1 INTRODUCTION

Utilities throughout the world are facingunprecedented change. Privatisation ofgovernment-owned utilities, competition forwholesale and even retail customers, and mergersand acquisitions have added new elements of risk tothe management of utilities today. No longer areenergy and water providers able to rely on regulatoryprotection of territory and customers; those who arenot affected by direct deregulation are facingincreasing awareness by their key customers of priceand service options. New tools and strategies arerequired for capital-intensive companies to staycompetitive in the marketplace. Better use of spatialdata is one of the key areas of focus for manyelectrical, gas, and water utilities.

Improved hardware, software, and networkingtechnology have created opportunities for the utilityindustry to build and benefit from morecomprehensive and sophisticated GIS. GISapplications have evolved from their foundation inmap production to advanced analysis tools forplanning and operations. GIS products arecommonly used by utilities for marketing, facilitieslocation, and engineering applications (Birkin et al,Chapter 51). There are two common objectives of

these applications: driving down costs andimproving customer service.

This chapter provides an overview of how GIS isbeing used in the utility industry to help meettoday’s needs. It begins with a brief discussion ofhow GIS has evolved in electrical, gas, and waterutilities to its current technological state. Theprincipal focus of the chapter is a discussion of thecurrent state of AM/FM/GIS, including adescription of GIS data sources, critical traditionaland non-traditional applications, and costs andbenefits. The chapter concludes with a brief sectionon how GIS implementation may impact uponcompetition and change in the utility industry.

2 EVOLUTION OF GIS IN THE UTILITIES

The primary goal of any utility is to plan andmanage the use of facilities to deliver acommodity such as water, natural gas, or electricityto its customers. The utility industry has alwaysrelied on hardcopy maps to manage facilities,so it was natural that electric, gas, and watercompanies should be among the first users ofdigital mapping software.

801

57GIS in the utilities

J R MEYERS

This is a time of major change in the utility industry. Privatisation, deregulation, and theintroduction of advanced technology are all having profound impacts on utilities. Thischapter examines how GIS has been used by electrical, gas, and water utilities. It focuseson automated mapping, facilities management, and geographical information systems(AM/FM/GIS). Critical utility applications include map editing, creation of map products,facility query and display, design/work order processing, network modelling, and executiveinformation systems.

The costs and benefits of implementing AM/FM/GIS in utilities are briefly examined. It isconcluded that any assessment of the impact of GIS on utilities must look at the widerpolitical and economic considerations of competition and change.

The earliest efforts at automation during the 1970sand early 1980s were focused around producing digitalmaps. Since this was largely a graphics-driven task,most utilities chose Computer Aided Design/Drafting(CAD) software as the basis for systems which cameto be called Automated Mapping/FacilitiesManagement (AM/FM). A number of approachesgeared to map reproduction were taken, based aroundthe CAD technology of the era. The original use ofthis technology was expensive, with costs ranging fromUS$ 50 000 to 100 000 per user and more. Long rangebenefits were projected, but rarely realised owing tothe limitations of the available technology, whichcentred on the problems of linking databases to setsof graphic primitives.

Among systems in use during this time frame, onemainframe-based system was developed whichfocused on a data modelling approach to theproblem. Called Geo-Facilities Information System(GFIS), this technology built an elegant andcomprehensive model of facilities networks, usingIBM’s IMS database technology to house the datamodel. Despite the expensive nature of the solution,it was popularised by several large utilities in NorthAmerica and Europe. Many observers credit thisIBM system with the first true network model usablefor tracing electric circuits, gas, and water networks,and for allowing performance analysis of networkmodels (Antenucci et al 1991).

As mid-range, minicomputer-based systemsbecame popular replacements for mainframe-basedtechnology, AM/FM systems which relied onhierarchical or network data models forrepresentation and storage of facilities data began toexploit this new platform. Relational databasetechnology came to the forefront in the mid 1980s.In a separate field of endeavour, GIS werepopularised in the early 1980s, largely in the purviewof land and natural resource management(Coppock and Rhind 1991). These systems weregeared towards abstracting different layers ofinformation about a geographical area andproviding a topological model (i.e. a representationin which spatial relationships among geographicalfeatures could be analysed: see Martin, Chapter 6).By the late 1980s, many utility users of AM/FMbegan to embrace the ideas of GIS, using thegeorelational technology to model and map facilitiesand land. In this model, spatial definitions of landand facilities are stored in terms of their coordinatesin binary flat file format, while attributes about those

entities are stored in relational databasemanagement system (RDBMS) tables (seeOosterom, Chapter 27). The ability to relatecustomer and accounting systems to spatialrepresentations of the facilities in a service area(area covered by a utility) transformed a mappingsystem into an information system.

Like many popular acronyms, both AM/FM andGIS have come to assume very broad definitions,perhaps even broader than their originatorsintended. Many use the term GIS to mean the entiresystem of hardware, software, and data (and evenpeople), rather than just the particular technologyfor building a digital map (Maguire 1991). Othersuse the acronym AM/FM to mean any and alltechnology used in the field. These two aresometimes used together (as AM/FM/GIS) and areoften used interchangeably. Most utilityimplementations of GIS, when they include themapping and modelling of networks, are referred toas AM/FM or AM/FM/GIS.

The wide use of AM/FM/GIS among utilities hasprovided the motivation for some key technologicalimprovements, driven by the different and in someaspects more difficult requirements of detailedfacilities models. The key improvements since themid 1980s include:

● open systems that have the ability to interconnectwith and manage commercial RDBMS(Sondheim et al, Chapter 24);

● ability to manage edits on individual modelentities (e.g. a single pipe, valve, or wire), knownas ‘feature locking’ database management;

● ability to provide long transaction and versionmanagement functions, to permit multiple usersto edit or modify the same geographical data atthe same time (Newell 1994);

● a move to Windows-based/desktop systems withno decrease in functionality or performance(Elshaw Thrall and Thrall, Chapter 23);

● use of object-oriented programming languagesand modelling techniques (Maguire, Chapter 25;Worboys, Chapter 26).

