Ask any of the people who were left idle at work and stranded on the hot and humid streets of New York City on the afternoon of August 14th, 2003 just how much electricity matters and one thing will be resoundingly clear: electricity powers our lives and if we do not take the assistance of usave and switch to a greener form of electricity, many non-renewable sources of energy will perish soon. The 2003 Northeast blackout left around 50 million people without power for four days and caused 61,800 megawatts to be taken off the grid. When all was said and done, eight northeastern states and Ontario, Canada had lost power. Between four and ten billion dollars and the loss of nineteen million working hours were among the blackout’s disastrous side effects, not to mention interruptions in transportation, communication, water supplies, and even sporadic cases of looting.

There was no warning before the chaos and no way to easily mitigate it once it began, but it could have been avoided had our electrical grid been supported by some form of adaptive intelligence to track and report on its operating conditions or, put simply, had the transmission grid been able to know what was happening over its system and why. Mary Ann Piette, a deputy staff scientist at Lawrence Berkeley National Lab (LBL) and research director of the California Energy Commission’s (CEC) Public Interest Energy Research (PIER) Demand Response Research Center, hopes to create the essential elements of a system that will do just that. Part of a larger “smart grid” movement aimed at modernizing the national electricity grid, her technology, dubbed “OpenADR,” forms the set of embedded instructions in a software package that will revolutionize the way the power grid transmits not only electricity but also information. As Piette points out, such a system could have averted the crisis in the northeast: “Had OpenADR been widely deployed during the 2003 electric grid crisis, the utilities could have issued automated demand response alerts to unload the electric grid and more quickly bring it back to stable operation.”

Do ask, do tell

The 2003 blackout demonstrated that excessive load on the transmission grid can cause electric faults that cut through the protective systems in place today, leading to large-scale disaster. Official government investigation into its causes established that overheating of transmission lines due to excessive load on the grid in Ohio caused the lines to expand and sag until they came into contact with adjacent trees and failed. Simultaneous monitoring system malfunctions eventually led to a cascade of rolling outages. Along with her colleagues at LBL and in industry, Piette has been working to resolve inadequacies in managing the load on the power grid to prevent future cascade events. Her goal is to create a system that is self-aware and can therefore avoid blackouts by fixing itself through demand management. The Automatic Demand-Response Open System Specification (OpenADR) is a piece of open-source software that implements a set of standardized price-modeling instructions and communicates them across the Internet to “smart” power management devices—like smart meters and programmable thermostats—in homes and buildings. OpenADR allows customers to play an active and responsive role in reducing their use of power and, in doing so, enables them to help control the large-scale movement of electricity through the power grid, reducing the peak load on the system in a way that was not possible in the past.

With control signals issued by OpenADR, consumers can be alerted to either reduce their unnecessary electricity usage at critical hours of the day or pay higher rates. Research indicates that when people are informed of grid overload conditions in advance, they tend to cooperate and take action to reduce their own consumption. This response translates into a significant relief of the peak load on the power grid, exactly the kind of relief that could have prevented the 2003 blackout. Moreover, from the viewpoint of utilities, standardized signaling from OpenADR can allow building and industrial power control systems to be pre-programmed, enabling a demand response event to be fully automated without human intervention.

As the load carried by our aging power grid continues to grow, more unpleasant surprises might be in store for consumers in the near future. However, in the next ten to 20 years our electrical power system will evolve such that the generation, transmission, distribution, and consumption of power will be governed by extensive, real-time communication among power system components. Software like OpenADR will be crucial to this self-aware, self-healing “smart grid.” But why is our current grid so vulnerable?

This figure shows an example of a typical “grid load profile” forecast, the actual power consumption over the course of a day, and the amount of power generation resources that are online to maintain reliability. Power system operators use this type of forecast to allocate resources. Technology like OpenADR can be used to lower the peak demand and flatten the overall shape of the profile, lessening the need for extra generation capacity.

All brawn, no brain

Although the size and complexity of the electric power system, as well as the total demand for electricity, have grown considerably over the past century, the operating principles of the power grid have remained more or less the same. Fundamentally, the power grid is a system that connects power generators with power users. Each individual power plant generates a huge amount of electricity, which must be distributed to millions of consumers. System designers usually locate power plants close to transmission lines that route electricity to the end-users’ distribution grid. To efficiently send electricity over long distances, voltage is increased to anywhere from 135 kilovolts to as high as 500 kilovolts, thousands of times the voltage of a residential electrical socket. These high-voltage transmission lines run across the country, transporting power from power plants located in remote locations to the edge of urban areas. Once within city limits, the high transmission voltage is reduced to a safe distribution level that can vary from 12 kilovolts to as low as 110 volts for ordinary usage in our homes. This complex chain of transmission and distribution grid-lines, along with the interposed controlling and protective equipment, is what connects our homes and offices to power plants.

