1. K-State home
  2. »Engineering
  3. »ECE
  4. »People
  5. »Faculty
  6. »Personal Webpage of Anil Pahwa
  7. »Chapter 2

Electrical & Computer Engineering

Chapter 2

Distribution Automation Functions

2.1 General Description

The distribution automation functions can in general be divided into two main categories, namely customer level functions and system level functions. The customer level functions are those functions which require installation of some device with communication capability at the customers' premises. These include load control, remote meter reading, time-of-use rates, remote connect/disconnect etc.. The system level functions are those functions which relate to system operations. The control and communications devices for these functions are installed at different locations in the system, such as substations and feeders. These functions include, fault detection and service restoration, feeder reconfiguration, voltage/var control etc.. In addition to system operation type functions, distribution automation systems can also be used for digital protection of substations and feeders.

This chapter will deal only with system operation related functions. Many people prefer to subdivide these functions into two groups, namely substation related functions and feeder related functions [1]. In this chapter such division of functions has not been considered. Each function selected may be applicable for both substation as well as feeders. In many situations, the functions at substation and feeder level are performed in a coordinated fashion, for example, the switching of capacitors on the feeders may be coordinated with the switching of capacitors at the substation. A list of functions considered here is as follows:

  1. Fault location and service restoration
  2. Feeder reconfiguration and transformer balancing
  3. Extension of transformer lifetimes
  4. Recloser/breaker monitoring and control
  5. Capacitor switching for voltage/var control
  6. Voltage control using regulators
  7. Substation transformer load-tap-changer (LTC) control
  8. Distribution system monitoring.

These function can be split into subfunctions as has been done in the EPRI report [2] . For the sake of simplicity we have decided not to take that approach. Since many functions depend on each other, we had to compromise with two conflicting objectives to finalize this list for determination of cost/benefit associated with these functions. If all the functions which depend on each other are merged into one function then the user has very little choice. On the other hand, if too much choice is given then programming becomes difficult and also the use of the program becomes very difficult.

Salient features of each of the selected functions are discussed in detail in the sections below. Under each function heading first the manual procedure for that function is discussed. Then the automated procedure using distribution automation systems is discussed. The benefits associated with automation of that function is also discussed.


2.1.1 Fault Location and Service Restoration

A distribution system, particularly an overhead type, is susceptible to various type of faults. In the event of a permanent fault, the protective devics are expected to operate and isolate the faulted section. However, if the fault is of a high impedance type, the protective devices may not operate to isolate the faulted section. In such situations, location of faults becomes more difficult. In both cases, some customers experience a power outage. Since no information on status of various devices in the distribution system is available to the distribution system dispatchers, there is no direct way to find out about the outages. Thus, the dispatchers depend on telephone calls from customers or a sudden change in power flow at a metered location upstream in the system to come to know of the outages. Customers' calls only provide an approximate location of the outage. Moreover, in case of a major storm the outages can be widespread and difficult to locate. Once the approximate location of outages is known, line crews are dispatched to drive along the lines to look for damage. After the damaged area is located, it has to be isolated from the rest of the system if the fuse protecting that line has not operated. This is done by first opening the substation breaker and then manually operating the switches or removing the fuses. Coordination between the line crews and dispatchers is maintained via portable radio to perform this task properly.

The next step is to restore power to those parts of the system which are undamaged but have lost power because of problems elsewhere in the system. The power to these parts may be provided from alternates routes. The dispatchers determine such possible routes and ask the line crew to operate the isolators. Most of the isolators cannot be operated under load, therefore, the substation breaker is opened before operating the isolators. Since the whole process is done manually, it may take a long time.

The automation of this function requires that sensors be installed at various locations in the distribution system. In addition to these sensors, remotely controlled sectionalizers are needed on the feeders at different locations. When the fault takes place, the affected sensors send the information to the central computer. The location of the fault is determined using the data received from the sensors. An alternate approach is to map the calls received from the customers automatically on the system map. From these maps inference can be obtained about possible locations of faults. Then the faulted section is isolated from the rest of the system with the help of remote controlled sectionalizers if the protective devices have not already isolated the faulted part. Subsequent to this, the switching needed to restore power to unfaulted parts of the system can be accomplished remotely. Moreover, since the location of the fault is known, the crew is sent to the precise location instead of asking to go in a general area. Thus the whole process of fault location and service restoration can be accomplished more efficiently by less people in much less time.