Today’s typical AM/FM/GIS is built using someform of client/server model with both geographicaland attribute data stored on a central server which isaccessed by clients in several different environments.It may serve from one or two to hundreds of users,and almost certainly has the capacity to share datawith applications which are external to the GIS(Gittings et al 1993). It will have a graphical user

J R Meyers

802

interface (GUI) which has been customised to suitthe specific utility’s business operations and a systemfor transaction management (that is, a procedure formanaging how multiple users interact with the samedata at the same time). The modern AM/FM/GISwill perform analyses based on topologicalrelationships, such as tracing lines to find customersaffected by a power cut, and data relationships, suchas isolating pipelines constructed from a specificmaterial or installed before a certain date.

New trends in AM/FM/GIS technology includemore sophisticated geographical data types andadvanced software tools. The overarching trend inthe industry is the convergence of technologies whichbegan as isolated tools for mapping and recordkeeping into integrated systems where it is difficult todistinguish geographical and other related attributedata. The second major trend is a move towards theavailability of desktop applications through the useof GIS servers which supply distributed clientapplications, without the overhead of large databasesand complex GIS models (Coleman, Chapter 22).There is increasing interest in object-orientedmethodologies for defining complex object classes(Worboys, Chapter 26; Martin and Odell 1995) andstandard development environments and analysistools (see Maguire, Chapter 25).

3 GIS DATA: SOURCES, MODELS, ANDAUTOMATION TECHNIQUES

Data are the foundation of a utility GIS. While it isdifficult to generalise, it is common for the cost ofdata to be 60 to 80 per cent of the total first cost ofan AM/FM/GIS solution (see Rhind, Chapter 56,for a general discussion of data costs and pricing).Utility data automation is complicated by thediverse quantity and quality of the sourcedocuments, the detailed facilities and base-mapdata that may be necessary to support applicationrequirements, the existence of segregateddepartmental databases containing existing facilitiesrecords, and even the lack of standards forcartography among different departments andgeographical regions of a company. Accordingly,any attempt to generalise about the cost of datacapture is nearly certain to be misleading. That said,the graph presented in Figure 1 gives a range ofexpected data capture costs in terms of US$ percustomer served from several utilities surveyed.

Data may be grouped into low, medium, and highaverage cost acquisitions, both for gas and electriccompanies. Not surprisingly, higher average costs areincurred by electric companies, since these tend tohave more detailed models. Yet costs incurred inrespect of low and medium cost sources tend to be

GIS in the utilities

803

Fig 1. AM/FM/GIS data capture costs.

Low

/med

ium

/hig

hav

erag

e co

st

$3 $4 $5 $6 $7 $8 $9 $10 $11 $12

US$/customer

$13

Electric

Gas

higher for gas companies, and in general the costvariance for gas companies is smaller than that ofelectric utilities.

The following section offers a brief review of thekey sources of AM/FM/GIS data, describes therelationship between data model and applications,and overviews some critical automation techniques.

3.1 Key sources of AM/FM data

A data conversion project at a utility consists of atleast three phases:

● identifying existing data sources and convertingthem to digital formats;

● defining a data model which fits bothgeographical and business process needs;

● designing screen displays, output maps, analysisfunctions, and reporting features that the GISshould produce.

By far the largest and most common source of dataused in AM/FM/GIS is existing paper records(Rogers 1996). Most utilities use a variety of maps ofvarying scales, content, and detail. Among thecommon paper map types are:

● inventory, distribution, or physical maps showingaccurate locations for facilities (typically large-scale, detail maps);

● electrical schematic/feeder (or network maps) orgas/water system pressure zone maps showingconnectivity (often not-to-scale);

● special facilities maps, such as cathodic protection(a technique used to save metal pipes fromcorrosion by passing a positive charge though thepipe) or valve testing location maps;

● leak survey maps;● rights-of-way or easements maps (containing

detailed cadastral information);● service record cards (usually with a sketch and text);● corridor or strip maps (usually on odd-sized, long

and narrow media) showing pipe runs, accesspoints, etc.;

● municipal/regional government and privatesources (base-map).

Dowman (Chapter 31) provides an overview of theproblems inherent in capturing digital maps frompaper sources. Existing paper maps are not just thesource of facility location, but also a link toattribute data which are often already in digitalform. Such digital databases offer a wealth of

information about plant and customers for utilities.Among the more common digital sources forcompanies are Customer Information Systems(CIS), Continuing Property Records (CPR) or AssetManagement Systems, and Plant Accounting andEquipment Databases. These corporate systems areoften supplemented by departmental databases forstorage of individual facilities data, like valverecords, or transformer or protective equipmentdatabases. These are often small, PC-based systems,built and maintained by a specific department orgroup within a company. The data in these databasesare often supplemented by personal knowledge ofthe staff, which is not recorded in any digitaldatabase, paper map, or other document.

Existing paper and digital records may provide thebest information available, but may still be inadequateto support AM/FM/GIS applications. Examplesinclude paper maps which do not depict the locationof facilities, are out of date, or do not show therelationships required. In these circumstances, fieldsurveys are often undertaken and increasingly globalpositioning systems (GPS) technology is employed(Lange and Gilbert, Chapter 33).

