Loading...
HomeMy WebLinkAbout20100301Qly Market Monitoring Report.pdfPotomac Economics, Ltd. 701 City Center Drive Carmel, IN 46032 POTMAC ECONOMICS Telephone: 317-249-5721 Facsimile: 317-249-5943 Februaiy 19, 2010 ~-c::s~ ,.. Members of Service List (EC05- 11 0) PAc-i: -oS-Ò~ Re: Quarterly Market Monitoring Report Dear Service List Member: Please find attached the public (redacted) version ofthe Fourh Quarter 2009 Quarerly Market Monitoring Report for PacifiCorp. Regards, ~~ Michael W. Chiasson, P.E. Vice President Enclosure ~ dJ.. ~ I RE!:~ 20iU Mp,R - i AM 8= 20 QUARTERLY MARKT MONITORIG REpORT ON PACIFICORP Fourth Quarter of 2009 Issued by: Potomac Economics, Ltd. Independent ~arket ~onnor CONFIDENTIAL MATERIL REDACTED PacifCorp Monitoring Report: Fourth Quarter of 2009 Table of Contents \ Table of Contents I. Overview .................................................................................................................. 1 A. Market Monitoring........................................................................................... 2 B. Summaiy of Quarterly Report ......................................................................... 3 C. Complaints and Special Investigations............................................................ 5 II. Wholesale Prices and Transactions ....................................................................... 6 A. Prices................................................................................................................ 6 B. Sales and Puchases ......................................................................................... 9 III. Transmission Congestion ...................................................................................... 11 A. Overview........................................................................................................ 11 B. Transm ission Operatig Procedures .............................................................. 12 IV. Transmission Access ............................................................................................. 13 v. Monitoring for Anticompetitive Conduct ........................................................... 15 A. Wholesale Sales and Puchases ..................................................................... 15 B. Generation Dispatch ...................................................................................... 18 C. Transm ission Outages.................................................................................... 25 D. Trasm ission Operations ............................................................................... 26 E. Conclusions on Monitoring for Anticompetitive Conduct ...................... ...... 29 Redacted Version ì PacifiCorp Monitoring Report: Fourth Quarter of 2009 List of Figures List of Figures Figue 1: East Wholesale Prices and Peak Load, Fourh Quarer of2009 ........................ 6 Figue 2: West Wholesale Prices and Peak Load, Fourh Quarer of 2009 ....................... 7 Figue 3: East Trends in Monthly Electrcity and Natual Gas Prices............................... 8 Figue 4: West Trends in Monthly Electricity and Natual Gas Prices ............................. 9 Figue 5: Summai of PAC Sales and Purchases ............................................................ 10 Figue 6: Disposition of Requests for Transmission Service on the PAC System.......... 13 Figue 7: Disposition of Transmission by Duration of Service....................................... 14 Figue 8: Prices Received for PAC Sales and Purchases................................................. 16 Figue 9: PAC Supply Curves ......................................................................................... 20 Figue 10: East Out-of-Merit Dispatch and Congestion Events ...................................... 22 Figue 11: West Out-of- Merit Dispatch and Congestion Events ..................................... 24 Figue 12: Curailment and Curilment Deviation ......................................................... 28 Redacted Version PacifiCorp Monitoring Report: Fourth Quarter of 2009 Overvew , I. OVERVIEW In connection with MidAerican Energy Holdings Company's ("MEHC's") acquisition of PacifiCorp ("PAC" or the "Company") in Federal Energy Regulatory Commission ("Commission") Docket No. EC05-l 10-000, the Commission accepted market monitoring plans for PAC and MidAmerican Energy Company ("MEC") and Potomac Economics was retained as the independent market monitor for both companies. The plans established that separate quarterly reports be produced for each company. This is the market monitoring report for the fourh quarer of 2009 for PAC. The market monitoring plan for PAC is designed to detect any anticompetitive conduct from operation of the company's transmission system, including any tranmission effects from the company's generation dispatch. As stated in the plan: The Market Monitor shall provide independent and imparial monitoring and reportg on: (i) generation dispatch ofPacifiCorp, and scheduled loadings on constrained trnsmission facilities; (ii) details on binding transmission constraints, such as transmission refuals, or other relevant information; (ii) operating guides and other procedures designed to relieve transmission constraints and the effectiveness of these guides or procedures in relieving constrints; (iv) information concerning the volume of transactions and prices charged by PacifiCorp in the electrcity markets affected by these companes before and after the companies implement redispatch or other congestion management actions; (v) PacifiCorp's calculation of Available Transmission Capability ("ATC") and Total Transfer Capability ("TTC") over transmission lines owned or controlled, in whole or in par, by PacifiCorp; and (vi) plans for constrction by PacifiCorp of expansions to its trnsmission facilities. To execute the monitoring plan, Potomac Economics routinely receives data from PAC that allows us to monitor generation dispatch, transmission system congestion, and the Company's operational and commercial activity durng periods of congestion. We also collect certain key data ourselves, including OASIS data and market pricing data. The purose of this report is to provide the results of our monitoring activities and significant events on the PAC system i during the four quaer of2009. As specified in the monitorig plan, a draft of the findings has been submitted to PAC prior to submission to the Commssion. PAC had no substative comments. Redacted Version Pagel -:, PacifCorp Monitoring Report: Fourth Quarter