TMCnet News
The Basic - What? Why? Who? How? [Chemical Engineering Progress](Chemical Engineering Progress Via Acquire Media NewsEdge) SMART GRID The electric-power grid is being transformed into a smarter network - one that could bring significant benefits to the chemical process industries. Learn what a smart grid is and how you can participate. The electric power industry is poised for an unprecedented change in the way it operates. Motivated by forecasted changes in its generation profile, including more-extensive generation by renewable sources, the power industry is looking to create a more interactive relationship with its consumers. Since the industrial community consumes about one-third of all generated power, the chemical process industries (CPI) are expected to play a major role in the transition. This article provides an overview of the smart grid: what it is, why it is being created, who stands to benefit, and how to participate in it. Examples are given to illustrate how the smart grid operates, as well as to explain how the smart grid differs from the existing grid. This article serves as a foundation for the subsequent articles in the energy supplement - introducing key players, terms, and concepts that will be discussed in more detail in the articles to follow. What is a smart grid? The notion of a smart grid is an evolving concept, with multiple, community-specific meanings. Many definitions of the term exist, but the consensus definition in Wikipedia seems to capture the intentions of most stakeholders best: "A smart grid is a modernized electrical grid that uses analog or digital information and communications technology to gather and act on information, such as information about the behaviors of suppliers and consumers, in an automated fashion, to improve the efficiency, reliability, economics, and sustainability of the production and distribution of electricity." While the definition calls for a modernized grid, it does not call for a replacement of existing transmission lines. Rather, a smart grid adds new technology and operating principles to the existing power network (Figure 1 ). Figure 2 illustrates this as the addition of two new layers, a cyber layer and a market layer, both of which overlay the existing power network (physical layer). The cyber layer consists of new hardware and software for the exchange of information about the current and future conditions of the existing physical layer among the various grid participants. The market layer uses data from the cyber layer to establish economic incentives and operating policy agreements among the participants. The intelligence created by these new layers is the core concept behind the smart part of the smart-grid definition. Through a greater awareness of grid conditions and the appropriate dissemination of this information, more intelligent decision-making processes can be implemented, i.e., having better information will lead to better decisions and more-efficient grid operation. This intelligence can be illustrated by considering a grid-connected wind farm. A weather forecast is predicting high wind speeds, and this collection of wind turbines is expected to produce large amounts of energy over a 6-hr period. The market layer will encourage many of the conventional generators in the area to reduce their power output by reducing the compensation rates during that period. However, if the forecast is incorrect and the actual wind speeds are much higher than expected - to the point that many of the wind turbines must shut down due to safety concerns - the grid will experience a sudden imbalance between power generation and consumption. Such an imbalance could lead to a local voltage collapse and in the worst case a system-wide blackout. While all consumers will be negatively impacted by such an outage, the CPI should be acutely aware of the process safety concerns associated with a loss of electric power. An obvious way to address this scenario is to have the conventional generators quickly increase their output. A less-obvious option is to ask consumers to reduce their consumption. Clearly, the most basic aspect of managing such an event is the existence of a cyber layer that allows the grid operator to receive information from the wind farm and then send a command signal to the appropriate generators and consumers. However, equally important is the market layer, which will have put into place a set of agreements with particular generators and/or consumers to be on call to respond to such an event. These agreements stipulate the compensation that each will receive for providing the service. Consider again the wind farm, but now assume that the average wind speed forecast is correct and that all of the wind turbines remain operational. Wind gusts, however, present a different challenge, since the forecast cannot predict minute-to-minute changes in wind speed. Because the power system must remain in balance, some entity within the network will need to respond. The first step uses the market layer to put in place a generator or consumer that will provide the necessary response in return for a predetermined compensation. Then the cyber layer collects data on the minute-to-minute fluctuations from the wind farm and communicates that information to the responder. A well-known example of smart-grid participation