3.2 AM/FM/GIS data and applications

There are three key aspects of the utility data modelwhich contribute to the success of an AM/FM/GIS.The first is the accurate (at least in relative terms)depiction of the locations of all physical facilities,including lines, devices, and structures, which makeup the network. Shown in conjunction with base-map and foreign facilities data (from other utilitycompanies), this model provides the basis forautomated map-making. When attributeinformation is added the data model can be used formany basic facilities management functions. Thesedata support the basic spatial and attribute queryoperations which are familiar to GIS users. Theyadditionally provide the foundation for manyexecutive information system, customer service, andmarketing functions, and also support some basicengineering and operations applications.

The second key characteristic is the ability todefine a model of the utility which provides anaccurate representation of the way the systemoperates. Integrity of topology is critical to theaccuracy of the model. By storing attributes whichdefine whether a pathway is open or closed (asmight be indicated by a valve in a pipe system),

J R Meyers

804

or if the magnitude of flow has changed(as might be indicated by regulators, transformers,and other devices in an electric model),sophisticated applications are possible, especiallyin the realm of operations (switching, valving)and engineering analysis.

The third important aspect of the utility datamodel is the ability to integrate a spatialrepresentation of facilities with other utilitydatabases such as customer, accounting, or workmanagement systems. A broad range of aspatialdata is required to enable the integrated functionswhich make up the full range of traditional andnon-traditional AM/FM/GIS (e.g. customer files,

geodemographics, and billing information). Birkinet al (Chapter 51) provide an overview of GISapplications in business and service planning.

Figure 2 shows the relationships among thespatial representation of facilities data, relatedattributes, and external utility systems. The core ofany AM/FM/GIS is the spatial model of facilitiesand land base features. Attributes of these entitiesare typically stored in a RDBMS and referenced bythe many applications of an AM/FM system whichperform analysis, respond to queries, and createmaps. Many of the applications which are part ofthe system also reference data from externalcorporate systems.

GIS in the utilities

805

Fig 2. AM/FM/GIS data and integrated functions (Press = pressure; Mgt = management).

AM/FM/GIS

Facilities

Foreignutilities

Land base

Oper

atio

ns

Emer

genc

y res

pons

e

Trou

ble c

all

Switc

hing

Valvi

ng

Netw

ork A

nalys

isLo

ad m

anag

emen

tPr

ess/

volt

stud

iesLo

ss an

alysis

Emer

genc

ypl

anni

ng

Customer Service

Service call analysis

Satisfaction

surveys

Dispatching

Managem

entBusiness geographics

Decision supportProjections

Valuations

Marketing Sales

Market analysis

Demographicanalysis

Reliability

Engineering Design

Work order sketch

Cost estimating

Materials mgt

As-builts

Externalcorporatesystems

RDBMStables

Satisfactionsurveys

3.3 Techniques for AM/FM/GIS data automation

This brief treatment cannot adequately identify allof the possible aspects and implications ofAM/FM/GIS data automation, and some furtherdetails are provided by Dowman (Chapter 31) andJackson and Woodsford (1991). However, nodiscussion of the topic would be complete without asurvey of the available technology and at least asummary discussion of the relative merits of eachtechnique. This section will seek to identify thecommon data capture techniques and summarise thestrengths and weaknesses of each.

3.3.1 Base mappingMany utilities have the opportunity to use base mapdata produced by a local, regional, or nationalgovernment agency, or in rare cases, by anotherutility serving the same area. In cases where simplestreets and land forms match AM/FM/GISrequirements, commercially available street files withaddress information may be an economical choice.When digital base maps are not available, companiesare forced to rely on other sources for base mapping.Many utilities simply digitise base map planimetricfeatures in vector format from existing paperfacilities maps. Others prefer to scan the existingutility or other paper base map source, and use theresulting raster image as the backdrop for the digitalfacilities map.

Digital orthophotography, a method in whichvery accurate aerial photographs are rectified(see Dowman, Chapter 31), has grown as a sourcefor base map information, although some users mayargue that photographs are prohibitively expensiveto update. One possible answer to the highreplacement cost of digital orthophotos is the use ofvery accurate (+/- 1m) satellite images, which arebecoming technically and economically feasiblealternatives as low-altitude commercial satellites’images begin production (see Barnsley, Chapter 32;Estes and Loveland, Chapter 48). The relative meritsof methods for base mapping data capture aresummarised in Table 1.

3.3.2 FacilitiesBy far the most common technique still in use in datacapture today is board (table) digitising (Dowman,Chapter 31; Flanagan et al 1994) in which sourcedocuments are placed on a digitiser table and tracedelectronically. Preparation of source documents,including verification of data and updating (oftencalled ‘scrubbing’), enhances the reliability of the data,

but adds to the time and cost involved. At the veryleast, digitising is expensive and requires intense labourfor completion. In addition, because of the diversity ofsource data available, this method is often inadequatefor producing the consistent format of spatial datarequired to support AM/FM/GIS requirements.

Another technique for acquiring digital vectordata involves scanning the source maps and usingsemi-automatic line tracing to ‘vectorise’ theappropriate facilities. This method can result insignificant savings, but difficulties can arise with thevectorisation accuracy, especially when sourcedocuments are cluttered or unclear. Commercialdata vendors have recently developed expert systemsto permit more accurate automatic vectorisation,with the potential to reduce costs greatly.

Some companies prefer to begin their utilityapplication implementation with a raster onlybase-map for their facility database, and thenconvert selected features (usually pipes and fittings,or wires and devices) to vector format in each areaas needs arise. Called ‘incremental conversion’, thistechnique allows the deferment of large-scale datacapture effort, focusing only on those areas wherecurrent work is ongoing. In some cases, digital datafrom corporate systems can be used to populateattributes of existing facilities, or to validate dataentered by other means.