2.1.2 Feeder Reconfiguration and Transformer Balancing

The load in a distribution system varies by hour, by day and by season. For every load level, the system has an optimal configuration of feeders in which the system losses are minimum. Moreover, the total transformer losses can be minimized is the substation transformers are loaded in proportion to their capacity instead of loading some transformers very heavily and others very lightly. In a manual system, the reconfiguration of system is done on a seasonal basis, perhaps at the most a few times in a year. Since such reconfiguration may require several manual switching operations, it is not feasible to do it more frequently. Moreover, the configuration achieved through manual switching is hardly close to optimal.

The reconfiguration of the system for loss reduction can be accomplished in an automated mode using the sectionalizers which are used for fault isolation and service restoration. Only extra need will be for application software. Since the operation of the sectionalizers is controlled remotely, system reconfiguration can be done as frequently as the dispatcher desires. From a practical point of view it is possible to reconfigure the system every hour. Thus the feeder and transformer losses can be minimized.

2.1.3 Transformer Life Extension

The substation transformers normally operate at loads lower than their capacity. However, during emergencies, such as failure of another transformer, they can be operated at loads higher than the rated capacity. But overloading can be done only for a limited time without jeopardizing the life of the transformer. Higher the overloading, lower is the time allowed for overloading. In a manual process, the dispatcher has to rely on trial-and-error to get proper level of loading. The dispatcher would close the switch to supply additional load with an expectation that the total load will be less than a certain value. But if the load after switching happened to be higher than expected, he will have to open the switch, drop a few feeders, and then close the switch again. The process will have to be repeated until the load is at a desired level. The switching on and off of load can stress the transformer significantly and thus reduce its total life. Using an automated procedure, this task can be performed without stressing the transformer.

Automation of this function requires equipment for monitoring the transformer including the oil temperature. Data and measurements from the feeders connected to the transformer are also needed. The oil temperature is used to determine the level of overloading possible under the given conditions. Then the feeders can be selected such that there is a balance between the desired loading and the loads of the feeders. Since overloading of the transformers can be controlled precisely and many switching operations are not required, stress on the transformers can be avoided. Thus life extension of transformers can be achieved.

2.1.4 Recloser/Breaker Monitoring and Control

In the manual mode, no remote monitoring and control is available. The settings of the relay and recloser timings can be changed only by going to the substation. Since no monitoring is available, the recloser and breaker contacts are refurbished at fixed intervals weather it is necessary or not. This maintenance frequency is usually based on the duty level the recloser or breaker is expected to perform. Generally, the maintenence interval is estimated conservatively (i.e., refurbishments are made, on the average, sooner than is necessary). Hence, in many cases the contacts are serviced before it is necessary.

In an automated scenario the relay settings and recloser timings can be set remotely. This will allow for better control of the system whenever the system configuration changes. Moreover, the labor needed to reset the relay and recloser timings can be saved because these settings can be done remotely instead of going to the substation. The monitoring of the energy interrupted by the recloser and breaker can provide a precise estimate of the health of the contacts. Thus, using this information refurbishing of the contacts can be scheduled whenever necessary. Hence, too early or too late servicing of the recloser and breakers can be avoided.

2.1.5 Capacitor Switching for Voltage and Var Control

Capacitors are used in the distribution systems for voltage and reactive power support. The capacitors are placed at strategic locations to improve overall system operation. These capacitors can be fixed type or switched type. The switched capacitors switch-on or switch-off upon receipt of a signal from a controlling device attached to the capacitor. This control device may be a timer, a temperature sensitive relay, a voltage sensitive relay, a current sensitive relay, a reactive power sensitive relay, or a combination of the above. The timers are set based on an assumed load curve. However, on a given day the load may not be the same as the assumed load curve. Moreover, the timer does not discriminate a working day from a holiday. The temperature sensitive device is set based on an assumption that the load is high when the temperature is high because the air conditioning demand goes up during hot weather. This type of control does not work very well because there is a lag of few hours between the outside temperature and air conditioning load because of thermal inertia of the houses. Other types of controls also have problems, which are discussed in the available literature[3].

To alleviate some of the problems associated with the above mentioned schemes of controlling capacitors many companies have introduced microprocessor based controllers. These controllers perform better than the conventional controllers, but do not provide most optimal system configuration. A major drawback of these controllers is that they respond to the local conditions existing at the location of the capacitor. They cannot possibly take into account the impacts of switching a capacitor on other parts of the system.

