators with “battle scars.”17 Looking back on the period, Charles Concordia, an eminent electrical engineer with General Electric, wrote to a retiring colleague, “when the frequency changed by a tenth of a cycle […] it was sufficient cause to establish a committee to study the problem for a year and a half.”18
Problems started on the transmission lines of the Interconnected Systems Group (ISG), one of the oldest and largest systems in the country. Eleven companies in western Pennsylvania, eastern Ohio, and part of West Virginia organized the interconnection in 1928. By the 1930s, when ISG formed the Test Committee to address frequency and load control problems, this portion of the system reached across nine states. In the early 1950s, ISG included seventy-five major companies, stretched from ← 285 | 286 → Canada to the Gulf of Mexico and from the Rocky Mountains to the Atlantic Ocean, and included the Tennessee Valley Authority (TVA). Through the work of the Test Committee, ISG issued voluntary operating standards and strongly encouraged participating power companies to adopt these standards. In 1951, the Test Committee established that “bias settings are to be increased on all systems … to reflect approximately 1 percent of system load per 0.1 cycle frequency departure.”19 In other words, ISG encouraged all interconnected companies to set the bias voluntarily to be one percent of the operating area’s governing characteristic.
On February 2, 1955, a relay at the TVA Shawnee power plant tripped, causing part of the Illinois-Missouri power pool to disconnect.20 This resulted in interruption of the scheduled generation across Iowa, Missouri, Illinois, and Kentucky. A fault on a generator started the trouble, triggering the relay to trip and disconnect the line carrying power between the Illinois-Missouri Pool and Electric Energy, Inc., two participants in the larger interconnected system that included TVA. The varying operator decisions, bias settings, and techniques for tie line control led to a small cascade of problems. Two weeks later, representatives from the largest affected utilities met to address how to operate during this and other types of emergencies. Participants agreed on the basic facts, but differed considerably about how to respond. Separately, three of the utilities, Union Electric Company, Central Illinois Public Service Company, and Illinois Power Company, met with individuals from Leeds & Northrup Company (L&N), the manufacturer of their automatic control instruments.21 Nathan Cohn, representing L&N, discussed the February trouble.22 He noted that the bias settings were lower than one percent of the governing characteristic, resulting in improper control responses and exacerbating the unfolding problems. Cohn recommended that the utilities adjust bias settings upward.23 ← 286 | 287 →
Six additional disturbances occurred on the ISG networks over the next six months, including one on the Ohio Valley Electric Corporation (OVEC) system in March 1955. Several utilities had earlier formed OVEC strictly to provide power to the federal government’s uranium enrichment plant in Piketon, Ohio, pictured in figure 9. Power trouble affecting the Piketon uranium plant undoubtedly raised concern.
The ISG Test Committee began to focus on these disturbances. Following the Piketon incident, Howard Stites, an electrical engineer with Central Illinois Public Service Company, reported to the ISG Test Committee that the trouble lasted only four minutes, and in this case, “demonstrated that the frequency bias [of one percent] was satisfactory.”24 At the May meeting of the Test Committee, Stites reported on all seven ← 287 | 288 → disturbances, which resulted in a less salutary view of the bias settings used by the affected companies.25 The Test Committee then told a larger Interconnected Systems committee “the present frequency bias of 1 percent per .1 cycle of deviation is inadequate, unrealistic and detrimental to Interconnected System operation.”26 This resulted in a decision to research the question more thoroughly.27 Despite this high level of concern, problems continued. Figure 10 indicates the record of a second disturbance on OVEC, this one occurring on June 21, 1955. The chart shows that the frequency jumped between 9:10 pm and 9:20 pm and it appeared to take about ten minutes to bring the system back to a stable 60-cycle frequency.
