Tanzania - Power loss reduction study (Vol. 2 of 2) : Transmission and distribution system - reduction of non-technical losses

E SKM20H voi. 2 Uc5nt Unea2d Masons DeveDopmeon Programme / WcOPd Mank m1LLpe ESMAAP nEol2g Sectr Management Msssince ProqogDranme Tanzania PowerLoss Reduction Study Volume 2: Transmission and Distribution System Reduction of Non-Technical Losses Report No. 204B/98 June 1998  JOINT UNDP /WORLD BANK ENERGY SECTOR MANAGEMENT ASSISTANCE PROGRAMME (ESMAP) PURPOSE The Joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP) is a special global technical assistance program run as part of the World Bank's Energy, Mining and Telecommunications Department. ESMAP provides advice to governments on sustainable energy development. Established with the support of UNDP and 15 bilateral official donors in 1983, it focuses on the role of energy in the development process with the objective of contributing to poverty alleviation, improving living conditions and preserving the environment in developing countries and transition economies. ESMAP centers its interventions around three priority areas: sector reform and restructuring; access to modern energy for the poorest; and promotion of sustainable energy practices. GOVERNANCE AND OPERATIONS ESMAP is governed by a Consultative Group (ESMAP CG), composed of representatives of the UNDP and World Bank, other donors, and development experts from regions benefiting from ESMAP's assistance. The ESMAP CG is chaired by a World Bank's Vice President, and advised by a Technical Advisory Group (TAG) of four independent energy experts that reviews the Programme's strategic agenda, its work plan, and its achievements. ESMAP relies on a cadre of engineers, energy planners, and economists from the World Bank to conduct its activities under the guidance of the Manager of ESMAP, responsible for administering the Programme. FUNDING ESMAP is a cooperative effort supported over the years by the World Bank, the UNDP and other United Nations agencies, the European Union, the Organization of American States (OAS), the Latin American Energy Organization (OLADE), and public and private donors from countries including Australia, Belgium, Canada, Denmark, Germany, Finland, France, Iceland, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Sweden, Switzerland, the United Kingdom, and the United States. FURTHER INFORMATION An up-to-date listing of completed ESMAP projects is appended to this report. For further information, a copy of ESMAP Annual Report or copies of completed projects, contact: ESMAP c/o Energy, Mining and Telecommunications Department The World Bank 1818 H Street, NW Washington, DC 20433 U.S.A.  TANZANIA POWER LOSS REDUCTION STUDY VOLUME 2: REDUCTION OF NON-TECHNICAL LOSSES November 1992   Table of Contents I. EXECUTIVE SUMMARY .................................1 Intoduction ................................. 1 Observations ..................................2 Service Drops ..................................2 Meters and Meter Management ..................................2 Metering Installations ..................................3 Meter Reading .4 Billing .5 Billing Administration .5 Recommendations .6 First Priority for Consumer Installations and Metering .6 First Priority for Meter Reading and Billing .8 Second Priority for Consumer Installations and Metering .8 Second Priority for Meter Reading and Billing .9 Disconnections .9 Fraud .10 Tariffs .1.... l0 Conclusions .10 I. METERING .11 Meters and and Installation Practice ...................................................................... 11 Energy Meters ......................... 11 Demand Meters ......................... 11 Meter Mounting and Installations: Existing, Dictly Connected Meters Mounted on Walls ..........................11 Meter Mounting and Installations: Directly Connected, Low-Voltage Meters on Metal Panels or Boxes .12 Meter Mounting and Installations: Low-Voltage Installations Using Current Transformers .13 Meter Mounting and Installations: Medium-Voltage Installations Using Current and Voltage Transformners .13 Meter Testing: .............................................. 15 Workshop .15 Field Testing Installed Meters .15 Meter Data Base .16 Recommendations .16 Meter Installations .16 Service Drops .17 Location of Meters .17 Sealing of Meters, Instrurnent Transfonners and Cutouts .18 Meter Boxes and Switchboards .18 Cutouts .19 Earthing .20 Demand Meters for Low-Voltage Consumers .......................................... 20 Demand Meters for Medium-Voltage Consumers .21 Meter Specifications .21 Meter Testing: Workshop .22 Meter Testing: Field .................................................................. 22 Meter Data Base ....................................................................................... 22 Conclusion .................................... 23  III. METER READING AND CONSUMER BILLING .............................................. 24 Introduction .24 Meter Reading .24 Meter Locations .24 Meter Reading Process .24 Reading Demand Meters .26 Reporting Anomalies .26 Accounting and Billing .26 Consumer Account Numbers .26 Detemiination of Zones in Dar Es Salaam .27 Computer Keying and Validation of Meter Reading .27 Recommendations. 29 Consumer Records .29 Suggested New Account-Numbering Scheme .29 Meter Reading Routes .29 Meter Reading Lists .31 Anomalies .32 Consumer Addresses .32 New Billing Program .32 Validations .33 Security of Meter Reading Sheets ............................ 34 Bill Preparation and Delivery .34 Disconnections .34 Conclusion .35 IV. TARIFFS .......................... ... 36 V. RESULTS .38 Introduction .3 Consumer Selection .38 Conclusion .40 Annex 1: Terms of Reference ...... ............ 41 Electric Utility Billing Software ..................... . 41 Annex 2: Electricity Tariffs .47  PREFACE This report was prepared in 1992 consequent to a study conducted during 1990 to 1992 in association with a distribution planning unit established in the Tanzania Electric Supply Company (TANESCO). The report is in two volumes; volume 1 dealing with the technical studies conducted on transmission and distribution systems and volume 2 dealing with non-technical loss issues. The Bank team carrying out the study consisted of Messrs. Winston Hay, Chrisantha Ratnayake, and Ms. Paivi Koljonen. Volume 1 was prepared by Mr. Ratnayake and volume 2 by Mr. Hay. The assistance and participation of the distribution planning unit in TANESCO in carrying out the detailed studies on which this report is based is readily acknowledged.   