The primary project focus is acquiring requisite data for understanding the complex corrosion and deterioration processes affecting Arizona’s hull, both internally and externally, and modeling and predicting the nature and rate of structural changes resulting from this corrosion. Developing reasonable and effective management alternatives, and deciding the most desirable actions, particularly those regarding intervention or rehabilitation, cannot be done without this information. The current research program is an important step in obtaining necessary scientific information upon which sound management decisions will be made. A central goal of this research is to develop and recommend short-term and long-term management plans for site preservation based on the results of the research program. This project addresses another important issue besides preservation of an important national shrine. USS Arizona contains several hundred thousand gallons of fuel oil, which has been slowly escaping since its loss in 1941. This oil, a potentially serious environmental hazard, is contained within the corroding hull. Catastrophic oil release, although by all indications not imminent, is ultimately inevitable. Understanding the complex hull corrosion processes, structural changes and oil release patterns offers the most efficient method of mitigating this potential hazard. This project will develop a research strategy for environmental impact risk assessment and abatement to address the oil issue.

Because of the particular national importance of Arizona, any solution to the oil issue must incorporate a minimum-impact approach so that long-term site preservation will not be compromised. All research operations are conducted with the respect due an American war grave and with minimum impact to the site. Unnecessary disturbance to Arizona’s hull is likely to be seen by many as more problematic than the limited oil release now occurring, although managers will ultimately have to face the possibility of a larger release. Addressing the oil release problem within a site-preservation framework provides the best balance between the competing social values of preservation and ecology, and it has the highest probability of arriving at the optimal solution for both issues.

METHODOLOGY
NPS-SRC is providing project principals who have been involved in Arizona research from the early 1980s (Lenihan 1989). NPS is also partnering with military units, researchers, academic institutions, commercial companies, research laboratories, professional societies and  other federal agencies in addressing the multifaceted questions confronting managers responsible for both USS Arizona’s preservation associated environmental risk. This research program is designed to be a cumulative progression of multidisciplinary investigative steps. Multiple lines of evidence are being pursued simultaneously, each directly or indirectly linked to the others and to the overall project objectives. Operationally NPS is following a two-fold strategy of research combined with monitoring.

Primary research is directed towards characterizing overall corrosion processes and determining internal and external corrosion rates. These data are required to develop a predictive model of how Arizona is deteriorating and when corrosion will reach the point where structural changes indicate imminent collapse . Archeologists and conservation specialists in Australia conducted pioneering research on iron and steel shipwreck deterioration and determined the major factors affecting shipwreck corrosion are metal composition and metallurgical structure, marine growth, water composition, temperature, extent of water movement, seabed composition and depth of burial beneath the seabed (North and MacLeod 1987:68). Collecting data necessary to characterize critical corrosion processes on USS Arizona will involve evaluating each of these factors, perhaps identifying additional processes, all of which are complex and interrelated, that affect corrosion in many different ways.

When attempting to evaluate the corrosion history of an object it must be considered individually—there are very few oceanographic and environmental parameters that are uniform between sites. In addition to corrosion research, related research is focusing on the oil that remains trapped within Arizona’s hull and on the geological substrate supporting the ship. Monitoring activities are aimed at collecting baseline data for inclusion in corrosion analysis that can also be used to assess changing conditions over time. These data are being used to quantify various on-site conditions such as physical movement of the ship and oil release amounts. Data collected during monitoring is incorporated into overall research domains that give researchers and managers an indication of overall site stability and rates of change.

Research and monitoring activities are broken down into individual research domains discussed below. Each research domain either directly contributes to primary research goals or plays a key supporting role in project objectives. All are interconnected on some level.