During the summer, the Test Committee focused on understanding the state of the art. The committee surveyed all seventy-five ISG member companies, most of which did not participate in the OVEC system. On October 25, 1955, the committee issued a report that contended that settings below one percent bias were too low.28 Although the committee members reached this finding, they felt that the survey results were incomplete and inadequate for fully addressing the bias setting problem. In December, the committee decided to bring the bias setting question to the full ISG membership.29 The “plan of attack” was “to invite a representative from each of four (4) industries, which industries are associated with the problems of system regulation, to meet with the Test Committee.”30
In mid-February 1956, the Test Committee convened for a two-day meeting. Engineers from L&N, General Electric Company, Westinghouse Electric Corporation, and Woodward Governor Company made their presentations. All four companies competed in the automatic control field. Speaking for L&N, Nathan Cohn discussed the theoretical fundamentals of system regulation, the basic concepts of automatic control and operation, the priority of customer service, and the status of present control techniques. He addressed a number of questions about bias control, including the effect of different bias settings on tie lines after a fault occurs on a system. He outlined various scenarios and likely outcomes based on settings below one percent, at one percent, and above one percent. Cohn then argued that the best outcomes occurred when the bias setting was equal to or greater than one percent of the system’s natural characteristic. Following the presentations, the utility members met in private session to discuss the talks and decide how to proceed. The Test Committee asked only Cohn to address the full ISG committee later in the spring.31
The elevation of the bias setting issue to the full ISG membership prompted industry-wide interest in the topic. The American Institute of Electrical Engineers (AIEE) asked Cohn to present his paper at the summer general meeting in June. While preparing for his spring and summer presentations, Cohn exchanged correspondence with Russ Purdy, Vice Chairman of the AIEE Committee on System Engineering. Purdy underscored the disagreement among system operators regarding ← 289 | 290 → increasing the bias setting. He wrote to Cohn, “I can state flatly that there is still a wide divergence of opinion within the group.”32 Although he had a clear preference for a higher bias setting, Cohn emphasized his intent to leave the final decision to individual system operators. “We are always glad to discuss the theoretical aspects of this problem, but in the final analysis it is the operating people themselves who will want to determine what operating practices they will use.”33
The AIEE distributed Cohn’s talk in advance of the June meeting. Fifteen well-known and well-respected engineers, most from very large utilities in the United States and Canada, provided comment. They concurred that the bias setting question, though seemingly minute, was of great significance to the industry. The comments also reflected the lack of agreement among these experts about the bias setting itself.34 Cohn presented his paper and addressed the commentary. He clarified that the controversy rested on two potentially incompatible priorities among different power companies: rapid restoration of system stability and economy. If rapid restoration of stability to guarantee steady and reliable service was the priority, then a bias setting greater than one percent had a significant advantage. For a company trying to maximize economy on its own system, and thus sell a commodity at the lowest cost, a bias setting below one would be preferred. Cohn pushed utilities, as they decided on bias settings, to determine whether their first responsibility was to meeting stability obligations or economy goals. Following the AIEE meeting, the discussion moved into trade journals and international publications.35
Beginning with the ISG full membership meeting in the spring, engineers at the power companies began to converge on a standard bias setting equal to or greater than one percent of the natural characteristic. One Test Committee member reported, “While there was much controversy last year on the suggestion of bias equal to system characteristic, I was very pleased to find that there is now apparently complete agreement after ← 290 | 291 → a year’s operation under that plan.”36 ISG disseminated the standard to all member companies in 1957 and again in 1960. The ISG recommendation stated, “Each individual operating company should set the bias setting of its tie line load controller equal to or as close as possible to its natural system characteristic as estimated to apply to its system peak load (for the current year). […] In no case should the bias be set at a value of less than 1 percent of estimated system peak load (for the current year) per .1 cycle change.”37 The North American Power Systems Interconnection Committee, formed in 1963 and including networks across the United States and Canada, adopted the same standard.38 The industry universally embraced higher bias settings to improve system frequency, achieve closer adherence to tie line schedules, and reduce the regulating burden on individual power plants.39
Implications of the New Standard
The informal process of developing a bias-setting standard reflected the nature of the American power industry. Without a government regulator, national standards group, or common owner to formally determine the solution, individual stakeholders pushed the discussion of bias setting through isolated companies, interconnection meetings, and professional associations. The problem, though confined to a small detail on a single piece of apparatus, posed a serious threat to expanding interconnections in the 1950s. The inherent instability in grid operations resulted in repeated failures, and, if left unattended, might have led to major power outages.40 The process of resolution matched the grid itself. Engineers and operators cobbled together information and ideas to produce a coherent body of ← 291 | 292 → knowledge. While many shared a consensus about the right approach, all agreed that bias settings ultimately rested in the domain of individual system operators. In this socio-technical system, both innovation and operating technique depended upon a fraternity of experts to delineate a problem and define the purposes and consequences of multiple solutions. The ideal solution provided interconnections with greater stability across the continent.
