Cost of supply study for the city of Ekurhuleni: a case study

Transcription

1 Cost of suppy study for the city of Ekurhueni: a case study by Stephen Deport, City of Ekurhueni and Dr. Johan Deport, Eectrica Energy Management Systems The City of Ekurhueni (COE) has been updating its set of tariffs on a yeary basis. The cost of suppy study project was initiated by Mark Wison, the head of the energy department. This was a ground breaking project in that it was one of the first detaied cost of suppy study competed for the COE. Wison recognised that, in order to be successfu, a modern eectricity utiity has to compete a cost of suppy (CoS) study in ine with the principes of NRS 058 which prescribes how to reguate energy prices to different customer casses whie being fair and equitabe and at the same time yieding the expected budgetary income per annum. To achieve this, the foowing specific objectives must be addressed: Understand the cost of eectricity, from purchase to consumption, incuding technica and non technica osses. Understand the cost of equipment, from new equipment to repacing existing equipment. Understand the cost of support (peope), from eectrica empoyees (maintain and expand), to support functions e.g. financia, ega empoyees (biing). Understand the cost of cross-subsidisation, socia responsibiities and prescribed price brackets by NERSA. Reference to NRS 058 The NRS 058 document is used as guide to this study. Therefore, there wi be constant referencing, quotes and direct text, tabes and pictures used from the document. By no means do the authors caim the NRS 058 as their own work. Referencing to the NRS 058 document wi be per paragraph number in brackets e.g. (NRS 058:3.4.2). As far as possibe, the aim of the mode is to keep it as simpe as possibe, without osing credibiity and accuracy to ensure the foowing objectives are met as documented in the NRS document. The tota cost of suppy is given by the equation: (see eqn. 1 beow). Where: Cost of suppy: The tota cost of suppy in per tariff group Power cost: The tota cost of energy purchases, incuding technica osses in Parts cost: Capita/maintenance cost apportioned per tariff group in Eqn. 1: Cost of suppy cacuation. Peope cost: Cost of the eectrica department cost in apportioned to tariff group Pain cost: The cost of non-technica osses apportioned per tariff group in Profit: Average contributing to the surpus is cacuated as the panned surpus amount divided by the energy sod NRS 058 assumptions NRS 058 provides cear guideines to conduct a cost of suppy study. Some trade-offs and assumptions however had to be made. Consumption categories (NRS 058: ) The foowing different consumption categories proposed by NRS 058 are used in the study: Residentia Commercia Industria Reseing Losses: - Technica - Non-technica Votage/geographica ocation of the suppy to a customer (NRS 058:4.1.4) Due to the size of the COE suppy area it was decided not to appy any geographic ocation factors or differentiation. Athough the votage eves used for this study are not the same as the NRS 058, it was agreed that the votage eves in Tabe 2 wi be used in the study. These votage differentiations are aso refected in the existing energy rates of the buk consumption tariffs. These percentages differentiation in energy rates due to votage eves were used as the first iteration of technica osses due to votage eves. To date COE have used a fixed vaue of 5,9% for technica osses based on previous high eve engineering studies. A need has arisen to test this vaue for adequacy using engineering modes. Technica osses are due to energy dissipated in the eectrica components that make up the eectrica distribution system and predominanty occur in the transmission/ distribution ines, cabes and transformers. Abbreviations Description COE FBE HV IMS LV MOL MV NAC NRS 058 POD SUPRIMA VENUS City of Ekurhueni Free basic eectricity High votage 22 kv 132 kv Inteigent metering system Low votage 400/230 V Metering onine Medium votage 6,6 kv to 11 kv No access charge Cost of suppy methodoogy for appication in the eectrica distribution industry Point of deivery (normay Eskom suppy point) Prepaid vending system used by the city of Ekurhueni Biing system used by the city of Ekurhueni Tabe 1: Abbreviations used. Auxiiary equipment used for metering and protection devices require energy that are not recovered by the direct biing. The network mode in the PowaMaster software was used to cacuate the theoretica technica osses reated to the COE network. A significant amount of work was required to enhance the PowaMaster mode to incude the compete medium votage network as we as a the MV/ LV transformer points. The Aberton network was seected as the network to be enhanced for detaied technica oss cacuation. The technica oss cacuation approach in PowaMaster was broken up into a number of steps: Prepare the mode in PowaMaster by adding eectrica oads to the network at each MV/LV transformer suppy point in the network by interconnecting the oads with sections of cabe. Setup and scae the oads in PowaMaster to refect the oads as recorded in the metering onine (MOL) data. Create an Exce integration too to automaticay extract the osses resuts from the PowaMaster cacuation resuts to report in the technica osses consumed in the eectrica network th AMEU Convention 2016