of 2009 Overview i A. Market Monitoring Potomac Economics performs the market monitoring fuction on a regular basis, as well as performing periodic reviews and special investigations. Our primary market monitorig is conducted by way of regular analysis of market data relating to transmission outages, congestion, and transmission access. This involves data on transmission outages, transmission reservation requests, Available Transfer Capability ("ATC"), and curailments or other actions taken by PAC to manage congestion. Analyses of these data aid in detectig congestion and determining whether market participants have full access to transmission service. In addition to the regular monitoring of outages and reservations, we also remain alert to other significant events, such as price spikes, major generation outages, and extreme weather events that could adversely affect transmission system capability and give rise to the opportty for anti competitive conduct. Ou periodic review of market conditions and operations is based on operating data PAC provides us, as well as other data that we collect on a routine basis. Our review consists of four pars. First, we evaluate regional prices and transactions to provide an assessment of overall market conditions. Second, we sumarze transmission congestion in order to detect potential competitive problems. Congestion is identified by schedule curailments on the PAC transmission system. Third, we evaluate the disposition of transmission service requests to analyze transmission access and to detect whether there are circumstances on the PAC system that require closer analysis. Finally, to monitor for anticompetitive conduct, we examine periods of congestion and evaluate whether PAC operating activities raise concerns that PAC appears to be behaving anti-competitively. The operating activities that we evaluate are wholesale purchases and sales, generation dispatch, transmission outages, and the curailment and reduction of schedules. In addition to our periodic reviews, we may from time-to-time be asked to or deem it necessai to undertake a special investigation in response to specific circumstances or events. No such events occured this quarter. Redacted Version Page 2 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Overview , B. Summary of Quarterly Report 1. Wholesale Prices and Transactions Prices: We evaluate regional wholesale electrcity prices to provide an overview of general market conditions. Over the course of the quartr, Eastern and Western control area electrcity prices remained correlated with load and natual gas prices. There was a pronounced price spike on December 9. Therefore, we focused particular attention on that day throughout the report to detect whether it may have been related to any anti-competitive behavior. Sales and Purchases: PAC engages in wholesale purchases and sales of power on both a short- term and long-term basis. We analyze sales and purchases in Section V.A. 2. Transmission Congestion We studied congestion on the PAC system by examining schedule curailments and reductions. In the period of study, PAC implemented 481 curailments and schedule reductions totaling 30,636 MWh across 22 paths. We utilize curailments as an indication of congestion. In addition, we analyze the accuracy of curilments because unjustified curailments can be used to foreclose competition. 3. Transmission Access We evaluate the patterns of transmission requests and their disposition to determne whether market paricipants have had difficulty accessing the PAC trnsmission network. If requests for transmission service are frequently denied, this may indicate an attempt to exercise local market power. The volume of approved requests was higher than the levels observed in the fourh quarer of 2008 and higher than the third quarer of 2009. The volume of refusals was higher than the preceding quarer and higher compared to the same quarer of the prior year. We see no evidence that these refusals were not legitimate. Our review of the disposition of transmission requests does not indicate anti competitive behavior. Redacted Version Page 3 " j PacifCorp Monitoring Report: Fourth Quarter of 2009 Overvew 4. Potential Anticompetitive Conduct Wholesale Sales and Purchases: We examined the transactions that PAC executed durng the period of study. We focus on real-time transactions because these best represent the spot price of electrcity and wil most closely reflect power prices that might arse on the PAC system under conditions most conducive to market power. Under a hypothesis of market power, we would expect high sales prices or lower purchase prices durng times when transmission congestion arises. Real-time daily average transaction prices ranged between t MW and _ MWh. We focused our evaluation of PAC's generation and transmission on days with congestion that may have benefited PAC's net sales position. Dispatch: To fuher evaluate competitive issues, we examine PAC's generation dispatch to determine the extent to which congestion may be caused or exacerbated by uneconomic dispatch. Congestion can result natually when PAC or any utility attempts to dispatch its units in a least- cost maner. Such congestion does not raise competitive concerns. If an unjustified depare from least-cost dispatch ("out-of-merit" dispatch) occurs, causing congestion, competitive concerns arse. Our investigation found that out-of-merit dispatch durng the study period that had significant effects on transmission constraints was justified. Hence, this analysis did not reveal evidence of anti competitive conduct. Transmission Outages: We also evaluate PAC transmission securty events and transmission outages in order to determine whether these events may have unduly caused congestion. We focused our analysis on seven outage events that were associated with curailments. We investigated these events and found no evidence of anti competitive conduct. Transmission Operations: We analyze PAC curilments to determine whether curailments are being properly implemented. PAC manages congestion, prioritization of schedules, and low voltage events with schedule curailments. We scrutinized 57 curailments that were at least 75 MW above what we estimate to be justified by net schedules and TTC. We were able to fully justify 45 of these 57 curailment deviations. Given that 481 curailments were implemented over the quarter, we find that actions taken to manage the system were accurate. We do not fmd evidence of anticompetitive conduct. However, improvements are needed in the process