The bias controversy is but a tiny sliver of a very large project. North America’s power grid has been called the world’s largest interconnected machine. Once power companies chose to interconnect, they were bound to consider the demands of the larger system, sometimes to the exclusion of their own interests. The people actually and physically controlling electricity on a day-to-day basis had to decide how to set the bias for each of their own plants. The vast majority of operators worked for profit-seeking companies. At the end of two years of bitter controversy, the operators voluntarily chose the setting that preserved the integrity of the system, rather than the setting that guaranteed greater energy efficiency at their own plants and delivered higher profits.
For utilities selling a commodity, energy efficiency defined another tension inherent in the North American power system, between natural resource protection and use. On the one hand, energy efficiency suggests that a power company can produce each kilowatt of electricity using less of the primary energy resource. On the other hand energy efficiency indicates that the cost of production is lower, hence the charge to the customer may be lower, and the customer may be inclined to use more power. Historically power companies increased energy efficiency at the very same time that they encouraged greater consumption, which in turn led to increased resource use rather than improved environmental protection. While the debate about bias setting ostensibly addressed system stability, it also spoke to the industry’s complicated position within movements to conserve resources and protect ecosystems.
This episode from the middle of the last century illustrates that many individuals in the power industry saw themselves as service providers, even when their employers sought to market a commodity. The bias setting controversy resulted in a leaning toward service on the part of the US power industry. For the majority of power system experts, the obligation to aid one’s neighbor on an interconnection outweighed the opportunity to marginally boost profits. Notably, by protecting the quality of electrical service, system operators also reduced the conservation benefits achieved through energy efficiency. In establishing a bias setting standard, the fraternity of experts navigated two central tensions of electrification. They made choices between general system stability and ← 292 | 293 → local economic goals, and between the good of the network as a whole and the management of primary energy resources. These same tensions will frame energy choices going forward, and will prove crucially important when Americans attempt to align environmental goals with energy wants and needs. ← 293 | 294 →
* Author’s note: I would like to thank the Espace Fondation EDF and the Comité d’histoire de l’électricité for organizing the Mondes Électriques conference, and for inviting participants to contribute to this volume. In addition, I would like to express my gratitude to Sarma (NDR) Nuthalapati, Ph.D. for providing technical comment on my paper.
1 The seminal work in the history of electrification, particularly for historians of technology, is Thomas Parke Hughes, Networks of Power: Electrification in Western Society, 1880-1930 (Baltimore: Johns Hopkins University Press, 1983). Networks of Power compares the early sociotechnical systems of electrification in the United States, England, and Germany. For an exploration of the lives of Edison, Westinghouse, and Tesla, see Jill Jonnes, Empires of Light: Edison, Tesla, Westinghouse, and the Race to Electrify the World (New York: Random House, 2003).
2 For examples of histories of electrification of cities and regions in the United States, see Paul W. Hirt, The Wired Northwest: The History of Electric Power, 1870s-1970s (Lawrence: University Press of Kansas, 2012); Harold L. Platt, The Electric City: Energy and the Growth of the Chicago Area, 1880-1930 (Chicago: University of Chicago Press, 1991); Mark H. Rose, Cities of Light and Heat: Domesticating Gas and Electricity in Urban America (University Park, PA: Pennsylvania State University Press, 1995).
3 To understand the social shaping of the American power system, see David E. Nye, Electrifying America: Social Meanings of a New Technology, 1880-1940 (Cambridge, MA: Massachusetts Institute of Technology Press, 1990). For an examination of federal programs and domestic electrification, see Ronald C. Tobey, Technology as Freedom: The New Deal and the Electrical Modernization of the American Home (Berkeley: University of California Press, 1996). For a gendered social history of electrification, see Ruth Schwartz Cowan, More Work for Mother: The Ironies of Household Technology from the Open Hearth to the Microwave (New York: Basic Books, 1983).
