Toward Infrastructure Improvement: An Agenda for Research (1994)

To be effective, infrastructure research efforts must strike an appropriate balance between need and capability. On the one hand, the research should seek to expand knowledge, fulfill demands, capture opportunities, and gain benefits of applying new technology to U.S. infrastructure. On the other hand, the research should take advantage of available research resources or seek to build new capabilities where needed.

Information on the overall level and scope of U.S. effort on infrastructure research, development, and technology adaptation activities is not well developed. Spending occurs in both public and private sectors, and in uncoordinated programs by municipalities and commercial enterprises, as well as large governments or corporations. Some spending occurs to solve problems on a specific project rather than to develop broadly applicable techniques, although lessons learned on a single project often diffuse to projects in other regions. As a part of this study, committee members and staff pieced together available statistics to estimate total U.S. spending for infrastructure research, development, and demonstration.

Since World War II, the U.S. government has become the primary sponsor of all R&D—by one estimate, currently providing about 43 percent of R&D spending in the United States (NSF, 1992b). The Federal government is estimated to be the principal source of infrastructure research spending as well, although only a small portion of the total government R&D funding is

devoted to infrastructure. A recent survey conducted by the Civil Engineering Research Foundation (CERF) found this spending to be between $1.026 and $1.386 billion in fiscal year 1992, approximately 1.6 percent of total federal R&D expenditures (CERF, 1993a).

In the CERF study, the NSF was reported to spend less than 10 percent of its $2.7 billion R&D budget on public works infrastructure research. Other reports place total 1993 NSF spending on infrastructure somewhat lower, at $78 million (ENR, 1993).

Drawing on NSF data, the U.S. Congress' Office of Technology Assessment (OTA) reported 1990 federal outlays for public works R&D by five key agencies¹ to have been some $1.4 billion (measured in 1982 dollars) (OTA, 1991), and noted that infrastructure R&D outlays had declined by nearly one-third over the decade from 1980 (in constant dollar terms), a trend that NSF data show to be continuing. Citing National Science Board² data, Cohen and Noll (1992) show infrastructure accounting for 1.8 percent of federal R&D appropriations. With such supporting evidence available, the committee adopted $1.026 billion (midpoint of the CERF range) as an estimate of total annual federal infrastructure R&D spending.

Within this total, the allocation of government research funding varies substantially among infrastructure's principal functional modes. At the one extreme, highway research has long been reliably supported by budget allocations from federal and state gasoline tax revenues and is conducted in a large number of federal and state laboratories, universities, and many private companies. At the other extreme, institutional buildings receive little explicit attention as infrastructure, and building research on the whole receives limited funding.

However, the federal government is not the sole source of infrastructure R&D spending. Spending by private-sector firms and not-for-profit organizations on infrastructure research does occur, some of it managed through industry groups such as GRI and EPRI. APWA, AWWA, ASCE, and other professional and trade groups also have research foundations or other units that sponsor and conduct infrastructure-related research. Much of the work of these latter groups is carried out by funding from membership contributions and government contracts and grants.

A variety of private companies is involved in infrastructure-related businesses, including construction, the manufacture of telecommunications and power generating equipment, wire and cable, and chemicals for water treatment, waste processing and disposal, transport and communication services, and instrumentation and software development. These companies spend for both new product development and process improvement.

The construction sector in the United States, a $410 billion industry, ³ is estimated to spend annually only about 0.25 percent of gross sales on research,⁴ or $1.025 billion. Much of this research is funded by commercial

building-product manufacturers. Only 20 to 25 percent of the total R&D spending is likely devoted to research related to construction of highways, power supply, and other infrastructure.⁵ These figures imply that annual infrastructure R&D spending by the construction industry is about $230 million (midpoint of the range).

Manufacturing related to infrastructure, but not to construction products and materials, includes primarily transportation, communications, and power generation and equipment. These industries may spend as much as $3 billion annually on R&D. However, some two-thirds of this amount is attributable to the motor vehicle and aircraft manufacturers, much of it for new model development. Perhaps 20 percent of the total, $600 million, may be directed toward achieving substantive infrastructure systems innovation.

The contribution of transportation and public utilities businesses to the U.S. gross national product (GNP) was approximately $440 billion in 1988 (DOC, 1991). If perhaps 0.02 percent of sales are devoted by such businesses to infrastructure-related R&D,⁶ then the estimated annual spending is $88 million.