of implementing changes to the list of active points of receipt and points of delivery. This is Redacted Version Page 4 r PacifiCorp Monitoring Report: Fourth Quarter of 2009 Overvew ~ evidenced by the fact that all twelve unjustified curailment deviations were associated with software issues related to this process. C. Complaints and Special Investigations We have not been contacted by the Commssion or other entities regarding PAC's market behavior. We also have not detected any conduct or market conditions that would warant a special investigation. There were no complaints lodged against PAC regarding transmission access durng the study period. Redacted Version PageS PacifCorp Monitoring Report: Fourth Quarter of 2009 Wholesale Prices and Transactions J II. WHOLESALE PRICES AND TRASACTIONS A. Prices We evaluate wholesale electricity prices in the PAC region in order to provide an overview of general market conditions. Examining price movements can provide insight into specific time periods that may merit further investigation, although they are not definitive indicators of the presence or absence of anticompetitive conduct. PAC is not par of a centralized wholesale market where spot prices are produced transparently in real-time. Wholesale trading in the areas where PAC operates is conducted under bilateral contracts. Because of its geogrphic expanse, we consider two sets of pricing points to represent the Western and Eastern portions of PAC's system. Figue 1 shows the bilateral contract prices for Four Comers and Mona (representing the East) and Figue 2 shows the bilateral contract prices for Mid Columbia and Mona2 (representing the West). Figure 1: East Wholesale Prices and Peak Load, Fourth Quarter of 200980 20,000 II PACE Peak Load - Four Corners Price - - - SoCal Gas Cost --Mona Price 70 60 15,000 I"~ 40 i~mæ i ~.~ ¡: 30 20 5,000 10 o o i:'""p !\ !\ !\ !\ ,. ,. ,. ,. -"- -"- i: -"-"ç; Oç; Oç; "ç; f:Ç) f:Ç) f:Ç) f:Ç) ~"'- ~"'- ~q; ~"'-~~-~~~~~~~~~~ Date 2 Mona is a relatively iliquid and lightly trded market point in central Uta. It is included in both figues to provide a baselie for comparson between them. Redacted Version Page 6 PacifCorp Monitorig Report: Fourth Quarter of 2009 Wholesale Prices and Transactions 1. Figure 2: West Wholesale Prices and Peak Load, Fourth Quarter of 2009120 9,000 li PACW Pea Load - Mid Columbia Price - - - PG&E Malin Gas Cost -Mona Prce100 7,500 80 6,000 ~ ~ 60 .~ i: 40 4,500 ~ ~ 3,000 20 1,500 o o ¡;'".."0"''' . Ol: p",,, pl: i!~ *'~ i!04 i!04 ".~e,'" ...:~e,'" A~"" .~""~ ",e.. ",'V .t '? ",'V ",0( ",'ó, ~ .;' ~ Date System load data is also shown because of the expected correlation with power prices. The Eastern control area load is shown on the East figue and the Western control area load is shown on the West figue. Natual gas prices are also shown because natual-gas-fired units are most often the marginal unit supplying the grd, and because fuel costs comprise the vast portion of a generating unit's costs. For the West analysis we use the daily price of natual gas deliveries at PG&E Malin (at the Northern California Border) translated to a power cost assuming an 8,000 btuWh heat rate. This number roughly corresponds to the fuel cost portion of the operating cost of a natual gas combined cycle power plant. For the East comparison, we use SoCal Border Gas (at the Southern California Border) price and apply the same power-cost conversion. Prices for bilateral contrct transactions are compiled and published by commercial pricing sureys. The bilateral pricing data shown in the figue above is published by Platts. The Mid Columbia pricing location includes a collection of hydroelectric units along the Columbia River in Oregon and Washington, and represents the value of electrcity in the Pacific Northwest. This is a liquid point in PAC's Wester control area. The Four Comers location is at the southern end of the PAC transmission system where New Mexico, Colorado, Arzona, and Utah meet. Prices at Four Comers represent the value of electrcity in the Desert Southwest. This is the liquid point that is closest to PAC's Eastern control area. Redacted Version Page 7 PacifCorp Monitoring Report: Fourth Quarter of 2009 Wholesale Prices and Transactions J Figue 1 and Figue 2 show that power prices at both Mid Columbia and Four Comers are generally correlated with fluctuations in natual gas prices and load, which is consistent with expectations in a properly fuctioning market. There was one price spike on December 9 in both the East and West regions, which was mainly caused by rising spot natual gas prices and unusually cold temperatues in the west region. The effect of the temperatue on load can be seen in Figue 2 by the sharp load increase. The next analysis compares the average Four Comers and Mid Columbia power prices for the period from October 2006 through December 2009 with average prices during the same period over the past thre years. These results are shown together with the average Platts SoCal Border and PG&E Malin natual gas prices discussed above. As the figues show, electrcity prices have generally been highly correlated with natual gas prices over longer timeframes. Figure 3: East Trends in Monthly Electricity and Natural Gas Prices Fourth Quarter, 2006-2009 120 12 II Four Corners 100 -- SoCal Border Gas 10 ~~~'-80 8 ~'"'".. 'C ~i:È '"'" ..60 6 .~ .C i:'t '";~ .;40 4 e'"::.S ~I:-=~ 20 2 0 0 October November December Redacted Version PageS r PacifiCorp Monitoring Report: Fourth Quarter of 2009 Wholesale Prices and Transactions \. Figure 4: West Trends in Monthly Electricity and Natural Gas Prices Fourth Quarter, 2006-2009 120 12 il Mid Columbia -- PG&E Mal Gas 100 ;¡ i 80'".... 'Ci: i'60 :s.. ~ ~40'" .;.=~ 20 0 Octber November 10 ~81 :i 6 .