The Electric Transmission Network: A Multi-Region Analysis
Footnotes
[1] Energy Information Administration, “An Exploration of Network Modeling: The Case for NEPOOL,” Issues in Midterm Analysis and Forecasting 1998, DOE/EIA-0607(98) (Washington, DC, July 1998), web site www.eia.doe.gov/oiaf/issues98/modtech.html.
[2] Strictly, out-of-order dispatching refers to operating a generator whose costs of production exceed those of another idle generator.
[3] As with the events of July 6 and July 19, 1999, in the eastern half of the Pennsylvania-Jersey-Maryland (PJM) interconnection. Voltages were severely reduced because of inadequate incentives for generators to produce reactive power. See U.S. Department of Energy, Report of the U.S. Department of Energy’s Power Outage Study Team (Washington, DC, March 2000), pp. 5 and 10.
[4] Buses represent points where major pieces of electrical equipment connect to the grid or where major transmission lines meet. In the model described herein, buses represent any point where power flow equations must be balanced.
[5] The analysis accounts for alternating current indirectly by checking to assure that voltage levels are maintained for the power flow solution. Because of the detailed level of this analysis, it was not possible to include alternating current explicitly due to the computational difficulties that occur when both of the components of alternating current (real and reactive power) are modeled.
[6] For a description of the EMM see Energy Information Administration, The Electricity Market Module of the National Energy Modeling System: Model Documentation Report, DOE/EIA-M068(2000) (Washington, DC, January 2000), web site ftp://ftp.eia.doe.gov/pub/pdf/ model.docs/m068(2000).pdf. For an overview of NEMS see Energy Information Administration, The National Energy Modeling System: An Overview 2000, DOE/EIA-0581(2000) (Washington, DC, March 2000), web site www.eia.doe.gov/oiaf/aeo/overview/index.html.
[7] Energy Information Administration, The Electricity Market Module of the National Energy Modeling System: Model Documentation Report, DOE/EIA-M068(2000) (Washington, DC, January 2000).
[8] The regions used in this analysis are part of the NERC regions that were formed to assure the reliability of the electricity transmission network. One of the NERC regions is the Northeast Power Coordinating Council. NYPP and NEPOOL are subregions of this Council.
[9] “Equivalencing” is a method used to reduce computations when analyzing electrical networks. It involves replacing a network with a simplified representation that has the same electrical properties as the original network.
[10] The distribution network, the system of lower voltage lines (usually less than 69 kilovolts) that distribute power locally, is not included because it imposes computational complexity and offers little additional information on the interregional transmission network.
[11] Federal Energy Regulatory Commission, Form FERC 715, “Annual Transmission Planning and Evaluation Report,” case nepp97s.raw.
[12] Significant Canadian imports are also delivered to both NYPP and ECAR.
[13] For June 15, 2000, ISO New England reported Canadian capacity deliveries of 2,508 megawatts, in order to meet a peak load of 16,675 megawatts. See web site www.iso-ne.com/power_system/morning_report_external.html.
[14] The assumption of constant fuel costs per unit output between minimum and maximum operating levels is an approximation. Operating costs are U-shaped, that is, they are relatively high at the extremes of the operating range and lower in the region between the extreme points.
[15] Industrial customers, however, are frequently forced to adhere to strict reactive power constraints and are charged explicitly for their inability to do so.
[16] Some configurations of generator outputs will meet real power demands but fail to provide enough reactive power to deliver useable power: voltages are either too high or too low. The procedure used here weeds out such seemingly feasible solutions. See the 1998 EIA study for a discussion of the consequences of ignoring the effects of reactive power in trade with Canada.
[17] For example, respondents to Form FERC 715 planned on a peak load for MAAC of 47,687 megawatts. Actual peak load for summer 1997, according to the PJM Independent System Operator, was 49,406 megawatts (July 15, hour ending 5 P.M.), which at the time constituted the all-time high demand in the region. Actual peak load was 44,302 megawatts in 1996 and 48,524 megawatts in 1995. See web site www.pjm.com.
[18] See web site www.powerworld.com for more information about the PowerWorld® software.
[19] All data are available on request from Tom Leckey (202-586-9413, thomas.leckey@eia.doe.gov).
[20] Heat rates are the quantities of energy, usually expressed in British thermal units (Btu), needed to produce 1 kilowatthour of electricity. It takes about 3 Btu of energy input to produce 1 Btu of electricity from a typical baseload fossil-fuel unit. In this analysis the heat rate was multiplied by the cost of the fuel to approximate the variable cost of producing electricity. Other costs that contribute to the total generating cost include the O&M costs, which usually are small relative to the fuel costs and were omitted from this analysis.
[21] The operating cost of a generator is calculated using the general formula (a + bg + cg2 + dg3) × unit fuel cost. The parameter a is a constant, and b represents a fuel use coefficient. In the formula, c and d are the curvature parameters and are assumed to be zero. The quantity in parentheses is the amount of fuel required to generate g kilowatthours of electricity. The product is the fuel cost for producing g kilowatthours of electricity. Capital costs and variable O&M costs are not included in the cost calculation. Capital costs were excluded because they are not expected to be included when generation services are bid into competitive markets. The variable O&M costs were excluded because they are small and it is difficult to obtain data for individual generators. Small fossil units may fall below the reporting threshold of Form FERC 1 and Form EIA-412. All costs are reported in nominal dollars. Because of these simplifying assumptions, especially with regard to the c and d coefficients, costs reported here are, presumably, low estimates.