Supervisory Control and Data Acquisition (SCADA) is the name of a large network of computers and automation devices that control and coordinate this supply chain. Using SCADA, human operators, like those working at the California Independent System Operator (CAISO), are continually watching the load on the grid and directing power plants to adjust their generation to meet power demand.

We are habitual creatures, and our use of electricity is no exception. When typical daily electricity consumption for millions of consumers is averaged, system operators can clearly see the growing cumulative demand on their computer displays. The load starts to build as people begin their day, eventually reaching a peak value that continues to hold from afternoon until late evening when it gradually begins to drop. This stable trend is called the “grid load profile,” and it tends to repeat predictably, allowing electric system planners to forecast how consumption patterns change on a day-by-day basis. It is here that human system operators play an essential role, balancing the amount of generated power supplied to the grid with the demand placed on it without overloading the system. The system operators are well trained to coordinate with power plant operators to continually accommodate the load profile with available power. However, on abnormally hot summer days or cold winter days, more people turn on air conditioning devices or heaters, respectively. This leads to a greater demand for power that can put a lot of strain on the power grid and, in extreme cases (like at peak demand hours) system operators are more exposed to the risk of electric faults. Lacking the ability to communicate real-time demand information to consumers, SCADA is unable to reduce the overall grid load profile from dangerous levels in such extreme situations, sometimes leading to cascading outages like the 2003 blackout.

It is worth noting that our electric grid is highly reliable (power is available up to 99.97 percent of the time in a typical year) despite its vulnerability to extremes in peak demand. This is due to a large reserve of excess capacity that is built into the system. US Department of Energy data shows that over the course of the year, our power plants utilize, on average, only slightly more than 50 percent of their capacity. Some of the remaining capacity only comes online during peak demand hours of the day while the rest is in place in case of forced outages and other contingencies. This means our assets are often sitting idle and simply waiting to be used for the most demanding hours. While this reserve capacity provides an element of reliability, its brute-force approach is economically inefficient and, as previous catastrophes indicate, technologically insufficient in responding to peak load extremes. Given that the US Energy Information Administration estimates that in the next 20 years, the country will consume 40 percent more electricity than in 2005, an upgraded and efficient power grid is becoming increasingly urgent.

One obvious and necessary upgrade to the current power grid is construction of newer, more efficient transmission lines to accommodate the projected increase in demand. However, such infrastructure is expensive to install: for a typical double circuit 220 kilovolt transmission line, the final installed cost can be as high as one million dollars per mile. A more effective, complementary strategy is to embed modern information management systems, like OpenADR and other smart grid technologies, in the architecture of the power system. These systems can sense and analyze performance characteristics of the power grid in real-time and quickly communicate the necessary actions to the devices that control its operation using the Internet. By making power use far more efficient from the demand side, smart grid systems can not only make the electric grid more reliable, but will also reduce the need for extra generation capacity and expensive new transmission lines.

Unfortunately there has been little investment in enabling the power grid with technologies that can intelligently direct the flow of electricity. In the current system, consumers have no knowledge of what is happening on the local grid: when it comes to using electricity, they simply plug in or flip a switch, with no regard for the time of day, current weather conditions, or total local demand for electricity. Furthermore, a lack of adequate investment in construction of new transmission lines has exacerbated the problem: over the past ten years, fewer than 700 miles of new interstate transmission were built. In order to achieve a more stable and uniform load profile, there is a clear need for new information technologies to improve the coordination between power supply and demand.

Enter the smart grid

The smart grid aims to remedy major demand problems by keeping electricity consumers informed of their usage. Power plants send electricity from remote locations out to a large network of users. Under the smart grid architecture, power will be routed through this network in the optimal way and consumed as efficiently as possible. LBL, in cooperation with utility providers, has conducted a series of technology demonstrations by installing their demand response software on utilities’ supply automation servers (part of SCADA). These demonstrations have shown that when a communication link exists between these servers and end-use control systems like power management hardware at participating residential and commercial buildings, consumers will choose—either by pre-programming automated settings or direct user control—to cut down on unnecessary consumption rather than incur higher prices or suffer blackouts.

As a results of such studies, “demand response” has become the mantra of the smart grid movement. Demand response is defined as the ability of the power delivery system to continually measure its performance and respond to load changes by communicating the appropriate control actions to power management devices installed at consumers’ homes and businesses. Demand-response technology works by generating an online signal that is communicated to consumers to inform them of the conditions on the local grid and the corresponding effect on their electrical bill. Consumers can then choose to alter their behavior in response to this information. Under this scheme, the two ends of the supply and demand chain establish an interactive link via the Internet and are consequently able to behave symbiotically: utilities supply people with a reliable, efficient, and inexpensive source of power while simultaneously directing them to adjust their habitual power consumption to spare the local grid from unnecessary load.