~i: ~ 4 f ~ 2 o December Overall, our evaluation of wholesale electricity prices in the PAC region indicates a price spike on December 9. Thoughout the report, we focus paricular attention on that day to detect whether it may have been related to any anti-competitive behavior. B. Sales and Purchases PAC engages in wholesale purchases and sales of power, both firm and non-firm transactions. Figue 5 summarzes PAC's sales and purchases activity for trades that delivered during the fourth quarter of 2009. We consider only short-term trades because we are interested in transactions made by PAC where they could have benefited from any potential market abuse durng this time period. Short-term transactions include all transactions that are less than one month in duration. Longer-term transactions generally occur at predetermined prices that would not be directly affected by transitoiy periods of congestion. Additionally, short-term transaction prices are good indicators of wholesale market conditions durg periods of congestion. Redacted Version Page 9 PacifCorp Monitoring Report: Fourth Quarter of 2009 Wholesale Prices and Transactions Figure 5: Summary of PAC Sales and Purchasesr rt transactions during congested periods in Section V.A to detect potential anti competitive conduct. Redacted Version Page 10 ,. PacifCorp Monitorig Report: Fourth Quarter of 2009 Transmission Congestion " III. TRASMISSION CONGESTION A. Overview PAC is a member of the Western Electrcity Coordinating Council (WCC). In WECC, regional congestion is primarily managed by ensurg that the scheduled flows do not exceed flow limits on specified paths.3 However, because actual flows sometimes exceed scheduled flows due to loop flow (or parallel path flow), additional congestion management procedures are employed. Power flows in the WECC follow a relatively predictable pattern. Most of the flows over the network occur on the high-voltage facilities that roughly correspond to the geographic perimeter ofWECC. The transmission system in the interior of the WECC boundaries operates at a lower voltage and carres less power. The topology of the trsmission network causes power to circulate around the perimeter of the system. Typically, power transfers from the Pacific Northwest are scheduled south to Californa. However, sometimes this nort-to-south power flow results in unscheduled increases in flow around the perimeter of the WECC system in the clockwise direction, passing through the PAC system and then on to California from the west though Arzona. The PAC system consists of two control areas: PACW in Nortern California, Western and Central Oregon and Southeast Washington, and PACE, which is in Wyoming, Southeast Idaho, and Utah. PAC extends across a broad geographical area, having a presence in six states. It has 16,400 miles of transmission lines and approximately 10,700 MW of owned or controlled net generation capacity. PAC operates a significant portion of the transmission facilities that provide north-to-south flow along the eastern perimeter ofWECC.4 These flows pass through a key intedace that is operated by PAC known as Path 20 (someties referred to as Path C). Path 20 was a "qualified path" in the nort-to-south direction under the UFRPs used by WECC.5 3 This is in contrast to how congestion is maaged in the Eater Interconnect where congestion management generally is focused on actual flows on flowgates as opposed to scheduled flows on contract paths. 4 While nort-to-south flow is common, patters of schedules and generation dispatch sometimes cause south-to- nort flow. 5 WECC uses UFRPs when actul flow exceeds scheduled flow on a "qualified path". There are a limited number of qualified paths identified based on cerin cnteria that include the path having a history of uncheduled flow. The UFRP consists of a senes of nine steps that are intended to relieve the congestion through the operation of equipment and, ultimately, the curilment of schedules. Redacted Version Page 11 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Transmission Congestion 1 However, effective September 15,2008, the path was disqualified by the WECC operating committee. In this section, we investigate congestion on the PAC system by examining curailments and transmission service request refusals. We also examine plans for constrction of expansions to transmission facilities and found cases where the planned expansions may reduce congestion in constrained areas. Nothing from our review of PAC's planned expansions raised competitive concerns. B. Transmission Operating Procedures Durng the period of study, PAC implemented 481 curilments (including cases when curailments were reversed) and schedule reductions totaling 30,636 MW across 22 paths. Curilments can be initiated when one of four conditions occurs: (1) the path is overscheduled (due to conditions on the transmission system causing a reduction in TTC); (2) a schedule with a higher priority reservation displaces a schedule with a lower priority reservation; (3) a low- voltage constraint is binding; or (4) actual flows exceed the capability of the path. The accuracy of these curailments and schedule reductions are evaluated in Section V. Redacted Version Page 12 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Transmission Access l IV. TRASMISSION ACCESS A main component of the market monitorig fuction is to evaluate trsmission availability on the PAC system. In this section, we evaluate access to the trmission network by analyzing the disposition of transmission requests. The patterns of transmission requests and their disposition are helpful in determning whether market paricipants have had diffculty accessing the PAC transmission network. In order to make this evaluation, we calculate the volume of requested capacity that spanned the time period under study. For example, if a request was approved in January for service in June, we categorize that as an approval for June. Because requests var in magnitude and duration, we assign a total monthly volume (GWh) associated with a request, which provides a common measure for all tyes of requests. Hence, a yearly request for 100 MW has rights for every hour of the month for which the request spans, just a like a monthly request. A request covering less than the entire month is assigned the hours between its sta and stop time. Figue 6 shows the breakdown of transmission servce requests in each month from October 2008 through December 2009 and sumarzes the disposition of the requests. Figure 6: Disposition of Requests for Transmission Servce on the PAC System October 2008 - December 2009 5,000 30,000 25,000 'i ~ 20,000 ~i i 15,000 OS ë i 10,000~ o 2008 2009 Redacted Version Page 13 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Transmission Access The figue shows that the total volume of approved requests durng the four quarer of 2009 was higher than the four quarter of 2008 and higher than the third quarter of 2009. The volume of refused servce requests durng the quaer was slightly higher than the preceding quarer, averaging 2846 GWHr. Hence, the approval rates for the fourh quarter 2009 and the third quarter 2009 were comparable, averaging 96 percent. We reviewed the refusals and found no evidence that these refusals were not legitimate or that PAC had uneasonably restrcted access to its transmission system. To fuher evaluate the disposition of transmission requests, we compare the volume of transmission requests over the study period by increment of servce to the requests from the corresponding period twelve months prior. This comparson is shown in Figue 7. Figure 7: Disposition of Transmission by Duration of Service 80,000 70,000 :::Refused 10,000 I~ Approved 'i 60,000 ~ 50,000 i 40,000i:..