In an automated system the switching of all the capacitors can be coordinated to achieve optimal system configuration under different load conditions. In such a scheme, meters are installed at different locations to measure real and reactive power, voltage and current. The metered data and status of capacitors is sent to the central computer via communication lines. The computer then determines the optimal switching configuration of the capacitors for the measured system conditions. Under the optimal configuration the system losses are at a minimum value. Since the system has real-time measurement capabilities, the switching configuration can be changed as frequently as desired. However, from a practical point of view it is not desirable to switch capacitors more frequently than perhaps once every hour.

2.1.6 Regulator Operation for Voltage Control

Voltage regulators are used in the distribution for finer control of voltage, particularly on long distribution lines where voltage drops are high. These regulators are set to maintain voltage within a specified band. When the voltage becomes lower than the low setting, the tap on the regulator moves to increase the number of turns of the output side. Similarly, when the voltage goes above the high setting, the tap moves to reduce the number of turns on the output side. Although the tap moves automatically it only uses the voltage near its location as the input. It does not have any coordination with other regulators in the system.

Automation of regulators provides centralized coordination of the regulator settings. Moreover, the regulators can also be coordinated with capacitor switching. Thus, the regulators can be set to reduce losses under different load conditions. Coordinated control of regulators will also provide better voltage profile on the feeders and, hence, lesser low voltage complaints from the customers.

Another advantage of having remote control of regulators is during emergencies. It is generally believed that the load on the substation can be reduced by reducing voltage. Thus, before an anticipated emergency, the crew go to the substation and manually override the load-tap-changing (LTC) control of the transformer. With remote control of regulators, voltage on some of the feeders of the system can be reduced, and thus reducing the load partially. Hence, the visits to substation to override LTC can be avoided.

2.1.7 Transformer LTC Control

The substation transformers have LTC control which changes the tap position in response to load. Since higher load results in higher voltage drop, the tap moves to higher position to maintain the voltage at proper level at the end of the distribution line. Similarly, under low load conditions, the tap moves to lower setting to compensate for the increase in voltage due to lesser voltage drop on the line. If the substation has two transformers that are operated in parallel, the LTC control devices on the two transformers coordinate with each other to maintain the same output voltage to prevent circulating current in the transformers. The existing control devices work quite well, therefore, there is little benefit from further automating this function using remote control. However, since the new control devices are digital, they may need lesser maintenance and diagnosis of malfunctioning devices will be easier. A major advantage of remote control of LTC becomes realizable under emergency situations. As mentioned in the previous section, load can be reduced by reducing the voltage. Hence, using remote control the tap on the transformer can be moved to a lower setting under emergency conditions to alleviate load. Thus, visits to substation which will be necessary to manually override the control and set the tap to a lower value will be avoided.

2.1.8 Distribution System Monitoring

The purpose of distribution system monitoring is very similar to SCADA in the traditional sense. Monitoring is necessary to acquire data for many of the distribution functions. Some of these functions require real-time data from the system to make control decisions. Real-time data is also useful in providing information to operators on abnormal system conditions in the form of alarms. In addition to the real-time data, system data can be gathered and archived for later use. Such data can then be used for forecasting and planning purposes. As defined in the EPRI report [2], there are three components of distribution system monitoring, namely, data monitoring, data logging, and analog data freeze. For the purpose of cost-benefit analysis, these functions are considered to be in addition to the measuring equipment needed for implementation of the other distribution automation functions.

The main purpose of data monitoring is to maintain system databases for alarms, user interface and logging. Thus, under abnormal system conditions alarms can be enunciated to alert the operator of those conditions. In addition to the alarm, operators are provided with the relevant data, which can be used by them to take corrective actions.

The main purpose of data logging is to prepare printed reports of the system operating conditions or events for future use. The data types that can be logged using this function are various [ ], for example, alarms and their summaries; periodic logs, such as off normal summary, substation bus voltage log, tagged and out-of-service equipment; on demand logs, such as present values of the variables, limits, settings, and status; log of operator control actions; one-line diagrams of substations and feeder real-time configurations; fault reports; sequence-of-event logs.

Analog data freeze function gives a ''snapshot" of the quantities of interest. This function can be set to start capturing pertinent data based on threshold value of certain variables. Thus, the system conditions prior to a disturbance can be obtained. This information can be used by the operators to restore the system to original conditions following a disturbance. The operators can also freeze the data during normal operating conditions. Such data can be used by the operators to study the system and for planning purposes.

2.2 Cost-Benefit of Distribution Automation Functions

Determination of costs and benefits associated with different distribution automation functions is very important for making decisions regarding implementation of these functions. A microcomputer program, PC-ADAM (Analysis ofDistribution Automation Methods) providea cost/benefit analysis of the eight functions presented in the previous section. The developement of the economic models for some of the distribution automation functions is based on the EPRI report [2]. However, most of the models were changed as a result of the different emphasis of the study for which PC-ADAM was developed. Moreover, some of the model parameters have been defined differently to allow a utility analyst to more easily estimate their values.

