Toward Infrastructure Improvement: An Agenda for Research (1994)

Management of infrastructure as a multifunctional system, expected to provide and support a wide range of services over an extended period of time, represents a radical departure from past practices that have focused on individual modes (e.g., highways or water supply) and on sequential stages of facility development. Research is needed to enable day-to-day decision making to reflect the complexity inherent in this system concept.

The complexity stems from the hierarchy of activities (e.g., from establishing basic goals setting the operating procedures for each functional element), uncertainties in demands placed on the system by users and neighbors, evolution in technologies, and relationships among institutions. The complexity stems also from gaps in knowledge. There is no generally accepted definition of good performance that spans the range of infrastructure.¹ There are few methods for exploring the tradeoffs among ways to meet service demands, through resource distribution among infrastructure modes or by influencing demands rather than service supply. There is little reliable guidance for how best to introduce new technology without unduly adverse social or environmental impacts.

Research to make this complexity more manageable is necessarily multi-disciplinary. Such topics as human settlement patterns; "opportunity costs"² of pre-emptive actions such as reserving rights-of-way for future systems; impacts of infrastructure on adjacent communities and lands; interactions among transport, telecommunications, and other infrastructures, as well as distributions of costs and benefits among groups of people in a region, among regions, and among present and future generations, call for research contributions from the social, political, economic, and physical sciences.

The committee proposes that the NSF should foster research to enhance fundamental understanding of the long-term role of infrastructure as a contributor to the quality of life in cities, suburbs, and other developed areas, one that can be more effectively managed in design, construction, and operations. Three primary topic areas offering high payoff potential are discussed in this chapter:

1. enhancing knowledge of the sources of infrastructure demand and how demand interacts with the system to influence service life to enable development of more effective infrastructure management tools and procedures;

2. data collection and information-analysis tools to support total-system inventory and coordinated management of infrastructure within a metropolitan area; and

3. analyses of standards, regulations, and other external factors (e.g., political and social values, and technological advances) that influence infrastructure performance and service life.

A basis for many infrastructure facility design decisions is that the facility will be in service for a period of 25 to 50 years. In practice, infrastructures are often kept in service much longer, and these aging facilities are frequently found to be poorly suited to current demands. It is frequently found that these long-lived facilities are poorly suited to current demands. The system generally lacks flexibility to adjust. This is particularly true for underground facilities, where initial costs are very high and repair or replacement are very difficult.

The demand for infrastructure depends on a wide range of economic and social activities, environmental constraints, technological options, and societal priorities. For example, the volume and characteristics of municipal solid wastes that ultimately are deposited in landfills or dumped in the oceans depend on patterns of human consumption and how wastes are processed at intermediate stages (e.g., volume reduction, segregation by type, and recycling). Incinerators produce air pollutants and ash that may be considered hazardous, which in turn create new disposal demands. Landfills present possible soil and groundwater pollution hazards.

Similar chains of demand, which may be cited for all infrastructure modes, change over time and are not generally well understood. However, over the course of the infrastructure's service life, the value of activities that depend on the infrastructure's services is very large. Seemingly small gains or losses in productivity caused by infrastructure performance can accumulate into very large impacts on an area's economy and quality of life.

Research on this set of issues concerns explorations of the factors influencing infrastructure demand and how demand interacts with the system to affect service life. The urgency of this research is high because infrastructure capacity limitations currently are being encountered in many areas. Government budgetary deficits that restrict new investment make capacity expansion difficult, increasing the need for management of demand, enhancement of capacity through operations, or both.

The benefits to be gained through research in these areas include more effective infrastructure management tools and procedures that can be used to enhance infrastructure performance at lower costs. Research on the life expectancy of infrastructures—through analysis and characterization of components and subsystems—an improve infrastructure managers' predictions and thereby foster better decisions about facilities operation, maintenance, repair, or replacement. Research on demand can improve managers' abilities to achieve a better match between existing systems and the demands placed on those systems.

Infrastructure's services fall generally into the category economists' term ''public goods,'' for which prices that can be charged seldom cover the costs of production. Understanding the allocation of public costs and benefits, excessive demand spurred by low prices, network economies,³ and distortions of markets for other goods and services that use infrastructure are among the areas that warrant continuing research. New measurement and communications technologies that facilitate continuous in-house metering of water and power consumption, automated vehicle tracking and toll collection, and other innovative charging mechanisms can enhance the manager's ability to compute and reflect full costs in infrastructure service pricing.