I. EXECUTIVE SUMMARY Introduction 1.1 The overall energy losses on the public electricity grid of Tanzania Electric Supply Company Limited (TANESCO) presently amounts to some 21 percent of net generation. System losses commonly divide into two sources (a) technical and (b) nontechnical. Technical losses result from a certain percentage of the power flowing through transmission and distribution systems being converted into heat. The amount of energy so converted is dependent on the physical characteristics of the system and the electrical load imposed on it. Nontechnical losses represent unbilled consumption resulting from inaccurate metering, errors in meter mading and billing as well as from consumer fraud (power theft). Losses result in economic burdens for the country, lower revenues for the utility and the need for unproductive investment in generation, transmission and distribution planL 1.2 Investigations indicated that the technical losses on the TANESCO system did not exceed about 11.5 percent of net generation annually. This implied that, with nontechnical losses being about 9.5 percent of net generation, more than 145 GWh of electricity was being consumed annually without producing revenue for TANESCO. In 1989 the company's average tariff level was 5 Tanzanian shillings (TSh) per kWh. At that rate, nontechnical losses amounted to TSh 725 million, or nearly US$4 million, in lost revenues annually. 1.3 TANESCO requested the Energy Sector Assistance Programme (ESMAP), a joint United Nations Development Programme (UNDP)/World Bank program, to help determine how much the various loss sources contribute to overall losses and assist development of a program to reduce losses to an economic optimum. ESMAP obtained funding for the execution of the study from the Swedish International Development Authority (SIDA) and began the study in late 1989. 1.4 This volume of the study report focuses only on the nontechnical loss reduction activities, conclusions, recommendations and results. The study of technical losss is dealt with in the first volume of the reporL 1.5 The primary sources of nontechnical loss are (a) unmetered supplies. These may be due to consumer theft of power. They can also result from direct connections the utility made during meter shortage and failed to correct when meters became available; (b) defective metering. This may be the result of tampering. Or the meter may have been connected incorrectly by TANESCO, may have been installed with defective mechanisms or may have developed defects after installation; (c) meter reading errors. Some meters may be read incorrectly or not at all; (d) billing deficiencies. For example, the billing departnent may not prepare some bills because it has not received meter readings; it may calculates some incorrectly; and it may not be able to send others because consumers are not properly recorded on the billing register. 1.6 Unlike technical losses, nontechnical losses can often be reduced with little investment of capital and in relatively short time. In addition, and again unlike technical losses,  -2- nontechnical losses can be reduced to levels very close to zero. Successful loss reduction of any nature, however, requires careful development of strategy, thorough implementation of the program decided on and scrupulous attention to detail. 1.7 In mid-1992 as the study neared its end, significant progress had been made in the reduction of nontechnical losses. Joint TANESCO-ESMAP efforts had led to the forrmation of inspectorate teams in all zones to, inter alia, carry out inspections of consumer service installations, rectify irregularities and prepare invoices for unbilled consumption where such was detected. By TANESCO's estimates the inspectorates have been successful in reducing the level of nontechnical losses by four percent of net generation by the end of June 1992. Invoices for more than 300 TSh had been issued for consumption previously unbilled and of that amount more than 30 million had already been collected The training which TANESCO staff have received as well as the institutional structures developed for detection and control of nontechnical losses augur well for continued reduction of losses on the Tanzanian grid. Observations 1.8 During the time spent in the field, the mission observed a number of practices and administrative weaknesses which contribute to the high level of nontechnical losses that TANESCO currently experiences. The more important of these involve service drops, meters and meter reading management, meter reading, billing, and billing administration. Service Drops 1.9 TANESCO apparently has no uniform standards for the design of service drops. (Service drops are the conductors and other hardware used to connect the utility's distribution system to a consumer's meter). Service crews installing the drops run conductors through the most convenient routes. The results are not only untidy but often present easy opportunities for illegal abstraction of power. The untidiness of the installations increases the difficulty of detecting illegal connections. Meters and Meter Management 1.10 TANESCO's meters are dust resistant but not weatherproof. They are therefore not suitable for outdoor installation where they would be directly exposed to rain. Electrical connections to all meters are made at the bottom. It is possible to seal the meter's terminal box, but this procedure is not rigorously followed. 1.11 Before installation, each new meter is tested in TANESCO's meter test workshop. There is no program to routinely test meters after installation. Meters are brought in for retesting only if they are known to be damaged or otherwise nonfunctional or in response to consumer complaints. Some of the testing equipment in the workshop (in particular, the rotating substandards) is of doubtful accuracy. 1.12 TANESCO does not assign its own numbers to meters but uses the manufacturers' serial numbers for identification. Since meters are purchased from a number of different manufacturers, each using a different approach to selecting serial numbers, no logical system of classification by meter number can be employed by TANESCO. 1.13 After testing, new and used meters are stored in rooms adjacent to the workshop. The number of meters observed in inventory indicated that lack of meters was not the cause of the relatively slow rate at which new consumers are added to the utility's system, as is sometimes suggested.  -3- 1.14 TANESCO's only central meter records are the notebooks in which the test results are registered. The volume of information and the random nature in which it is stored make access to data on specific meters extremely difficult Metering Installations 1.15 There appears to be no design standard for metering installations. The untidiness noticed in the installation of service drops is often repeated in metering installations, especially in locations, such as apartment buildings or office blocks, where a number of meters are grouped together. Most meters are unsealed, and the wiring around the meters is generally exposed, again making fraudulent abstraction of power a relatively easy operation. 