2 Create a oad sweep automation in Exce to automaticay generate a set of technica oss resuts for changes in network oad. To adjust for higher osses at ower votages Tabe 3 was used to distribute osses at different votage eves. Using this tabe, the tota osses is equa to the 4,46% cacuated theoretica technica osses. A the technica osses are differentiated by votages ony. From Tabe 3 it can be seen that ower oss factors are appied at higher votages. The geographica ocation and the distance from the point of suppy is not part of the mode. It is assumed that COE is in one distance zone. A 55 of the point of deiveries (PODs) are summated and used as one big POD. The foowing adjustments were made when the cost of the suppy of eectricity was cacuated: Number of PODs for administration and service charges. Diversity between PODs when cacuating the demand cost. Constructing oad profies per tariff group One of the main focus areas was the construction of a disaggregated oad profie for the COE energy baance. To cacuate the cost of suppy for each tariff component it is important to have the contributing (disaggregated) oad profie of each tariff group. This profie wi be used to cacuate each tariff group s specific cost components based on the profie. Avaiabe data Eskom data was received and oaded into a MySQL database. Queries were run against the data to aggregate the tota suppy at various Eskom PODs. This was compared to interna COE summaries and agreed we. Metering onine (MOL) provided a arge data dump of a interva meters. Data was anaysed and a the tariff groups were identified. Queries were run against the database to obtain the oad profie for each of the tariffs groups. Data for the study have been extracted from IMS and SUPRIMA, oaded in the MySQL database. Queries were run against the data to obtain tota monthy energy consumption and FBE figures. The data of the foowing consumers/ consumer group s data were received from COE as monthy figures: Conventiona residentia and conventiona commercia as a group. The initia idea was to run queries against the Venus system data. A CCC's streetights and masts ighting and traffic ights. A spreadsheet with equipment counts were used to derive a oad profie. Departmenta civic centre use. Construction of oad profies and energy baance The main purpose of this anaysis is to produce haf-houry oad profies per tariff group and votage. There are currenty ±45 variations on the current ±8 COE tariffs. At this point the avaiabe data is used to do a first iteration of comparing the haf houry data avaiabe with the summary of COE. The data compared favouraby with the COE summaries and therefore the oad profies obtained from the raw anaysis have been used in the study. The process foowed was to summate a the Eskom PODs and obtain: Haf-houry data. Sum of the maximum demand of each POD. Contribution of each PODs maximum demand to the maximum demand of Eskom. This diversity was used when cacuating the tota cost of suppy. The maximum demand was compared with MOL data: Use a the meters of the PODs. Add the haf houry data of a the PODs together and get the monthy maximum demands (MD sum of PODs). The MOL data was anaysed and the foowing was obtained per tariff group: Haf-houry data per tariff group. Maximum demand of each tariff group. Sum of the maximum demand of each tariff group. Add the monthy maximum demand of a the PODs together (sum of MDs). This is from the COE data suppied in sheet Eskom. Contribution of each tariff group s maximum demand to the maximum demand of Eskom. The origina objective was to obtain the conventiona residentia and commercia readings from Venus. Due to the form of the information in Venus, this was not possibe. This tariff group is spit into two casses with information received from COE: Residentia: 77% of tota conventiona meter readings. Commercia: 23% of tota conventiona meter readings. The foowing information is used to construct the haf houry oad profies of the streetights and high masts: From the data in the oad is cacuated as kwh per 30 minutes. In other words, a maximum demand of kw. The power factor used, from the sheet, is 0,87. 65th AMEU Convention

3 6 5thA M E U C on v e n t i o n Leve Low votage (LV) is generated, adding the reevant osses to each current oad profie. Description Votages between 230 V, singe phase and 400 V, three phase and 1 kv (incuded), three phase. From POs: COE aso has a tariff differentiation on the 400 V eve, Tariff C and Tariff E 400 V substation suppy - Use a the meters of the Eskom and City Power PODs Medium votage (MV) Votages higher than 1 kv up to and incuding 11 kv High votage (HV) Votages higher than 11 kv. - Add the haf houry data of a the PODs together. Tabe 2: Votage differentiation. From MOL data: - Use a the meters of customers % differentiation in energy rates due to votage eves Previous years % was 400 V 8,65% 12% 400 V and in SUB 5,60% 10% >400 V and 11 kv 3,50% 8% >11 kv 0,00% 0% Votage eve Tabe 3: Appied oss factors per votage eve. - Add the haf houry data of a the customers together, according to each tariff group and appied osses factor per votage eve. From COE data: - Where there was no haf houry data, the constructed oad profies, described above are used. Resuts of the energy baance Disaggregated oad profie For interest sake an exampe view of the disaggregated oad profies for the peak week in a winter month is shown beow. A 42 tariff variants are show together with the technica and non-technica osses and therefore are very difficut to show. This adds up to the tota purchased. Fig. 1: Grand tota profie for a PODs within COE. Winter on time for the ights is from 18h00 to 06h00. Summer on time for the ights is from 19h00 to 05h00. These assumptions sometimes created spikes, or dips in the non technica osses profie. The resut is then scaed to get the a CCC's prepaid + FBE tota kwh per month. This assumption created spikes, or dips in the non technica osses profie. If this assumption created a spike, or dip, that day was fitered with the previous week, same day, data, or next week, same day, data in the origina Greenfied Aberton feeder data from MOL. The foowing information is used to construct the haf houry oad profies of the traffic ights: From the data in the oad is cacuated as 1358 kwh per 30 minutes. In other words, 2716 kw. The power factor used, from the sheet, is 0,87. On time for the traffic ights is taken as 24 hours a day. The departmenta civic centre etc. haf houry profie is buit up with the foowing tariff group: The information suppied is used to cacuate the technica osses per tariff group, according to the suppy votage. The direct suppy from Eskom group was put into the suppy votage group <=11 kv. The foowing considerations were appied when cacuating the osses: Ekurhueni tariff B bus 150 A 3 phase. The resut is then scaed to get the departmenta civic centre etc., tota kwh per month. The tota osses is the difference between the tota purchases and the sum of the constructed oad profies. The a CCC's prepaid + FBE haf houry profie is buit up with the foowing tariff group: The tota osses is the sum of the technica osses and the non technica osses. Greenfied Aberton feeder data from MOL. The more the technica osses, the ower the non technica osses and vice versa. For each tariff group, a new oad profie 14 In the winter osses comparison above the eft axis indicates the scae of osses and the right axis the Eskom purchase totas. It can be seen that the non-technica osses are more than the technica osses. The technica osses foow the form of the Eskom graph. The whoe equation baances on the determination of the non-technica osses. This means that when a the various oad profies incuding cacuated technica osses is subtracted from the Eskom purchase totas the non-technica osses vaue is determined. Zooming in coser the osses profies can be compared with the tota Eskom purchase. High-eve energy baance The resuts of the mode (using raw data) for the purchased energy per month and for the year when compared the energy summaries that was provided by COE is ess than 0,1%, which indicates that there is a high eve of confidence in the interna COE energy totas. The resuts of the mode (when using raw data) for the purchased demand per month and for the year when compared to the interna COE totas is 0%. Athough there are monthy differences the average resut for the year is the same for the mode and the COE summaries. This is expected, because the average cacuated diversity is used by the mode. There are some differences per month. This can be addressed by using the monthy diversity factors in the mode. 65th AMEU Convention 2016

4 Premset MV switchgear Because a reiabe network depends on safe, fexibe, and maintenance-free switchgear Switchgear designed for enhanced peace of mind. Because business reies on the avaiabiity of eectricity, buidings need their medium votage distribution systems not ony to be reiabe, but aso to be energy efficient, durabe, and abe to adapt to changing business needs. But the operators of these systems require more. Peace of mind is paramount, and can ony be achieved with ow-maintenance switchgear that heps ensure the safety of both peope and assets. Switchgear that enabes monitoring and owers the tota cost of ownership is critica.

5 6 5thA M E U C on v e n t i o n The annua average cost is the sum of the monthy costs divided by the sum of the monthy consumptions, or energy sod, per tariff group. It is important to remember that this average cost per month is cacuated as if the tariff group received its energy at the intake point. In other words, technica osses incuded. The purchase cost is the average cost per kwh that COE pays Eskom for each kwh that is procured to se on to a specific tariff group and incudes the group s contribution to maximum demand as we as osses. It is very important to remember that for the annua average cost per tariff group the energy sod is used. The amount that the tariff groups costs COE is cacuated (energy sod and technica osses incuded). This amount is then divided by the energy sod (ony), in other words, the technica osses energy is excuded. Now this amount can be used to cacuate the cost of energy sod to a tariff group. Fig. 2: Winter peak week profie disaggregation. Capita cost per tariff group (parts cost) The asset ist was received from COE. This was summarised into the votage categories as discussed previousy. The category a other was added, consisting of non HV, MV and LV reated equipment but is needed to run and operate the system. Using this cassification, the tota asset vaues were distributed to votage eves as indicated beow. A methodoogy was deveoped to distribute a portion of the tota assets to the cost of suppy study. The study aims to cacuate at each votage eve a that needs to be recovered to cover capita costs. The foowing assumptions were used: A tariff groups need the a other. They wi share in the cost. A tariff groups need the HV part. It is apportioned per kwh consumed per tariff group in each votage eve. Fig. 3: Losses and purchase comparison: peak week in Juy High-eve osses comparison per tariff group can be cacuated as foow: With the modeed oad profies per tariff group, the osses can be cacuated. The technica osses are apportioned with regard to the votage eve, according to the tariff differentiations. It is recognised that there are more ways in which technica osses can occur, but for this study, ony the votage eves are used. The energy figures in kwh resuts are in the next tabe. The maximum demand of each month is determined. For each tariff group, the haf houry consumption, incuding the technica osses, is used to cacuate the eectricity cost. A MV and LV tariff groups need the MV part. It is apportioned per kwh consumed per tariff group in each votage eve. A LV tariff groups need the LV part. It is apportioned per kwh consumed per tariff group in each votage eve. For cacuating the contribution to cost of assets, a tota energy is used excuding technica and non technica osses. For a first iteration 7% of the asset repacement vaue wi be used for the cacuation. That is R Each tariff group s contribution to the maximum demand of each month is determined. The Eskom eectricity tariff is appied on the oad profie of each tariff group, with the maximum demand vaue as the contribution to the tota maximum demand. With higher technica osses one can expect a ower non technica oss percentage. This is given in the next tabe, with the tota osses for the year. Contribution to eectricity cost per tariff group (power cost) With the disaggregated oad profies per tariff group, the average purchase cost () The tota cost per month is divided by the tota consumption per tariff group to get the monthy average cost in per tariff group per month. This is the consumption ony, in other words, kwh that was sod. 16 When a figure of 7% is appied to the tota asset vaue tabe above the foowing vaues are obtained. What this means is that a tota asset vaue of R1,703-biion woud have to be accounted for per year as a for every unit sod. 65th AMEU Convention 2016

6 Month Purchases (kwh) Technica osses (kwh) Technica osses (%) ,58% ,39% ,55% ,48% ,66% ,70% ,67% ,81% ,64% ,88% ,84% ,61% Tota ,64% Month Purchases (kwh) As each ower votage eve needs the higher votage eve network the totas for each votage eve was apportioned to ower votage eves in the ration of the energy sod at that eve. This is aso done as a proxy for eves were equipment is upgraded more frequenty (LV) than other (HV). So athough the tota asset vaue is captured in the tabe above it is re-apportioned using this methodoogy to refect repacement. Using the above apportionment, the tota rand vaue per votage eve can be used together with the tota kwh sod at that eve to cacuate the average that each votage eve need to make to recover 7% of Tabe 4: Technica osses comparison. Non-technica osses (kwh) Nontechnica osses (%) Tota osses (kwh) Tota osses (%) ,61% ,19% ,11% ,50% ,80% ,35% ,45% ,93% ,50% ,16% ,32% ,03% ,63% ,30% ,04% ,84% ,32% ,96% ,15% ,04% ,23% ,07% ,14% ,75% Tota ,86% ,50% Tabe 5: Non-technica osses and tota osses comparison. Votage eve Tota rand vaue of assets HV R MV R LV R A other R Tota R Tabe 6: Asset vaue per votage eve. the tota asset vaue in a given year. The resuts are summarised in the tabe beow. Energy department cost per tariff group (peope cost) The cost to operate the eectricity department needs to be added to the cost of suppy, incuding maintenance cost. The aim is to distribute a cost per kwh sod to each votage eve that refects the cost to operate and run the department. The energy department cost used is the tota budget, excuding the foowing: Buk purchases eectricity Depreciation existing assets. This is assumed to be part of the assets cacuation Grants and subsidies paid free basic services indigents Operating surpus The amount for 2014/2015 was R 3,24-biion. It woud have been easy to apportion this cost to each tariff group if COE kept record of a cost per tariff group. That is not avaiabe. This cost can be apportioned according to energy sod. This wi penaise the arge consumers, because normay it is not the arge consumers that is causing a the peope cost. It is the residentia and sma commercia consumers. The apportioning of the energy department cost wi be done on the same basis as the asset cost, using the same percentages. It is very important to remember that for this cacuation of annua average cost per tariff group for the assets, the energy sod is used. Surpus The average contributing to the surpus is cacuated as the panned surpus amount divided by the energy sod, excuding: Technica osses Non-technica osses Own consumption, ike a CCC's street and masts ighting and traffic ights Own consumption, ike departmenta civic centre, etc. This gives an average of 9,5788 sod. Tota cost of suppy The next tabe summarises the resuts of this study in terms of CoS for power, parts, peope, pain and profit per tariff group as cacuated as described. Reseers investigation This section provides brief comments reated to the potentia discount that can be provided to the reseers tariff. To obtain a view of a potentia discount the advantages and disadvantages needs to be weighed. Some of the cost savings opportunities introduced by reseers are discussed as we as some of the disadvantages introduced by the system. According to the NERSA guideines on eectricity resae: 8.3. The tariff rates and tariff structure according to which eectricity is resod must be identica to the approved tariff rates and tariff structure that woud have been appicabe had the customer been suppied with eectricity by the suppying icensee. This is in ine with eectricity pricing poicy [Position 43(a)], which states that: non-icensed traders of eectricity sha provide the eectricity at terms, tariffs and services not ess favourabe than that provided by the icensed distributor in the area. This cause ceary states that customers may not be worse off procuring energy via a third party than if directy transacting with the suppier (municipaity). The municipaity may give the reseers a more favourabe tariff, but this may be to the cost of other customers. This reduced tariff coud be inked to the potentia reduced costs the municipaity incurs due to 65th AMEU Convention

7 Votage eve 7% Rand vaue of assets HV R MV R LV R A other R Tota R Tabe 7: 7% of asset vaue per votage eve. Votage eve HV MV LV A other Tota per votage eve HV R R R MV R R R R LV R R R R R Tota R R R R R Tabe 8: Contribution of votage eves to asset vaue according to assumptions. the services the reseer is providing. Some of these potentia cost savings components are: Appication of TOU tariff to reseers: It enabes the icensee to charge a time of use (TOU) to residentia end user oad. Administration of accounts: A the account administration of the end users is undertaken by the reseer. The municipaity ony deas with one account of the reseer. Meter reading services (conventiona or AMR): There is one meter to read instead of many end user meters. That burden is shifted to the reseer. Reduced risk: The coection of monies due is simpified to