by an industrial company is that of an Alcoa aluminum-smelting plant (2). The aluminum-smelting process is seen by many in the industrial smart-grid community as an easy target because it uses large amounts of electric power (electric power accounts for 30-40% of aluminum production costs). Alcoa's initial participation in the smart grid (Figure 3, center) was akin to the first wind-farm example, in that the company obtained a contract to be on call to reduce a portion of its power consumption in the event of a grid emergency. In recent years, the demand for aluminum has fallen and Alcoa decided to set production levels below the maximum throughput at this particular plant - creating an opportunity for smart-grid participation similar to the second wind-farm example. That is, the company obtained a contract that allowed the grid operator to specify how much power Alcoa would consume in order to maintain the minute-to-minute balance between grid generation and consumption (Figure 3, right). The dumb grid: Not so dumb To understand the need for the smart grid, it is helpful to appreciate the operation of the existing electric-power grid. Consider what happens every time you switch on a light. For electricity to flow and allow this light to go on, a power plant must respond appropriately to meet this change in power demand. How does a coal plant know that it should bum just a little bit more coal? What mechanism is used to communicate this information? More importantly, how does the system respond to the 5 p.m. surge in power demand that occurs when millions of flat-screen televisions and air-conditioning units are turned on, almost simultaneously? What will happen when plug-in electric vehicles are added to this demand surge? This section describes the classic feedback mechanisms grid operators use to respond to somewhat predictable, but ultimately unknown, consumer demands. At the most basic level, virtually all power systems communicate using an amazingly simple, century-old technology - the power transmission lines, which also serve as the communication lines. The frequency of alternatingcurrent electric power in the U.S. is 60 Hz, but that 60 Hz is more of a target than a rule. In fact, detection of small changes in the AC frequency of the grid is how generators know to increase or decrease power output. Specifically, when consumers draw more power, the frequency of the entire system decreases by a small amount. In response, generators increase power output until the system frequency returns to the 60-Hz target. This type of frequency-based feedback is usually referred to as frequency control or automatic generation control (AGC). Unfortunately, relying exclusively on frequency control results in somewhat arbitrary power outputs from the various generators, as well as operation of the network outside of its safety limits. For example, if a transmission line attempts to deliver too much power, the slightly resistive wires may overheat. To avoid these problems, only a few generators actually operate under frequency control. The remaining generators are given power-output setpoints, which are determined by an optimal power flow (OPF) calculation. OPF is an optimization problem that minimizes the operating costs of the overall system based on the operating cost curve of each generator, the generator and transmission line operating limits, a model of the transmission network, and, most importantly, measurements and/or forecasts of consumer demand. A shortcoming of the OPF calculation stems from the fact that there is a non-zero lower limit to the power output of an online generator. While a generator can be taken offline, power output within the gap (between zero and the non-zero lower limit) is impossible. If one were to include this gap in the decision space of the OPF optimization problem, this discontinuity would lead to a multifold increase in computational effort. Because the OPF calculations must be completed in a matter of minutes to be sufficiently responsive to changes in consumer demand, the assumption of a fixed set of online generators must be employed. To address larger changes in consumer demand over larger time scales, a second level of optimization, known as the unit commitment problem (UCP), is performed. The UCP determines which generating units will commit to being online during each hour of the upcoming day. The UCP's ability to incorporate the discontinuities associated with generator startup and shutdown stems from the fact that it need not be solved as frequently as the OPF calculation. A critical element of the UCP calculation is a fairly sophisticated forecast of consumer demand over the ensuing 24 hours. These forecasts are not perfect, so the OPF software makes corrections closer to the time of implementation. Frequency control is then used to make the final corrections required to achieve a balance between consumption and generation. Interestingly, this power system feedback structure is similar to the time-scale decomposition structure commonly used to control a chemical plant (Figure 4). The UCP is also responsible for calculating contingency plans in the event of a loss of any generator or transmission line - denoted as the N-l security constraint. If one generator trips offline (for instance, due to a transmission line fault), the frequency of the