Various private firms provide design services, construction management, operations research, and other professional services related to infrastructure. United States firms annually perform some $30 billion in domestic and overseas design and related work, of which approximately half is related to infrastructure (ENR, 1992). Such firms typically conduct informal in-house R&D for new product development and may compete for government-sponsored research funds, spending in aggregate less than 0.05 percent of sales on R&D, or an estimated $15 million annually.

Much of this spending is associated with civil engineering and related professions. U.S. government 1991 budgetary obligations for civil engineering research totaled $55 million, some two-thirds of which was to be spent under NSF programs (NSF, 1992b; CERF, 1993c). Perhaps $22 million, less than half of the total amount, is targeted on areas specifically related to infrastructure. This amount, probably included in the previously mentioned CERF and OTA estimates of government-funded infrastructure research, may be representative of levels of private spending, reinforcing the previous estimate.

Combining these several estimates, the total annual U.S. R&D spending on infrastructure-related technology may be approximately $2.1 billion, as summarized in Table 2. This amount is approximately 1.6 percent of total estimated U.S. R&D spending.

Members of the infrastructure professional, academic, and research communities have expressed concern that these spending levels are too low and that evidence is strong that declining research underlies in part a purported lag in technological innovation in the field. However, the data

are poor, the topic is complex, and such relationships are difficult to demonstrate. Further, many analysts have pointed out that R&D activity does not necessarily lead to innovation.⁷

Nevertheless, the committee agreed that low research effort signals an environment less likely to yield innovation, and evidence suggests that the United States may be lagging in comparison to other countries. For example, Cohen and Noll (1992) noted that the United States spends comparatively less on civilian technology than other advanced, industrialized nations. Taken as defined by the reporting government agencies, infrastructure R&D priority is similar in the United States, Germany, Japan, and Britain, but is a notably higher priority in France (see Table 3).

However, the definitions of what comprises infrastructure R&D in these countries is not necessarily similar. Taken together, environment, energy, and "infrastructure" R&D spending account for approximately 6

percent of R&D spending in the United States and Britain, and about 10 to 14 percent in France and Germany. In Japan, which has no significant domestic energy resources, the total is over 25 percent.

Table 3 data seem to match observers' impressions of the infrastructure R&D activities in these countries. Participants in reconnaissance visits sponsored by CERF and the NSF have returned with tales of extensive research facilities and far-ranging programs (CERF, 1991b; NSF, 1992a). Others have pointed to such examples as the French telecommunications system, Japanese precision construction, and Swiss bridge design as demonstrations of the payoffs of greater effort on R&D and willingness to apply new technology to infrastructure.

The pattern of who does research and how their results are applied differs substantially among countries, and this variation clearly has impact on the productivity of R&D spending. In the United States, R&D funding is spent at government laboratories, universities, quasi-governmental laboratories,⁸ and private research organizations; and much of this spending is directed at investigating specific problems encountered in practice.

The federal government has developed several R&D centers to support agencies' infrastructure-related missions. The U.S. Army Corps of Engineers, for example, is responsible for navigable waterways, flood control, hydropower, and other water resources, as well as for military programs, operates the Construction Engineering Research Laboratory in Illinois and the Waterways Experiment Station in Mississippi. The Department of Transportation has research facilities in Virginia (highways), New Jersey (air transportation), and Massachusetts (transportation systems). The Bureau of Reclamation conducts and manages research from its Denver headquarters. Similarly, the Department of Energy, the Environmental Protection Agency, the Department of Agriculture's Forest Products Service, and others maintain R&D operations that work on infrastructure topics.

Several of the laboratories operate at least partially as contracting organizations, entering into project agreements to investigate specific problems faced by the program elements of their agencies. These laboratories depend on funding provided by the operating programs rather than direct budget allocations for research, and thus have limited ability to set their own research agendas.

Within academic institutions, government and private research laboratories, ⁹ and private companies, research and development activities applied to infrastructure are performed largely under the auspices of civil engineering programs or units associated with development and management of public works. Programs in electrical, mechanical, water resources, and construction engineering, architecture, urban planning, urban design, and landscape architecture also readily include topic-related infrastructure systems and technology. In addition, the fields of economics, public finance, public administration, geography, and regional planning contribute regularly to enhanced understanding and methodology for infrastructure development and management.