The EPRI methodology is designed for planning new systems, i.e., the EPRI component models are for the analysis of conventional versus automated functions when building a new distribution system. Thus, they compare automated case with base (conventional) case over a number of years to determine costs and benefits. In PC-ADAM, by contrast, the cost-benefit methodology is applied to the retro-fitting or upgrading of an existing distribution system for which conventional equipment has previously been purchased and installed. This latter approach is a much more severe test of the economic feasibility of using distribution automation, since the cost of automated equipment is not offset (as in the EPRI study) by the avoidance of conventional equipment purchases. At best, replaced conventional equipment contributes salvage value in this study. Moreover, since expansion and planning of system are not considered, benefit for deferring installation of substation equipment, such as transformers are not included. However, proper credit for generation capacity release has been included wherever pertinent.

Standard engineering economic methods based on a utility's discount rate and constant fixed charge rate are used to estimate the present worth of all costs and benefits associated with each function. The program, PC-ADAM, calculates benefit/cost ratios for each selected function as well as an overall benefit/cost ratio for the entire mix of selected functions. Also, a program option allows the analyst to perform a sensitivity analysis to quickly see which parameters influence most strongly the overall benefit/cost ratio.

2.2.1 Assumptions Used in The Economic Analysis

In the evaluation of any proposed distribution automation strategy, many different costs and benefits must be considered. The description of the cash flow associated with the many costs and benefits over the study period can be treated in several ways. The procedure and simplifying assumptions are outlined below.

(1) Retrofit of an Existing Distribution System

It is assumed that the distribution system to be analyzed already exists. PC-ADAM then determines the economic feasibility of upgrading the system to include any combination of eight automated distribution functions. Consequently, new automation equipment will have to be procured and will replace some conventional equipment which, at best, will have only salvage value. The present methodology is not designed to analyze the construction of a new distribution system for which either conventional or automated equipment could be purchased.

(2) Study Period

The analysis of the economic feasibility of the selected distribution automation options is for a fixed time period, i.e., all costs and benefits occurring within some specified number of years are estimated. The number of years N is arbitrary, varying from 1 year to infinity for a perpetual analysis (amortization) period.

(3) Time Line for Payments and Benefits

All capital and initial installation costs are assumed to occur at the beginning of the study period. Subsequent costs and benefits are then calculated on a year-by-year basis with the costs or credits being accounted for at the end of the year in which they occur.

(4) Present Worth Approach

To evaluate the costs and benefits of any particular group of distribution automation functions, the present worth (at the beginning of the study period) of all costs and benefits associated with the selected functions over the duration of the study period is computed. From these present worth calculations, a benefit/cost ratio is then computed so as to give a single quantifiable measure of the economic feasibility of the selected distribution automation options.

(5) Discount Rate

To obtain the present worth of a expense or benefit, the discount rate D for the utility is used. This discount rate is assumed to remain constant over the study period.

(6) Capital versus Expense Costs

When evaluating different costs for utility planning, it must be realized that some costs are capitalized (i.e., financed over a period of years from revenue income) while others are expensed in the year in which they occur. To put all costs on an equivalent basis, this study converts all capital expenditures to an equivalent revenue requirement by using the concept of a fixed charge rate (FCR). This fixed charge rate is assumed to remain constant over the study period.

(7) Constant Failure Rates

The frequency with which various types of distribution equipment fail is an important parameter for determining maintenance costs. In this study, the failure rate of distribution devices is assumed to remain constant over the study period so that annual O&M costs (correcting for inflation) remain constant.

A more realistic failure model would allow for increased unit failures during the initial burning period and towards the end of life as the units wear out. However, failure data for the distribution automation devices considered in this study are relatively scarce and, consequently, parameter values for such elaborate failure models are unavailable. Moreover, such detailed failure models are hard to justify for use with the relatively simple economic assumptions (e.g., constant discount rates, constant fixed charge rates, constant escalation rates, etc.) employed in this study.

(8) Inflation

The many costs and benefits associated with distribution automation can be expected to escalate over time. While different escalation rates are used for different costs and benefits, the escalation rate for any particular cost category is assumed to remain constant over the study period.

(9) New Equipment Costs

All the distribution automation functions considered in this study require some new equipment. Generally, the initial cost of such equipment is assumed to include the installation cost. Further, these costs are assumed to be capital expenditures.