When maps are out of date or non-existent, fieldsurveys are often undertaken to verify the state andlocation of lines and equipment. Several advancedtechniques combining the use of GPS and otherlocational devices like laser range-finders haveautomated the process (Lange and Gilbert, Chapter33), but field survey is still regarded as expensive andlabour-intensive.

Table 2 provides a summary of the strengths andweaknesses of the various techniques available for dataautomation in AM/FM. Variations on these techniquesare often used in AM/FM data capture, withcombinations of two or more methods employed.

4 CRITICAL APPLICATIONS

The benefits of AM/FM/GIS are realised throughthe implementation of applications. This sectionreviews the requirements and functions of thecritical applications that determine the success ofAM/FM/GIS applications through meeting businessneeds and improving processes. Traditional applications such as mapping, ad hoc query, work

J R Meyers

806

GIS in the utilities

807

Table 1 Base-map data capture methods: advantages and disadvantages.

Capture Advantages DisadvantagesTechnique/Source

Base-map – – Low initial cost – Accuracy/content may vary fromShared data from agency/utility – Access to non-utility data owned by utility requirements

others (e.g. cadastre, foreign utilities) – Update responsibility must be clarified/negotiated– May eliminate maintenance – Facilities data from existing source maps

may have to be adjusted to fit base

Base-map – Commercial street centreline files – Very low/low initial cost1 – Accuracy and planimetric content may

– Address, demographic information be too limited for utility requirements.often already included – Update process (for original address,

– Good support for dispatching and demographic data) may be cumbersomerouting applications

– Can more easily be adjusted to fitexisting facilities source maps

Base-map – Raster backdrop from existing – Low initial cost – Non-intelligent; no support for facilities maps – Good fit to geometry of existing facilities spatial operations, address matching, or

tracing functions– Ongoing maintenance may be awkward;

re-scanning of updated maps required

Base-map – Raster backdrop from – Very accurate base – High initial costdigital orthophotos – Low labour requirements – Non-intelligent; no support for spatial

– More planimetric features available operations, address matching, or tracing functionsfor use – Moderately high replacement or update cost

(high-resolution satellite photos may reduce)

Base-map – Vector digitised from facilities – Best fit to geometry of existing facilities – High initial costdetail maps – Can be used for spatial analysis, – Ongoing maintenance for all vector

depending on detail available planimetric features– Familiar ‘look-and-feel’ for users – May be hard to share common data across agencies

1 Based on current US data pricing practices: see Rhind (Chapter 56) for a discussion of international data pricing policies.

Table 2 Facilities data capture methods: advantages and disadvantages.

Capture Advantages DisadvantagesTechnique/Source

Facilities – Board digitising – Most straightforward, direct method – High initial cost– Produces maps with familiar – Typically requires expensive ‘scrubbing’,

‘look-and-feel’ which inflates in-house staff efforts

Facilities – Scanning/automatic – Low initial cost – May require manual intervention tovectorisation – Fast turn around produce quality data

– May not be suitable for cluttered, old, or unclear source documents

– Large volume storage requirements

Facilities – Scanning vectorisation – Very low initial cost – All normal drawbacks of using incremental conversion – Vectorisation confined to focused raster-based systems

or important areas – Slow development of intelligent– Fast turnaround (for raster base) facilities models for analysis

Facilities – Field survey using GPS – Produces accurate facility locations – Can add significantly to cost of capture– Helps to validate facility locations and – May result in locational data

connectivity relationships (from GPS or laser range-finder) whichfits poorly with existing data

order design, and network analysis are reviewed(Meehan 1994), along with more non-traditionalapplications such as marketing, customer service,and executive information systems.

4.1 Mapping applications

Mapping applications perform three basicrequirements:

● map editing tools provide the ability to create andmaintain the facilities database;

● standard map products can be applied to utilityoperations;

● ad hoc mapping tools allow creation of maps inresponse to specialised requests.

Following this intensive period of activity, there isusually a much less intensive, but equally importantand much longer term, spatial and tabular datamaintenance requirement. As described previously,bulk data conversion is often used to create the initialdatabase, using specific tools geared for efficient dataloading. Once a database exists, editing tools are usedto create new map features and make updates toexisting components of the utility network.

4.1.1 Map editingAlthough many applications use the data in anAM/FM/GIS, maintenance typically comes throughfour applications: a map sheet editor, a desktopdesign editor, a field design editor, or a tabular dataentry system associated with legacy data. The mapsheet editing application shown in Figure 3 is typicalof many AM/FM editing applications used to createor maintain a utility distribution system.

The fundamental goal of any AM/FM editingsystem is to locate facilities (e.g. pipes or electricalconductors and the associated devices) so that theywill function correctly, within the context of thepublic they serve. Figure 3 shows a screen display ofgas mains (solid thick lines) through aneighbourhood and their relationship to the streetnetwork. It also identifies the customer serviceconnections (dotted lines) and building locations(shaded polygons). Symbology shows the presenceof fittings (white), indicating that there is a change inpipe material or size, and a regulator which signals achange in gas pressure. The topological integritywith which these features are placed, frequentlyusing sophisticated coordinate geometry, establishes

J R Meyers

808

Fig 3. Typical gas distribution map sheet editor.

the gas distribution system model. Associatingdatabase attributes of pipe size, material, and pressuretransforms a map of the gas system into a functionalmodel of the network, providing the ability toperform many types of analysis.

Feature placement refers to the set of tools usedto build and maintain a utility’s database. The termincludes both the cartographic representation offacilities and their incorporation into an underlyingdatabase. The process of feature placement isundertaken using tools selected from a user interface(see the toolbars across the top of the display andthe menu in Figure 3). Through these tools, businessrules are built into the application to ensure modelintegrity. The key rules or relationships in anAM/FM/GIS include:

● enforcement of topological connectivity –features which need to be linked in theAM/FM/GIS are actually connected in thedatabase, rather than simply appearing to beconnected graphically;

● maintenance of referential integrity between boththe spatial (graphical) and attribute (tabular) datacomponents – for example, if a transformer bankis deleted, all transformer units associated withthe bank are also deleted;

● validation of feature combinations – for example,only a limited number of pipe diameters are validfor a section of plastic pipe.

4.1.2 Map productsProducing the standard, time-tested hard copyoutput from an AM/FM/GIS is typically the job ofa map products application. Utilities rely on maps formany operational tasks; it is not uncommon forutilities to produce and update seven to ten or moredifferent map products. These may vary fromdetailed distribution and transmission facilitiesmaps, to property and rights-of-way maps, toschematised network views. Part of the benefit ofAM/FM/GIS is the ability to drive these differentviews from a single integrated database.

The feeder map in Figure 4 is an example of astandard map product which is used by many electricutilities. A feeder is the path of electrical current fromits source (typically a substation circuit breaker), to its many destinations at residences, schools, hospitals,and other commercial locations. Cartographicdecisions, which specified that the feeder shouldappear as a bold linear feature against a backgroundof street centrelines and that only significantbuilding locations are identified, were pre-defined as

GIS in the utilities

809

Fig 4. Electric distribution feeder map product.

part of this map product. Note that buildings, poles,or devices which affect the flow of electrical currentare symbolised on the map display.

Many electric companies produce schematic viewsof electric networks. Other common map products,such as pole inventory or physical maps, rights-of-wayor corridor maps are very different depictions of thedata in common integrated databases. In the pipelineutilities (gas, water, waste water) there are analogousmap products for different operating purposes.

Map scale is one of the key challenges to be facedin a map products application, and is reviewed byWeibel and Dutton (Chapter 10). A wide variationin scales (1:1200 to 1:24 000 or greater) makes theplacement of entities and annotation very difficult.For example, a map symbol for a valve may beperfectly appropriate at a detailed map scale such as1:1200. However, even with symbols scalingcontinuously, its location may obscure other featureswhen it is displayed at scale 1:24 000. Modern GIStypically include tools to manage ‘whitespace’, andto mitigate over-posting of text and symbols.

4.1.3 Ad hoc mappingAs the above discussion suggests, map products aredefined by a specific format which includes scale,map surround or marginalia, contents (facilities andmap features), and symbolisation. Ad hoc maps arealso widely used in utilities. Their content andappearance are defined entirely by users, includingwhich facilities to display at which scale. Often suchmaps are the result of a particular query or analysis.Figure 5 shows a user-defined section of a wastewater application, together with the results of aquery for pipe material in pie-chart form.

Many desktop systems, with functionalityreduced from full-blown professional GIS, have beenperfected for use as ad hoc mapping tools, makingthe technology available to a wider audience on lessexpensive hardware.

4.2 Facilities management applications

Facilities management applications moveAM/FM/GIS beyond the mapping to the analysis andmanagement of the operation of a utility network.

J R Meyers

810

Fig 5. Ad hoc sewer system map. Source: BaySys Technologies, Inc.

The goal of the FM application is to place spatial andattribute data in the hands of those that makedecisions about how a utility can maximise itsresources and minimise its costs. FM applicationsmay be grouped into the following general categories:

● query and display;● design/work order processing;● equipment maintenance;● network analysis;● customer service/service call analysis.

Two examples, query and display, and design/workorder processing, should serve to illustrate the rangeof FM systems.

4.2.1 Query and display applicationsThe simplest form of FM application is a query tool,used by a broad range of utility personnel fromexecutives to customer service representatives. Byproviding a link between spatial and attribute

facilities and even customer data, query applicationshave provided tremendous benefits for companies.The combination of spatial and logical queries isamong the most powerful functionality inAM/FM/GIS. Simple questions like ‘How manystreetlights do we operate in the town of Jonesville?’may result in great labour savings over traditionalmanual methods.

Figure 6 is a view of a desktop query application.It shows the results of a simple spatial query whichlists the attributes of a single electrical customer,including the criticality of the customer foroperating purposes, the unique identifiers for thecustomer, and the type of metering used.

Geospatial relationships are the basis for anumber of utility applications which may include adhoc queries, such as the one represented above.More periodic or repeated queries may be used forapplications such as maintenance or tracking ofwork orders.

GIS in the utilities

811

Fig 6. Desktop spatial and logical query tool.

4.2.2 Design/work order processing applicationsUtility companies generally manage additions ormodifications of outdoor plant with a work order, adocument or set of documents which describes thework to be accomplished. A work order packageoften includes a sketch made from a map, or drawnfrom scratch, illustrating the work area, togetherwith construction notes and instructions. Followingcompletion of the work sketch, it may be issued to acrew for construction. After amending the sketch toinclude any field changes (the so-called ‘as-built’representation) the sketch is used to update thepermanent record of a database (Figure 7).

This design process is often labour intensive, and atmany companies results in delays of record completionand duplication of effort. For these reasons, manyutilities seek to automate this process through anAM/FM/GIS application. Figure 7 shows the sketchfor a new electrical line extension. The screenrepresents the output product from the design,including construction notes and a vicinity map.

