Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 75
Shiphandling Simulation: Application to Waterway Design 8 Research Needs Shiphandling simulation is a high-level technology that is emerging as an important tool in waterway design. Previous chapters identified that many aspects of the present state of practice in the development, use, and interpretation of results for shiphandling simulations are, however, less than rigorous and scientific. As reflected in the case studies, the benefits of shiphandling simulation for visualization of waterway design problems and of consensus building are nonetheless great. GAPS IN THE STATE OF PRACTICE The committee's review of the use of shiphandling simulators for waterway design revealed an overall concern for validity and five specific technology areas that could benefit from substantial research. These areas, which are dependent on one another in an approximately sequential fashion, are the following: the level of accuracy required for the mathematical model, procedures for identifying and validating the mathematical model for ship behavior in restricted and unrestricted shallow water, information and procedures for determining the effect of fidelity of the pilot's visual and physical interface with the simulator on results of real-time simulations,
OCR for page 76
Shiphandling Simulation: Application to Waterway Design a framework and standards for interpreting the results of simulation, and guidelines for the level and scope of simulation required in relation to the type of waterway design process. A research program to fill these gaps is not presently being undertaken in the United States. Moreover, from a review of the state of practice of shiphandling simulation for waterway design in foreign countries, the committee found no evidence that such a comprehensive research program is being conducted abroad. Apparently, fundamental research on shiphandling simulation is rather moribund worldwide (for example, the privatizing of the Computer Aided Operations Research Facility (CAORF) at Kings Point, New York, has resulted in a shift in focus from fundamental research to contracted applied research and shiphandling training), although practical use of simulators for waterway design is growing. It is not clear whether or not congressional interest in research of marine simulation for operator training (generated by major tanker disasters) will result in a resurgence of basic operations research. An original goal of this study was to develop guidelines regarding the appropriate level of simulation. Because substantial gaps remain in the five research areas necessary for developing such guidelines, the committee could not attain this goal. Substantial improvement in knowledge and capabilities in each of the preceding areas holds promise for improving the confidence of practitioners and waterway designers in the results of simulations and, ultimately, for achieving the full potential of simulation. Although this study does not address use of shiphandling simulation for operator training, the basic questions concerning fidelity of simulation also apply where port-specific ship behavior is an element of the training regimen. FUTURE RESEARCH The committee has identified five specific areas for further research that would address the five technology areas defined above: 1) accuracy requirements for mathematical models, (2) identification and validation of the mathematical models, (3) effect of fidelity (visual and behavioral) on real-time simulation results, (4) interpretation of the results of simulation, and (5) guidelines for the required level and scope of simulation. Fidelity Requirements for Mathematical Models As discussed in Chapter 6, shiphandling involves intelligent feedback to available cues (either in real time using a human pilot or in fast time
OCR for page 77
Shiphandling Simulation: Application to Waterway Design using a sophisticated pilot model). In either case, the pilot or autopilot corrects for errors, whether they are due to real effects or errors in the mathematical model. This situation appears to reduce the demands for accuracy on the mathematical model for ship behavior (over that required for an open-loop, dead-reckoning model), but there is little or no information in the literature to document this conjecture or to indicate what level of model accuracy is required. Research could be conducted to determine the sensitivity of the results of simulation for waterway design to either the framework for the mathematical model or the accuracy of the coefficients used in conjunction with this framework. The mathematical model frameworks in typical use differ little, if at all, in their linear terms. The differences, where they occur, exist in the number and arrangement of higher order terms. A better understanding is needed about the requirement for accuracy of the various coefficients in typical mathematical frameworks, in particular, the coefficients that characterize the effects of small under-keel clearances and the interactions with varying banks and currents. Identification of these coefficients is exceedingly difficult and therefore expensive. Sensitivity studies constitute a necessary preliminary for the remainder of research opportunities, but need not represent a significant investment. Fast-time simulation is ideal for this purpose because it is repeatable and does not include the variability inherent in human pilots. Examples exist of complete mathematical models for ships in restricted waterways. Investigations of the adequacy of the framework can be based on the use of different known models for one ship type and models for several representative waterways. Investigations of the accuracy of the coefficients probably will involve systematic perturbations of the mathematical models for several different ship types and for several representative waterways. Identification and Validation of the Mathematical Models If, presumably, the above-mentioned research determined the level of accuracy needed in coefficients of a mathematical model, the problem would remain of identifying these coefficients. Chapter 5 revealed that considerable weaknesses exist in the identification of hydrodynamic coefficients for use in a mathematical model. Although scientific means are available for performing accurate identification, the expense would be prohibitive for tests to characterize the behavior of just one ship in all possible small under-keel clearance and bank situations. New, less costly approaches are needed to overcome this gap. Developing this information by computational fluid dynamics (CFD) is beyond present capabilities, although the use of CFD methods is rapidly expanding. Physical modeling techniques and the scaling relations neces-