(10) Salvage Values

In this study no consideration is given to the value of used but functional equipment at the end of the study period, i.e., salvage credit is not considered. Generally, such credits are small and can be neglected, especially if a long (>15 years) study period is used. Also when considering a distribution automation option that requires new equipment to replace existing equipment (e.g., new metering), no credit is explicitly given for the replaced components (although salvage value for the replaced components can be approximately accounted for by taking the new equipment cost to be the actual cost less the salvage value for the old components).

2.3 Benefit/Cost Ratio Analysis

If many distribution automation options are considered for simultaneous implementation, the large number of different costs and benefits often make it difficult to see the overall economic attractiveness of the selected functions. To quantify the degree to which the options are economically feasible, it is usual to calculate an overall benefit/cost ratio.

In the calculation of a benefit/cost (B/C) ratio, there is always some ambiguity as to whether a particular expense or capital item is to be classified as a "benefit" or a "cost". For example, the replacement of a conventional piece of distribution equipment by a device that fails less frequently but has higher operational costs may result in greater or smaller overall maintenance costs. Thus the economic impact of many economic components may be negative or positive. Rather than classify all values as benefits or costs a priori, the present study defines any positive savings to be a "benefit" while all revenue losses or expenditures by a utility are considered "costs". In this way, there are no negative benefits or negative costs. With this convention, a particular item may be a benefit under one set of conditions and a cost under another set.

This definition for costs and benefits has two principal advantages. First, no arbitrary classifications of economic items need be made. Second, it can be shown that treating negative values as "negative benefits" gives B/C ratios than are artificially farther from the break-even B/C=1 value than would be obtained by treating such negative items as costs. In other words, when B/C > 1, negative benefits yield larger B/C ratios thereby exaggerating the economic attractiveness of the option under study.

To compute a benefit/cost ratio for any combination of distribution automation options, the present worth of all annual cost components associated with the options are computed over the specified time period. Cost components which represent capital investment or savings are first converted to revenue requirements using the fixed charge rate concept and then added to those components that are treated as expenses. Thus for a study period of N years, the benefit/cost ratio is computed as



ES(n) = total savings in expensable costs in year n resulting from distribution automation options (n > 0); ES(0) represents immediate initial benefits (if any)
RS(n) = total savings in capital costs in year n (n > 0); RS(0) represents initial savings in capital costs (if any)
REV = revenue-to-expense conversion factor
E(n) = total expenses in year n necessary to implement all distribution automation options (n > 0); E(0) represents any initial expenses
R(n) = capital investment in year n needed for all the distribution automation options (n > 0); R(0) represents any initial capital costs

To estimate the total expenses and capital requirements for a given year of the study period, all savings and cost components for each distribution automation option are evaluated. An example of various economic components and the models used for their computation are described in the following section.

2.4 Economic Models For DA Functions

For each of the eight distribution automation functions considered in this study, there are generally many different costs associated with each option as well as several potential economic savings. Besides costs directly attributable to each option, there are also planning and initial base-line equipment costs which are incurred no matter what mix of automation options are selected. As an example, economic models of benefits and costs associated with fault location and service restoration are presented in this section. Models for other functions are available in a report [4].


2.4.1 Initial Planning

An important cost component in implementing any of the distribution automation options considered in this study is the cost of technical and management personnel needed to select particular automation functions and to plan for their installation. Moreover, many of the automation functions in this study can use the same basic central control and communication equipment, and only incremental specialized or add-on equipment costs need to be ascribed to particular automation options.

Thus in the present approach, planning and base-line equipment costs are usually included in the upfront or initial costs, while incremental central and site-specific equipment are then assigned to the various automation functions. However, the analysts may elect to give zero value to these upfront planning and central equipment costs and to assign a pro rata share of these planning and central equipment costs to each automation function.

The advantage of separating out general planning and central-equipment costs is that the different automation options can be more fairly compared since no one option will be unduly burdened by assigning all upfront general costs to it. Of course, the overall benefit/cost ratio for all the selected options will be unaffected by how these initial planning and base-line equipment costs are allocated.

Planning Costs

An important expense involved with initiating any distribution automation policy is the initial planning by management and technical personnel. Often these initiation costs are overlooked or are assumed to be a normal part of utility operations; however, they represent real manpower commitments and should not be ignored in a comprehensive benefit/cost study.