= Ë 30,000i;;.20,000 o ~i ~ ~g ~ ~ ~~i ~ ~ i Fourth Quarer 2009 ~ ¡: Fourth Quarter 2008 Figue 7 indicates an increase in the volume of approvals for all categories of service except for weekly and monthly. There was an increase in the volume of refused yearly requests, all of which were submitted prior to March 2007. They appear in this figue because the requested service spans the period of study. For these cases in general, the customers did not continue with the application and study process needed to ultimately perform system upgrades to make the transmission available. As a result, our review of the disposition of transmission requests does not raise any anti competitive concerns. Redacted Version Page 14 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct V. MONITORIG FOR ANTICOMPETITIV CONDUCT In this section, we evaluate the available market and operating data to identify any evidence of anti competitive conduct or market manipulation. The market monitoring plan calls for identifying anti competitive conduct, which includes conduct associated with the operation of either PAC's generation assets or its transmission assets that can create transmission congestion or erect barrers to rival suppliers, thereby raising electrcity prices. To identify potential concerns, we analyze PAC's wholesale sales in the fit subsection below, its dispatch of generation assets in the second subsection, trsmission outages in the third subsection, and PAC's transmission operations in the four subsection. A. Wholesale Sales and Purchases We examine sales and purchase data to determine whether the prices at which PAC transacted power may raise concerns regarding anticompetitive conduct that would warrant fuher investigation. We are parcularly interested in periods when transmission congestion arises. If PAC were engaging in anticompetitive conduct to create the congestion, it could potentially benefit by making sales at higher prices in constrined areas or purchases at lower prices adjacent to constrained areas. We examined the real-time bilateral transactions made by PAC using PAC internal records. We focus on real-time transactions because they best represent the spot price of electricity. Competition is facilitated by the ability of rivals to reserve and schedule trnsmission service. This ability wil be limited if ATC is unavailable, tranmission requests are refused, or schedules are curailed. Curailments are also an indicator of congestion because they can be made when a path is over scheduled. If PAC's abilty to curail schedules is being abused, we would expect to see systematically higher prices for sales or lower prices for purchases coincident with curilments. Figue 8 shows the daily average prices received by PAC for real-time bilateral sales and purchases. The figue also indicates days when curilments occured that could have potentially benefited PAC's position in the real-time bilateral markets. A curilment may impact system Redacted Version Page 15 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct 1 flows at market deliveiy points to the benefit of PAC's net position at those deliveiy points.6 The maximum daily effective market position (labeled as "Max Effect" in the figue) is also displayed. This is the impact of PAC's sales and purchase transactions on the congested paths, calculated as the sum of the products of the volume of each market position and the shift factor of the delivery point to the curailed path. "Max Effect" identifies periods when PAC is actively buying or selling in constrained areas and, therefore, could benefit itself by restrcting other suppliers' access. The figue displays this value for the path and hour that has the maximum value for each day. Figure 8: Prices Received for PAC Sales and Purchases The volume weighted average daily sales prices ranged from ti MWh to tI/MWh and the average was tI/MWh. We say a day has a "beneficial curtailment" if PAC is a net seller at a deliveiy point where the curtailment restricts supply or PAC is a net purchaser where the curtailment increases supply. On days when potentially beneficial curtailments occurred, the 6 The relationship between constrained paths and market delivery points is determined through shift factors, which are the portion of power injected at the market delivery point that flows over the constrained transmission path. Redacted Version Page 16 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct t weighted average daily sales prices average _ MW. The volume weighted average daily purchases prices ranged from t MW to _ MW and the average was _/M. On days with potentially beneficial curtailments, the weighted average purchase price was _ MWh. These prices do not show a pattern of PAC benefiting from curilments. Though the overall price patterns do not raise concerns, we selected seven days for closer examination. On six of these days, the maximum daily effective market positions were greater than or equal to 80 MW. The seventh day, December 9, 2009, was included due to the price spike on that day shown in Figue 2. . November 5, 2009: The curailment was on the Northern Uta (NUT) to Path C path. At the time of the high Max Effect, there were several purchases and sales. The transaction that may have benefited from the curilment 7 was This is a low purchase price when compared to the suroundig days. . December 3, 2009: The curailment was on the PACE to NUT path. At the time of the high Max Effect, there were several purchases and sales. These purchase prices are not low for that portion of December. . December 7, 2009: The curailment was on the PACE to NUT path. At the time of the high Max Effect, there were several purchases and sales. This purchase prices is not low for that portion of December. 