1.16 Many meters were seen to be stopped, although the premises continued to be supplied with power. TANESCO's own records indicate that in December 1989, the utility had about 7,000 active accounts whose meter readings had indicated zero consumption for six months or more. The majority of these accounts are thought to have defective meters. Also, several hundred meters have been reported defective but have not yet been investigated and corrected. Some or all of these defective meters may be associated with the 7,000 accounts mentioned above. 1.17 Energized but unused and unmetered service conductors were observed in a number of premises. This situation is hazardous and invites fraudulent use of electricity. 1.18 Early in 1990, TANESCO initiated a program to relocate all meters (except those with instrument transformers) to the outer walls of buildings. The objective is to increase the rate of meter reading and reduce fraud. The program is progressing slowly, primarily because of inadequate resources committed to the task. 1.19 Metering installations with instrument transformers are not adequately secured. The transformers are not sealed, and conductors between them and the meters are not run in a tamper- proof manner. The meters themselves are often not sealed. 1.20 Meters for a wide range of installations, from residential to industrial consumers, am often located in locked areas whose keys are kept by the consumer. TANESCO, therefore, cannot gain immediate access to these locations without cooperation from the consumer, who will thus know when TANESCO staff are coming to inspect the metering. Unscrupulous consumers may use this opportunity to remove evidence of fraudulent power abstraction. 1.21 The meters are also frequently installed in areas that make meter reading uncomfortable, difficult and sometimes dangerous. Common problems are (a) the placement of meters so high that a stool or ladder is needed for reading them, (b) the placement of meters so low that they cannot be conveniently read, (c) inadequate space in front of meters for proper reading, (d) stored materials that impede access to meters, and (e) exposed and energized conductors nearby that endanger the safety of personnel. 1.22 TANESCO's general practice for supplies that do not involve metal switchboard panels is to install a fused cutout upstream from the meter. Some, but not all, cutout covers have provision for security seals to fasten the covers to the body of the device. Several different cutout models are in service, and their covers are not always interchangeable. In any case, it is not unusual to find cutouts with the cover unsealed, without a cover, or with fuses removed and replaced by wires. Many installations were observed where the cutout had been completely removed and the wires at each end spliced together. Unsealed cutouts and spliced conductors make it easy to bypass the meter when supplying power to the premises. 1.23 Industrial consumers are required to provide revenue metering equipment in the main supply switchgear of their premises. There are two important disadvantages to this practice.  -4- First, TANESCO cannot provide a level of security which will ensure that only its employees have lawful access to the metering installations. Second, the metering equipment is not standardized, and defective units therefore cannot be readily replaced from TANESCO's inventory. As a result, the energy consumption of a consumer who has defective meters may remain unmetered until spares have been procured from the original supplier. Meter Reading 1.24 TANESCO's service area is divided into several meter reading zones, with each zone normally being read on a specific day of the month. Individual zones are subdivided into meter reading routes; a route comprises the accounts assigned to a meter reader to be read in a full working day. Meters with demand registers are read by special crews. 1.25 At the beginning of the workday, meter readers are issued hard cover books with a sheet for each customer whose meter is to be read that day. When new, each sheet has space for about two years' readings. The meter reader is expected to take the book along the route and enter meter readings directly onto the appropriate sheets. However, it was observed in Dar es Salaam that in practice, meter readers leave the book in the office, write the readings on a sheet of paper and transcribe them to the book toward the end of the workday. This practice is apparently followed because the sheets are not arranged in the books in the sequence in which the route is followed. 1.26 Because up to two years' readings are maintained on any given sheet, the meter reader will know the previous month's reading for each meter to be read. 1.27 The meter reader also records the reading on a card that is kept near the meter. This card is another source where the reader can obtain the previous month's meter reading. 1.28 In the so-called "unsurveyed areas" where street names and numbers do not exist, routing information is unreliable. In these areas, several meter readers go house-to-house and read all the meters in a specified area Afterward, the group assembles; one person reads from a computer printout the numbers of the meters that were assigned to be read that day. When a meter rader hears the number of a meter he has read, he supplies the reading, which is then recorded on the sheet. This process is time consuming, especially given the unsystematic assignment of meter numbers, and inaccuracies are common. 1.29 Meter readers are supposed to report the meters they find that are not on the meter reading schedule. Despite this, many meters that are listed as having been issued cannot be located in the field. The previously mentioned lack of a central data base for meter information (paragraph 20) increases the difficulty of solving this problem. 1.30 Sample checks indicated that the accuracy of meter reading is well below what ought to be considered acceptable. This observation is supported by the relatively high number of validation requests ("query reports") 'that are issued after meter readings are entered into the computer. A validation request is generated whenever the computer's initial calculations indicate inconsistencies between a new meter reading and previously recorded data. Not all validation requests, however, result from meter reading anomalies. 1.31 Meter readers are expected to report irregularities observed during the course of their work. A special form is supposed to have been developed for this purpose, but no example was seen. Some meter readers stated that no action is being taken to conrect irregularities that have been repeatedly reported, as a result, they have lost the motivation to continue making reports. TANESCO's records indicated in March 1990 that a backlog of 500 defective meter reports from meter readers was awaiting investigation. (See also paragraph 1.16.)  -5- Billing 1.32 In current practice, meter readings are transferred from the hard cover books to a list that is used for data entry to the billing computer. This means that a reading is transcribed four times--three by writing and one by keying--before it is finally recorded in the computer. Such repetitious transcription increases the likelihood of errors. 1.33 The billing program now in use suffers from a number of serious inadequacies. Important among these are situations where (a) a consumer's average consumption is not calculated from previous records but from an estimation. An estimate of average consumption is made when the account is first established and remains in use unless manually changed; (b) no limit exists on the number of consecutive bills that can be issued on the basis of estimated consumption. At the end of 1989, TANESCO had more than 11,000 accounts that had been billed for six or more consecutive periods on this basis; (c) consumptions higher than average are flagged, but those that are lower are not. Consumption that is lower than average is far more likely to adversely affect utility revenues than consumption that is higher, (d) the computer is not programmed to recognize when a meter has exceeded the limits of its register and the register has returned to ze. 1.34 It is possible for any staff member with access to the computer and knowledge about data entry to change meter reading records without prior authorization from a designated manager. No explanation needs to be given about why the change is made. Such changes affect only current billing (to reduce the amount invoiced in all instances observed) without any corresponding adjustments being made to previously issued bills. 1.35 A large number of consumers (7,000 in December 1989) are being billed for zero energy consumption over extended periods of time. Since these consumers continue to make the minimum payments stipulated in the service contract, it may be safely assumed that the majority, if not all, of them are in fact consuming energy that is not being recorded by the meters. (See also paragraph 1.16.) 1.36 Computer problems at the time of the field work for this report resulted in irregular bill preparation. It was therefore not possible to detemmine how much time normally elapses between reading a meter and billing the corresponding consumer. Billing Administration 1.37 No group within TANESCO is charged specifically with investigating and eliminating the causes of revenue loss. As information about nontechnical losses is transferred from department to departnent, the sense of urgency becomes diluted. A department tends to assign priority to activities it considers to be its more substantive duties. Task forces (such as the one that recently started investigating irregularities) formed to address specific issues often do not have the expected effect because they depend on other groups, which have different priorities, to act on the findings of the investigation. 1.38 The laws of the Republic of Tanzania expressly include theft of electricity among the criminal offenses punishable by imprisonment. However, no instance was found where TANESCO had taken a consumer suspected of power theft before the court TANESCO's own penalties for fraud require an offending consumer to pay a fine, the cost of inspection, expenses for disconnection and reconnection, and the cost of repairing any damage done to the utility's  - 6 - equipment. But even these official company procedures are not followed consistently, and the popular perception may well be that TANESCO does not consider unauthorized use of power to be a serious offense. 1.39 The bills sent to consumers are imprinted with the following statement: "Pay promptly within 21 days otherwise power will be disconnected without further notice" This warning is manually stamped on the bill. It would be more efficient to have it printed as a standard part of the bill form or to have it printed by the computer when the bill is prepared. Bills are also folded and inserted into the envelopes by hand, again an exercise that seems unnecessarily labor intensive. 1.40 Despite the warning printed on the bill, disconnections are not consistently undertaken. The head office in Dar es Salaam stated that when each new bill is issued, disconnection orders are issued for accounts that are one month's payment in arrears. The Morogoro office reported that disconnections are undertaken on accounts that are three months in arrears. The observed facts did not support either assertion. Many cases were found where active accounts were much more than three months in arrears. 1.41 If a consumer requests his supply to be disconnected, one person is sent to obtain a final reading and another is sent later to disconnect the supply. The procedure is not only labor intensive, but it also increases the probability that a consumer will continue to use energy after a final bill has been issued and his account has been removed from the billing register. Recommendations 1.42 A wide range of actions is required if TANESCO is to reduce nontechnical losses to a level close to what is economically achievable. The presented recommendations here focus on those actions that can be taken in the relatively short term and that should produce more or less immediate benefits. Succeeding chapters deal with the fundamental issues in greater detail and, while repeating the recommendations of this chapter, provide additional advice on organizational and procedural changes that will further reduce nontechnical losses and keep them at low levels. First Priority for Consumer Installations and Metering 1.43 TANESCO should establish a task force charged with developing and executing a countrywide program of reducing nontechnical system losses. The program must be directed by a senior level officer with no other substantive duties. The task force must have the authority and resources to * disconnect supplies where warranted, such as in case of fraud, correct irregular installations, such as defective meters; reseal meters and cutouts where required; record supplies to premises for which there are not registered accounts and specify the location of each; calculate lost revenues and penalties for back-billing-,  - 7 - record the serial number, type (e.g., kVA demand, directly connected, three-phase, single-phase) and location of all meters in the field, whether they are registered or not. The task force must also monitor the execution of those remedial actions that fall outside its jurisdiction and inform higher management of actions that are not addressed with the required degree of timeliness. 1.44 The task force's first priority in loss reduction must be to ensure that all consumers, connected to TANESCO lines are properly (a) registered on the company's consumer records, (b) metered and (c) invoiced for their energy consumption. It is therefore essential that all consumers are properly recorded and that consumer metering installations be rigorously inspected for metering and other irregularities. These two tasks may proceed concurrently but should be undertaken by different work crews. 