one customer. The reseer is burdened with coecting the amounts due from the end users. Maintenance: Uness there is a maintenance agreement with the municipaity that burden is carried by the reseer as a contractuay appointed buiding management agent. Capita cost: Normay the capita cost of suppying a point with eectricity is paid by the customer or deveoper, not the Reseer. This cannot be seen as a benefit to either the municipaity or the reseer. The resuts of the CoS study have shown that the percentage of cost that wi be for the reseer s account is very sma. This means the benefit that the reseer can get because of reduced cost to the municipaity is very sma. What are the disadvantages of reseers to municipaities? Compaints, dispute resoution, quaity of service, tariffs and pricing principes, and biing: Municipaities wi receive compaints from reseer s customers. Reseers operate outside the energy reguator s radar screen: It is amost impossibe for the end user to receive the benefits of programmes such as Free Basic Eectricity even eigibe, to the detriment of poor customers. Activities of reseers are currenty unreguated: This means that the end user is at the mercy of the reseer. It aso means that the energy reguator has no sight of their business, the customer and tariff data. Votage eve Tota per votage eve kwh sod excuding osses contribution to assets annua average HV R ,2077 MV R ,6406 LV R ,6221 Tota R ,9179 Votage eve Tabe 9: Contribution in per votage eve for assets. % apportioning per votage eve Rand for energy department Energy per votage eve sod excuding osses (kwh) Peope per votage eve HV 18,95% R ,1036 MV 75,67% R ,5398 LV 5,38% R ,4915 Tota 100,00% R Tabe 10: Contribution in per votage eve for energy department cost. Pricing structure/excessive tariffs: Their customer base is captive and vunerabe to being charged exorbitant prices with minima prospects for recourse. Some reseers charge tariffs above the approved tariffs by NERSA for a municipaity. With the information currenty avaiabe, it is not possibe to recommend what discount or more favourabe tariff a reseer is entited to. What needs to be quantified is how much the reseers is savings the municipaity. To ensure that the benefits accruing to the municipaity is derived it woud be necessary to poice a of the above. This wi add another burden to the municipaity. It is proposed that a study be competed to quantify the actua cost per customer per tariff group to the municipaity. That wi incude things ike meter reading (manua or AMR), bis, stamps, enveopes, chasing of payments, non-technica osses, etc. The outcomes of this study wi hep to quantify the benefits that coud be passed on to the reseer segments, if any. Possibe changes in tariffs and motivations It is not meaningfu to propose ways to address the reasons why the CoS per tariff group is more than the income derived for some tariff groups. It was aready shown that the changes in the number of customers per tariff group, oad factors, etc., is the main drivers of the cost of eectricity per tariff group. This means that every time there is a CoS study, there wi be tariff groups that costs more than what the group s income is. Of course one may ook at a tariff group individuay and evauate the different eements of the tariff, but that wi aso change during a year. The biggest differences may be evauated annuay, especiay where there are arge cross subsidisations. It may not be possibe to change some of the tariffs, but the crosssubsidisation may be quantified. One aternative is that one may reook the methodoogy used to cacuate the contribution of each cost of suppy component (power, parts, peope, pain, and profit) to the tota cost cacuated. Some of the options/ questions that can be expored are: Is the use of the votage eve the ony way to cacuate the contribution to the cost of the parts and the peope per tariff group? Shoud the anaysis aso take the number of customers per tariff group into consideration? Houd parts and peope cost be added as a or as a fixed charge per customer or a combination? How wi the financia system be abe to hande different options? 18 65th AMEU Convention 2016

8 Arc-proof air-insuated switchgear for our environment? Absoutey. ABB s new medium-votage LeanGear ZS9 primary distribution switchgear is designed to meet the specific needs of emerging economies. The design boasts key features such as a compact construction, ow maintenance design and an interna arc withstand. LeanGear ZS9 provides the quaity, reiabiity, safety, performance and high eve of technica support provided by ABB wordwide. For additiona information: new.abb.com/medium-votage ABB South Africa (Pty) Ltd. Eectrification Products Te E-mai: LP@za.abb.com

9 Tariff group name Power Parts Peope PAIN (nontechnica osses) Profit (surpus) Tota, excuding surpus Tota, incuding surpus Income in Tariff A business 78, , ,4915 7,9690 9, , , ,2372 Tariff B residentia 114, , ,4915 7,9690 9, , , ,0023 Tariff C <=11 kv 74, , ,5398 7,9690 9, , , ,6891 Tariff C <=11 kv No NAC 82, , ,5398 7,9690 9, , , ,124 Tariff C 400 V 79, , ,4915 7,9690 9, , , ,3356 Tariff C 400 V (substation suppy) 75, , ,4915 7,9690 9, , , ,91 Tariff C off-peak 11kV 78, , ,5398 7,9690 9, , , ,5502 Tariff C off-peak 400V 87, , ,4915 7,9690 9, , , ,9479 Tariff D <=11 kv 68, , ,5398 7,9690 9, , , ,7715 Tariff D > 11 kv 64,1788 4,2077 6,1036 7,9690 9, , , ,1888 Tariff D 400 V (substation suppy) 101, , ,4915 7,9690 9, , , ,0008 Tariff E <=11 kv 72, , ,5398 7,9690 9, , , ,9694 Tariff E > 11 kv 73,9145 4,2077 6,1036 7,9690 9, , , ,2018 Tariff E 400 V 78, , ,4915 7,9690 9, , , ,6043 Tariff E 400 V (substation suppy) Tariff E 400 V No NAC 75, , ,4915 7,9690 9, , , ,727 77, , ,4915 7,9690 9, , , ,2398 Res reseers <=400 V 79, , ,4915 7,9690 9, , , ,0775 Res reseers >400 V 76, , ,5398 7,9690 9, , , ,2634 Conv. residentia 76, , ,4915 7,9690 9, , , ,4181 Conv. commercia 79, , ,4915 7,9690 9, , , ,3437 A COE street and masts and taffic ighting Departmenta, civic centre, etc. 61, , ,4915 7,9690 9, , , , , , ,4915 7,9690 9, , , ,9752 A COE's prepaid and FBE 79, , ,4915 7,9690 9, , , ,7169 Non-technica osses 89, , , Tabe 11: Summary of the resuts of the study in terms of power, parts, peope, pain and profit per tariff group. One may aso reook at the actua budgeted figures used for the study. For exampe, the study showed that the cost of eectricity (sod incuding technica and non-technica osses) is ony 61,0% and the combined cost of parts and peope the remaining 39,0%. Currenty the budgeted surpus is 6,6%. NERSA s guideine aows for eectricity cost to be 75% and other cost 25% of the tota. It is recommended to change parts and peope cost contribution to proposed NERSA aowance of 25% by adjusting assets to 2,45% and peope to 60% of the current figures. By adjusting these figures, the cost of suppy mode wi resut in 75% purchase cost and 25%, other cost as stipuated by NERSA s benchmarks. If this is done, the cost of eectricity (sod and technica and non-technica osses) is 75,3%. The combined cost of parts and peope is the other 24,7%. Then the cacuated surpus is 8,1%. This is more in ine with NERSA s contribution to eectricity cost and other cost, but not yet with the possibe surpus. It may aso be that some of the parts and peope s cost can be funded from the surpus, but that wi ony baance the books to come coser to NERSA s benchmarks. Concusion It is recommended that a municipaities execute a detaied CoS study. The good data records of the COE assisted in making this study somewhat easier as far as that component goes, but the sheer size of a city made the study fairy compex. A monthy energy baance that entais comparison between kiowatt-hour units bied (saes) and kiowatt-hour units purchased from Eskom wi provide insight in terms of possibe energy osses. This information may then be used to execute meter audits to ensure tampering and other non-technica osses are kept as ow as is practica. Any municipaity s eectricity tariffs must be derived from a sound basis, taking into account the expected income and expense figures. Soid statistica figures must be gathered to guide this process. Further progression of the tariff structure, and especiay tariff eves, wi ony be gained foowing a CoS study by a speciaist tariff consutant with adequate experience in the fied of arge scae eectricity distribution. Acknowedgement The author wishes to acknowedge and thank Dr. Freddie Fryer (Ekurhueni), Tobie Nortje (EON Consuting) as we as Dr. Johan Deport for their comments and editoria contributions. References [1] data base [2] Actua Ekurhueni metering data via VENUS. [3] Dr. Johan Deport, Tariff Speciaist and Cost of Suppy study for Ekurhueni Metropoitan Municipaity 2012/13 and 2014/15. Contact Stephen Deport, Ekurhueni Metropoitan Municipaiity, Te , stephen.deport@ekurhueni.gov.za 20 65th AMEU Convention 2016