7 The other trsactions could not benefit from the curilment because either the deliver points are electrcally distat from the curiled path, or the effects of the curilment on the trsaction would not be to PAC's advantage. An example of the latter would be a purchase position immediately down-stream of the congested path. The curlment would put upward pressure on the delivery point price by reducing import into the congested region of the delivery point. Redacted Version Page 17 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct . December 8,2009: The curailment.was on the PACE to NUT path. At the time of the high Max Effect, there were several purchases and sales. These purchase prices are not low for that portion of December. . December 9, 2009: The curailment was on the PACE to NUT path. At the time of the high Max Effect, there were several purchases and sales. These prices are not unusual for that portion of December. . December 10,2009: The curailment was on the PACE to NUT path. At the time of the high Max Effect, there were several purchases. This is a not low purchase price when compared to the surounding days. . . December 25, 2009: The curilment was on the PACE to NUT path. At the time of the curailment, there were sales at multiple points. high compared to surounding days. Our primaiy concern is whether PAC anticompetitively created the congestion though generation and transmission operations. Accordingly, we focus particular attention on these days when we evaluate PAC's generation dispatch and transmission outages in the remainder of this section. We also review the accuracy of all curailments in Section V.D below. B. Generation Dispatch To fuer evaluate whether PAC's conduct raises any anticompetitive concerns, we examine the company's generation dispatch to determine the extent to which congestion may have been the Redacted Version Page 18 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct result of uneconomic dispatch of generation by PAC. Therefore, we first examine PAC's dispatch durg the study period to determine whether it was consistent with the least-cost use of its resources. Congestion can result natually when PAC or any utility dispatches its units in a least-cost manner, and does not raise competitive concerns in such circumstances. If a depare from least-cost dispatch ("out-of-merit" dispatch) occur unjustifiably and it causes congestion, this effect can raise potential competitive concerns. We consider a unit to b.e out-of-merit when it is dispatched, but could have been replaced by lower-cost generation that was not dispatched. The PAC system is made up of two control areas: PAC West and PAC East. PAC is the balancing authority for both of these control areas. The movement of power between the two systems is limited by both transmission capabilty and contractual rights. Efficient merit order dispatch is practical within each control area, but not necessarly between them due to these limits. To account for this, we evaluate out-of-merit dispatch for each control area separately. To identify out-of-merit dispatch, we first estimate each control area's marginal cost cures or "supply cures".8 We used incremental heat rate cures, fuel costs, and other variable operations and maintenance cost data provided by PAC to estimate marginal costs. This allowed us to calculate marginal costs for PAC's units. We ordered the marginal cost segments for each of the units from lowest cost to highest cost to represent the cost of meeting varous levels of demand in a least-cost manner. For our analysis, the cure is re-calculated daily to account for fuel price changes, planned maintenance outages, and planed deratigs. Figure 9 shows the estimated supply cures for a representative day durng the time period studied. As the figue shows, the marginal cost of supply increases as more unts are required to meet demand, as expected. 8 We use the term marginal cost loosely in this context. The value we calculate is actually the incrementa production cost and does not include opportity costs, risks, and other factors not reflected in the incremental production cost. Redacted Version Page 19 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct Figure 9: PAC Supply Curves We used each day's estimated marginal cost curves as the basis for estimating each control area's least-cost dispatch for each hour in the quarter. In general, this will not be the exact level of least-cost dispatch because we do not consider all operating constraints that may require PAC to depart from our estimate of the least-cost dispatch. The analysis is limited to peak hours to avoid times of ramping and commitment issues which prevent achievement of the theoretical least-cost dispatch. This analysis does not model generator commitments, assuming instead that all available generators are online. While market monitoring resources could have been expended refining the estimated generator commitment and dispatch to make it correspond more closely to actual operating parameters (i.e., start costs, ru-time and down-time constraints, etc.), we believe this simplified incremental-operating-cost approach is adequate to detect instances of significant out- of-merit dispatch that would have a material effect on the market. When a unit with relatively low rung costs is justifiably not commtted, our least-cost dispatch wil overstate the out-of-merit quantities because it wil identify the more expensive unit being dispatched in its place as out-of-merit. This may result in higher levels of out-of-merit dispatch durng low-load periods when it is not economic to commit certain units. Redacted Version Page 20 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct Other justifiable operating factors that cause the out-of-merit dispatch to be overstated are energy limitations and ancilaiy services. An example of an energy limitation is a governental regulation limiting the number of hours a plant may ru in a year. Since the unit is physically capable of producing, the limitation does not result in a planed outage or derating. The necessity to limit the hours of plant operation can cause the out-of-merit values to be overstated. Ancilai services requirements such as sping reserves, system ramp rate limitations, and AGC control requirements can make it operationally necessai to dispatch a number of units at part load rather than having the least expensive unt fully-loaded. These operational requirements can cause the out-of-merit values to be overstated. The out-of-merit quantities include units on unplanned outage since a sudden unplaned outage may be an attempt to uneconomically withold generation from the market. Hence, it wil tend to overstate the quantity of generation that is trly out-of-merit. For our analysis, the accuracy of a single point is not as importt as the trend and any substatial depares from the tyical levels. Figue 10 and Figue 11 shows the daily maximum "out-of-merit" dispatch for the peak hours of each day in the study period for the Eastern and Western control areas, respectively. Also shown in the figues are days when PAC curailments were made on paths that were also loaded as a result of out-of-merit dispatch. These days are represented as blue bar. For these days when potential generation-induced curailments occured, the out-of-merit dispatch displayed corresponds to the hour when the impact of the out-of-merit dispatch on the congested path was at its daily maximum. The figues also show "Path Impact" (red bars). This is a calculation of the power flow change on the curiled paths as a result of the out-of-merit dispatch. In other words, if dispatch had been "in-merit", flow on the curailed path would have been lower by the amount shown. All curailed paths are tested for impact from generation dispatch from generators in both control areas. The impact of out-of-merit dispatch was determined using generation shift factors. 