Of course, the graphic output is only a part of thebenefit to be gained from automating the work orderprocess. Several companies have integrated the GIS

design environment with external digital workmanagement systems, to provide integrated costestimating for rapid customer response and acapability to update a corporate plant accountingsystem. In the most advanced systems, the ‘as-built’representations of outside plant are fed back intothe AM/FM/GIS database and thus feed into themaintenance process.

4.3 Network modelling applications

Network analysis tools can be categorised into twomajor groups on the basis of the time frame overwhich they operate. Engineering analyses are generallyperformed on an irregular basis usually in response tosome known condition, while operations applicationsare driven by near real-time operating conditions.

An example of an engineering analysisapplication is shown in Figure 8. Using the base GISdatabase as its source, this electrical networkanalysis application employs a detailed energy usagemodel to assess system performance. Like mostengineering analysis packages, it includes functionsfor evaluating load flows, voltages, and short-circuit

J R Meyers

812

Fig 7. Work order design sketch.

currents, as well as some switching and othermiscellaneous functions. Unlike other such tools,however, it operates directly on a GIS database,rather than a separate dataset prepared andmaintained specifically to support network analysis.Figure 8 shows, in the top half of the screen, theresults of an analysis of a single circuit or feeder,traced from a point of interest back to the sourcesubstation. Voltages for each node along the feedpath are computed and displayed in a line graph inthe bottom portion of the screen.

In the water and gas industries, analogousapplications are available for providing hydraulicanalysis, and determining pressure and flow basedon pipe network characteristics and demands.Techniques like this permit engineers andtechnicians to undertake analyses with current dataand make models in seconds or minutes thatformerly took days or weeks. In addition to

increased productivity, GIS-based applications likethis can be based on more detailed models, and canproduce results in ad hoc mapping or report form.

The second network modelling applicationdiscussed here is oriented towards operations. It usesan AM/FM/GIS database for real-time or near-real-time trouble call or outage management. In suchapplications, an operator uses a model of a network,along with data from customer calls or monitoringequipment, to determine the location of a problem,possible consequences and even the best method forrestoring service to affected consumers.

A trouble call management application is shown inFigure 9. In this desktop viewer, trouble calls whichwere originally entered into the CIS on a mainframeare relayed to distributed district office databases, basedon an identification tag. Once loaded, the address ofthe caller is used to associate the call through addressgeocoding to the nearest point in the network.

GIS in the utilities

813

Fig 8. Electrical distribution network analysis.

In normal outage events, a number of customerscall to register complaints. After locating theavailable calls, traditional network processing is usedto determine the probable outage device, or thepoint at which the network has been isolated. Oncethe outage device has been found, a reverse networktrace can be used to identify all points in thenetwork downstream of the device. When thisinformation is combined with address information,engineers can be dispatched to investigate the deviceand affected customers can be notified.

Integration with the customer system is oneobvious requirement for efficient trouble callmanagement. Modern systems are also beginning tobe integrated with telemetry or supervisory controland data acquisition (SCADA) systems for outageand problem information. Traditionally constrainedto the transmission or bulk delivery level, SCADAtechnology has moved into the distribution systems’domain in the past few years, providing data fromkey points in a network. This real-time data can be

used to anticipate problems in gas, electric, andwater delivery systems, and to provide networkoperators with data about the nature and location ofa problem in all networks.

The trouble call system is a valuable tool forutilities, especially energy companies (electric, gas),for responding to customer complaints during anevent. However, these applications also provide greatbenefit for compiling statistics about networkperformance, and reporting on the nature andduration of outages. These data have proved to beextremely valuable in identifying and correctingrepetitive system failures based on faulty equipmentor poor maintenance practices.

4.4 Executive information systems

The technical nature of GIS applications hastraditionally restricted their use to technicians,drafters, and engineers responsible for mapping andanalysing the operation of a given utility network.

J R Meyers

814

Fig 9. Desktop trouble call management interface.

With the advent of desktop GIS (see Elshaw Thralland Thrall, Chapter 23), the capability to analyse anduse spatial data can be put into the hands of thosewho make decisions about how a utility canmaximise its resources and minimise its costs, so thatthey are able to survive in a competitive environment.These ‘non-technical’ GIS applications have simpleIT industry standard interfaces and run on lower costplatforms (typically PCs). They can be consideredspatial executive information systems (EIS) for use bythe traditional executives of a utility.

Figure 10 is screen output from a desktopGIS/EIS showing a combined spatial and logicalquery. The user has asked the application to make amap of all ‘critical’ customers (such as policestations, fire stations, and life-support dependenthouseholds). Furthermore, the query has beengraphed on the basis of the classification of thecustomers who are noted as critical, based onwhether they fall into the residential, smallcommercial, or industrial rate tariffs. Note the

Windows-style standard user interface, designed toprovide a non-GIS user with familiar functions.

This simple query is only one of many possibleexamples which illustrate the power of GIS to addintelligence about location to managementapplications. Marketing questions such as ‘Show meall the customer locations within 100 feet of our gasmains, now classified as Residential Electric Heatingcustomers’, or ‘Make me a map of every customerwho has called to report a problem in theHamptonfordshire District within the past 13months’ are now at the direct disposal of managerialstaff. Operating queries such as ‘Show me thelocation of all current work orders’ are just as easilyperformed. Birkin et al (Chapter 51) provide anassessment of how such functionality fits with themanagement and strategic functions of business andservice organisations.

While not all AM/FM systems currently performthese functions, they are among the critical applicationsthat will determine which companies will set the pace in

GIS in the utilities

815

Fig 10. Desktop GIS executive information system.

an industry driven by the market economics ofderegulation. These functions also emphasise the needto integrate GIS with other corporate functions thathave historically operated independently.