4 Bureau of the Census, Central Electric Light and Power Stations, 1902 (Washington: US Government Printing Office, 1905).
5 Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1970 (Washington: US Government Printing Office, 1975).
6 “Station Efficiencies,” Electrical World 52/22 (1908): 1158.
7 For a discussion of this “utility consensus,” see Richard F. Hirsh, Power Loss: The Origins of Deregulation and Restructuring in the American Electric Utility System (Cambridge, MA: Massachusetts Institute of Technology Press, 1999).
8 See Platt, Electric City for an examination of how Samuel Insull made the utility company, Commonwealth Edison, both profitable and essential to consumers in Chicago at the turn of the last century.
9 This tension is even more evident in the twenty-first century with a disaggregated electricity market across North America. In some regions, companies trade wholesale power much like any other commodity and in others government-owned power systems persist. See Hirsh, Power Loss, for a discussion of industry restructuring in the late twentieth century.
10 To be entirely accurate, today there are three major operating units that comprise the grid in the United States: the Eastern Interconnection, the Western Interconnection, and the Texas Interconnection. Frequency is held steady within each, but may differ slightly from the other two.
11 As a point of further clarification, the change of frequency would likely pass unnoticed if the bulbs are incandescent because the filament, once heated, remains bright; but a frequency change might be more noticeable with other types of bulbs, such as fluorescent bulbs, which flicker.
12 R. Bailey, “Fundamental Plan of Power Supply in the Philadelphia Area,” American Institute of Electrical Engineers – Transactions 49/2 (1930): 605-620; R. Brandt, “Automatic Frequency Control,” Electrical World 93/8 (1929): 385-8; Lloyd F. Hunt and Hydraulic Power Committee Subcommittee on Automatic Frequency Control, “Automatic Frequency Control in Hydroelectric Plants,” Electrical West 64/6 (1930): 337-354; Philip Sporn and W.M. Marquis, “Frequency, Time and Load Control in Interconnected Systems,” Electrical World 99/11 (1932): 495, 618.
13 Telemeters gathered data and transmitted it from one location to another, totalizers aggregated and summed the data transmitted by multiple telemeters, network analyzers modeled the behavior of power systems under different scenarios.
14 Nathan Cohn, “Recollections of the Evolution of Realtime Control Applications to Power Systems,” Automatica 20/2 (1984): 148.
15 S.B. Morehouse, “Frequency-Load Control on Interconnected Power Systems, Atlanta Meeting, February 13, 1935” presented to the Interconnected System Operating Meeting, in papers of the North American Power Systems Interconnection Committee, courtesy of the North American Electric Reliability Corporation, Atlanta, Georgia, 7. This collection is now housed at the Hagley Library, Wilmington, Delaware, where it is currently being processed. Hereinafter, the collection will be cited as NAPSIC/NERC.
16 R.T. Purdy Discussion in Nathan Cohn, “Some Aspects of Tie-Line Bias Control on Interconnected Power Systems,” Transactions of the American Institute of Electrical Engineers: Power Apparatus and Systems, Part III 75/3 (1956): 1429.
17 E.S. Miller, “Letter to Members of Northwest Regional Committee,” 24 June 1955, NAPSIC/NERC.
18 Charles Concordia, “Letter to Cohn,” 28 March 1972, Nathan Cohn Papers, MC 317, Box 1, Institute Archives and Special Collections, Massachusetts Institute of Technology Libraries, Cambridge, MA. Herinafter this collection will be cited as Nathan Cohn Papers, MC 317.
19 “Memorandum from J. R. Smith, Chairman, to All Members of the Northwest Regional Committee of the Interconnected Systems Committee,” 21 October 1952, NAPSIC/NERC.
20 “Memorandum of Conference Held in D.H. Cameron’s Office February 14-17, 1955,” Nathan Cohn Papers, MC 317, Box 44; Untitled “factual report of seven Interconnected System disturbances as they affected the Illinois-Missouri and the United Pools,” undated, Nathan Cohn Papers, MC 317, Box 5.