9 9 Generation Shift Factors are defined as the incremental increase or decrease in flow on a flowgate divided by an incremental increase or decrease in a Generation Resource's output. Redacted Version Page 21 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct Figure 10: East Out-of-Merit Dispatch and Congestion Events Fourth Quarter of 2009 150 3201200 li Days with Curtments _ Path Impact -+MaxOOM 2801050 ~ 900 l ==0' 1 § ~..co ~ 240 750 200 ~ 6 600 "t160 L! fl 120 ¡f450 300 80 40 o ('c," 06- . oc," rt Ol) .:.... .:.... .:.... .:.... of of of of¥" 'l \."f~ .."" 'f' f1" \."1" ~F ..fo" .,~ \.f:~ ~~ ,.~ Date o As the analysis of the East control area in the figue shows, there were five days when out-of- merit dispatch was at least 600 MW and contrbuted at least 150 MW of increased flow over congested paths durg the study period. We inquired fuher into these days in addition to November 5, 200910 and found the following: . November 5: had a four-day forced outage to repair a_in while the more expensive Combined Cycle unit was generating signficant amounts of power. At the time of the outage, schedules on the NUT to Path C path were being curailed. The . have little impact on this path, but the combined cycle unit replacing the energy adds to the congestion on the path when output is increased (raise-hur). . December 7 through December 9: repair had a thee-day forced outage to while the more expensive 10 November 5, 2009 was added because it was identified as a date of interest in the "Wholesales Sales and Purchases section" and it had 474 MW of out- of-merit dispatch with 107 MW of path impact. The other dates from the "Wholesales Sales and Puchases section" that are otherwise not included had minimal out-of-merit dispatch and path impact. Redacted Version Page 22 r PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct _ Combined Cycle and Combined Cycle unts were generatig significant amounts of power. At the time of the outage, schedules on the PACE to NUT path were being curailed. The are a strong raise-help (raising output relieves congestion) on this path by providig counter flow. The combined cycle unts add to the congestion on the path when their output is increased (raise-hur). . December 10: and more expensive_ Combined Cycle and were generating significant amounts of power. were at part load while the Combined Cycle units time of the outage, schedules on the PACE to NUT path were being curiled. . December 16: trpped off-line while down, it stayed off-line for eight days to _Durng this time, the more expensive_ Combined Cycle and_ .Combined Cycle units were generating significant amounts of power. At the time of the outage, schedules on the NUT to Path C path were being curailed. Redacted Version Page 23 ", PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct ~" Figure 11: West Out-of-Merit Dispatch and Congestion Events Fourth Quarter of 2009 800 !I Days with Curtaments 700 !I Path Impact --MaxOOM ~600e~'==500..=0'..~400.. ~ ~300.. :a.. .!200=0 100 200 175 150 125 ~ 'G100 :.! 75 ~ 50 25o 0 r. o~ ,o~ . O~ ,,~ ~a" ~a" ~a" ~a" ,..:~r§ .....;~r§ ..,.~r§ ~~~r§" 'b ,,"' ..V~ "" 0/ "V ~' ..Id ., ~ ,,\ :' Date As the West control area figue shows, there were three days when the out-of-merit dispatch was at least 200 MW and contrbuted at least 35 MW of increased flow over congested paths durng the study period. We evaluate these days below and include December 9,2009 due to the price spike in Figue 2. was in a one-day forced outage to At the time of the outage, schedules on the Wyoming East to Wyoming Central path were being curailed. are a raise-help (raising output relieves congestion) on this path by providing counter flow. . December 3:was in a thee-day forced outage to At the time of the outage, schedules on the PACE to NUT path were being curailed. The path by providing counter flow. are a raise-help on this . December 9: was in a three-day forced derating due to_ _At the time of the outage, schedules on the PACE to NUT path were being Redacted Version Page 24 r PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct curailed. The are a raise-help on this path by providing counter flow. The maximum out of merit for this day was 276 MW and the Path Impact was 24 MW. . December 12: was in a 37-hour At the time of the outages, schedules on the PACE to NUT path were being curiled. Days identified in the Whole Sales and Puchases analysis above that are not included in the above out-of-merit dispatch analysis were found to not have signficant out-of-merit dispatch that affected constrained paths. Based on our review of the generation outage information in the operating logs, and information garered from discussions with PAC personnel, we conclude that the aforementioned out-of- merit dispatch from both control areas were justified and did not constitute attempts to engage in anticompetitive behavior. C. Transmission Outages We evaluate PAC securty events II to determine whether PAC's operation of transmission assets may have contrbuted to the congestion events that occured durng the study period of the report. We also evaluate transmission outages recorded in PAC's "Compass" system, its transmission outage logging system. Between the two systems we found 23 trsmission outage events that were associated with schedule curailments and were planned less than thee weeks in advance. This includes six transmission outage associated with curilments that coincided with the six days when PAC had purchase and sales positions that may have benefited from congestion as presented above. We reviewed these six outages plus one associated with a curtailment deviation identified in the next analysis and found the following: . II Securty events are defmed as trsmission security/reliability events that may impact the Provider's ability to schedule transactions. Redacted Version Page 25 . PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct ! . outage. .This four-week outage commenced on_Curailments on th~ . 