This initial planning cost could be set to zero, and the planning costs ascribed to each separate automation function selected on a pro rata basis in the initial incremental cost provided for each function. Generally, however, it is easier to enter a single initial planning cost. The initial one-time planning cost is described in this study by a single parameter



$PLANNING = cost of initial management and technical planning for implementing all selected automation functions. Generally this cost will increase with the number of options. ($)
Acquisition of Base Equipment and Software

Regardless of the mix of automation functions selected, there will be some basic central monitoring, control, and communication equipment that is needed and which can be shared by several of the automation options. For example, the data collection computer used for distribution-line monitoring can also be used for fault detection and service restoration. Similarly, the communication equipment used to receive sensor data from the field and to transmit control signals can be shared by several automation options. In addition to base hardware costs will be associated costs for software needed to operate the monitoring and control equipment and for training personnel in this use of the automated distribution functions.

Rather than try to apportion the base equipment costs to specific options, one initial cost for central equipment may be used. Moreover, each specific automation function generally will also require additional specialized equipment and software whose costs are considered as part of the function's costs. The one-time base line equipment and training cost is


Cost= ($BASE)(REV)


$BASE = purchase and installation costs of base-line equipment, software, and training. This cost generally will increase slightly with the number of automation functions selected ($)
REV = expense-to-revenue conversion factor for equipment and software that is to be capitalized
O&M of Central Base Equipment

Associated with the central base-line equipment will be annual operation and maintenance expenses. Such costs include expenses for operating and supervisory personnel, software updates, testing and repair, report generation, performance evaluation, training, etc. For year n, these costs are estimated as


Cost(n) = ($OM.BASE)(ESCOM(n))



$OM.BASE = annual operating and maintenance expense for the base-line equipment. Software updating costs should also be included in this item. ($/y)
ESCOM(n) = escalation factor for O&M costs in year n
Unanticipated Benefits From Automation

There are many intangible benefits to be gained from an automated distribution system that cannot rigorously be accounted for in a benefit/cost study. For example, unanticipated system problems of varying severity arise routinely and it is reasonable to expect that many can be solved more quickly and inexpensively with the help of the automated distribution functions considered in this study. The automated capability will also improve the quality of power, an issue of contemporary importance whose benefits are difficult to quantify. To account for these annual intangible and unforeseen benefits, a single benefit value in year n is introduced as

Benefit(n) = ($INTG.DA)(ESCLAB(n))


$INTG.DA = the annual value of intangible or unforeseen benefits from having an automated distribution system. This value depends on the mix of automation functions selected. ($)
ESCLAB(n) = escalation factor in year n for labor (here assumed equal to that for intangible benefits)
2.4.2 Fault Location and Service Restoration

With the help of sensors installed on distribution feeders, faults can be detected and isolated quickly. Then those parts of the system which have lost power but are unfaulted are supplied power through alternative paths. The system may experience only one fault or may have multiple faults such as might occur, for example, during a major storm.

As an alternative to using sensors to check the status of various distribution lines, fault detection can also be done by interfacing the distribution automation system with a customer call-in and mapping system thereby avoiding the need to purchase and install feeder sensors. In this case, the cost of central equipment will be increased while the expense for sensors on the feeders can be eliminated.

In either approach, the detection of faults can be done very quickly with an automated system. Moreover, service can be restored rapidly if additional transducers, which provide the central computer with information on the feeder loads, are also installed. This rapid restoration can yield considerable savings in labor costs compared to conventional field restoration. Faster restoration of service also provides more revenues and higher goodwill of customers.

To implement this function, fault detecting sensors, remotely controlled sectionalizers, feeder load transducers, and software to determine switching sequence must be purchased and installed.


Reduced Labor for Fault Location

With centralized monitoring of the status of distribution lines, fault locations can be more rapidly determined. The potential benefit for reduced crew time to locate feeder faults for year n is given by

                  + (OP.MHR/FLT)(OPER$/MHR)](ESCLAB(n))


FAIL/CK.MI = average number of faults per year per distribution circuit mile that require restoration effort (faults/y/mi)
CKT.MILES = number of linear circuit miles under automation (mi)
MHR/FAULT= man-hours required to locate distribution feeder fault without the aid of distribution automation (man-hours)
LC$/MHR = cost of labor for a line or substation crew ($/man-hour)
OP.MHR/FLT = decrease in the number of dispatcher man-hours needed to locate a distribution feeder primary fault using automated fault-location technology compared to conventional location methods (man-hours)
OPER$/MHR = dispatcher cost to operate the automation system ($/man-hour)
ESCLAB(n) = escalation factor in year n for labor costs (assumed the same for line crews, substation workers, dispatchers, engineering services, etc.)
Reduced Labor for Isolating Faulted Feeder

The potential annual benefit for reduced crew time needed to switch and isolate a primary feeder fault is given by

                   x (LC$/MHR) + (MHR/ISOL)(OPER$/MHR)](ESCLAB(n))



FAIL/CK.MI = average number of faults per year per distribution circuit mile that require restoration effort (faults/ y/mi)
CKT.MILES = linear circuit miles under automation (mi)
LOC/FAULT = number of isolation points per primary fault
MHR/SWT = man-hours needed for one manual switching operation of the feeders (man-hours)
LC$/MHR = cost of labor for a line or substation crew ($/man-hour)
MHR/ISOL = decrease in the no. of dispatcher man-hours needed to isolate a distribution primary feeder fault using automated isolation technology compared to conventional isolation methods (man-hours)
OPER$/MHR = dispatcher cost to operate the automation system ($/man-hour)
ESCLAB(n) = escalation factor in year n for labor costs (assumed the same for line crews, substation workers, dispatchers, engineering services, etc.)
Reduced Labor for Service Restoration

The potential annual benefit in year n for reduced crew time for service restoration is given by

                  x (LC$/MHR)+ (MHR/REST)(OPER$/MHR)] (ESCLAB(n))


FAIL/CK.MI = Average number of faults per year per distribution circuit mile that require restoration effort (faults/y/mi)
CKT.MILES = number of linear circuit miles under automation (mi)
SWTS/FAULT = number of switching operations needed for service restoration per primary feeder fault
MHR/SWT = man-hours needed for one manual switching operation of the feeders (man-hours)
LC$/MHR = cost of labor for a line or substation crew ($/man-hour)
MHR/REST = decrease in the number of dispatcher man-hours needed to perform service restoration with automated distribution technology as compared to conventional restoration methods (man-hours)
OPER$/MHR = dispatcher cost to operate the automation system ($/man-hour)
ESCLAB(n) = escalation factor in year n for labor costs (assumed the same for line crews, substation workers, dispatchers, engineering services, etc.)
Increased Sales From Faster Restoration

The potential benefit for increased revenue in year n due to faster service restoration is given by

Benefit(n) = (NET$/KWH)(AVG.DMD)(%SYS.LOST)(NO.CUST)
                   x (FAIL/CK.MI)(CKT.MILES)(HR/OUT)(ESCP(n))/100


NET$/KWH = average electrical energy cost to the utility customer minus the average generation cost ($/kWh)
AVG.DMD = average diversified customer demand (kW).
%SYS.LOST = average percentage of the distribution system lost during a typical outage (%)
NO.CUST = number of customers served from feeders under study
HR/OUT = average decrease in the outage hours per event afforded by automated system restoration (i.e., difference in outage duration between non-automated and automated system (h)
FAIL/CK.MI = Average number of faults per year per distribution circuit mile that require restoration effort (faults/y/mi)
CKT.MILES = number of linear circuit miles under automation (mi)
ESCP(n) = escalation factor in year n for average energy costs (sales or fuel costs)
Increased Customer Satisfaction

The potential benefit in year n for customer goodwill due to decreased outage time and reduced complaint cost is given by

                   x (NO.CUST)(ESCGW(n))/1000



CMPL.RAT = ratio of the number of customer complaints for service interruption with restoration automation to the number without automated restoration
$/COMPLNT = cost per customer interruption complaint for either an outage or low voltage complaint ($)
INT.COMPL = average number of customer interruption complaints per year per 1000 customers served (without automation)
NO.CUST = number of customers served from feeders under study
ESCGW(n) = escalation factor in year n for the value of customer goodwill, or, equivalently, for the cost of a customer complaint for service interruption or low voltage ($)
Purchase of Feeder Sectionalizing Equipment and Software

New remote controlled feeder sectionalizing equipment must be purchased and installed to permit automated fault isolation and service restoration. Manual switching equipment will be replaced, and any salvage credit is assumed to be reflected in reduced capital costs of the new equipment. There may also be specialized software costs for this automation function as well as for additions to the base-line equipment. Thus the initial cost of the automated sectionalizing devices and software is

Cost = [(AFS/CKT)(NO.CKTS)($/AFS) + ($AFS.PGM)
              + ($BASE.AFS)](REV)