1.45 If proper accounting is to be made for all consumers connected to the system, TANESCO must institute a program to patrol the full length of its distribution lines and record every consumer supplied from them. Although recognition of fraud and other irregularities is also important, this must assume second priority if a choice of priorities must be made. Therefore, the work force identifying consumer connections, meter types and serial numbers must not be required to search for irregularities, although they must report the obvious ones. If workers are required to carry out detailed investigations of the connections, the accounting program will proceed with unnecessary slowness and will probably never be completed To be effective, such a program must be initiated and completed within a reasonable time--not more than six months. The work force involved will need to be trustworthy and diligent and will need some training or experience in what to look for. It is possible that TANESCO cannot provide sufficient staff with the qualifications needed to complete the program in the stipulated period without severely disrupting other important functions. In that event, persons outside the organization will have to be hired temporarily. The potential increase in revenues will justify major expenditures on such a program. 1.46 The program should begin in areas of high consumer density and gradually extend over the whole system. 1.47 The second area needing urgent attention is the investigation of irregularities. Mnother set of work groups should be established to undertake these investigations. The staff will need to be experienced meter readers and installers or persons given thorough special training. The installations to be examined first should be those with known or suspected irregularities that have been reported but not yet addressed. The work crew must be empowered to take such actions as disconnecting services, installing new meters and resealing meters and cutouts. When the backlog of outstanding inspection requests has been completed, the work groups should investigate other accounts, giving priority to the largest consumers as well as to irregularity reports from the groups identifying connections or from other'sources. Special attention should be given to accounts that (a) have indicated zero consumption or have been billed on estimated consumption for several months, (b) have indicated zero or drastically reduced kVA demand (where applicable) and (c) have shown significant reduction in average monthly consumption. 1.48 The investigation teams will need to be able to check on the accuracy of the meters. For directly connected meters, this check can be performed adequately by using a clip-on ammeter and timing the meter for three to five minutes. The teams may use a portable load of a known constant demand, such as a one kilowatt resistive load, to test where premises are unoccupied and there is no consumer demand at the time of inspection. The overall approach requires at least a close approximation of the power factor. That parameter can also be measured by certain types of modem clip-on ammeters. Larger metering installations with instrument transformers should be checked with portable kilowatt and kilowatt-hour measuring devices.  -8- 1.49 TANESCO should carefully consider introducing new meter seals when the loss reduction program starts. It appears that control has been lost over existing sealing tools, and the utility should probably change the ":ethod of sealing. Tight control will make it possible to know who is responsible for sealing a specific meter or to hold staff accountable for the use of seals issued to them. New seals may also be more difficult to break and can be color coded to identify, for example, accounts with previously reported irregularities. First Priority for Meter Reading and Billing 1.50 The existing billing program suffers from many inadequacies that make it unsuitable for use by a modem utility. TANESCO recognizes this shortcoming and plans to purchase new billing hardware and software. Annex 1 is a draft specification for a new billing program. It will be some time, however, before a new computer and program can be functioning, and there are at least three areas, discussed in the following paragraphs, where urgent action is needed to begin controlling nontechnical losses. 1.51 At present, any staff member with access to the computer and knowledge about data entry can change meter reading records in the computer. This situation must be changed immediately. Only a restricted number of responsible management staff should be able to change meter reading records that have been entered and validated. The easiest way to accomplish this control is to have a program that requires a password that can be changed frequently and is known only to the authorized persons. Any change in meter reading records, as well as reasons for the change, must be documented. In addition, appropriate modifications to billing already issued should be made after such changes to meter readings. 1.52 The method of calculating estimated consumption needs to be revised. The program should be modified to calculate estimated consumption from actual meter readings stored in the computer, which currently maintains data from the previous six months. 1.53 The program must also be modified to flag abnormal decreases in consumption. Abrupt reductions in the average consumption of a single account often indicate metering problems or fraud. Abnormal increases in average consumption (currently identified in the billing process) are less important to the utity. 1.54 The program changes recommended above may require TANESCO to seek outside assistance. The changes are relatively simple. Consulting costs to have them carried out would be relatively small compared with potential benefits. Second Priority for Consumer Installations and Metering 1.55 TANESCO must develop, publicize and rigidly enforce standards for service drops and metering installations. Single-phase service conductors should be run from the distribution line to a bracket on the building to be served and then through a metal conduit, attached to the outer surface of the building, to a meter (or meters) mounted at a height where it can be conveniently read. The service line should preferably be of concentric cable to reduce the possibility of making fraudulent connections upstream from the meter. 1.56 Meters must be installed in locations protected from the weather or in weatherproof compartments. The latter must have windows that allow the meters to be read without opening the compartment. It is also recommended that TANESCO consider using weatherproof, socket-type meters like those in normal service in North America. 