2009. T e outage was ta en to C path coincided with this outage. : This twelve-hour outage commenced on_e outage was ta . en to Curailments on th~ NUT path coincided with this outage. .: This l8-da T is was a constrctIon outage taken to to Path C path coincided with this outage. outa e commenced on_Curt~T . . Through our review of the records and conference calls with PAC staff, we find that all the outages were justified and the events raise no competitive concerns. D. Transmission Operations Under PAC operating procedures, path flows can be managed by curailing transactions scheduled over the path. This can provide the opportity for anti competitive conduct by initiating curailments when they are not necessai. By selectively initiating these procedures, PAC may have the ability to influence power prices in the region to its benefit. Accordingly, we analyze the transmission schedules to determine whether curilments are being initiated properly. PAC initiates curailments when one of four conditions occurs: (1) the path is overscheduled (due to conditions on the transmission system causing a reduction in TTC); (2) a Page 26Redacted Version PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct schedule with a higher priority reservation displaces a schedule with a lower priority reservation; (3) a low voltage constraint is binding, or (4) actual flows exceed the capability of the path. To be over-scheduled, the net schedules (the sum of firm and non-firm schedules minus the sum of schedules that provide counter-flow) must exceed the TIC (less the scheduled amount of capacity reservations where applicable). 12 We analyzed the 22 paths where curailments were initiated by PAC. We compare aggregated ex post net schedules and real-time flows to TTC. Ex post net schedules are the net schedules actually realized at the end of the operating hour. PAC makes the curilment decision twenty minutes prior to the operating hour. However, NERC standards also allow schedules (referred to as "etags") to be submitted up until twenty minutes prior to the hour. Because it takes ten minutes to evaluate a submitted schedule, the resultig net schedule can change from what it was when PAC initially made the curailment decision. There may also be emergency etags submitted later than twenty minutes prior to the hour. Yet, this ex ante data is not available. Thus, utilzing ex post data provides only an approximation. The curailment deviations calculated and reported in the analysis below equal the TTC value minus the aggregated ex-post net schedules, except for the ''Northern Utah (NT) to Path C" path. On the "NUT to Path C", loop-flow is significant, so we calculate these curailment deviations as the TTC value minus the maximum of either the real-time flows or the aggregated ex-post net schedules. Using real-time flow allows us to captue the loop-flow on this path. The curailment deviations are limited to a ceiling equal to the curailment amount and a floor of zero, since we are less concerned with under curailments. In the absence of emergency tags or tags otherwise submitted after PAC makes its curilment decision, if a path is over-scheduled and the curailments are accurate, this value should be close to zero. 13 Figure 12 shows the results of this analysis. 12 Effective April 28, 2008, PAC utilizes forecasted values for Path C capacity when makg its curilment decisions. Accordigly, when evaluating curilments on the path "PACE to Path C", we utilize the forecasted capacity value rather than TTC. 13 The other reasons for curailments aside from the path being over scheduled wil not necessarly result in a curilment deviation close to zero. Redacted Version Page 27 PacifiCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct Figure 12:Curtailment and Curtailment Deviation Fourth Quarter of 2009 700 . Curtilment Deviation o Curilment 600 500 400 ~0 0.0 300 0.0 . 200 0 .0 0 0 0.0. 100 0 ~i/i/i/ rY",i/~~q i/i/i/~q i/#'~i/#'~i/ ,,~~.t .."~ o$~~~~i.~.t ~~r.'o~.N.,~....~t.~.f ~~..~.."~~~.N ,,'V " Date Over the quarer, 481 curailments were implemented. Of these, 57 curtilments had at least a 75 MW deviation. We reviewed all 57 for accuracy. Twenty-six of the curilments were of schedules between Furher review finds that all 26 curilments are justified by physical flows exceeding schedules and reaching the path limits. Two of the curailments were of schedules on the on November 6,2009. This was caused by a set-up error that was identified and corrected that same day. When PAC made modifications to their active points of receipt and points of delivery, the softare had to be set-up to recognize the schedules so that the schedules that load the paths could be identified and curailed if the path limits were reached. Set-up errors can cause unelated schedules to be included or related schedules to be excluded. Twenty-eight of the curailments were of schedules on the Eighteen of these are justified by physical flows exceeding schedules and reaching the path limits. The remaining ten curailment deviations on this path occured from December 8 through December 10,2009, Redacted Version Page 28 PacifCorp Monitoring Report: Fourth Quarter of 2009 Potential Anticompetitive Conduct and were caused by a set-up error that occured on December 3,2009. This was corrected on December 11, 2009. One curailment deviation was on the on December 16, 2009. This curailment was needed and justified because of an emergency outage on the . This outage is discussed on the Transmission Outages section above. Of the 57 curailments that we reviewed, twelve were found to be inaccurate or unjustified. All twelve were related to setting up softare. PAC had made modifications to their active points of receipt and points of deliveiy. There is room for improvement in the process of implementing these tyes of changes, but we do not view these events as evidence of anticompetitive conduct. Other than these software issues, we fid that actions taen to manage the system were appropriate. E. Conclusions on Monitoring for Anticompetitive ConduCt Based on our analysis of PAC's conduct and the market outcomes, we find no conduct by PAC that raises potential competitive concerns durg the period of study. Redacted Version Page 29