Oklahoma Gas & Electric P.O. Box 321 Oklahoma City, OK, Main Street, Suite 900 Cambridge, MA 02142

Transcription

1 Final Report Oklahoma Gas & Electric System Loss Study Submitted to Oklahoma Gas & Electric P.O. Box 321 Oklahoma City, OK, Submitted by Stone & Webster Management Consultants, Inc. 1 Main Street, Suite 900 Cambridge, MA August 14, 2008

2 LEGAL NOTICE This report was prepared by Stone & Webster Management Consultants, Inc. ("Stone & Webster Consultants") expressly for Oklahoma Gas & Electric. ( OG&E ). Neither Stone & Webster Consultants, nor OG&E, nor any person acting in their behalf, (a) makes any warranty, express or implied, with respect to the use of any information or methods disclosed in this report; or (b) assumes any liability with respect to the use of any information or methods disclosed in this report. Any recipient of this report, by their reliance on, acceptance or use of this report, releases Stone & Webster Consultants and its affiliates from any liability for any direct, indirect, consequential or special loss or damage whether arising in contract, tort (including negligence) or otherwise. Nothing expressed in this report should be construed as a legal opinion as to compliance with law or regulation. ELECTRONIC MAIL NOTICE Electronic mail copies of this report are not official unless authenticated and signed by Stone & Webster Consultants and are not to be modified in any manner without Stone & Webster Consultants' express written consent. i

3 Electric System Loss Study Table of Contents Section Description Page 1 EXECUTIVE SUMMARY TECHNICAL LOSSES LOSS ALLOCATION LOSS FACTOR DEVELOPMENT RESULTS INTRODUCTION TRANSMISSION SYSTEM LOSSES TRANSMISSION LINE LOSSES TRANSMISSION TRANSFORMER LOSSES CORONA LOSSES TRANSMISSION SYSTEM LOSS SUMMARY DISTRIBUTION PRIMARY LOSSES TRANSFORMER LOAD LOSSES TRANSFORMER NO- LOAD LOSSES SUMMARY OF DISTRIBUTION PRIMARY TRANSFORMER LOSSES DISTRIBUTION PRIMARY DISTRIBUTION LINES PRIMARY 4-KV LINES PRIMARY 12-KV LINES PRIMARY 24-KV LINES PRIMARY 34.5-KV LINES SUMMARY OF PRIMARY DISTRIBUTION LINE LOSSES DISTRIBUTION SECONDARY SYSTEM DISTRIBUTION SECONDARY TRANSFORMERS DISTRIBUTION SECONDARY LINES AND SERVICE DROPS CUSTOMER METERS NON-TECHNICAL LOSSES HOURLY LOAD ALLOCATION PROCEDURE CONCLUSION HISTORICAL LOSSES LOSS STUDY COMPARISONS METHODOLOGY ATTACHMENTS...30 ii

4 Electric System Loss Study Tables in Report LIST OF TABLES IN OG&E SYSTEM LOSS STUDY Table 1 Calculated Losses and Non-Technical Losses...2 Table 2 Adjusted Calculated Losses...3 Table 3 Losses by Service Level...5 Table 4 Monthly Losses and Loss Factors...6 Table 5 Categories of Load and No-Load Losses...9 Table 6 Corona Losses...13 Table 7 Transmission System Losses...13 Table 8 Distribution Primary Transformer Losses...15 Table 9 Modeled 4-kV Circuits...16 Table 10 Modeled 12-kV Circuits...18 Table 11 Modeled 24-kV Circuits...20 Table 12 Modeled 34-kV Circuit Loads and Losses...21 Table 13 Losses in Primary Distribution Lines...22 Table 14 Secondary Transformer Losses...24 Table 15 Secondary Lines and Service Drops Losses...24 Table 16 Customer Meter Losses...25 Table 17 Comparison of Previous Loss Studies...28 LIST OF FIGURES IN OG&E SYSTEM LOSS STUDY Figure 1 Sales to Ultimate Customers...7 Figure 2 Losses as a Percent of Total System Load...8 Figure 3 Transmission System Line Losses...11 Figure 4 Circuit Loss Results and Exponential Equation for 4-kV Circuits...17 Figure 5 Circuit Loss Results and Exponential Equation for the 12-kV Circuits...19 Figure 6 Circuit Loss Results and Exponential Equation for the 24-kV Circuits...20 Figure 7 Circuit Loss Results and Exponential Equation for 34-kV Circuits...22 Figure 8 Historical Losses as a Percent of System Load...27 iii

5 Electric System Loss Study TABLES IN ATTACHMENT Table A-1 Table A-2 Table A-3 Table A-4 Table A-5 Table A-6 Transmission Transformer No-Load Losses Distribution Primary Transformer Losses (4 kv Class) Distribution Primary Transformer Losses (12kV Class) Distribution Primary Transformer Losses (24.9 kv Class) Distribution Primary Transformer Losses (34 kv Class) Secondary Transformer Losses iv

6 Electric System Loss Study 1 Executive Summary This report documents the results of an Electric System Loss study performed for the Oklahoma Gas & Electric (OG&E) system for the 2006 test year. OG&E, a subsidiary of OG&E Energy Group, is a regulated electric utility company that serves about 755,000 retail customers in Oklahoma and Western Arkansas and a number of wholesale customers throughout the region. OG&E has nine power plants capable of producing about 6,903 MW. The company delivers electricity across an interconnected transmission and distribution system spanning about 30,000 square miles. Electric power system losses are a consequence of doing business for a full service electric utility. The electric system is dynamic and decisions are made every day that affect losses and thus the efficiency of the system. In addition, the losses that do result from the operation of the electric system must be properly charged to the customers that are responsible for those losses. To enhance the operational decision making process and fairly allocate losses to customers, it is necessary to understand the electric losses in detail as a function of where they occur on the system 1.1 Technical Losses The technical losses were investigated and calculated for the following categories: Transmission lines, corona, transmission transformers, sub-transmission lines, distribution transformers, distribution lines, distribution secondary transformers, service drops and meters. Technical losses within the electric system are a function of both currents in the system and voltage. Load losses or copper losses are the normal terms used to describe the current-related losses. Excitation losses in transformers, meters and reactors are called no-load, iron or excitation losses and are a function of the voltage at these different components. Corona losses are also in this category. Coronal losses depend on the voltage level and the length of the transmission circuits. The corona losses are not very significant in the OG&E system. Generator step up (GSU) transformers, depending on meter location, can contribute to the load and no-load losses of the system. The results of the technical calculation of system losses by the voltage levels on the OG&E system are shown in the Table Loss Allocation The total calculated losses for all categories, including the load and no-load components and the nontechnical losses, must be consistent with the FERC Form 1 reported energy losses. When there is a difference between the reported loss and the calculated study loss, as shown in Table 1, an allocation process must be conducted. The difference in losses was allocated to the calculated losses. The resulting adjusted losses are shown in Table 2 on the next page. 1

7 Electric System Loss Study Table 1 Calculated Losses and Non-Technical Losses System Function Calculated Losses for Year 2006 Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Calculated Technical Losses Transmission System Line and Transformer (Load) 161, , ,117,832 Line Corona ,811,152 Transformer No-Load 7, , ,396,512 Distribution Primary System Primary Transformer (No-Load) 4 kv Class ,322, kv Class 6, , ,136, Class ,144, kv Class 1, , ,622,768 Primary Transformer (Load) 4 kv Class 1, , ,945, kv Class 32, , ,801, Class 1, , ,699, kv Class 3, , ,518,719 Primary Lines 4 kv Class 6, , ,857, kv Class 88, , ,584, Class 4, , ,999, kv Class 7, , ,444,218 Distribution Secondary System Secondary Transformer (No-Load) 33, , ,826,995 Secondary Transformer (Load) 31, , ,948,473 Secondary Lines & Service Drops 134, , ,043,758 Customer Meters ,723,066 Non-Technical Losses Energy Diversion Unmetered Company Use 5, , ,523,977 Total Calculated Losses 530, , ,741,469,357 FERC Form 1 Reported Losses 1,957,267,000 2

8 Electric System Loss Study Table 2 Adjusted Calculated Losses Adjusted Calculated Losses for Year 2006 System Function Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Calculated Technical Losses Transmission System Line and Transformer (Load) 161, , ,117,832 Line Corona ,811,152 Transformer No-Load 7, , ,396,512 Distribution Primary System Primary Transformer (No-Load) 4 kv Class ,273, kv Class 9, , ,381, Class ,015, kv Class 1, , ,869,691 Primary Transformer (Load) 4 kv Class 2, , ,177, kv Class 46, , ,278, Class 1, , ,918, kv Class 5, , ,815,801 Primary Lines 4 kv Class 9, , ,565, kv Class 128, , ,330, Class 5, , ,062, kv Class 10, , ,867,712 Distribution Secondary System Secondary Transformer (No-Load) 33, , ,826,995 Secondary Transformer (Load) 46, , ,268,480 Secondary Lines & Service Drops 134, , ,043,758 Customer Meters ,723,066 Non-Technical Losses Energy Diversion Unmetered Company Use 5, , ,523,977 Total Adjusted Losses 614, , ,957,267,002 FERC Form 1 Reported Losses 1,957,267,000 3

9 Electric System Loss Study 1.3 Loss Factor Development For this study, Stone & Webster Management Consultants Inc. in partnership with Siemens Power Technologies (Siemens PTI), International developed several techniques to provide both the system loss study and the service level loss factors. Using Siemens PTI software and a representative sample of circuits, both transmission and distribution systems were modeled in detail to determine the technical losses. Energy losses were calculated using the load factor / loss factor methodology based on the peak losses and 8,760 hourly load research profiles. 1.4 Results The final work products for this loss study project included the following deliverables to OG&E: Technical loss study using engineering calculations based on OG&E voltage levels and system configuration Hourly demand factors based on 8,760 hourly load research data Monthly typical energy and demand loss factors by service level Seasonal energy and demand loss factors by service level Annual energy and demand factors by service level Final report and results with supporting tables Tables 3 and 4 summarize the results of the loss study by month and by service level, respectively. The remainder of this document describes the methodology utilized to develop losses and loss factors for the Oklahoma Gas and Electric System. 4

10 Electric System Loss Study Table 3 Losses by Service Level OG&E Losses by Service Level Test Year 2006 Energy kwh Losses kwh Loss Percent 1 TOTAL SYSTEM 27,942,194,067 Load No-Load Total 1,953,471, % Service Level 1 3,722,173,097 Load 80,341,184 No-Load 8,581,978 Total 88,923, % Service Level 2 3,867,856,408 Load 106,193,001 No-Load 27,049,148 Total 133,242, % Service Level 3 2,280,896,124 Load 101,155,581 No-Load 15,951,031 Total 117,106, % Service Level 4 716,913,655 Load 35,534,225 No-Load 16,789,041 Total 52,323, % Service Level 5 17,354,354,783 Load 1,149,742,760 No-Load 412,133,198 Total 1,561,875, % 1 Loss percentage is calculated as loss / energy in this table 5

11 Electric System Loss Study Table 4 Monthly Losses and Loss Factors OG&E Monthly Loss Factors Energy (kwh) Energy Losses (kwh) Energy Loss Percent Demand (kwh) Demand Losses (kwh) Demand Losses Percent Months January 2,220,507, ,198, % 3,732, , % February 2,128,846, ,504, % 4,002, , % March 2,157,171, ,860, % 4,892, , % April 2,177,976, ,347, % 4,892, , % May 2,544,458, ,010, % 5,649, , % June 2,876,975, ,147, % 5,500, , % July 3,329,420, ,892, % 6,418, , % August 3,433,204, ,111, % 6,473, , % September 2,410,606, ,054, % 5,168, , % October 2,202,006, ,853, % 4,940, , % November 2,093,776, ,021, % 4,402, , % December 2,372,261, ,272, % 4,257, , % OG&E Seasons Summer Jun-Oct 14,252,211, ,059, % 6,473, , % Winter All Other 15,694,995, ,214, % 5,649, , % Calendar Seasons Summer Jun-Aug 9,639,599, ,151, % 6,473, , % Fall Sept-Nov 6,706,388, ,929, % 5,168, , % Winter Dec-Feb 6,721,614, ,975, % 4,257, , % Spring Mar-May 6,879,605, ,218, % 5,649, , % 6

12 Electric System Loss Study 2 Introduction The OG&E peak demand in August of 2006 was 6,473 MW with an annual energy requirement of 28,492,417,000. Sales to the ultimate customers, excluding requirements and non-requirements sales, have been growing at a rate of about 2 per cent a year, compounded annually, from 1994 to Figure 1 shows this growth. Figure 1 Sales to Ultimate Customers 30,000,000 25,000,000 20,000,000 Energy MWh 15,000,000 10,000,000 5,000, Year Electric system losses as a per cent of the energy through the system have been relatively constant over the same period. In 2006 the energy loss was about 6.9 per cent. Figure 2 shows the energy losses over the same period. 7

13 Electric System Loss Study Figure 2 Losses as a Percent of Total System Load Percent Year Total energy losses are determined by taking the difference between all the known inputs to the system and all the known outputs. Inputs and outputs at the control area boundaries and generator outputs are normally recorded in KW every hour. These values can then be determined on a year-end to year-end basis. Most customer meters are recorded each month by meter reading methods that can only record about a twentieth of the meters each day in the month. Although each customer has an annual energy recorded, it is not based on a true calendar year consumption period. This usually results in meter reading billing cycling errors. The loss factor methodology was used in this study. This is a time proven methodology that should be based on accurately recorded energy. The issue with this is that not all the energy consumption can be measured by the end of the year as some energy used is not metered. Most residential and commercial meters measure continuous energy which must be manually read, usually once a month. The billing cycling errors distort the energy used in the calendar year. When the December and January weather patterns are similar from one year to the next the billing cycle error is minimized. There is another potential difficulty in that every load must be metered and recorded. A large error is introduced when one load is not metered or when there is an error in recording the energy. Capturing all the loads becomes very important. Additionally, energy diversion (stolen energy) and unmetered substation station light and power use must be estimated. 8

14 Electric System Loss Study OG&E rates are based on the voltage level of service to the customer. There are five service levels. Service level 1 (SL-1) is large industrial and full requirement customers. These customers have their own transmission transformers and are connected directly to the transmission system. The OG&E transmission voltages are: 500-kV, 345-kV, 161-kV, 138-kV, and 69-kV. Service level 2 (SL-2) customers are also connected to the transmission system but the connecting transformer is owned by OG&E and the meter is on the low voltage side. Losses become part of the OG&E transmission system in this situation. Service level 3 (SL-3) customers are connected directly to the distribution primary lines and the customer owns the transformer. In this case the customer is responsible for the transformer losses. The distribution primary voltages classes are: 34.5-kV, 24-kV, 12-kV, 4-kV, and 2.4-kV. Service level 4 (SL-3) customers are the same as SL-3 customer except that OG&E owns the transformer and is responsible for the losses in the transformer. These customers are normally small industrial and commercial. Service level 5 (SL-5) customers are residential and commercial. They are connected to the distribution primary lines by a transformer owned by OG&E with secondary lines to a meter. Secondary service voltages are: 480-V, 277-V, 208-V, 240-V, and 120-V The methodology used to allocate electric system losses to the system voltage levels and to the voltage service classes is based on the difference between the meter readings in and out at each voltage level. Total system energy losses are reported on page 401 line 27 of the FERC Form 1. There are eight categories that can be established to define and calculate specific losses that occur on the electric system. Within these categories there may be load and no-load losses. These categories, and the presence of load and/or no-load losses, are provided in Table 5. The demand and energy losses for the OG&E system were calculated for these eight categories and were allocated to the five service levels. Table 5 Categories of Load and No-Load Losses Category Load Losses No-Load Losses Transmission Lines Yes No Corona No Yes Transmission Transformers Yes Yes Distribution Substation Transformers Yes Yes Distribution Lines Yes No Distribution Secondary Transformers Yes Yes Service Drops Yes No Meters No Yes Finally, the non-technical losses were estimated to complete the calculation process. Energy diversion and substation station power and light are the two main non-technical losses. Loss factors for each hour of the year for each of the service levels were also calculated. 9

15 Electric System Loss Study 3 Transmission System Losses The OG&E transmission system consists of lines and transformers with voltage levels between 69 KV and 500 KV. Both the high and low voltage sides of transmission transformers must be within this range. The losses of those transformers with voltage levels on the low side below this range are considered part of the distribution system. Power and energy losses within the transmission system are due to the resistive component of the transmission lines, the resistive and admittance components of the transformers, and corona losses. Transmission system losses can be grouped into three types: transmission line losses, transformer losses and corona losses. 3.1 Transmission Line Losses Power flowing through transmission lines results in demand and energy losses. These losses are a function of the resistive component of the transmission line and the square of the current. The current flows through the transmission system from generators to the five service level loads and through tie lines, to and from OG&E s neighboring utilities. There are over 100 tie lines to nine control areas in place or planned in the next few years. The transmission line loss analysis was performed using Siemens PTI s computer program, Power System Simulator for Engineering (PSSTME) Revision 29. PSSTME is an integrated program for simulating, analyzing, and optimizing power system performance that uses the most advanced and proven methods for performing power flow studies, unbalanced fault analysis, and dynamic stability simulation. Five power flow cases representing different power flow conditions of the Eastern Interconnected System, which includes the OG&E control area and all of its neighboring utilities, was provided by OG&E. The five cases were for the year 2006 and represented the summer peak, summer shoulder, winter peak, spring peak, and fall peak load conditions with representative generation and tie flows for each case. The power flow cases represent only the transmission system. The control area hourly load data for the OG&E s transmission system in 2006 was also provided. From the original five cases, twenty-one power flow cases were developed by modeling system loads in increments of about 225 MW from the minimum to the maximum system load. In developing the cases, the original cases that most closely matched the new load were used. From the original cases, the system load was scaled up or down to match the new or desired case load. The generation was also scaled accordingly. Five extra cases were also created using the five power flow cases to create an overlap among the five cases. In total, thirty-two cases were run and the OG&E transmission losses were recorded for each of the cases. The resistances of the transmission transformers and the corresponding no load admittance data of the transformers were already included in the power flow cases. Therefore, the load and no-load losses of the transmission transformers were automatically included in the transmission losses. The power flow simulations performed for the thirty-two cases studied provided the transmission demand losses for each of the corresponding system load levels. A mathematical relationship between the 10

16 Electric System Loss Study transmission system losses and the system demand was found using regression analysis. The resulting relationship was used to calculate the transmission losses corresponding to the system load for each hour of the year. The plot of the transmission loss vs. system demand and the curve for the estimated mathematical relationship are shown in Figure 3. The correlation coefficient R2 was found to be indicating a reasonable curve fit. The data points at the higher end of the demand scale show the variances in the provided power flow cases. These variances correspond to the control area boundary flow differences and the internal generation differences that swing the losses up or down from the regression curve. The curve as plotted shows that, in some instances, the near same load levels produce somewhat different losses. The differences occur as a result of transitioning from one case to another (i.e. winter peak to summer peak). The transition cases have different generation dispatches, different import and export schedules and loop flow differences. Figure 3 Transmission System Line Losses y = x R 2 = Losses (MW) Demand (MW) 3.2 Transmission Transformer Losses Transformers have two main types of losses. The first is the resistive loss (load-losses) which is a function of the transformer resistive component and the current squared. This loss depends on the loading of the transformer. The second is the no-load loss (excitation loss), which is a function of the transformer 11

17 Electric System Loss Study admittance and the operating voltage squared. All transformers on the electric system have these two types of losses. The no-load loss is constant and it is always present as long as the transformer is energized. The loss is in the form of heat energy and noise. Both the load and no-load losses were included in the power flow case, so there was no need to calculate them separately. Generation step-up (GSU) transformers are only included in the transmission category if the metering is located on the generation side of the transformer. The meter location for the OG&E generation step-up transformers are on the high side, which is the industry standard. Therefore, the losses in the GSU transformers are already taken into account in the generation plant efficiency, so they were not calculated. In this case, the GSU transformer losses, both load and no-load losses, are part of the cost associated with the net energy output from each generating unit having this metering arrangement. 3.3 Corona Losses Corona loss is an electric discharge to the air that surrounds a conductor. Under the proper weather conditions, the air surrounding the conductors of high voltage transmission lines becomes ionized and conducts electricity to a limited extent. As a result, a small part of the electric energy flowing in the transmission line leaks into the air resulting in electric loss. The amount of the corona discharge depends on the voltage level, the diameter of the conductor and the weather conditions. Other factors affect the corona discharge, such as, adverse weather conditions, elevation, conductor spacing, and the presence of a shield wire. Rain also increases corona loss substantially. Corona demand losses are calculated separately for the 500-kV, 345-kV, 161-kV, 138-kV, and 69-kV transmission lines using the Bonneville Power Administration computer program, CORONAII, Corona and Field Effects. Corona loss is negligible for voltages that are 69-kV and below in fair weather conditions. OG&E had about 4,576 miles of lines at 69-kV and above within its control area in Table 6 shows the length of the lines of the transmission system and the resulting demand and energy corona losses by voltage level. The losses are based on normal conditions for most of the hours in a year and for an average of one-half inch of rain for the 137 hours per year according to local publicly available sources. The demand is assumed at time of normal conditions. 12

18 Electric System Loss Study Table 6 Corona Losses Lines kv Length of Circuits Miles Loss No Rain kw/mile Loss with Rain kw/mile Hours of Rain Hours Demand Losses kw Energy Losses kwh 69 1, , , , ,536, ,194,386 Total 4, ,811,152 Coincident Demand Transmission System Loss Summary The transmission losses are summarized in Table 7. The calculating techniques used in the calculation of transmission losses are very straight forward and are performed with a high degree of accuracy. Therefore, there is great confidence in the accuracy of these calculations. As a consequence, during the allocation process that was performed to match the reported losses to the calculated losses, the transmission system losses were not modified. System Function Table 7 Transmission System Losses Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Line and Transformer (Load) 161, , ,117,832 Line Corona ,811,152 Transformer No-Load 7, , ,396,512 13

19 Electric System Loss Study 4 Distribution Primary Losses Demand and energy losses have been calculated for the distribution primary system. Included in this category are the load and no-load losses in the distribution primary transformers and the losses in the distribution primary lines. Customers with rates associated with service levels 3 and 4 are connected to the distribution primary system. Distribution primary transformers have low side voltages in the following voltage classes: 34.5-kV, 24-kV, 12-kV, and 4-kV. The actual low side voltage varies within a small range around the class voltages with the exception of the 2.4-kV low side voltages which were also included in the 4-kV class. Distribution primary lines operate at the same voltages mentioned above. The non-coincident peak losses were calculated for transformer load losses and primary line losses. Energy peak losses were calculated form the peak losses using the loss factor methodology. On the other side, transformer no-load losses are voltage dependent and, as a result, are relatively constant. No-load transformer losses were calculated from the nominal voltages and the transformer no-load parameters. OG&E maintains a sophisticated load research program that enables the load and loss factors to be calculated directly from the data without having to use the empirical formula methods. The load factor is calculated by taking the average of the 8,760 hourly loads in the test year. If the loads are in per unit the loss factor is calculated from the square of the per unit loads for the same 8,760 hours. The per unit (pu) load is any individual load divided by the peak load in that study period. For example, at the time of the peak, the per unit load would be 1.00 pu, the highest value in the set. 4.1 Transformer Load Losses Transformer load losses are associated with the current flowing through the transformer. The peak load losses are a function of the square of load current through the transformer at the time of the noncoincident peak. The OG&E SCADA system (System Control and Data Acquisition) provided the peak load for many of these transformers. These peak loads were used to calculate the load losses using typical transformer resistance in per unit values. The average loading for those transformers having recorded load information was also determined. This average loading was then used as the peak load for those transformers with no historical loading information. The calculation of the load transformer losses is a simple per unit calculation that uses the square of the load multiplied by the resistance, thus providing a non-coincident peak loss. Non-coincident peak loss is used with a loss factor and 8,760 hours in year to determine the annual energy loss. It was assumed that a diversity factor at this level was 5 percent. Tables A-1 through A-4 in the Appendix show the non-coincident load losses for each individual transformer. 4.2 Transformer No- Load Losses The no-load losses are associated with the winding excitation and are a function of the square of the applied voltage. The voltage was assumed to be constant for the calculation of the no-load losses. Transformer no-load losses have a relatively small variance when converted to per unit based on the Oil to Air (OA) transformer rating. Therefore, if manufacturer test values were not available, typical values were used. The OA rating is the lowest rating given to a transformer. The OA rating is the most basic cooling rating, as there are no oil pumps to circulate the oil, and cooling fans are offline so only natural convection occurs. 14

20 Electric System Loss Study The no-load losses for the distribution primary transformers are shown on Tables A-1 through A-4 in the Appendix. The demand no-loss that is calculated is the coincident peak loss. To calculate the no-load energy losses the coincident peak loss is multiplied by the 8,760 hours in the period. 4.3 Summary of Distribution Primary Transformer Losses The distribution primary transformer losses are summarized in Table 8 below for every voltage class. Table 8 System Function Distribution Primary Transformer Losses Non-Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Primary Transformer (No-Load) 4 kv Class ,322, kv Class 6, , ,136, Class ,144, kv Class 1, , ,622,768 Primary Transformer (Load) 4 kv Class 1, , ,945, kv Class 32, , ,801, Class 1, , ,699, kv Class 3, , ,518, Distribution Primary Distribution Lines The procedure used to calculate the primary distribution line losses is described below. Given the high number of primary lines in the OG&E system, it was not practical to perform detailed loss calculations for each of the primary distribution circuits. Therefore, the detailed loss calculations were performed for a representative sample of circuits. Typical circuits for each of the four voltage classes were selected by OG&E. The following number of circuits was studied: Eight 4-kV circuits, sixteen 12-kV circuits, six 24-kV circuits, and ten 34.5-kV circuits. The circuits selected were modeled using the Siemens PTI computer model PSS/E Adept. The data for these circuits were taken from previously modeled circuits that OG&E had set up on their own distribution program Cyme. The data included: Conductor length, type, phasing (A, B, C, AB, BC, AC, and ABC), loads by phase, and capacitors. Total circuit load was scaled for each circuit to match the SCADA system recorded non-coincident peak loads on that circuit. The circuit losses were calculated by the PSS/E Adept computer program. A regression analysis was performed on the losses as calculated for each distribution voltage classes. The resulting equations were used to estimate the losses for any particular circuit loading. The equations were applied to all circuits in their class to determine each circuit s non-coincident demand loss. Energy losses were calculated from the non-coincident demand using the loss factor. The coincident demand was determined using the calculated loss at the system peak hour. 15

21 Electric System Loss Study 4.5 Primary 4-kV Lines The 4-kV voltage class customers have a population of 125 circuits at 4.16-kV and 4 circuits at 2.4-kV. Eight of these circuits were used and modeled as a representative sample of the total population. Table 9 shows the resulting losses from these 8 circuits and the percentage based on the circuit peak demand. Table 9 Circuit Number Circuit Loading KVA Modeled 4-kV Circuits Circuit Losses KW Percent Losses % , , , The regression analysis was performed using linear, power, and exponential equations. The equation selected from these three choices was the exponential equation even though the calculated correlation coefficient R 2 was relatively low. The low correlation coefficient results from the wide variation in calculated losses as a function of the circuit peak load. Figure 4 shows the graph of the loss results and the exponential equations curve. 16

22 Electric System Loss Study Figure 4 Circuit Loss Results and Exponential Equation for 4-kV Circuits y = e x R 2 = Losses KW ,000 1,200 1,400 Circuit Loading KVA 4.6 Primary 12-kV Lines The 12-kV distribution primary lines are the typical voltage with 742 circuits. Sixteen of these circuits were selected to represent the total population. The selected circuits were modeled and the losses were calculated. Table 10 shows these circuits, the losses and the percent losses. The losses range from 1.00 percent to 2.44 percent. 17

23 Electric System Loss Study Table 10 Modeled 12-kV Circuits Circuit Number Circuit Loading KVA Circuit Losses KW Percent Losses % , , , , , , , , , , , , , , , , A regression analysis of the losses calculated using the computer model was performed to determine an equation that could be used to calculate the losses for the entire population of 742 circuits. Circuit loss as a function of circuit peak load in kva was used. The correlation coefficient was again low (though a better fit than for 4-kV) because of the variation in the calculated losses as a function of circuit load. Figure 5 shows the losses for the sixteen circuits and the equation that represents the 12-kV circuits. 18

24 Electric System Loss Study Figure 5 Circuit Loss Results and Exponential Equation for the 12-kV Circuits y = e x R 2 = Losses KW Circuit Loading KVA 4.7 Primary 24-kV Lines There are 53 distribution circuits at 24-kV on the OG&E system. Six of these circuits were selected for detailed study. Losses were calculated as a function of the circuit peak load. The results of these six circuits are shown in Table 11. Note that circuits and have very similar loads and losses. This similarity results in two data points on top of each other on the graph making it appear to have only five entries. 19

25 Electric System Loss Study Table 11 Circuit Number Modeled 24-kV Circuits Circuit Loading KVA Circuit Losses KW Percent Losses % , , , , , , A regression analysis of the losses calculated using the computer model was performed to determine an equation that could be used to calculate the losses from the 53 circuits. Circuit loss as a function of circuit peak load in kva was used. This correlation coefficient was also affected by the variation in the calculated losses as a function of circuit load. Figure 6 shows the losses for the sixteen circuits and the equation that represents the 24-kV circuits. Figure 6 Circuit Loss Results and Exponential Equation for the 24-kV Circuits Losses KW y = 6.924e x R 2 = Circuit Loading KVA 20

26 Electric System Loss Study 4.8 Primary 34.5-kV Lines The 34.5-kV distribution circuits have the heaviest loads. Ten of these circuits were selected to be modeled to represent the total population of 57 circuits. Table 12 shows the modeled circuits. Table 12 Modeled 34-kV Circuit Loads and Losses Circuit Number Circuit Loading KVA Circuit Losses KW Percent Losses % , , , , , , , , , , The exponential equation was selected to represent the 34.5-kV distribution circuits. The variation of the losses as a function of load was the reason why the correlation coefficient was relatively low. Figure 7 shows the losses for the ten circuits and the equation that represents the 34-kV circuits. 21

27 Electric System Loss Study Figure 7 Circuit Loss Results and Exponential Equation for 34-kV Circuits y = e x R 2 = Losses KW Circuit Loading KVA 4.9 Summary of Primary Distribution Line Losses Table 13 summarizes the losses in the primary distribution lines. Table 13 System Function Losses in Primary Distribution Lines Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Primary Lines 4 kv Class 6, , ,857, kv Class 88, , ,584, Class 4, , ,999, kv Class 7, , ,444,218 22

28 Electric System Loss Study 5 Distribution Secondary System Distribution secondary transformers, distribution secondary lines, distribution service drops and customer meters comprise the distribution secondary system. Distribution secondary transformers have a load and no-load loss component. The meter losses are considered to be caused by excitation losses and, therefore, are quantified as no-load losses. SL-5 customers are connected to the distribution secondary system. This service class has the most customers and their demand and energy contribution to the system is the most significant. 5.1 Distribution Secondary Transformers Distribution secondary transformers on the OG&E system range in size from 1 kva to 5,000 kva. In 2006 OG&E reported 176,530 installed units on the system with a total capacity of 10,904,204 kva. The average size of secondary transformers was 71 kva in 2006 and the most frequently installed size was the 25 kva unit for the overhead system and the 50 kva unit for the underground system. As with all transformers, there are load loss and no-load loss components. No-load losses were developed using typical loss characteristics for each size of transformer. These typical no-load per unit values are shown in Table A-6 in the Appendix. The no-load demand was calculated by using the typical no-load per unit value for each transformer size and multiplying it by the number of transformers in that size category. The energy no-load losses were found by multiplying the demand no-load losses by 8,760 hours for the year. The coincident demand load loss is a function of the current squared (current is proportional to load) at time of system peak, but this is not necessarily the maximum demand on the transformer. Based on the peak demand from the load research information from service levels SL3, SL4, and SL5, it was found that the average distribution secondary transformer is loaded to 32 percent. This is the loading from which the demand must be calculated. This is an average loading and because the calculation of loss is a square function, it is not correct to simply take the 32 percent loading to calculate the losses. The proper method is to calculate each transformer s demand using the loading of each transformer but this individual loading is not known. An approximation method is to create a frequency distribution that will average to 32 percent loading for each transformer size, but will capture the various loadings above and below the average. This was accomplished by using the average load on 60 percent of the transformers, a loading of 120 percent of average on 20 percent of the transformers and a load of 80 percent on 20 percent of the transformers. The energy load losses were calculated from the non-coincident demand load losses by using a loss factor of and 8,760 hours. The load factor was increased to 0.3 during the allocation process to balance against total energy losses. A coincidence factor of 85 percent was used to calculate the coincident peak loss. The secondary transformers load and no-load losses are summarized in Table

29 Electric System Loss Study Table 14 System Function Secondary Transformer Losses Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Secondary Transformer (No-Load) 33, , ,826,995 Secondary Transformer (Load) 31, , ,948, Distribution Secondary Lines and Service Drops Losses that occur on the secondary lines and service drops are the most difficult to calculate because of the lack of data and the sheer number of secondary lines and service drops. Information as to the configuration, conductor size, and length of each of the services to customers are not kept on a drawing because a large number of drawings would be required and because a distribution standard is used. Each customer s electric service installation is slightly different than a standard. Based on the OG&E standards, 15 different secondary and service drop configurations were used with 5 different loads each. The customer load was assumed to be un-balanced for the 240/120 volt configurations with 50 percent of the load on one leg, 40 percent on the other leg and 10 percent on the neutral. The non-coincident demand losses were calculated based on these 75 different loads and configurations (15 service drops, 5 loadings, with unbalance configurations). It was assumed that the coincident factor was 85 percent. Energy losses were then determined using a loss factor of and 8,760 hours. The loss factor was increased to 0.3 during the allocation process that was performed to match the calculated values to the recorded energy loss. The losses in the secondary lines and service drops are summarized in Table 15. Table 15 System Function Secondary Lines and Service Drops Losses Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Secondary Lines & Service Drops 134, , ,043, Customer Meters Losses can be attributed to each meter on the electric system. The standard residential meter takes just under one watt of energy for each hour of operation. The demand loss for electric meters is calculated by taking the number of meters times the hourly losses. The coincident and non-coincident demands are the same. The energy losses are calculated by multiplying the demand loss by 8,760 hours in a year. The resulting electric meter energy losses are shown in Table

30 Electric System Loss Study System Function Table 16 Customer Meter Losses Non- Coincident Peak Losses KW Coincident Peak Losses KW Energy Losses KWh Customer Meters ,723, Non-Technical Losses There are two main components that make up energy that is unaccounted. These two components are energy diversion and Company unmetered use. Energy diversion is the term used to describe energy that is stolen by customers tampering with the meter or bypassing the meter. Energy diversion in the United States is very small. In discussions with OG&E, it was determined that energy diversion was not a problem in their service territory. The only company unmetered use that is believed not to be in the calculated losses is the substation station, light and power. An estimate of the non-coincident demand for this use is based on a 25 kva transformer at the 362 substation loaded to 60 percent of capacity. This yields a non-coincident peak load of 5,430 kw. The coincident peak is estimated at 90 percent of the non-coincident value and is 4,887.0 kw. Energy is calculated using a 45 percent utilization factor yielding 21,523,977 kwh. 25

31 Electric System Loss Study 6 Hourly Load Allocation Procedure Technical losses were calculated using OG&E data for eight categories including, transmission lines, transmission corona, transmission transformers, distribution substation transformers, distribution lines, distribution secondary transformers, service drops, and meters. Summing the energy on these eight independently calculated systems should approximate the total energy loss determined by taking the difference between the inputs to the system and the sales. The calculating methods use statistical approaches for the calculation of the losses of these eight systems. This usually results in differences between the recorded annual energy losses and the annual calculated values. Therefore, the loss difference needs to be allocated back to the calculated values so that the sum of the eight categories is equal to the recorded difference. After the sum of the annual losses has been allocated to match the recorded annual losses, it is necessary to determine the losses at each hour for each service level. The loss factor at each service level is a function of the entire load that passes through the service level or system element such as the transmission system. This is accomplished by using the load research service level 8,760 load data. Losses at the transmission level are the sum of all five service levels. The combination of these five service level load shapes is determined by adding each specified hour. The sum in any one hour is the load that results in losses on the transmission system. The loss factor is the per unit square of this load data set. If the demand loss, which is a function of the square of the current, is multiplied by each hour, the annual energy loss is determined. No load losses that are constant in each hour are added separately. SL-4 losses are determined using the data sets of SL-2, SL-3, SL-4, and SL-5. This is the load that goes through the distribution transformers. The same process as described above is used to determine the hour by hour losses. Likewise the service level 3, the primary distribution lines follows the same procedure as is that used for service level 4, distribution secondary transformers and service level 5, distribution secondary lines and service drops. 26

32 Electric System Loss Study 7 Conclusion The loss study presented in this document relies on standard industry approaches and the best available data produced by OG&E for use in the analyses. The resulting losses are in line with previous studies for the OG&E system and reflect changes that occur over time: differing service territory characteristics, changes in standard equipment in use by the utility, degradation of equipment not yet ready for replacement and weather conditions. 7.1 Historical Losses A comparison of historical system losses (using percentages) is provided in Figure 8. Figure 8 Historical Losses as a Percent of System Load Percent Year 7.2 Loss Study Comparisons There are two previous studies that calculated losses on the OG&E system, one with a test year of 1992 and the other with a test year of The present study has a test year of The 1992 study used a different methodology to determine losses. The 2001 and 2006 studies used similar methods for the most 27

33 Electric System Loss Study part, but the data was, of course, five years newer and had more advanced collecting and recording techniques. Table 15 shows the comparison of the latest study with previous loss studies. Table 17 Energy Loss Percent Comparison of Previous Loss Studies OG&E System Losses Comparison of Loss Studies 1992 Study 2001 Study 2006 Study Demand Energy Demand Energy Loss Loss Loss Loss Percent Percent 2 Percent Percent Demand Loss Percent Service Level % 1.10% 2.48% 2.47% 2.39% 2.73% Service Level % 2.54% 3.14% 3.30% 3.44% 6.23% Service Level % 3.88% 4.67% 5.81% 5.13% 10.14% Service Level % 5.79% 6.95% 7.25% 7.30% 10.29% Service Level % 7.52% 8.41% 9.29% 9.00% 11.11% The difference in the demand losses has been progressively higher from the first study to the last study as load is attached to the electric system farther away for the generators. This is because the hour by hour method of calculating losses can more accurately determine losses, both peak and energy. 7.3 Methodology For this study, total energy losses are determined by taking the difference between all the known inputs to the system and all the known outputs. Inputs and outputs at the control area boundaries and generator outputs are normally recorded in KW every hour. These values can then be determined on a year-end to year-end basis. Most customer meters are recorded each month by meter reading methods that can only record about a twentieth of the meters each day in the month. Although each customer has an annual energy recorded, it is not based on a true calendar year consumption period. This usually results in meter reading billing cycling errors. The loss factor methodology was used in this study. This is a time proven methodology that should be based on accurately recorded energy. The issue with this is that not all the energy consumption can be measured by the end of the year as some energy used is not metered. Most residential and commercial meters measure continuous energy which must be manually read, usually once a month. The billing 2 Loss factor in this table is calculated as Loss / Energy 28