FINITE ELEMENT ANALYSIS
Finite Element Analysis (FEA) is a principal research method that will produce the primary predictive tool used during USS Arizona research. A Finite Element Model (FEM) is a computer model used to calculate theoretical stresses and shape changes in a structure under load using experimental variables based on observational data. The FEM divides a complex solid into many small components called elements, each of which can be one of numerous simple shapes. Properties for the material of each element are input to describe the element’s behavior between its end points (for example, mechanical properties, heat flow, density, etc.). The end points of each "finite" element are called nodes. Conditions are set regarding how nodes connect to one another and loads (known as boundary conditions) are added to the model. As each individual element changes under different boundary conditions, it transmits a slightly changed boundary condition to neighboring elements, which then repeat the process. The result are plots of displacements of nodes and calculated stresses in the structure at all points—taken in the aggregate, the displaced nodes and stresses of all the elements in the FEM offers a predictive model of stress and change under different conditions for an entire structure.

For historical shipwrecks such as USS Arizona, an FEM will allow manipulation of multiple variables, such as corrosion rate and hull thickness, to analyze loads and stresses on hull structure for prediction of probable collapse rate, their nature, sequence and consequent impact on structures containing fuel oil. In addition, the FEM will provide a fundamental tool to evaluate consequences of proposed management alternatives involving structural intervention or preservation strategies. There are particular difficulties in applying an FEM to shipwrecks, however. Geometry is constantly changing due to ongoing corrosion, loads can be very complex, and load and corrosion interact in such a way as to increase the complexity of the model (for example, stress corrosion cracking). There are ways to overcome these difficulties, but accurate data based on direct measurements and observations are of primary importance. For the model to be representative of the real world, input data such as structural dimensions and connections, corrosion rates and loads must be as precise as possible.

Baseline FEM development is being conducted by the National Institute of Standards and Technology (NIST) and is focused on modeling the Arizona hull structure in its as-built original state for an 80-ft. cross-section, amidships from frame 70 to 90. The 80-ft area selected for initial modeling represents an area affected by the blast that sank the vessel and the ensuing fire. Because this is pioneering research in the sense that FEA has not been applied to corrosion and deterioration of a historical shipwreck before, this preliminary model is a necessary step to refine and test methodologies for developing the overall model required for predicting present and projected future structural strength. It is important to note that the great majority of the work in creating a finite element model of a structure is in the generation of the model and mesh in the computer. Remediation scenarios can then be tested and further stability studies can be made by simply changing the inputs and accounting for new measurements or ideas.

The next development stage of the FEM will focus on incorporating structural effects of the blast and fire that sank the vessel. Modeling the structural changes to Arizona resulting from the explosion and subsequent fire that sank the ship is the logical starting point for understanding the vessel’s present condition and projecting its future condition and rate of deterioration.

The final stage of FEM development will incorporate external and internal corrosion and thickness measurements to complete the model of Arizona’s present condition and to allow researchers to extend the model into the future. Predictions about current status and future collapse will vary in accuracy depending on the detail of the input data, crafting the correct boundary conditions, and by minimizing simplifying assumptions. For the first issue, the greatest deficiency in data in this case is knowledge of the actual thickness and conditions of hull features both internally and below the mudline. All other assumptions and simplifications have a much smaller effect on the results than these data. The boundary conditions are similarly difficult, as the hull is being supported by a soft, water saturated semisolid that moves relative to the hull.

As the primary "product" of the current research program, much of the data collected during fieldwork and as a result of the ongoing monitoring is designed to be fed directly into revising and refining the FEM to make it as accurate as possible. When combined with corrosion rates and other variables, the model will provide predictability required for evaluating timing, necessity and long-range consequences of management actions.

If monitoring of change in Arizona’s structure over time conforms well with changes predicted by the FEM, researchers will have confidence in extending the model’s predictions to areas of the ship (such as the lower decks) that are difficult to access directly for monitoring purposes. If monitoring of changes does not accord well with the predictions of the FEM the disjuncture between real and predicted behavior will cause researchers to modify the FEM, gather new data that may have been overlooked in the initial model, or both. Over the course of this investigation we anticipate a dynamic give and take between the FEM and ongoing research.