Most of this work and the centers where it occurs are mode-specific (e.g., transportation or electric power). However, in recent years, several universities and other groups have established research centers under the title of "infrastructure." These include a consortium of New York City area institutions led by Columbia University; a research center at Northwestern University;¹⁰ a joint venture of Harvard's Taubman Center for State and Local Government and the Graduate School of Public Affairs of the University of Colorado at Denver; separate and unrelated programs at the University of Nebraska, the Massachusetts Institute of Technology, George Mason University, and the New Jersey Institute of Technology. However, it is the committee's sense that the absence of either a national institutional focus or reliable funding for research on infrastructure as a whole has limited the ability of such centers to develop broad capabilities. Their programs remain generally small, unrelated to one another except through collegial communication, and opportunistic in their pursuit of research funding and their consequent research agendas.

Several professional associations and industry groups have formed research institutes that sponsor or undertake infrastructure-related research. Some of the more prominent of these organizations are listed in Table 4. Most of these organizations depend for funding on membership fees and voluntary contributions, and contracts or grants from government agencies, foundations, and private corporations.

The electric power and natural gas industries have established EPRI and GRI to conduct research. EPRI is funded through voluntary contributions from member utilities (NRC, 1989a). GRI is funded by surcharges on the rates charged by interstate pipelines for gas for resale and transportation. These surcharges would be passed on through the distribution com-

panies to the ultimate consumers (NRC, 1989b). Seeking to emulate the federal gasoline tax set-aside that supports much of the nation's highway research, membership fees are based on the annual revenues of the member companies, which are generally proportional to the kilowatt-hours of electric power or cubic feet of gas produced and sold.

As in the United States, most infrastructure research in other countries is directed at specific modes and concentrated in government agencies, but is funded directly by the national governments to a much greater extent than in this country. However, a few foreign programs are multimodal in scope. Most notable is Sweden, where the University of Umea, the Royal Institute of Technology, and other universities have been involved in research on infrastructure related to regional development, primarily with Swedish government funding. Research on technological innovation, focused primarily on transport infrastructure, and on global climate change (with some attention to the influence of infrastructure) has been conducted at the International Institute for Applied Systems Analysis in Austria. Japanese industry, universities, and government have been pursuing several initiatives to explore new urban infrastructure systems, under the guidance of the Ministry of Construction and Ministry of International Trade and Industry. The National Research Council of Canada, a government agency, has also undertaken programs of infrastructure research.

The World Bank, with lending programs operating in some 60 nations and 200 cities worldwide, has sponsored a wide range of research on infrastructure, emphasizing technologies appropriate for application in regions with low income or at early stages of industrialization. The Bank seeks to define more clearly the links between infrastructure investment and economic development to provide sound bases for their lending programs.

These various research sponsors and researchers have, from time to time, undertaken to define infrastructure research needs statements, to direct their own activities, or to have broader influence within the research establishment. The committee found that few of these past efforts were comprehensive, aimed at infrastructure as a multi-functional system of facilities and services. Most such studies were developed to deal with one mode alone, such as highways or water supply. In view of the structure of infrastructure, such a modal focus is not surprising.

For example, a study funded by the Federal Highway Administration and conducted by the Transportation Research Board produced an agenda for highway research focused on what might be termed ''hard'' technology, such as asphalt behavior, highway deicing technology, and bridge deck performance. The report, America's Highways: Accelerating the Search for Innovation (TRB, 1984) recommended a program of such research that became the basis for the Congressionally-funded Strategic Highway Research Program (SHRP), a five-year activity administered by NRC. The SHRP, ending in 1993, did in fact produce many useful results, and the Federal Highway Administration is establishing a national training center for state officials, intended to speed transfer of SHRP-produced technology into practice in the 50 states.

The committee did review some studies that were based on a broader perspective of infrastructure, although not necessarily aimed explicitly at developing research agendas. One such study, exploring the development of a national urban policy (NRC, 1984), held a symposium entitled The Adequacy and Maintenance of Urban Public Facilities, that identified an extensive research agenda, including the implications for technology development for new and existing infrastructures. The cross-cutting infrastructure research topics addressed primarily management and economic needs, such as the role of standards and risk analysis in infrastructure decision making, life-cycle cost analysis strategies for infrastructure maintenance and repair, methods for evaluating benefits and costs of various levels of system performance, and effective diffusion of innovation into practice. One paper presented in the report calls for research on condition assessment methods and the value of information systems in condition assessment.

Another NRC study (NRC, 1985), not directly related to this earlier work, included a workshop that attempted to identify technological alternatives—new systems, materials, and devices—for future replacement of present urban infrastructure systems. What emerged from the workshop was a call for further research on the integration of various infrastructure elements. Workshop papers described cross-cutting research areas such as communications, broadly used to describe the use of new electronic technologies (e.g., computerized mapping and modeling, telematics, remote measurement technologies, automated decision and control systems, and integrated computer systems). Others considered ways to combine infrastructure functions.

The National Council on Public Works Improvement (NCPWI) 1988 report, Fragile Foundations, was a broad assessment of the nation's infrastructure systems, that included considerations of the needs and opportunities for research (NCPWI, 1988). An NRC-conducted study supporting the NCPWI's work advocated integrated or cross-cutting infrastructure research and innovation as well as modal research. While the report does not present an agenda for such cross-cutting research, it did target four areas of opportunity, in materials science, information technology, nondestructive evaluation, and urban and regional planning (NRC, 1987).

The National Academy of Engineering sponsored a workshop in 1986, on The Evolution of Future Infrastructure. Workshop papers were 'subsequently published as a report reviewing the technology, history, and theory of infrastructure. A paper by Ibbs and Echeverry (1988) concentrates on cross-cutting technologies for more efficient construction, referring particularly to new materials technology, new monitoring and sensing technologies, nondestructive testing, new construction methods, construction robotics, new management technologies, and data bases for management.

An extensive discussion of infrastructure research is contained in an OTA report on the federal role in infrastructure policy (OTA, 1990). This report includes a chapter on cross-cutting technologies for infrastructure, identifying several technologies on which research would be appropriate:

• measurement, instrumentation, and nondestructive evaluation;

• information and decision systems, including maintenance management information systems, geographic information systems, artificial intelligence, and simulation models;

• communications and positioning systems, including signal systems and the use of satellites;

• field construction technologies, including trenchless construction technologies, tunnels, soil improvements and stabilization, dredging technology, and rail track construction and rehabilitation;

• materials and corrosion, including continued research on concrete, asphalt, steel, geotextiles, plastics, advanced composites, and research on protecting materials against corrosion; and

• technology management, including project design, procurement options, standards and specifications, risk and reliability, proprietary technologies, and personnel training, education, and recruitment.

Revitalization of public works infrastructure was one of five thrust areas at the 1991 Civil Engineering Research Needs Forum held in Washington, D.C., by CERF. Participants in that forum identified and ranked 10 high-priority research topics (CERF, 1991a):

1. developing tools to make intelligent management decisions;

2. finding new ways to finance infrastructure investments;

3. extending the useful life of the infrastructure;

4. protecting bridges from natural hazards;

5. identifying structural problems by diagnosis;

6. removing institutional barriers to innovation;

7. economic benefits from public works investments;

8. improving water-resource systems data through new technology;

9. mitigating coastal damage from natural hazards; and

10. protecting dams against earthquakes and floods.

Estimates were made of appropriate levels of spending on each topic. The third item, extending the useful life of the infrastructure, was estimated to warrant some $100 million in research spending, more than 10 times the allocation to any other topic. More recently, CERF has proposed establishment of a research program in high-performance materials (CERF, 1993b).

These various statements of infrastructure research needs bear great similarities, particularly in their emphasis on materials, condition assessment, and monitoring, and applications of newly developed electronics and information management technologies to traditional infrastructure. A shift in emphasis from construction of new facilities to maintenance and repair of existing ones pervades the final decades of the twentieth century, and is reflected in these studies. The committee considered this shift and these previous research needs studies in their own deliberations.

The NSF's support of research is meant to build a knowledge base of engineering and science to support technological innovation throughout the nation's economy. Because the NSF's research spending is concentrated in the nation's colleges and universities, its impact extends beyond immediate research results to the next generation of infrastructure professionals.

The NSF's independence and freedom from association with any particular infrastructure mode or interest group means that the Foundation's spending can be directed in a neutral and nonproprietary manner. In the

past, this spending has been allocated and administered primarily along disciplinary lines, matched to the structure of academic institutions that are the NSF's primary grant recipients.

However, from time to time, special programs have been undertaken that encourage collaboration across academic disciplinary boundaries or that address needs not easily placed within traditional organizational structures. For example, over the past several years the NSF has funded initiation of several research centers, focused on earthquake engineering, Portland-cement-concrete technology, and large-scale structures. These centers typically bring together faculty and students from several parts of the university—or even from several universities—to work on projects housed within the center's administrative structure. The housing may be physical as well, with offices, equipment, and meeting areas providing a base of operations and sense of identity for the center's activities.

Within the context of the NSF's traditional role and the complexity of U.S. infrastructure needs, a CIS (Civil Infrastructure Systems) ¹² Task Group recommended (NSF, 1993) a "Research and Knowledge Transfer Program... to enhance system performance and the longevity for existing and future infrastructure systems." The recommendation encompassed research in three broad topic areas—deterioration science, assessment technologies, and renewal engineering—intended to support this aim, and knowledge-transfer efforts to encourage practical application of research results (Table 5). Subsequent planning has added a fourth topic, institutional effectiveness and productivity. These NSF task-group recommendations were part of the context for the committee's work.

The NSF's CIS program is being developed for application over the next several years to address currently well-recognized needs for better understanding of facilities aging, maintenance, and repair. The research niches presented in this report are intended to extend beyond the CIS pro-

gram as it is being defined, to look toward needs emerging over the coming 10 to 20 years and to consider the full range of services and facilities that comprise the nation's infrastructure.

Many needs for infrastructure research have to do with processes (e.g., for waste treatment, energy transmission, and control of leachates) outside the scope of NSF programs that sponsored the committee's work. The committee generally sought to maintain a focus on the structures and technologies that support or house such processes, rather on than the processes themselves.¹⁴

Both the CIS program and this broader scope of research build on current practice and previous research. The task facing the committee was not simply to judge what are significant questions for which research might find answers, but rather what are the questions that particularly warrant research that might be funded by the NSF, alone and with other agencies. Such questions will comprise the niches, within the broad scope of infrastructure research, that will be the most appropriate application of NSF's inevitably limited resources.

These niches represent broad cross-cutting areas for research wherein the NSF's sponsorship is likely to be particularly productive. Research topics within these niche areas would represent an intersection of current and anticipated infrastructure need and the NSF's mission and capability, within a broader context of the U.S. infrastructure research establishment.

The scope of the committee's deliberations was set to address, if not resolve, a tension between committee members' tendencies toward broad systems generality and reductionist specificity. The former tendency, growing out of recognition of the complexity of infrastructure and its urban settings, encouraged broadly stated research questions and expansive programs. The latter tendency reflected desires to assure that the research is practical and aimed at feasible targets. Both stances have merit, and striking the appropriate balance attracted continuing discussion as the committee defined and refined its guidance in these seven broad infrastructure research niche areas:

• Systems life-cycle management, including plant or network operations, asset deployment, maintenance practices, system performance assessment, and control, management, renewal decisions, quality of life and environmental management, throughout all stages of the life cycle from initial materials production to final facility demolition and waste disposal;

• Analysis and decision tools, for planning and design, needs assessment, dealing with capacity issues;

• Information management, including data collection, storage, assessment, and retrieval in forms that support decision making;

• Condition assessment and monitoring technology, for facilities and service performance (e.g., flow metering and effluent content);

• The science of materials performance and deterioration, including mechanical and chemical behavior, changes with time and use, and so on;

• Construction equipment and procedures, including construction management methodology; and

• Technology management, such as selection of treatment process or transport mode, bases for decision making, and environmental or social consequences.

The committee found that boundaries among broad niche areas are inexact and difficult to define. There is sometimes overlap or duplication among specific topics and research questions presented in different niche areas. Such overlap results not only because infrastructure technology is complex and multidisciplinary, but also because there is generally more than one way to look at each problem. Similar topics may be stated from several perspectives, each of which may lead to different research approaches, different likelihoods of success, and possibly different results. One approach to a problem may yield useful results where another does not. Also, because research sometimes yields unexpected and serendipitous results, variety increases the potential scope for discovery. Hence, variety may be a key characteristic of a robust research agenda.