1.57 Wiring upstream from the meters should not be exposed or easily accessible, and meter termninal boxes and cutouts must be sealed. It is recommended that the cutouts be relocated downstream from the meter to make sealing these devices unnecessary.  -9- 1.58 TANESCO must develop and implement a systematic program of assigning and affixing a proprietary company number to each of its meters. Meter suppliers can be required to affix this number, in accordance with TANESCO instructions, to each meter purchased by the company. The number should be displayed inside the meter case where it is easily visible through the glass window. The numbers will have to be affixed on existing meters by TANESCO, itself. This can be best accomplished by replacing existing meters with ones that have been tested and had numbers affixed to them in the company's meter workshop. 1.59 A computerized data base of all company-owned meters, including those in inventory, should be developed. These records will encompass all relevant data, such as manufacturer's name, serial number, company number, phases, range and location history. Ideally, this data base would be maintained on the mainframe computer used for consumer billing. But microcomputers can be used if that avenue is quicker. 1.60 All revenue metering equipment should be TANESCO property and conform to uniform standards. The current practice of requiring large industrial consumers to purchase metering equipment with their supply switchgear should be discontinued. Second Priority for Meter Reading and Billing 1.61 Meter reading sheets should be arranged in the order in which a reader encounters the meters as he walks the route. The sheets should be modified so that a reader will not know the value of previous readings. Also, the practice of maintaining a reading card near the consumer's meter should be discontinued. 1.62 The sheets on which meter readers enter the readings should be used, without transcription, for keying the data into the computer. This routine will not only expedite bill preparation (and therefore revenue collection) but will also reduce the number of errors that currently develop between reading the meter and keying the data into the computer. A carbon copy of the readings can be made when they are first recorded. 1.63 TANESCO needs to develop a method for describing a service drop in an "unsurveyed area" so that it can be located by TANESCO staff without much difficulty. Persons unfamiliar with local conditions cannot make many useful recommendations for developing such a method. But this task, if given adequate priority, is certainly within TANESCO's capabilities. One approach might be to display the location on maps drawn to an appropriately large scale. Another possibility is to number the poles in the unsurveyed areas and use the pole number as an identifying address. This pole-numbering scheme should be made generally applicable to the TANESCO-wide system, even if the scheme is initially implemented only in unsurveyed areas. 1.64 Bills for industrial consumers in Dar es Salaam should be prepared and hand- delivered within one day of the meter reading. For zones outside Dar es Salaam, readings should be forwarded to the head office the day the readings were made. The Head Office, in turn, should forward the bills calculated from those readings to the originating office within one day of receipt for prompt hand delivery. 1.65 TANESCO should investigate the feasibility of contracting meter reading, billing and collecton to one or more private organizations. Disconnections 1.66 The disconnection policy needs to be reviewed, revised if necessary, publicized and consistently enforced. Disconnected supplies need to be periodically reinspected if regular supply is not reestablished.  -10- 1.67 The disconnection warning that is currently hand-stamped on each bill should be printed on the bill form or printed automatically by the computer. TANESCO should plan to purchase an automatic bill-inserting machine along with other new billing hardware. Fraud 1.68 Penalties should be rigorously applied where fraud is discovered. They should include charges for all costs incurred in investigating and rectifying the irregularity as well as for revenue losses resulting from the theft of energy. A substantial deposit on account should be required before service is restored, and the metering installation must be conspicuously identified as one where fraudulent abstraction of power has occurTed. 1.69 TANESCO is understandably reluctant to bring consumers charged with fraud before the courts, as permitted under the laws of Tanzania Nevertheless, if theft of power continues to be a problem, prosecution of serious offenders will discourage the practice, especially if successful prosecutions are well publicized in the press. Staff who investigate power theft will need careful training in the collection and preservation of evidence for presentation to the courts. Tariffs 1.70 Tariff schedules do not impact losses directly, but they may provide consuners with incentives to defraud. TANESCO's tariffs (those in force in January 1992 are shown in Annex 2) show a number of inconsistencies which need to be addressed in a study to review and restructure tariffs on a more rational basis. The feature most likely to encourage fraud is the steep increase in the incremental unit cost of energy to residential, commercial and light industial consumers with increasing aggregate consumption. The intention is to encourage energy conservation but the effect may be more likely to increase losses. Conclusions 1.71 The study has already produced tangible results in reducing non-technical losses on the TANESCO system, as shown in the previous paragraph. The further actions recommended above should bring further reductions in losses and increases in revenues and deserve TANESCCYs urgent attention. The following sections deal with the fundamental issues in greater detail and make additional recommendations which enhance achieving the objective of reducing losses to an acceptable minimum.  II. METERING Meters and Installation Practice Energy Meters 2.1 The majority of TANESCO meters are of British manufacture. The meter bearings are of good quality; older meters run on jewels, and newer ones use magnetic attraction or suspension. Not many cases of movement seizure due to bad bearings have been reported with these meters, but some magnetic bearing models seem to have more defects than the rest. Japanese meters have been used in the recent Dar es Salaam distribution system rehabilitation project. These meters do not have magnetic bearings. A larger than normal proportion of these meters have been reported to have stopped because of bearing seizure. 