Typical questions for research might include the following examples:

• What are the full costs of infrastructure services, and how are these costs distributed?

• Can depletion allowances, extraction subsidies, and other new price-modification mechanisms be adapted to modify demand for municipal solid waste disposal services?

• How can infrastructure prices be adjusted to respond to variations in demand (e.g., by time of day, day of week, or season) to achieve more efficient utilization of facilities?

Much of the demand for infrastructure is derived from other economic and social activities (e.g., municipal disposal of wastes from consumption,

and transportation as a means to get to work or deliver goods for manufacturing). Influencing the patterns of demand for goods and services, from which demands for infrastructure are derived, is potentially a way of influencing infrastructure demand levels and operating efficiency. The "energy crisis" of the 1970s motivated increased interest in car-pooling and "flexi-time" among workers as ways of managing highway usage. Droughts in western U.S. states have spurred increased water recycling and limitations on water use for domestic irrigation (e.g., watering the lawn). Research on this topic could encompass both regulatory and pricing strategies, as well as explorations of the sociology and psychology of personal habits and cultural patterns that are nonprice determinants of demand.

Typical questions for research might include the following examples:

• Are regulatory tools effective in the long term for suppressing or redistributing peak-period infrastructure demand?

• How have land-use policies influenced infrastructure investment and performance levels?

• How does infrastructure investment influence land-development intensity and subsequent demand for infrastructure?

• How do the values underlying public expressions of environmental concerns influence use and willingness to pay for infrastructure?

While infrastructure in use provides services that support other activities, the preponderance of infrastructure research concerns development and management of facilities. Further, these facilities require substantial inputs of materials that in turn are major consumers of energy and other resources. Research is warranted to characterize more effectively the processes for production of infrastructure services, viewed on a basis of the total life-cycle of the inputs. Inputs would include clean air and water and other environmental quality elements, as well as labor and construction materials. Research would include modeling of technologies and tracing the chains of input and output from initial materials production through demolition or redeploymerit of obsolete facilities. New data collection methods (e.g., geographic information systems [GIS] and aerial and satellite imaging), applied to monitoring of operating systems, could enable the methods of statistical inference to be applied more broadly and could yield valuable results.

Typical questions for research might include the following examples:

• What are the relative life-cycle supply resource efficiencies of alternative infrastructure technologies (e.g., rail transit versus guideway bus versus conventional highway, water supply by greywater⁴ collection, and recycling versus conventional extraction and treatment)?

• How do infrastructure marginal costs respond at the system level to changes in land-use management practices and subarea abandonments?

• Can renewable biological materials (e.g., wood products) be more widely and advantageously applied in infrastructure?

• Are service-life assumptions being used in planning and design (e.g., materials durability, economic discount rates, and demand levels) distorting decisions in consistent ways?

R&D on new materials—products of molecular-level engineering, composites, modified Portland cements, and others (see Chapter 8)—with high strength-to-weight and strength-to-cost ratios offer opportunities for radically improved infrastructure designs. Increasing incentives for recycling and waste minimization may require new reprocessing technologies and materials selection to facilitate reprocessing. The potential consequences of such changes, for infrastructure service lifetimes, reliability, capacity to meet unexpected high demand, and maintenance practices, could be far-reaching and warrant research.

Typical questions for research might include the following examples:

• What are likely ranges of service lives for recycled or reprocessed materials, components, and facilities, considering both compatibility of these materials with existing facilities and application in new facilities?

• How might the maintenance needs of future infrastructure differ from today's, in terms of skilled labor, special equipment and procedures, new materials, and other resources?

No comprehensive database exists on the composition and condition of U.S. infrastructure systems; nor do many metropolitan areas have such data. Some modes, such as highways, have been relatively well inventoried and records are regularly updated. Other systems (e.g., urban water supply) in many areas contain elements that were constructed decades ago, and historical desigo and construction records (sometimes including even their locations) have been effectively lost.

The lack of such data presents obvious challenges to infrastructure managers, as evidenced by accidental damages and unexpected high costs when infrastructures are being repaired. However, the same lack of data also poses substantial barriers to development of effective and reliable tools for predicting infrastructure service life.

Data alone do not enable more effective analysis or decision making. Methods are needed to gain easy access to data and convert them into useful information. These methods must be able to deal with data and information at various levels of detail and aggregation (e.g., facilities and networks, local and regional), and referring to different ages and geographic areas. Data access, presentation, and security must be considered to make infrastructure information usable by a wide range of people in a wide range of private and public ownership and management settings. Highlighting and dealing with tradeoffs among conflicting objectives, a central part of infrastructure decision making, must be facilitated by good information.

