Comment On Article About The Author More Articles By This Author

Comment On Article About The Author More Articles By This Author

Comment On Article About The Author More Articles By This Author

Several electricity markets in the U.S. are implementing nodal pricing of transmission losses to more accurately assess the cost differences among various generation commitment and dispatch alternatives. In some Northeastern markets, particularly the systems operated by the New York ISO (NYISO) and the ISO New England (ISONE), nodal loss pricing has already been implemented. Since the marginal price of electricity at a location is actually an aggregate of three cost components (energy, congestion and losses), it is critical for market participants to understand how loss charges impact the market clearing prices. The nodal price component due to losses can be a significant portion of the aggregate marginal price for electricity even for systems with relatively small geographical reach such as the NYISO. In geographically extended systems, such as the PJM Interconnection or Midwest ISO (MISO), losses can become even more significant under certain operating conditions. A Survey of Locational Loss Prices in New York.
Between January and May 2003, the average real-time energy-only price across New York was approximately $48/MWh. By definition, this price is the same for all NYISO zones. (The average real-time price for each zone for the five months from January to May 2003 was calculated by averaging the hourly integrated zonal energy price published on the New York ISO’s website. The New York ISO calculates zonal prices using nodal generator prices and a pre-determined weighting factor for each generator that represents the contribution of that generator to the zonal spot price.) Losses and congestion, however, can cause delivered energy prices to differ across the system’s zones. Figure 1a shows the average real-time price for the marginal loss and congestion components of the location based marginal prices (LBMP) for the 11 NYISO load zones (see Figure 1b for a definition of the NYISO’s load zones) for the same period. The marginal loss component of the LBMP can be a significant part of the aggregate LBMP, particularly in Zones J (NYC) and K (Long Island) where losses add nearly 9% to delivered energy prices. The range of the marginal loss component for the first five months of 2003 are shown in dollars in Figure 2a and again as a percent of the total LBMP in Figure 2b. (A 98% confidence interval was used in calculating the minimum and maximum marginal loss values from the five months of data collected from the NYISO’s website.) These figures further illustrate the significance of the marginal loss component on the LBMP.

Figure 1a: Losses added up to 9% to the zonal wholesale price of energy in the January – May 2003 period in New York.

Figure 1b: A map showing the 11 load zones comprising the New York ISO.

Figure 2a: While the cost of losses was typically less than $5/MWh on average, during some hours, the hourly loss price component across the state could range from a little as -$8/MWh to as much as +$10/MWh.

Figure 2b: Expressed as a percentage of the aggregate zonal electricity price (including energy, losses, and congestion costs), the loss price component contribution to the aggregate hourly price ranged from –11% to +10%.

Understanding Transmission Losses:
Generally speaking, power flowing across a conductor or through a transmission circuit incurs losses in proportion to the square of the power delivered. Conductors and the other materials that comprise an integrated transmission network possess a natural resistance (or impedance) to the free flow of power. While this relationship varies somewhat in an AC circuit due to VAR (Volt Ampere Reactive) interchange, the magnitude of losses can be roughly approximated by the equation:

where ‘r’ is the resistance (impedance) of the conductor, ‘I’ is the current flowing through the conductor, ‘V’, is the voltage difference between either ends of the conductor, and ‘P’ is the power delivered. The constant ‘a’ is determined by the line voltage and resistance.

Effect of System Load on Losses:
The power flowing through the transmission network, in general, is a function of the load in the system. This can be seen in Figure 3a, which shows the loss component of the real-time LBMP for Zone J (NYC) of the New York ISO for April 19, 2003. The hourly loads for New York City are shown as the solid black line in the figure. The variation in the loss component of the LBMP (shown in yellow) due to the change in the load can be seen from the figure. Figure 3b shows the relationship between load and marginal losses on an X-Y plot. Generally, as loads increase, the amount of power flowing on the transmission system also increases. Consequently, the increase in losses that apparently corresponds with increasing loads is to be expected. Losses, however, are a direct consequence of transmission flows on a system and not the loads, per se.