21 By this date, L&N supplied automatic frequency control apparatus to over 90 percent of the market. I. Melville Stein, “Measuring Instruments, A Measure of Progress,” Newcomen Publications in North America (Princeton: Princeton University Press, 1958), 19.
22 Nathan Cohn, 1907-1989, was a systems control expert in the power industry and also the author’s father.
23 “Some Aspects of Bias Control – Getting the Most from It During and Following Periods of System Disturbance,” 13 April 1955, Nathan Cohn Papers, MC 317, Box 5; “Report on the Current Status of Load – Frequency Control Methods and Equipment by the System Controls Subcommittee of the Committee on System Engineering,” American Institute of Electrical Engineers Fall General Meeting, Chicago, Illinois, 1-5 October 1956: 2.
24 Howard E. Stites, “Letter to W.T. Pavely,” 21 March 1955, NAPSIC/NERC.
25 L.V. Leonard, “Letter to Nathan Cohn,” 28 November 1955, Nathan Cohn Papers, MC 317, Box 4.
26 Untitled “factual report of seven Interconnected System disturbances as they affected the Illinois-Missouri and United Pools,” 3.
27 Leonard, “Letter to Nathan Cohn.”
28 L.A. Mollman, “Report of Bias Analysis Survey for June 21, 1955,” 26 April 1956, Nathan Cohn Papers, MC 317, Box 4.
29 L.V. Leonard, “Letter to Test Committee Members, Interconnected Systems,” 6 January 1956, Nathan Cohn Papers, MC 317, Box 4.
31 L. P. Julian, “Minutes of the Test Committee Meeting, February 15-16, 1956, Cincinnati, Ohio,” Nathan Cohn Papers, MC 317, Box 4.
32 Russ T. Purdy, “Letter to Nathan Cohn,” 9 April 1956, Nathan Cohn Papers, MC 317, Box 4.
33 Nathan Cohn, “Letter to Russ L. Purdy,” 23 April 1956, Nathan Cohn Papers, MC 317, Box 4.
34 Cohn, “Some Aspects of Tie-Line Bias Control on Interconnected Power Systems.”
35 M.D. Leighty, “Memo to Nathan Cohn, Subject: Your Paper on Tie Line Bias Control,” 26 July 1956 (Photocopy of unnumbered page from Electrical West 117/1 attached), Nathan Cohn Papers, MC 317, Box 41; Nathan Cohn, “Memo to D. E. Moat, Subject: AIEE Paper 56-670. ‘Some Aspects of Tie Line Bias Control on Interconnected Power Systems,’” 14 January 1957, Nathan Cohn Papers, MC 317, Box 41; Nathan Cohn, “A Step-by-Step Analysis of Load Frequency Control Showing the System Regulating Response Associated with Frequency Bias,” Meeting of the Interconnected Systems Committee, Des Moines, Iowa, 1956 (Philadelphia: Leeds & Northrup Company, 1956).
36 L.A. Mollman, “Memorandum from Mollman to Howell, Subject: Test Committee,” 22 May 1957, NAPSIC/NERC.
37 “Operating Recommendations for the Interconnected Systems Sponsored by the Test Committee and Approved by the Main Committee, May 1, 1957,” NAPSIC/NERC.
38 “North American Power Systems Interconnection Committee Minutes of Meeting January 15-16, 1963 – New Orleans, La,” NAPSIC/NERC.
39 O.A. Demuth, “Letter to The Lamme Medal Committee of the American Institute of Electrical Engineers: Appendix III,” 7 September 1962, Nathan Cohn Papers, MC 317, Box 3.
40 Notably, the North American power system experienced its first major cascading power failure on November 9, 1965. A faulty relay setting initiated the blackout, and the magnitude of the mismatch between supply and demand far exceeded the capacity of automatic control devices to bring the system back into synchrony. This problem was different in both kind and scale from the problems addressed during the bias setting debate. Federal Power Commission, Northeast Power Failure, November 9 and 10, 1965: A Report to the President (Washington: US Government Printing Office, 1965); Nathan Cohn, L&N and the Control of Electric Power Systems (Philadelphia: Leeds & Northrup Company, 1966).