grid will suddenly drop, which will cause other generators to trip offline. This cascade of generators tripping offline is a typical precursor to a widespread blackout. Thus, while determining which units will be online, the UCP must consider whether this collection of units can withstand the loss of any one of its members. In many cases, this security-constrained UCP (SCUCP) will require more generators to be online than would be required if only economics were considered. In summary, the "dumb" grid is actually a collection of rather sophisticated decision-making algorithms aimed at achieving a balance between economic efficiency and network reliability. Deregulation changes the players but not the game Historically, all of the feedback mechanisms (frequency control, OPF, and UCP) were performed by one company (i.e., the electric power utility), which owned and operated all of the major generators and transmission lines and interfaced with consumers. However, over the last few decades, the electric power industry has been moving toward deregulated operation. This new deregulated system consists of individual generation units, usually owned by new generation companies (GENCOs), and new consumerinterface companies, which are typically referred to as load-serving entities (LSEs). Central to this deregulated configuration are the independent system operators (ISOs), which manage interactions among the various participants and are responsible for implementing the feedback mechanisms required to balance generation and consumption. Fundamental to ISO operation is the creation of various electric-energy markets, which are used to determine the appropriate compensation from the LSEs to the GENCOs. Day-ahead market. The most visible of the new markets is the day-ahead market. This market is an auction held by the ISO that sets the compensation rate (in $/MWh) for each hour of the next day (Figure 5). In contrast to a conventional commodity market, only the GENCOs submit bids, since the LSEs are constrained to meet the consumer demand forecasted for the next day. Another difference is that the auction results must satisfy the operational limits of the power network, i.e., the ISO must perform a UCP-type calculation in parallel with the auction result calculations. This market arrangement brings a degree of economic efficiency to the grid. If a GENCO bids higher than its oper- ating costs, it runs the risk of that bid not being accepted by the auction and it loses that revenue opportunity. The impact of a lost opportunity is compounded by the fact that these markets typically pay all accepted bids at a rate equal to the highest accepted bid. While such a scheme may seem counterintuitive to the goal of lowering costs to the LSEs, the alternative of "paying as bid" actually gives generators a greater incentive to inflate their bids (5, 6). Real-time market. This market attempts to make price corrections based on updated forecasts of demand, typically an hour ahead. This market can be thought of as analogous to the OPF module. The much shorter lead times make this market much more volatile than the day-ahead market. During extreme events, such as a major generator unexpectedly tripping offline during a heat wave, the price of electricity in the real-time market can spike to upwards of 200 times the typical price. Regulation market. The regulation market serves a role similar to frequency control, but with elements of the OPF calculation. If a generator's bid is accepted by the regulation market, then during the period of the contract, the generator is put on automatic generation control (AGC), Le., the generator agrees to follow power-output command signals sent by the ISO. The contract dictates the scope of allowable command signals (the maximum and minimum amounts of power that the ISO is allowed to request). Once the ISO sends one of these commands, the generator must respond within a predefined period of time. Contracts requiring faster response provide greater revenue to the generator. Emergency capacity market. This market focuses on ensuring that sufficient additional generation capacity will be available in the event of an outage. If a generator's bid is accepted in the emergency capacity market, the generator must guarantee that it will be able to quickly increase power output if called upon. For some generators, this requires the plant to be online during the period of the emergency capacity contract. These generators are typically referred to as spinning reserves. Generators that are offline but can be started quickly are referred to as nonspinning reserves. For both the regulation and the emergency capacity markets, generators with accepted bids receive the contracted payment even if they are never called upon to provide the service. However, if the generator is called upon but does not provide the agreed-upon service, the generator is subject to stiff penalties. The term ancillary services is often used to denote both types of service - regulation and emergency capacity. The physics of each type of power generation make it more or less attractive for each of the energy markets. The low operating costs of nuclear and coal-fired power plants allow for both to submit low bids to the day-ahead market. As a result, both are usually included in every hour of the day-ahead market - the so-called baseload plants. These plants also incur the largest penalty for a shutdown/startup cycle. However, during times of very low demand, it is possible to shut down some of the smaller coal plants, especially if it is for a short period of time and the startup is a warm start. A combined-cycle gas turbine is more expensive to operate (due to the more-expensive fuel), but it is a bit easier to start up. Thus, these plants are rather active in the dayahead market - they are cycled on during periods of high demand and cycled off when demand is low. Single-cycle gas turbines are among the most expensive to operate due to their low efficiency, but they can be cycled on and off rather quickly, in many cases in less than 15 min. These plants, therefore, play a large role in the emergency capacity market (as nonspinning reserve) as well as in the real-time market on days of high demand. Coal-fired and combined-cycle gas turbine plants are the best options for the real-time and regulation markets, though neither was designed to take this burden, especially the older coal plants. If it were not for the large economic incentives of participating in these markets, the operators of these plants would likely avoid offering their services. In places where hydroelectric plants are available, they can (and do) serve markets of all time scales, but baseload operation is likely avoided due to a limit on the amount of available fuel. Renewable-based power plants are distinguished from other power generators by their lack of dispatch capability (7) - that is, unlike conventional power plants, a wind farm cannot be put into service to meet an increase in consumer demand. While reasonably accurate forecasts of power from renewable power plants can be made, the ISO has virtually no control over how much power they will produce. One may even designate renewable plants as consumers of dispatch capability, since they require additional dispatch capability from the rest of the system. The volatility of the markets is expected to increase as more renewable power is brought online. Consumer demand Is asked to respond A large increase in renewable generation is predicted to leave the grid with insufficient dispatch capability (8). Building more combinedand single-cycle gas turbine plants or investing in massive energy storage facilities (9) would likely provide sufficient dispatch capability. However, these options require large capital investments and therefore are not considered viable options. Another option is to use the economic incentives created by deregulation to unlock the dispatch capabilities of consumers. These economic incentives for consumers that are willing and able to provide these demand-response (DR) services are expected to be significant. The three types of demand response are economic response, contingency response, and regulation response (Figure 6). Economic response. This type of demand response corresponds to the day-ahead and real-time markets. Consumers participating in this type of demand response will be encouraged to change their electricity usage at particular times of the day, week, or year based on changes in the price of electricity. As a result of deregulation of the power industry, many large electricity consumers already participate in one or both of these markets. Many industrial facilities have an energy manager who coordinates the purchase of electricity. However, the actions of an energy manager are often disconnected from the process side of the plant. Increasing collaboration between energy managers and process engineers is a critical first step toward unlocking the DR capabilities of industrial consumers. Contingency response. This type of demand response corresponds to the emergency capacity market. In the event of a generator outage, consumers providing contingency response quickly reduce their power consumption to prevent a collapse of the grid. A particularly attractive feature of contingency response is that an industrial facility can receive revenue with virtually no change in plant operation, except for the few times a year that load shedding is required. Regulation response. Regulation response occurs in the regulation market. Consumer participation is similar to that of a generator, in that the grid operator continually sends signals to the participant indicating the desired level of power consumption. These relatively small but ever-changing requests create the minute-to-minute balance between generation and consumption that is required to maintain grid integrity. Benefits of DR to the CPI For those in the CPI with sufficient operational flexibility to implement demand response, the benefits can be substantial. The benefit of providing economic DR is a reduction in operating costs by aligning energy consumption with the price of electricity. Those that can provide regulation and contingency response receive a new revenue stream. For Alcoa, the revenue stream provided by DR contributed to the company's decision to not shut down that particular site. However, it should be emphasized that participation in a DR program is not a small undertaking, and sufficient due diligence is required before taking any serious steps. Reference 2 highlights the considerations and trade-offs that Alcoa made before entering the DR program. In addition to direct benefits, DR offers several indirect benefits to all consumers of electric power. For example, the GENCO community suggests that greater participation in economic DR will help defer the massive capital investment required to build new power-generation plants. This argument stems from the fact that the power-generation fleet must have enough capacity to meet the demand required during the maximum hour of the maximum day of each year. If economic DR can be used to lower the peak during these critical time periods, then the total capacity of the generation fleet can be lower and fewer generation plants will remain idle for large portions of the year. Of course, one may ask why a consumer should be concerned with the capital investments of a GENCO; the answer is that costs incurred by the GENCOs are passed on as higher costs to the LSEs and eventually higher costs to consumers. Greater participation in regulation and contingency DR programs is expected to lead to greater economic and environmental efficiency of the overall electric-power system. Shifting these ancillary service responsibilities to consumers gives generation plants more operational freedom, and the ISO day-ahead market is more likely to find a gridoperating condition that requires fewer units to be online and fewer units to be operated in a low-efficiency state, both of which ultimately result in lower operating costs and reduced emissions. The emergency response time of an industrial consumer can be much faster than that of a generator. Recall the Alcoa example. Since the smelting process has virtually no moving parts, its electrical response time is likely much faster than the inertia-limited response of a coal-fired power plant. Thus, one can argue that greater participation in contingency DR programs will create a more reliable grid with fewer outages - which clearly benefits all stakeholders. Additional Resources Farhangi, H., "The Path of the Smart Grid," IEEE Power and Energy Magazine. 8(1), pp. 18-28 (Jan-Feb 2010). Rahimi, F" and A. Ipakchi, "Demand Response as a Market Resource under the Smart Grid Paradigm," IEEE Transactions on Smart Grid. 1(1), pp. 82-88 (June 2010). Walawalkar, R., et aL, "Evolution and Current Status of Demand Response (DR) in Electricity Markets: Insights from PJM and NYISO," Energy. 35 (4), pp. 1553-1560 (Apr. 2010). Literature Cited 1. U.S. Dept, of Energy, "Benefits of Using Mobile Transformers and Mobile Substations for Rapidly Restoring Electrical Service: A Report to the United States Congress Pursuant to Section 1816 of the Energy Policy Act of 2005," eneigy.gov/oe/downloads/ benefits-using-mobile-transformers-and-mobile-substationsrapidly-restoring-electrical, DOE, Washington, DC (accessed May 27,2014). 2. Todd, R. D" et aL, "Providing Reliability Services through Demand Response: A Preliminary Evaluation of the Demand Response Capabilities of Alcoa Inc.," Office of Electricity Delivery and Energy Reliability Transmission Reliability Program, U.S. Department of Energy, Washington, DC (Jan. 2009). 3. Todd, R. D., "Process Focused Dynamic Demand Response in Organized Markets," presented at the AIChE Workshop on Smart Grid for the Chemical Process Industry, Chicago, IL (Sept. 25-27,2013). 4. Pennsylvania, New Jersey, Maryland Interconnection, "Day-Ahead Historical Data," www.pjm.com/markets-and-operations/ energy/real-time/historical-bid-data.aspx, PJM, Audubon, PA (accessed Jun. 25,2013). 5. Kahn, A. E., et aL, "Uniform Pricing or Pay-as-Bid Pricing: A Dilemma for California and Beyond," The Electricity Journal, 14 (6), pp. 70-79 (July 2001). 6. Federico, G., and D. Rahman, "Bidding in an Electricity Pay-as-Bid Auction," Journal of Regulatory Economics, 24 (2), pp. 175-211 (Feb. 2003). 7. Lee, B. S" and D. E. Gushee, "Renewable Power: Not Yet Ready for Prime Time," Chem. Eng. Progress, 105 (4), pp. 22-25 (Apr. 2009). 8. Lindenberg, S., et aL, "20% Wind Energy by 2030: Increasing Wind Energy's Contribution to U.S. Electricity Supply," Office of Energy Efficiency & Renewable Energy, U.S. Dept, of Energy, Washington, DC (July 2008). 9. Chen, H" et aL, "Progress in Electrical Energy Storage System: A Critical Review," Progress in Natural Science, 19 (3), pp. 291-312 (Mar. 2009). Donald J. Chmielewski Illinois Institute of Technology DONALD J. CHMIELEWSKI is an associate professor in the Dept, of Chemical and Biological Engineering at the Illinois Institute of Technology (IIT; Phone: 312-567-3537; Email: [email protected]), where his research focuses on economic-based control system design with applications ranging from chemical processes and hybrid fuel cell vehicles to building HVAC and electric-power networks. A senior member of AlChE, he Is active in the Computing and Systems Technology (CAST) Div., and recently served as the chair of the AlChE Workshop on Smart Grid for the Chemical Process Industry. He is a past chair of the Chicago Local Section and helps to organize the Midwest Regional Conferences, and he has also served as general arrangements chair, conference chair, and programming chair. He received the Excellence in Teaching Award from the Armour College of Engineering at IIT. He received a BS from IIT and an MS from the Univ. of California, Los Angeles (UCLA), both in electrical engineering, and a PhD from UCLA in chemical engineering. (c) 2014 American Institute of Chemical Engineers |