With the knowledge of load and pricing conditions on the local grid, a power-management device installed in homes, commercial buildings, or industrial facilities is able to shave off unnecessary consumption (i.e. by turning off lamps, resetting thermostats or stopping some equipment) to help the local grid reduce its peak load. Alternatively, consumers can pre-program their load-management devices to run energy -intensive appliances like washing machines during off-peak hours, when the price of electricity will be lower. Consequently, the total load profile seen by the grid is smaller and more evenly spread throughout the day.

Future changes to power pricing schemes should encourage the adoption of such smart-grid technologies. As a recent report by LBL indicates, electricity markets in California are making a transition toward dynamic pricing in response to the fact that the price of electricity and the stress on the electric grid are much higher during specific periods like hot summer days. These changes could significantly affect the cost of electricity for many facilities. In reducing usage during these peak periods, the smart grid can lower costs for customers, providing a strong incentive for its adoption.

A schematic of the OpenADR communication architecture and power transmission. Using the Internet, supply and demand information and demand response actions are communicated between a utility’s demand response automation server (DRAS) running OpenADR software and end-users’ power management devices. An intermediary intelligent relay box (CLIR box) facilitates this communication. Based on real time instructions delivered by this system, power is routed through a series of transmission lines and substations to be used by consumers in their homes or commercial buildings.

Putting the pieces in place

Over the past six years, with funding from the California Energy Commission, the Department of Energy, and major public utility companies, LBL’s PIER Demand Response Research Center has played a lead role in creating both the concepts and technology that will enable implementation of the smart grid. During this period, the center’s focus has been creating communication architectures, developing system specifications, and establishing operational and planning leadership for a consortium of academic and industry partners involved in the smart grid movement.

Coordination of millions of end-use electrical customers with utility providers is no easy endeavor; in reality, it is nothing short of sheer complexity. To tackle this task, Piette and her colleagues decided to develop the OpenADR specification, which serves as a facilitator for Internet communication of demand response instructions within the smart grid. Think of OpenADR software as an intelligent online middleman between supply and demand in the electric grid.

In technical terms, OpenADR is a proposed standard for exchanging information between a demand response automation server (DRAS) and a client that wants to receive and act on this information at the point of end-use. As such, OpenADR works in an Internet-based communication architecture and is intended to specify the various functions that must exist in a DRAS. These functions calculate the price of electricity based on power grid conditions for a particular time of day and historical information from the utilities and subsequently determine the effect on customers’ electrical bills. The DRAS then gives the customers’ power management devices a set of options to manage electrical loads. OpenADR also permits devices and software from external third parties such as utilities, facility managers, and hardware and software manufacturers to interface with and use the functions on the DRAS in order to customize demand response programs for specific applications.

Price signals in data models embedded in OpenADR are communicated to the DRAS via hardware like Ethernet cables or using wireless communication technology like Wi-Fi or ZigBee, a low-power wireless technology designed for monitoring and device control. The DRAS then communicates with the end-use load controllers either through the Internet or over electric power lines, instructing them to change settings on air conditioners, refrigerators, and other appliances and equipment. All of these communications with a home or commercial building are automatic and pre-programmed.

Field trials conducted at LBL have already shown that OpenADR is able to help utilities deal with enhanced peak loads on cold winter mornings or hot summer days. In the summer of 2004, OpenADR was used to manage the electrical demand of several commercial buildings in California. Information provided by the local utility, Pacific Gas & Electric (PG&E), signaled the approach of peak load on the grid to power management devices in the buildings. These devices then automatically reset the building thermostats over a six-hour period and consequently shed approximately 100 kilowatts of load. This amounted to a 20 percent decrease in peak load, twice the expected result. Much larger tests, with over one hundred large commercial buildings in California, have since been conducted and have demonstrated even larger load sheds. Due to these successes at LBL, the National Institute of Standards and Technology (NIST) has already adopted OpenADR as a key standard in the national smart grid initiative.

The adoption of OpenADR technology in California is critical to the successful implementation of the smart grid on a national scale. According to a report authored by Piette in 2009, California registered about 53,000 megawatts of peak electric demand on the hottest summer day of that year. Large buildings account for about 5,000 to 7,000 megawatts or five to ten percent of this peak load, while small commercial buildings account for 10,000 to 12,000 megawatts or 20 to 25 percent of the load. Much of the peak load is due to easily regulated processes like air conditioning, lighting, appliances, and other facility uses. This high concentration of demand in commercial buildings coupled with prior successful demonstrations using local utility companies makes California the ideal arena for a large-scale proof-of-concept for OpenADR. In playing this role, the state would not only gain the practical benefits of effective demand response technology but would also put itself at the forefront of the smart grid movement.