Figure 3a: Figure shows the variation in the loss component of LBMP due to the change in load

Figure 3b: An X-Y plot showing the relationship between load and losses

Effect of Transmission Flows on Losses:
The impact of transmission flow on the loss component of LBMP can be best seen in Figure 4. Case I, shown in Figure 4a, represents the transmission flows and marginal loss prices that existed in the NYISO system during Hour 14 on January 22, 2003 when New York was exporting power to Hydro Quebec. (Marginal losses are transmission losses associated with each additional MWh of generation injected at a particular location. Marginal loss price is the component of LBMP at a location that accounts for the marginal losses, as measured between that location and the reference bus.) In this case, the power flows on the interface between Zone D and HQ are in a northerly direction. Case II, shown in Figure 4b, represents the transmission flows and marginal loss prices that existed in the NYISO system during Hour 15 on June 24, 2003 when New York was importing power from Hydro Quebec. In this case, the power flows on the interface between Zone D and HQ are in a southerly direction. Figure 5 shows the marginal loss prices of several western New York ISO load zones and interconnecting ISOs for the two cases. The impact of the directional flow of power on marginal losses is particularly evident for Zone D (North) and Hydro Quebec. As the flow on the interface from Zone D to HQ reverses, the marginal loss component changes sign (and magnitude) as shown in Figure 5. The change in load for Zone D (887MW for Hour 14, January 22, 2003 and 656MW for Hour 15, June 24, 2003) does not adequately account for the change in the marginal loss component. In summary, the magnitude and sign of the marginal loss component depends on the magnitude and direction of transmission flows in the system.

Figure 4a: Case I – New York exporting to Hydro Quebec: A map of the NYISO showing transmission flows and LMBP loss components for Hour 14 on January 22, 2003.

Figure 4b: Case II – New York importing from Hydro Quebec: A map of the NYISO showing transmission flows and LBMP loss components for Hour 15 on June 24, 2003.

Figure 5: Comparison of western NYISO load zone marginal loss prices

A Case Study Performed Using the Loss Model in GE MAPS^TM
A simulation of New York electric system for the first five months of 2003 was performed using the new marginal loss model in MAPS Version 12. The results of the simulation were compared with the data obtained from NYISO’s website. GE PSEC’s standard 2003 Northeast database (NYISO, ISONE, PJM and ECAR) with actual loads from the first five months of 2003 was used in the study.

Comparison of Loss Factors Calculated by MAPS with those reported by New York ISO
MAPS was used to estimate the hourly incremental losses for the transfer of 1MWH of energy from the West zone (Zone A) to New York city (Zone J). The marginal loss for the transfer of 1MW between Zone A and Zone J is simply the difference in the marginal loss factors at the two zones. (Marginal loss factor for a node (zone) is defined as the incremental change in system losses for a 1MW increase in injection at that node (zone) withdrawn at the reference bus). Using the differences in the marginal loss factors between the two zones eliminates the influence of the reference bus selection. The values obtained from MAPS were then compared to similar values obtained from the NYISO website.

Figure 6 shows the marginal loss factor difference between the West zone (A) and NYC (J) for the first week of 2003 as calculated by NYISO and MAPS. As seen in the figure, the marginal loss factor difference between Zones A and J as calculated by MAPS using the standard Northeast database compares well with those values calculated by the NYISO. The marginal cost of losses at a node is calculated using the marginal loss factor and the cost of replacement energy at that node, i.e.,

Marginal cost of losses = Marginal Loss Factor * Cost of Energy

Figure 6: Comparison of marginal loss factor difference between Zones A and J as calculated by MAPS and the NYISO

Comparison of MAPS results with AC Loadflow
In order to check the accuracy of the marginal loss factors calculated by MAPS, the dispatch and corresponding load from MAPS, for an hour, were input into an AC load flow program, GE PSLF. Loss factors were then calculated from the solved AC loadflow. Figure 7 shows a comparison of the loss factors as calculated by MAPS and PSLF at 467 generator nodes in the NYISO control area for the hour. As can be observed, MAPS is able to accurately calculate loss factors using a linearized AC solution.

Figure 7: Marginal loss factors calculated by MAPS compared to those calculated by PSLF AC loadflow program for a single hour at 467 generator nodes in the New York ISO

Conclusions:
Electricity markets in the U.S. are implementing nodal pricing of transmission losses to more accurately assess the cost differences among various generation commitment and dispatch alternatives. Nodal loss prices could constitute a significant part of the locational marginal price under certain operating conditions. Therefore, it is critical for market participants to be able to correctly predict the impact of transmission losses on location marginal prices. The new marginal loss model in MAPS accurately models the impact of losses on the locational marginal prices.