OCR for page 78
Shiphandling Simulation: Application to Waterway Design sary for their performance exist and have been known for some time. However, physical modeling is limited in three ways. First, the number of tests necessary to characterize the hydrodynamic forces on a single ship in a restricted waterway with a shallow bottom and banks is very large, and as a result, the costs would be large. Second, only a few facilities in the world have flat-enough bottoms to perform model tests in shallow water comparable to realistic under-keel clearances. None of these facilities are in the United States. Finally, it is impossible to scale viscous effects in smallscale model tests, and there is reason to believe that this factor is important in the case of small under-keel clearances. Even if extensive model tests were performed, validating the resultant mathematical model would be even more challenging. To date, validation by comparison with full-scale measurements of ship trajectories in restricted waterways has been limited to only a few cases. Even so, the most extensive and most scientific of these (the Esso Osaka trials) did not involve small under-keel clearances comparable to typical waterway situations or the influence of banks. Validation methods of deepwater maneuvering predictions that are based on full-scale maneuvering trials have often been incorporated in the delivery trials of new ships. However, these trials usually have been aimed at simple turning performance and steering stability. Similar trials in shallow or restricted waters common in waterway design have not been performed for reasons of safety. This constitutes a significant gap in the validation tools for waterway design. Associated research opportunities include the development of new, efficient techniques that could reasonably be expected to identify numerical coefficients for a mathematical maneuvering model for restricted waterways to the level of accuracy required, and the development of techniques for safely conducting full-scale tests in typical waterway situations and for analyzing the results to calibrate and validate the mathematical models. The committee anticipates that these techniques would be tailored specifically to the needs of waterway design. In particular, it is anticipated that the required accuracy of the coefficients may be less than that achievable by classical model tests, and that new, innovative, and more economical techniques would result from exploiting this requirement. The committee further anticipates that the identification and validation would likely be a combination of theoretical results and model tests (perhaps involving systems identification methods or CFD approaches).
OCR for page 79
Shiphandling Simulation: Application to Waterway Design Effect of Visual and Behavioral Fidelity on Real-Time Simulation Results Conventional wisdom states that the higher the visual fidelity of the simulator, the more useful will be the results of real-time simulation. Although some studies have addressed individual aspects of fidelity, proof of this conjecture does not exist. The importance of developing more information concerning visual fidelity is driven by two considerations. First, with the cost of computation now decreasing dramatically, visual fidelity is no longer the principal determinant of the cost of a simulator facility. Second, new shipboard instrumentation is developing at such a pace that some existing studies of fidelity may no longer be relevant. Bird's-eye view displays have often been the only displays available in low-fidelity simulators, a feature that is considered by some to be a defect because the displays provide the pilot with more information than would be available on a real bridge. The development of differential GPS (global positioning system), digital chart data, and inexpensive on-board computer graphics equipment (relative to the cost of the ship and cargo) have made accurate bird's-eye view displays a reality. Electronic chart systems can include all the aids to navigation and other waterway information. In the future, integrated bridges, some with piloting expert systems (that is, artificial intelligence decision aids), together with normal shipboard sensors may produce other displays that communicate real-time decision-making information to the pilot. Integrated bridges are presently available on only a small number of vessels worldwide. An associated research opportunity is to determine the presence and fidelity of such systems on real-time simulations used for waterway design. The effort would investigate the potential role and efficacy of new instrumentation available to pilots and the extent to which this needs to be represented in marine simulations. Interpretation of Simulation Results Chapter 3 stated that shiphandling simulation is based on the assumption that a small sample of simulations using one or two ships, a few environmental conditions, and a few pilots will provide meaningful information for use in waterway design. Because the accuracy of the mathematical models of ship behavior is in question, the setup of a simulator facility is so costly, and the collateral benefits of simulation (for example, consensus building) are often an important objective, less emphasis has been placed on developing a formal framework for interpreting marine simulation results than on simulation in other industries. Elaborate frameworks (usually statistical measures) have been developed for quantifying the performance of many engineering simulations in