Smart grid = green grid

A bright bustling evening in Times Square (left) and the darkness of the 2003 blackout (right).

Beyond a more reliable power grid and reduced need for excess generation and transmission capacity, demand response technology has the additional benefit of easing the implementation of renewable energy sources like solar cells and wind turbine generators. Intermittency presents an imposing barrier to the adoption of these technologies because the current power grid is designed for dispatchable sources of electricity that allow energy to be delivered within preselected hours based on advanced predictions. The output of solar and wind energy can vary drastically on short time scales throughout the day and thus, as no cost-effective means of energy storage currently exists, these technologies are difficult to dispatch in a reliable way. The valleys and peaks in power output due to shifting wind speed, cloud cover, and other environmental effects effectively prevent the integration of renewable energy technologies in the current grid.

In the smart grid, however, the distribution of load on the grid can be adjusted in a way that maximizes the usage and transmission of renewable energy sources. Demand response software, like OpenADR, can establish a real-time interface between consumer demand and output from solar and wind power plants, match them in an efficient way, and dispatch the energy they produce to the consumers that need it most. With the current push to slow global climate change and improve energy security by developing renewable sources of electricity, this added benefit of smart grid technology should be a strong motivator for policy makers in shaping our future energy landscape. The broad reach of OpenADR is not lost on Piette: “Our hopes for the future of this technology is that it facilitates a low-cost, low-carbon future to help provide demand side load flexibility, reduce peak loads, and allow more intermittent renewables on the electric grid.”

Looking forward

Despite significant progress in proving the efficacy of OpenADR in small-scale demonstrations, several key challenges need to be overcome to facilitate its integration into the electric grid at large. To begin with, OpenADR needs to be tested across a wider spectrum of building types, both commercial and residential, to ensure that it is versatile enough to adapt to the myriad properties and functions of buildings today and the disparate performance behaviors (in terms of energy use and efficiency) of buildings of different sizes. Equally important is the development of infrastructure that can respond to signaling from OpenADR and operate at high enough speeds along the entire signaling chain to take advantage of OpenADR’s price communication capabilities. While the Internet provides a high-speed interface, more homes and buildings need to be equipped with smart power management devices and the DRAS needs to be updated to run OpenADR software. Finally, price modeling by the electric utilities and the OpenADR software developers at LBL still needs to be refined and expanded to reflect real world data as accurately as possible across a broad array of scenarios. To begin addressing these needs, OpenADR is currently being implemented in a variety of programs in California, the Pacific Northwest, and Canada, and it is in development for a number of other demand response programs around the United States and abroad. The researchers at the PIER Demand Response Research Center are also constantly seeking new partnerships with utilities and updating their models in response to new experimental data.

Once the challenges facing OpenADR have been addressed and the combination of smart communication and smart devices has found its proper place across the power sector, the impact of this technology could be astronomical in scale. As much as the Internet has revolutionized communication, an Internet-enabled smart power grid will completely redefine the way people use electricity. This fusion of two immensely powerful systems will usher in an era where blackouts are a thing of the past, electric power is more affordable, renewable resources play a central role, and, sometimes, the lights will go out in order to keep the power on.

The following are inset boxes with supplemental background information. They are also found in the print edition of the article.

CLIR box

Credit: LBL

To facilitate machine-to-machine communication of demand response instructions, researchers at LBL have developed an intelligent relay box known as Client & Logic with Integrated Relay (CLIR). CLIR is a secure and self-configuring communication relay that sits between the demand response automation server running OpenADR software and the intelligent load controller in the end-user’s building. It operates via Internet encryption protocols that are used for secure data transactions like online shopping or banking. This ensures that hackers do not jeopardize information exchange between a utility’s servers and the computerized controller at the site of end-use.

Once the CLIR box is powered on after installation, its status is visible via LCD display. Internet connectivity, time since last successful communication with the server, event modes, and other relevant data are shown on the display. An integrated keypad allows installers to set all relevant configurations without the use of a laptop or remote terminal. During an automated demand response event, the intelligent load controllers at the customers’ building receive a signal from CLIR that causes the facility to automatically enter pre-configured low energy modes by adjusting air conditioning, lighting, and other energy intensive processes.

Through several major field tests with the participation of large commercial buildings, LBL has demonstrated that CLIR is capable of communicating OpenADR instructions to a load controller and reducing a building’s total power demand. Relay boxes like CLIR will play an important role in the updated infrastructure of the smart grid.