5 COSTS AND BENEFITS

This section explores the nature of benefits and costsof AM/FM/GIS, beginning with an assessment of thefactors in the costs of implementation. A high-levelview of the principal benefits is also presented, withexamples which identify key areas for reducing costs:as such, this section complements that of Obermeyer(Chapter 42) who conducts a more general review ofthe benefits and costs of GIS. Finally, the factorswhich will affect competitive advantage for utilities inthe coming years are discussed, in the context of theimpact of AM/FM/GIS.

5.1 Cost assessment and examples

There are a number of factors which affect the costof building an AM/FM/GIS and integrating itwithin the information systems of a utility. The costof implementation typically varies depending on theapplications required, the size and scope of the datato be captured, and the number of users to besupported. The legacy of existing systems andavailable procedures for data automation also affectthe cost. Any analysis of system development costsis also complicated by the fact that there are manypolitical and historical factors to consider.Unfortunately, because of the current competitiveenvironment, many utilities are reluctant to sharespecific information about the costs of theirAM/FM/GIS development. However, it is hopedthat the following summary of factors affecting eachof the six critical cost components for utilities(Rector 1993), along with a few examples, willprovide a useful background.

5.1.1 Data automationData capture costs are affected by the portability ofexisting data sources, the geographical size of theutility, and the scope or detail of the data to beautomated. Commercially available sources of digitaldata for the coincident territory are also a factor.

5.1.2 Hardware and software acquisitionSystem acquisition costs vary widely depending onthe hardware selected, the network requirements, the

type of base GIS and commercial database softwareselected, and number of user seats implemented.

5.1.3 Project managementProject management costs consist of the direct costsof utility employees and consultants hired to helpmanage AM/FM/GIS system development. Theremay also be hidden project management costsassociated with out-sourcing data conversion andthird party GIS software development. It should beobserved that project management costs occurthroughout the project life-cycle, and may also arisefrom migrating existing AM/FM/GIS applicationsto new platforms (Rogers 1996).

5.1.4 Applications developmentThe scope and cost of applications developmentvaries with the nature and complexity of theapplications to be implemented. In particular, theuse of commercial off-the-shelf software versusentirely customised solutions plays a major role (seeBirkin et al, Chapter 51; Elshaw Thrall and Thrall,Chapter 23; Maguire, Chapter 25).

5.1.5 Support staffDevelopment of internal support for utilityapplications may include the cost of staffing for asystem administrator and one or more applicationsupport specialists.

5.1.6 TrainingTraining is an often overlooked component of thecost of building an AM/FM/GIS. Costs varymostly with the level of involvement in dataconversion and quality assurance/quality control,and with the scope of applications and number ofusers to be trained.

5.2 Tangible benefit assessment

In assessing benefits, many of the same variableswhich affect costs come into play, along with otherfactors associated with the current state of facilitiesand business processes. The benefits to be realisedfrom AM/FM/GIS must be carefully andconservatively analysed for each prospective utility.A number of sources have identified areas of benefit(Meehan 1995; Meyers et al 1996). The five mostcommon high-level categories of tangible benefitnormally cited are addressed below, along withbrief examples.

J R Meyers

816

5.2.1 Increased productivity in mappingVirtually every utility investing in data automationcan increase the efficiency and accuracy of mapproduction and deliver maps which serve a broaderrange of business needs. Utilities with productionsystems have reported savings from one to threeperson-years of effort per year, based on eliminatingthe need for one or more drafters, or reducingoutside contract labour for mapping.

5.2.2 Increased productivity through information accessNetwork models for engineering and operationsanalysis require data that is time consuming andcostly to produce and maintain. Creation ofintelligent network models is typically highlylabour-intensive and normally requires a highlyskilled technician or engineer. Driving the analysismodel from the AM/FM/GIS database can actuallyresult in reduced labour costs and improved analyticalmodels. Utilities surveyed by Meyers et al (1996)report tangible savings that amount to reducedtechnician or engineering labour of one or moreperson-years, to the benefits of better, moreconsistent plans based on accurate analysis.

These benefits also extend to maintenance,especially where gas and water utilities invest largesums of money. An accurate data source with built-in query tools makes it possible to perform anumber of maintenance analysis functions whicheither could not be accomplished using paper maps,or would be much less efficient (Cotting andDaniels 1994; Marshall 1996). The savings inmaintenance may amount to hundreds ofthousands of dollars per year, both in reducedengineering labour costs and more focused use ofcapital for improvements.

5.2.3 Improved engineering and operations efficiencyThe opportunities for improving the efficiency ofdesigning and operating a utility network range frombetter maintenance and corrosion protection toimproved subdivision service design. Cost savings forthis category are normally measured in terms of moreefficient scheduling, or reduced costs in plant throughbetter design and maintenance. Gas and electricutilities with production AM/FM/GIS have estimatedannual savings in capital outlay of US$200000 toUS$1500000 through better design, reduced plant andinventory, and deferment of unneeded improvements.

5.2.4 Decreased use of mainframe computersFor some utilities, the cost of maintaining andexpanding mainframe computing to support anexisting AM/FM/GIS system can be avoided bymoving functions to a distributed microcomputingenvironment. The costs avoided are usuallycalculated in terms of reduced computing cycleswhich are then considered to be available for otherexpanding requirements such as CIS or workmanagement, or in one notable case, permitting theutility to eliminate one of two mainframecomputers. In one example the reduction resulted ina net monthly operating reduction of US$120 000for the five-year life of the new system.