2.2 All the meters are dust resistant but are unsuitable for exposed outdoor installations. All the meters are bottom connected. Most have cyclometer (digital) registers with five or six digits, not including the decimal register. 2.3 Nearly all the meters have a decimal register, but many do not conform with the latest construction standards that require noninteger digits to be identified by a color different from the white-on-black numerals used for integer digits. Decimal registers, especially when located on the same line with the other registers, may lead to meter reading errors. Demand Meters 2.4 When kVA demand readings are required, a combination of two or three meters is used The three-meter alternative is no longer installed for new supplies, but many of these devices are still found in consumer metering. In the two-meter combination, the instrument that measures kWh also has an unused kW demand register, the meter with the kVA demand register also has an unused display for reactive energy. The kVA demand mechanism (30 minute integrated demand) is very complex. Its driving motor and reset coil are not very reliable. 2.5 TANESCO does not own any meters that read kWh and kVA demand in one enclo- sure. These meters have been available in North America for at least 30 years. 2.6 Most of the meters are bottom connected. Some of the three-meter combinations are switchboard type, with the temiinals at the back of the instrument. 2.7 The cyclometer indication for meters used with current or current/voltage transformers is already adjusted by the application of the multiplier factors for the respective instrument transformers. However, many of the cyclometer indications are not in kilowatt hours (the only units of energy accepted by the computer for billing purposes) and must therefore be corrected by a multiplier before being entered on the meter-reading sheets. Meter Mounting and Installations: Existing, Directly Connected Meters Mounted on Walls 2.8 An outdoor service line terminates at a bracket on the outside wall of a consumer's premises. Wiring from that bracket to the cutouts is usually by a two- or four-core cable with vinyl protective coating. In three-phase installations, two two-core cables in parallel are often used instead of four conductor cable. These cables are frequently inaccessible to inspection because they run under roofs or are embedded in the waUs.  -12- 2.9 Energized but unused and unmetered service conductors were observed inside several premises. This practice is hazardous and invites unauthorized abstraction of electricity. 2.10 The supply voltage to directly connected meters is 230/400 V. In meter installations not involving metal panels, TANESCO standards require installation of single-pole, fused cutouts on the live lines on the supply side of the meter. Most cutouts have a compartrnent for the neutral link. Several models of cutouts are employed, and their covers are not always interchangeable. Some, but not all, cutout covers have provisions for a lead seal to fasten them to the body of the device. Most connectors in each pole of the cutouts have provisions for two conductors, but usually one of these spaces is unused. 2.11 It is not uncommon to find either a cutout without a cover and with a bridge between its terminals or signs that a cutout had burned, was removed and the wires spliced together. 2.12 Many meters are mounted inside customer premises, and permission must be obtained from the occupants to gain access to them. Occasionally, one or more meters are mounted inside a wooden box with a small slit, which is presumably sufficient to allow normal meter reading. Frequently these boxes are closed with a padlock, the keys being kept by the customer. One case was observed where the keys could not be located when requested by TANESCO employees. 2.13 Unwritten standards require that new installations have an earthing electrode to which all the earthing conductors are connected. This specification is not always followed, resulting in a variety of eanhing practices that sometimes conflict with each other and are not suited for the equipment being used. Earthing is important for protection of personnel against electric shock. 2.14 Early in 1990, TANESCO initiated a program in Dar es Salaam to relocate meters (except those with instrument transformers) on the outer walls of buildings in order to reduce opportunities for fraud and to expedite meter reading. TANESCO supplies all materials. The work comprises supply and instaUlation of * insulated cables from the incoming bracket or an existing service cable to an area where several meters are located, insulated cables from the meter load temiinals to the conductors that were previously connected to the load side of the original meter, * mounting boards or wooden boxes or other necessary equipment; * cutouts, junction boxes or other necessary equipment; and * new meters to replace those previously installed. Meter Mounting and Installations: Directly Connected, Low-Voltage Meters on Metal Panels or Boxes 2.15 These installations are found primarily in multistory buildings. The meters (and cutouts when provided) are mounted on the front of a panel or box. The main portion of the wiring is usually hidden inside the panel or box, and only a short length of the lead to the meter or cutout is accessible from the front 2.16 No identification of circuits that serve particular offices or apartments in the building is readily available.  -13- 2.17 The space provided in front of the meters is generally restricted, and other equipment is often placed directly in front of the panel or box. The rooms where the meters are located are sometimes used for storing tools, other objects and even trash. Panels are frequently covered with dust and cobwebs that could contribute to insulation failure. 2.18 The wiTing at the back of these panels lends itself to fraudulent installations. Meter Mounting and Installations: Low-Voltage Installations Using Current Transformers 2.19 Current transformners are used when the ampere rating of directly connected meters is less than the consumer's load. When current transformers are needed, the load is high enough to place the consumer in a tariff category that requires a demand meter. 2.20 Most of these installations are located in a room or cubicle provided by the consumer. Frequently the available space is shared by non-TANESCO equipment, and the space in front of the meters is insufficient to allow them to be read or checked safely. Doors to the meter cubicles are often locked by the consumers. TANESCO employees lose much time trying to obtain the keys and often have to make a second visit because the keys were not made available on the first. 