OCR for page 80
Shiphandling Simulation: Application to Waterway Design other disciplines and for use in design. Typical among these are simulations of the behavior of telephone networks, traffic congestion on highways, and flow of products in oil refineries. In these fields, the characteristics of the system elements are well known, but the system is subjected to random demands that stress it in complicated ways. Statistical measures are used to relate the design parameters (which reflect construction costs of the system) to risk of failure and its consequences (the contingent costs of the system). Other types of simulations use quite different frameworks for analyzing performance. For instance, simulators designed to train personnel in the art of aircraft handling include both objective measures, for instance calibration against measured performance characteristics (such as turning circles) and subjective measures, such as pilot confidence, that the simulator behaves like a real airplane (Appendix F). Unfortunately, the use of shiphandling simulators for waterway design does not fit comfortably in any of these molds. Shiphandling simulation is in many ways more difficult and complicated than the examples cited. As with aircraft simulators, marine pilots must feel confident that the ''feel'' of the simulated ship is like a real ship. However, information to calibrate the model, such as performance in very shallow water or near banks, is not known with scientific accuracy for any commercial ship. As with road traffic simulations, the quality of pilots and the number of different ship types and their performance vary greatly. However, an equivalent to the considerable data that characterize vehicle behavior on a highway does not exist for ships. The highway problem is also simpler in another way: there is no analog for the changes in steering performance needed in ships due to changes in under-keel clearance or banks. The basic problem in shiphandling simulation arises because the sample size in the simulation is so small restricting the application of classical inferential statistics. Needs for future research include the following: The development of a framework to interpret results of a small sampling of simulator runs in terms of the quantities that affect waterway design. This framework could include, for example, a numerical estimate of the significance of results and confidence bounds on the predictions of swept paths measured in the simulation. The development of a procedure for estimating risk of accident associated in a particular waterway design and for estimating the consequences resulting from potential accidents. Guidelines for the Required Level and Scope of Simulation A synthesis of the results of research suggested above could yield considerable insight into the level and scope of simulation required for the
OCR for page 81
Shiphandling Simulation: Application to Waterway Design waterway designer. However, simulation is a sophisticated technical discipline that relates to only one aspect of waterway design, and it is outside the typical focus and training of the waterway designer. The planning representatives of local project sponsors who must approve the expense of simulation are even further removed from this expertise. Therefore, an associated research opportunity exists for synthesizing information from the previously mentioned research and transforming it into a set of guidelines that could be used by the waterway designer and the sponsor to select appropriate simulation studies for a particular waterway design. The committee found that such guidelines do not exist. As a result, waterway designers and their sponsors have little basis for selecting one simulator over another and for selecting the scope of simulation studies to be performed. Further, a set of guidelines based on a firm scientific footing could permit more rational decisions regarding when and to what extent simulations should be performed for given waterway projects. STRATEGIES FOR IMPLEMENTING A RESEARCH PROGRAM Support for basic research on shiphandling simulation has withered within the past decade. Only the U.S. Army Corps of Engineers (USACE) has a modest, project-oriented program in this area. As a consequence, the number of persons within the interested federal agencies experienced with shiphandling simulations has declined. However, the services of a substantial number of ship hydrodynamicists, both internal and external to the federal government, might be applied to fundamental research. Several experimental facilities exist worldwide that could be used to conduct elements of the research. (A facility catalog of ship hydrodynamic facilities is maintained by the International Towing Tank Conference.) One notable limitation is the inability of existing facilities to scientifically validate the scaling of mud behavior from model scale to full-scale (that is, reproducing on a model scale a material that would emulate behavior of bottom sediment) when testing ship maneuverability in situations with very small under-keel clearances. The research program necessary to put shiphandling simulation for waterway design on a firmer scientific basis, thereby greatly increasing the confidence of the entire maritime community in the usefulness of the technique, would be ambitious, expensive, and long range. The committee believes such a program would require about 10 years of dedicated effort and about $15-30 million in research funds. However, costs in this range are modest relative to the annual investment in port facility capital improvements and in waterway construction, operations, and maintenance. Developing a strategy for the research program would entail addressing sponsorship as well as the resources necessary to conduct the research, including skills and facilities.