AFS/CKT = number of automated feeder sectionalizing devices installed per feeder circuit
NO.CKTS = number of feeder circuits in the system under study
$/AFS = cost of each automated feeder sectionalizing device (including installation and less any salvage value of replaced equipment) ($)
$BASE.AFS = cost of additions to the central base-line equipment needed for fault detection and distribution system restoration ($)
$AFS.PGM = incremental cost (above base system) of software for implementing the fault isolation and service restoration option and any associated training ($)
REV = expense-to-revenue conversion factor for equipment and software that is to capitalized
Purchase of Sensors and Transducers

To perform automated fault detection it is necessary to purchase and install fault sensors on the distribution system to indicate which of the feeders are active. On the other hand, if the phone call-in mapping approach is used then fault sensors are not needed. With either approach for detecting faults, it will be necessary to install load transducers on the feeder lines to perform centralized restoration of the distribution system. The initial cost for this equipment is




SENS/CKT = average no. of fault detection sensors per feeder circuit
$/SENS = purchase and installation cost of each fault detection sensor ($)
TRDC/CKT = average number of load transducers per feeder circuit
$/TRDC = purchase and installation cost of each load transducer ($)
NO.CKTS = number of feeder circuits in the system under study
REV = expense-to-revenue conversion factor for equipment and software that is to capitalized
O&M of Fault Sensors and Load Transducers

The replacement of failed fault detection sensors and general maintenance and testing of the sensors will result in an expected annual cost in year n of


               + ($OM/SENS)(ESCOM(n))] + (TRDC/CKT)(NO.CKTS)
               x [(FAIL/TRDC)($RPR/TRDC)(ESCRPR(n))
               + ($OM/TRDC)(ESCOM(n))]


SENS/CKT = average no. of fault detection sensors per feeder circuit
NO.CKTS = number of feeder circuits in the system under study
FAIL/SENS = failure rate per fault detection sensor (no. failures per year)
$RPR/SENS = cost of repairing or replacing a failed fault detection sensor ($)
$OM/SENS = annual O&M cost per fault detection sensor ($/y)
TRDC/CKT = average number of load transducers per feeder circuit
FAIL/TRDC = failure rate per load transducer (no. failures per year)
$RPR/TRDC = cost of repairing or replacing a failed load transducer ($)
$OM/TRDC = annual O&M cost per load transducer ($/y)
ESCEQUIP(n) = escalation factor for equipment costs in year n
ESCRPR(n) = escalation factor for equipment repair costs in year n
O&M for Feeder Switches and Controllers

There will be annual costs incurred for repairing failed automated feeder switches as well as O&M costs for these devices. These cost will be partially offset by the elimination of similar costs for the conventional (non-automated) feeder switches. For year n these costs are


Cost(n) = (AFS/CKT) [(FR.ASWT)($RPR/ASWT)(ESCRPR(n))
               + ($OM/ASWT)(ESCOM(n))] (NO.CKTS) 
               - (MSWT/CKT)[(FR.MSWT)($RPR/MSWT)(ESCRPR(n))
               + ($OM/MSWT)(ESCOM(n))] (NO.CKTS)


AFS/CKT = number of automated feeder sectionalizing devices installed per feeder circuit
FR.ASWT = annual failure rate per automated feeder (failures/y) switching system (failures/year)
$RPR/ASWT = repair or replacement cost of a failed automated feeder switching system ($)
$OM/ASWT = annual O&M cost for each automated feeder switching device ($/y)
NO.CKTS = number of feeder circuits in the system under study
MSWT/CKT = number of non-automated feeder switching devices per circuit
FR.MSWT = annual failure rate per non-automated feeder switching system (failures/year)
$RPR/MSWT = repair or replacement cost of a failed non-automated feeder switching system ($)
$OM/MSWT = annual O&M cost for each non-automated feeder switching device ($/y)
ESCRPR(n) = escalation factor for equipment repair costs in year n
ESCOM(n) = escalation factor for O&M costs in year n


  1. D.L. Brown, J.W. Skeen, F.A. Rahimi, and P. Daryani, "Prospects for Distribution Automation at Pacific Gas & Electric Company," IEEE Transactions on Power Delivery, Oct~91, pp.~1945-1954.
  2. J.B. Bunch, Guidelines for Evaluating Distribution Automation, EPRI Report EL-3728, Nov 84.
  3. B.W. Coughlan, D.L. Lubkeman, and J. Sutton, "Improved Control of Capacitor Bank Switching to Minimize Distribution System Losses," The Proceedings of the Twenty-Second Annual North American Power Symposium , Oct 90, pp. 336-345.
  4. J.K. Shultis and A. Pahwa, Economic Models for Cost/Benefit Analysis of Eight Distribution Automation Functions, Report No. 234, Engineering Experiment Station, Kansas State University, June 1992.