The scope of research in this area can be very broad: measurement devices, computer hardware and software; systems modeling, optimization methods, statistical analysis; database development and management, information display; historical studies, decision rules; and institutional structures. These are only a few of the sources of research questions. Research in this area, aimed at the needs of decision making, is closely related to research in the general area of information management (Chapter 6) and condition monitoring and assessment technologies (Chapter 7) adapted to infrastructure. Because the results of such work can contribute to more effective decision making, the payoffs of research in this area are closely allied to those concerning demand and service-life management. In contrast to that area, however, research directed at program elements, such as those discussed in this section, will typically involve substantial amounts of field observation of facilities in service.

Complete inventories of a region's infrastructure, even when they exist, are often in forms poorly suited to technical analysis and comprehensive management. Integrated databases are needed to support development of predictive models, estimations of reliability and risk levels, and in-service performance monitoring. In many cases, compilation of existing data will require assembly of diverse historic documents from several government agencies and jurisdictions particularly at local government levels. Development of such inventories may not warrant NSF support as a singular objective, but would form important elements of a broader infrastructure research program in a region. Advancement of data collection and management capabilities such as three-dimensional computer graphics and GIS, satellite imaging, and remote sensing technologies could be pursued.

Typical questions for research might include the following examples:

• What protocols should be adopted for data collection and storage for integrated infrastructure databases?

• How can three-dimensional and seasonal geographically based data be most effectively displayed to support integrated infrastructure analysis and decision making?

• What historical data (e.g., socioeconomic and hydrologic) should be maintained in disaggregated form to support future research in other fields (e.g., urban development, history, and natural hazard assessment)?

Descriptive statistics on age, composition, and other characteristics of infrastructure systems are important indicators of likely future problems and needs. Research to determine performance benchmarks and trends can help to establish investment and repair priorities and to estimate program costs. To some extent, such work has been partially done for highway bridges and to assess dam safety. For example, plastics have been used extensively in water and gas distribution systems for several decades, but there has been no comprehensive assessment of their in-service performance, nor are there reliable predictive models for their characterization and long-term behavior.

Similarly, acidic rainwater may be causing corrosion and reduced expected service lives of infrastructures, but statistics on the scale of the problem are not available. Research on this topic could include field sampling of infrastructure performance and statistical estimation of service parameters.

Typical questions for research might include the following examples:

• What portions of the infrastructure system are being exposed to service conditions that are changing substantially and that differ from design assumptions?

• Can performance of the several infrastructure modes be characterized on common scales to determine whether costs and effectiveness are in balance throughout the system?

Comprehensive data and benchmarking of performance will facilitate effective monitoring of infrastructure service conditions and, in combination with new detection and communications technologies, enhance ability to respond to deviations that signal threats to public health or safety or to service continuity. Research on this topic might explore early epidemiological warning systems for environmental hazards as well as alarm systems for infrastructure operators. Improving accident or hazard management might involve closer linking of police, fire, health, and infrastructure operations and maintenance personnel through an active sensor network,

perhaps supported by on-line knowledge-based expert system programs to assist in determining best response measures.

Typical questions for research might include the following examples:

• Are there generic rules that should be followed for optimal evacuation or congestion clearance and subsequent flow rerouting within infrastructure networks, when accident or failure has taken links out of service?

• Can improvements be made in matching infrastructure demands to network capacities during periods of peak demand, such that overall performance is improved?

• How should users' and operators' perspectives on hazard and risk influence appropriate response when new potential hazards are identified?

The allowable time interval for management to respond to changes in system conditions is shrinking. Research is needed to explore fundamental changes that may be necessary in infrastructure system organization and management to avoid increasing instability. Such research may advance technologies for SCADA.⁶ Technical innovations (e.g., imbedded sensors linked to a pro-active control center), innovative cost-sharing methods and other management tools, and new institutional arrangements for infrastructure management would be appropriate to this topic.

Typical questions for research might include the following examples:

• Do reductions of redundancy and dispersion (e.g., from increased use of fiber optic cables, rail transit, or other high line-capacity technologies) require changes in system management to maintain infrastructure performance?

• Are current response times and strategies for dealing with infrastructure disturbances appropriate and optimal, in terms of costs and potential avoidable losses?

• How do infrastructure management practices, including recurring maintenance deferrals, influence system reliability and expected service lives of facilities?

With increasing coordination of infrastructure planning and rehabilitation, junctions between subsystems become increasingly important. Significant advantages might be achieved through relocation of compatible elements in common-use corridors, within public or private rights-of-way (e.g., power line corridors or streets). Rapid transit on a highway median strip is a frequently cited example that may be extended to a wider range

of new and current infrastructure services. Research is needed to assess intermodal linkages and benefits and costs of space allocated to infrastructure. Typical questions for research might include the following examples:

• What would be the life-cycle costs and impacts of consolidating separate systems in common-use underground or elevated corridors?

• What are the merits or disadvantages of alternative strategies for use or preservation of easements associated with obsolete infrastructure elements (e.g., abandoned rail lines, underused highway rights-of-way)?

• What role does infrastructure play in the obsolescence of developed areas (e.g., urban precincts, small cities and towns, and economic regions)?

• How does development intensity influence land requirements for infrastructure?

• Can recreation parks and other environmental infrastructure (e.g., linear parks, wildlife corridors, and green strips) be integrated effectively with other infrastructure?

• Are current infrastructure systems excessively vulnerable to terrorist acts?

• What risk-management strategies will provide effective protection against multiple hazards that may cause interactive disruptions of infrastructure (e.g., water line failures that cut off power supplies, transportation accidents that disrupt telecommunications)?

• How should current systems be modified to facilitate adoption of distributed or decentralized infrastructure technologies (e.g., changes in public buildings to accommodate decentralized power generation)?

Research is needed into costs and benefits of various management structures and forms of infrastructure ownership. The often ideological discussion of private versus public roles and the relationships among private and public elements of the system hinders careful examination of alternative management methods. Research on this topic could include analyses of historic examples and legal reviews, and comparative studies of different subsystems (e.g., solid waste collection, transportation, and fire protection).

Typical questions for research might include the following examples:

• Can meaningful international comparisons of infrastructure management structures be made?

• How do public or private ownership and management influence rate of adoption of new technology (e.g., cellular communication and highway and street traffic information systems)?

• Do new infrastructure technologies necessarily entail bias toward new land development rather than toward developed areas where retrofit would be required?

Many factors such as current engineering standards and practices, governmental regulations aimed at environmental protection and public safety, general economic conditions, and political forces influence the way infrastructure is developed and used. These factors are external influences on decision making, shaping the options available to designers and operators, public perceptions of infrastructure and its impacts, and the likelihood that assumptions made at each stage of a facility's life cycle will subsequently remain valid.

While such factors thus play a key role in determining infrastructure performance and service life, understanding how they operate is limited. Research is needed to extend this understanding, drawing on basic principles in economics and the social sciences (e.g., multi-attribute utility theory, hedonic pricing, systems modeling, and public opinion research methods), as well as technological forecasting and engineering systems analysis. Payoffs from research in this area include help for infrastructure developers and managers seeking to mitigate adverse environmental, safety and health, social, and economic impacts, and to avoid unexpected public resistance to infrastructure development and operations.

Infrastructure design and management objectives change over time as a result of new technical knowledge, new regulations, or other factors, often as a result of shifts in public values and perceptions. For example, the Americans with Disabilities Act requires retrofitting barrier-free access in all public buildings. Emerging concerns about cancer risks of exposure to electromagnetic radiation may lead to changes in electric-transmission-line design standards. Research to estimate the full benefits and costs of such changes could encompass technical assessments of past and potential future changes and the social and political forces involved in the occurrence of change.

Typical questions for research might include the following examples:

• How susceptible is current infrastructure to obsolescence caused by changes in design objectives?

• Do current regulatory processes use available information effectively to assess life-cycle costs of new regulations and design standards?

• Can rational models be developed for assessing regulatory needs based on variables such as risk, affordability, and quality of environment?

• How have federal, state, and local infrastructure subsidies and regulations influenced urban development patterns?

Decisions about infrastructure siting and applications of technology (e.g., nuclear power) are typically made in a public forum and depend on availability of adequate and credible information. Research is needed to support improved procedures for estimating first-, second-, and lower-order effects of the installation and operation of infrastructure systems and their facilities and for presenting this information in forms usable in the public decision-making process. This research could include a broad range of technologies, such as new simulation modeling, technology assessment, information display procedures, interactive information-query systems, and public voting and joint-decision-support systems.

Typical questions for research might include the following examples:

• What economies in infrastructure development and operations can be achieved by means of effective public education and participation in decision making?

• How can network managers and users gain more effective practical understanding of network operations and management?