OCR for page 82
Shiphandling Simulation: Application to Waterway Design The USACE is the government agency charged with primary responsibility for waterways development, including waterway design, permitting, dredging, and disposal of dredged materials. With this leadership role comes implied responsibility for organizing and coordinating research needed for waterways development in the United States. USACE operates a modest computer-based simulator, has a small pool of technical expertise, and has used these resources in about 40 waterway design studies. With regard to basic simulation research, a technically broad scope of effort would be required. USACE, by practice, principally conducts and has good experience with limited-term, project-oriented research for civil works. Present USACE technical resources do not appear sufficient to undertake or guide the multi-year research program that is needed to improve the scientific basis of simulation technology. In the committee's opinion, USACE would need to augment its technical base with experts from industry, especially in the areas of ship hydrodynamics simulation-based research, and human factors. Other entities that have interest in waterway design (and in some cases operator training) and are potential beneficiaries of improvements in the national simulation capability include the U.S. Coast Guard (USCG) and U.S. Maritime Administration (MARAD) (both of which have previously been involved in simulation research), project sponsors, shipping companies, and operators of port facilities. Resources of these entities vary widely. The committee found that none of the non-federal entities appeared to have either sufficient focus or in-house capability to independently direct even a small part of the research program. Two implementation strategies appear feasible. The federal government could fund the research in support of a national interest in maintaining competitiveness in international commerce. USACE could undertake, plan, and coordinate a government research program, which includes participation (and perhaps cost sharing) by other involved government agencies such as MARAD, the USCG, and perhaps the U.S. Navy (USN), in a supporting role. An alternative would be to establish a government-industry research consortium that included key components of the U.S. maritime industry in both sponsorship and technical support capacities. Nongovernmental participants could include waterway designers, port operators, shipowners and operators, and pilot associations. This approach would have the advantage of involving all the direct beneficiaries to support the research, although coordination of such a body could prove cumbersome. This research program could take full advantage of available expertise and capabilities at existing simulation and research facilities throughout the United States to ensure that the selected research plan is focused on the areas of greatest need, is sufficiently comprehensive, and is cost-effective.
OCR for page 83
Shiphandling Simulation: Application to Waterway Design In an environment in which available funding is likely to remain limited, it is essential to ensure the maximum cost-benefit ratio of all research conducted. SUMMARY The development of meaningful guidance for waterway designers and sponsors on the use and applicability of shiphandling simulators for waterway design is inhibited by gaps in knowledge and capabilities in several critical areas. Because of the complex scientific basis for simulation and the hardware associated with it, the research required to fill those gaps is essential for full utility and acceptability of the technique, albeit its expense. Currently, no government agency, commercial enterprise, or research organization has undertaken or appears ready to undertake a dedicated research program on shiphandling simulation. Such a program could be a joint government-industry initiative, perhaps dovetailed with research that may be needed to establish shiphandling simulation as a fully accredited shiphandling training aid. USACE would be an appropriate organization to coordinate needed research because of that agency's primary role in waterway development.
Representative terms from entire chapter: