15 2004 March • JOM Materials Prognosis Overview Editor’s Note: Presentation of this paper is supported by the Air Force Research Laboratory, under agreement number F33615- 01-D-5801. The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright notation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the Air Force Research Laboratory or the U.S. Government. Performance and life limits for structural materials in complex mechan- ical systems are often established based heavily on a fear of failure. Conventional approaches for avoiding structural failure often involve extensive periodic inspections, lengthy maintenance processes, and highly conservative “go, no-go” operational decisions, all of which may significantly impair system readiness. This article summarizes a typical present-day life-management process for an advanced system and then presents the key elements of an alternative life-management approach known as materials damage prognosis. INTRODUCTION Platforms such as aircraft and helicop- ters must satisfy a wide variety of operational needs. To assure the safety, reliability, and adaptability of a given fleet of assets, it has been common practice to manage the fleet life based on the statistically expected behavior of the worst-case assets in the fleet. This is a well-justified response to the fear of failure of the assets, but this approach often imposes considerable conserva- tism in terms of initial design, deploy- ment, and operation of these systems. Figure 1 depicts a typical asset- management process in use today. At the center of the figure is a decision process for determining the capability of the asset to perform future missions. The decision is based on relatively basic knowledge of initial design, prior use and maintenance, and results of nonde- structive inspections. To assure safety Using Materials Prognosis to Maximize the Utilization Potential of Complex Mechanical Systems Leo Christodoulou and James M. Larsen and reliability, the system is approved for service only if there is a very low probability of failure (the book life), which is normally defined as the inability to perform the needed mission. Based on this scenario, the vast majority of the retiring assets would actually have a great deal of remaining lifetime or capability, but it is impossible or impractical to distinguish between the short-life and long-life assets. A new initiative 1–3 by the U.S. Defense Advanced Research Projects Agency (DARPA) will develop and demonstrate science and technology for materials damage prognosis to enable the manage- ment of materials and structures based on a real-time knowledge of evolving damage. This approach integrates three primary areas of technology: physics- based models of material damage and failure; noninvasive interrogation techniques to provide real-time quantita- tive knowledge of material damage state; and signature analysis, data fusion, and reasoner technology to predict future performance, including ultimate failure of individual systems. When imple- mented, this prognosis capability will revolutionize the ability for commanders to adaptively deploy and manage specific assets to meet rapidly evolving combat needs. The prognosis system will also maximize safety and readiness, substan- tially reduce the logistics burden and operational costs, enable new operational scenarios, and provide concepts for greatly optimized designs of future systems, such as piloted aircraft and uninhabited combat air vehicles. Figure 2 presents a schematic of the process of asset management based on materials damage prognosis. This approach integrates improvements in a variety of complementary sciences and technologies, which include the follow- ing areas of new or improved capability: (1) multiscale physics-based models of multiple and interacting damage and failure mechanisms in materials and structures subjected to complex and varying usage conditions; (2) quantita- tive and predictive understanding of stochastic and probabilistic effects in materials, emphasizing life-limiting or capability-limiting behavior; (3) innova- tive sensor technologies to detect, quantify, and characterize incipient and advancing damage; (4) techniques for establishing state awareness in large, complex machinery and structures and for detecting, tracking, and predicting evolution of materials damage using all available techniques (i.e., autonomic system interrogation, local and global sensors, signature deconvolution, and feature extraction); (5) a highly inte- grated database, data-management, and data-fusion system to provide historical and timely information on the pedigree, maintenance, and use of individual assets; (6) means for evaluating and quantifying uncertainties and error estimation for both model predictions and sensor data; (7) reasoning method- ologies for future capability prediction; (8) and a robust prognosis system architecture that limits false positive and false negative predictions and presents the appropriate data and predictions to appropriate individuals (i.e, system operators, local/unit commanders at various levels, fleet commanders, and maintenance engineers). MATERIALS DAMAGE PROGNOSIS FOR TURBINE ENGINES One of the most pervasive and essential of military assets is the gas turbine engine, which powers fixed-wing aircraft, rotorcraft, ground vehicles, and