2.21 The current transformers now used are of the "window" or "doughnut" type, which requires the power conductors to run through a central hole. If one of the transformers needs to be changed, the power conductor must be cut, the transformer replaced and the conductor respliced together. Service must be interrupted for several hours for this task. Since the transformer windows are quite small, passing the power conductors through them is often difficult, particularly if more than one conductor per phase is used. 2.22 The terninals of the secondary circuits of current transfonners are easily accessible to unauthorized persons who can thereby alter the meter registrations. Most existing installations lack the means to temporarily short-circuit the secondary windings of the current transformers when wiring to the meter must be opened. If the secondary circuit of a current transformer is left open, high voltages will develop across its terminals, jeopardizing the safety of personnel as well as the transformer itself. Meter Mounting and Installations: Medium-Voltage Installations Using Current and Voltage Transformers 2.23 These installations are for consumers supplied at 11 kV and 33 kV. All of these consumers are billed for energy (kWh)-and demand (kVA). TANESCO requires the consumer to supply indoor switchgear with a main incoming breaker equipped with overcurrent and ground protection, three ammeters, one voltmeter and switch, three current transformers and three voltage transformers. Consumers also purchase and install the meters - a two-element kWh meter and a kVA demand meter. The indications of these meters should already have been adjusted to reflect turn ratios of the voltage and current transformers. 2.24 The specification for this equipment is based on British standards and does not appear to have been updated for many years. The switchgear purchased by the consumers is not usually of British manufacture and may follow other construction standards.  -14- 2.25 The quality and maintenance of the installations vary widely, but all display the following weaknesses: * There is very little space in which to work when inspecting or testing the meters. * Most installations are in locked rooms, and the consumer keeps the key. Much time is lost gaining access to the meters. * The current and voltage transformers are used for metering and protection. The control and meter wiring is readily accessible to the consumer, making it easily possible for the meter wiring to be tampered with. Instrument transformers, which are the draw-out type, can be easily disconnected. Evidence was seen of voltage transformers having been "racked out" by the consumer, thereby preventing the meter from registering. * Instrument transformers form part of the switchgear and are probably not interchangeable between the installations of different consumers. Therefore when transformers fail, replacement units with the appropriate ratios may not be readily available and must be ordered from the manufacturer, with consequent high cost and long delivery time. * Current transforrners used for metering will not necessarily have winding ratios appropriate for use for protection and vice versa. The switchgear usually does not have separate current transformers for metering and for protection. Readings from several demand meters suggest that the transformer ratios are often too high for good metering, indicating that the transformer was probably chosen to satisfy protection requirements. * Very few installations have switches to accommodate meter testing. When test switches are provided, they are not always of the same design, and the installations use several wiring schemes. The lack of standardization prevents a large number of the plugs from being used for testing, which is one of the features intended to justify the use of test switches. The wiring connecting the meter to the instrument transformers is usually composed of individual conductors, and no attempt is made to color-code them. The conductors are often installed in a disorderly manner that makes following their connections difficult. The cabling between instrument transformers and meters is seldom shielded. Good practice requires these cables to be shielded by a metallic conduit or outer metallic sheath to minimize erors or damage due to external electric or magnetic fields. The meters are preprogrammed for specific instrument transformer multipliers, a situation which reduces the interchangeability of meters and increases the proba- bility that a suitable meter will not be available to substitute for one which must be removed from a consumer's premises for repairs.  - 15- Meter Testing Workshop 2.26 TANESCO's meter workshop in Dar es Salaam calibrates new meters before installation and inspects, repairs and recalibrates used meters as needed. 2.27 The workshop has three test benches: one modem single-phase unit with a 1 10230- V electronic substandard and two adequate three-phase units with the following rotating substandards: (a) three-phase rotating standards 110,230 V (b) three-phase rotating standard 110 V (c) Several three-phase rotating standards 230 V Some substandards are also used for field testing. The workshop has no high-precision "transfer standard" that could be used to check the accuracy of the substandards. 2.28 Part of the workshop area is used to clean and repair meters, which are then sent to the calibration benches. Ultrasonic equipment to clean the meters has been received but not yet installed- This equipment should accelerate the repair process because it cleans the meters without requiring disassembly. It is claimed that 120 single-phase meters can be calibrated in a day. 2.29 The testing routine for new meters is unnecessarily thorough and therefore time consuming. New meters are calibrated at the factory using equipment much more accurate than TANESCO's, and their construction is sturdy enough to withstand the relatively minor transportation shocks usually encountered between the factory and TANES CO's meter workshop. Compliance testing, for instance, at 100 percent load and unity power factor should be sufficient for meters that show no obvious sign of damage during transportation. If the meter passes this test, it will probably be within acceptable limits of accuracy throughout its operating range. Meters that show evidence of rough handling during transportation (damaged cartons, for instance) and used meters should undergo the complete testing routine. 2.30 The three-phase rotating substandards used for meter testing appear to be far out of calibration. An electronic single-phase substandard (considered the most accurate in the shop) was used to separately test each of the rotating substandards' phases. Such tests are not the most precise but provide reliable indication of the accuracy of the substandards. The results strongly indicated that the substandards have large errors, which may be negative or positive. The error indicated for two phases of three substandards ranged from -1.20 to +3.31 percent 2.31 The acceptable error for these substandards should not exceed