5.2.5 Eliminating costs in map and recordsreplacement/consolidation

At utilities where diverse and non-standard mapsexist, because of historical practices which led todiffering mapping standards, or through mergersand acquisition of other companies, the cost ofredrawing and maintaining duplicate sets of recordscan be significant. This cost is also present whenhardcopy map (e.g. paper, mylar, vellum, or linen)records need to be replaced because of degradation.For example, a gas company calculated that it wouldsave US$485 000 over ten years by avoidingconsolidation and replacement of existing strip maps.

5.3 Competition, change, and GIS

Utility restructuring, privatisation, and open energymarketing will soon have affected nearly every regionof the world. Electric, gas, water, and waste waterutilities from Australia to South America are beingprivatised (that is, government-run utilities are beingsold to private companies for operation).Deregulation of utilities is a present or fast-approaching reality (Rector 1993). Throughout theglobe utilities are under increasing pressure to reducecosts and improve the value of service.

Against this backdrop, the following fourkey issues will affect utility performance in thecoming decade.

5.3.1 Coping with changing governmental andregulatory policies

Without question, utility restructuring andreregulation is the single most significant challengefacing the modern energy utility. Policy changes will

GIS in the utilities

817

affect aspects of rate making, expansion, andobligation to serve. AM/FM/GIS will play a key rolein providing the integrated spatial view of plant andcustomers for changing policy environments.

5.3.2 Attracting and retaining customersUtility restructuring means competition, especiallyin urban areas where facilities in place can beeconomically utilised. The ability to attract new orretain existing customers in a competitive marketwill depend on efficient operations that deliver high-quality service at reduced costs. The accurate andefficient methodologies for engineering analysis andcreative design strategy offered by AM/FM/GIStechnology will be key to fulfilling this objective.Further, encroaching utilities will be required to paya transportation or ‘wheeling’ charge for using in-place facilities to deliver a commodity; accuratecosts for plant in service, aided by AM/FM/GIS, willbe critical in building a defensible cost base forwheeling charges.

5.3.3 The link between environmental responsibility andfinancial capability

Like many other businesses, the utilities industry hasbecome increasingly aware of the capital markets’interest in environmental responsibility, and thepossible risks associated with investment in entitieswith unsound or risky environmental practices.Beyond the realm of social responsibility to protectthe environment, utilities are now faced with theeconomic reality of projecting a ‘green’ image ofgood stewardship in order to attract and retaininvestment capital. AM/FM/GIS can and will play akey role in helping utilities mitigate or eliminateenvironmental risks.

5.3.4 The ability to promote sustainable developmentMany utilities have been actively involved ineconomic development at local, regional, andnational levels. Investment in industries that useenergy and water commodities is a rapidly growingfacet of utility marketing and management.Partnerships with sound, resource-wise businesseswill be one key to sustained growth in markets.AM/FM/GIS can provide the integration ofcommodity usage, network capacity, and reliabilitynecessary to attract these sound partners to theservice area of a utility.

References

Antenucci J C, Brown K, Croswell P L, Kevaney M J 1991Geographical information systems: a guide to the technology.New York, Van Nostrand Reinhold

Coppock J T, Rhind D W 1991 The history of GIS. InMaguire D J, Goodchild M F, Rhind D W (eds) Geographicalinformation systems: principles and applications. Harlow,Longman/New York, John Wiley & Sons Inc. Vol. 1: 21–43

Cotting C C, Daniels M J 1994 Boston Gas Company – AMMSproject. In Kunze A E (ed.) The best of AM/FM/GIS: a specialcollection of past conference proceedings for the gas industry.Aurora, AM/FM International

Flanagan N, Jennings C, Flanagan C 1994 Automatic GISdata capture and conversion. In Worboys M F (ed.)Innovations in GIS. London, Taylor and Francis: 25–38

Gittings B M, Sloan T M, Healey R G, Dowers S, Waugh T C1993 Meeting expectations: a review of GIS performance issues.In Mather P (ed.) Geographical information handling – researchand applications. Chichester, John Wiley & Sons: 33–45

Jackson M J, Woodsford P A 1991 GIS data capture hardwareand software. In Maguire D J, Goodchild M F, Rhind D W(eds) Geographical information systems: principles andapplications. Harlow, Longman/New York, John Wiley &Sons Inc. Vol. 1: 239–49

Maguire D J 1991 An overview and definition of GIS. InMaguire D J, Goodchild M F, Rhind D W (eds) Geographicalinformation systems: principles and applications. Harlow,Longman/New York, John Wiley & Sons Inc. Vol. 1: 9–20

Marshall W F 1996 Sound plan key to making cost effectiverehabilitation decisions. Pipeline and Gas Industry 79(8): 21–5

Martin J, Odell J J 1995 Object-oriented methods. Afoundation. New Jersey, Prentice-Hall: 11–24

Meehan W J 1994 Automated distribution dispatching usingAM/FM/GIS. In Kunze A E (ed.) The best of AM/FM/GIS:a special collection of past conference proceedings for theelectric industry. Aurora, AM/FM International

Meehan W J 1995 Process reengineering using GIS: anAM/FM example. ESRI White Paper Series. Redlands, ESRI

Meyers J R, Daniels M J, Killpack C C 1996 Costs andbenefits: implementing an ARC/INFO-based AM/FM/GIS.ESRI White Paper Series. Redlands, ESRI

Newell R G 1994 The why and the how of the longtransaction. In Green D R, Rix D, Cadoux-Hudson J (eds)Geographical information: the source book for GIS. London,Association for Geographical Information: 237–41

Rector J M 1993 Utilities. In Castle G H (ed.) Profiting from ageographical information system. Fort Collins, GIS WorldInc.: 193–208

Rogers C 1996 A project manager’s guide to data conversion(or six things your mother never told you). UtilityAutomation May/June: 21–3

J R Meyers

818