[22] Ideally, it is desirable to have costs for every generator in the system.
[23] Form EIA-867 has recently been redesignated as Form EIA-860B.
[24] The Millstone units were shut down by the U.S. Nuclear Regulatory Commission in 1996 because of design configuration issues and safety concerns. Units 2 and 3 returned to service in 1998.
[25] The 20-percent reduction results in loads of 38,169 megawatts in MAAC and 17,089 megawatts in NEPOOL. Load duration data reported by the two independent system operators indicate that median (4,380th greatest) hourly load for 1999 was 29,319 megawatts in MAAC and 13,532 megawatts in NEPOOL.
[26] Nationally, the average monthly demand in July and August 1997 was 16 percent higher than the average demand in the other months. Energy Information Administration, Monthly Energy Review, DOE/EIA-0035(1999/12) (Washington, DC, December 1999), Table 7.1.
[27] The engineering solution was validated by Dr. Tom Overbye, Professor of Electrical Engineering, University of Illinois. A few lines in NYPP exceeded their limits, a result attributed to several small generators that were reported to the FERC as being out of service.
[28] Marginal costs are computed using the average of the marginal costs at the buses in the transmission network. Detailed data are available upon request from Tom Leckey (202-586-9413, thomas.leckey@eia.doe.gov).
[29] Some of the lower voltage buses are excluded from the contour.
[30] Marginal cost is reduced $0.05 per megawatthour, too small an increment to show as a reduction in Table 4.
[31] Of the four NERC regions at issue, only NEPOOL’s marginal cost remains above that of the super region, the hypothetical aggregation of all four regions, which is $19.4 per megawatthour.
[32] The hourly system cost is a measure of all the operating costs of all the generators for a single hour.
[33] System cost in the four NERC regions declines by 0.1 percent.
[34] Exports increase even more at the peak, to 936 megawatts. NYPP does not, however, become a net exporter under any of the trade scenarios. There are 800 megawatts of power imported from Hydro Quebec, specified as an exogenous input.
[35] Transactions with Canada are included at a fixed level based on Form FERC 715. In these cases, Canadian generators have no costs assigned and are not dispatched. They play no role in the optimal power flow solution.
[36] The first 43,000 megawatts of supply in MAAC are available at 33 mills per kilowatthour or less; the next 1,000 megawatts of load raises the cost to nearly 43 mills per kilowatthour.
[37] One of these, Norwalk Harbor 138, is reported as “open” on Form FERC 715 and is not available to the model. Interestingly, when this line is closed, the model indicates that low-cost NEPOOL generators in Southern Connecticut are able to supply NYPP with nearly 250 additional megawatts, thereby reducing the marginal cost in NYPP slightly, increasing the marginal cost in NEPOOL by 4 percent, and increasing the marginal cost over the entire super region from $19.4 to $19.7 per megawatthour. Source: Office of Integrated Analysis and Forecasting, PowerWorld® model run FREE80W8NPTIES.D061500.
[38] Pleasant Valley-Long Mountain 398.
[39] Much of this power comes from the nuclear units James Fitzpatrick and Nine Mile Point in western New York.
[40] The New York ISO reports the same available transmission capability for that interface. See web site http://mis.nyiso.com/public/htm/atc_ttc. NEPOOL ISO reports slightly more capacity.
[41] Energy Information Administration, Annual Energy Outlook 2000, DOE/EIA-0383(2000) (Washington, DC, December 1999), AEO2000 National Energy Modeling System run AEO2K.D100199A (reference case).
[42] The EMM allows for “economy trades” between NEPOOL and Canada, with opportunities for trade arising from cost differences, but the bulk of the imports in EMM are fixed “contract” trades.
[43] Imports in the shoulder demand super region case were 995 megawatts.
[44] The competitive pricing algorithm sums four components: reliability, tax, transmission and distribution, and energy.
[45] Thus simulating PowerWorld®, as NEPOOL turns to regional sources of generation.
[46] See web site www.iso-ne.com.
[47] Another 700 megawatts comes from New Brunswick through Maine.
[48] ISO New England, Inc., Monthly Market Report (May 1999), p. 20, Figure 17, web site www.iso-ne.com.
[49] See also web site http://mis.nyiso.com/public/pdf/atc_ttc/, where the New York ISO reports transmission capacity of 1,600 megawatts.
[50] See web site www.iso-ne.com/economic_and_load_forecasting/monthly_1999.txt.
[51] Estimated as total monthly net interchange less estimated Canadian monthly imports of 1.8 gigawatthours × 730. This yields an estimate for Canadian imports of 15,811 gigawatthours, basically consistent with the NEMS estimate.
[52] ISO New England, Inc., Monthly Market Report (February 2000), p. 6, Figure 3, web site www.iso-ne.com.
[53] Another possibility is that EMM yields a price for NEPOOL that is roughly consistent with power flow models but underestimates the NYPP price. PowerWorld® model runs that raised NYPP exports to 1,377 megawatts (the hourly equivalent of 12,060 gigawatthours of imports in the AEO2000 reference case) produced negligible marginal cost reductions in NEPOOL but increased the marginal cost in NYPP by nearly 6 percent. Source: Office of Integrated Analysis and Forecasting, PowerWorld® model run REF80-HINYPPIMP.D061500.
[54] Losses are reflected as increased generation.