MSE Colloquium: Sridhar Narasi, Long Term Performance Assessment of Materials

Director, Materials and Sensors Program, DNV GL

All dates for this event occur in the past.

264 MacQuigg Labs
105 W. Woodruff Ave
Columbus, OH 43210
United States

Abstract

Structural materials are required to operate over long periods of time, often well past their original design lives. Failures of materials occur due to the accumulation of almost imperceptible changes in the materials, interfaces, and the surrounding environment. Much research is done to analyze the mechanisms of the failures, design test methods, build models, and develop standards to prevent future failures of the types that have already been observed. We have developed and honed our analytical skills.

However, we have been spectacularly poor in anticipating failure modes that have not already happened (the so called “unknown unknown” problems). For example, near-neutral pH SCC of carbon steel was not known until the 1990’s after such failures occurred in Canadian pipelines. SCC of steels in ethanol was not anticipated until failures occurred in the 20000’s. The list is long and depressing. What the decision makers need to know is not only what our research results mean today, but what it will mean tomorrow. This requires not only analytical, but also synthesis skills (and a lot of courage). We need to know how to systematically synthesize diverse sources of knowledge in a predictive framework. Since both knowledge of mechanisms and data are uncertain, the predictive framework is necessarily probabilistic.

This talk will focus on methods to develop probabilistic assessments of failure of materials using a hierarchical approach of fundamental, process level, and abstracted models. The talk will first discuss long-term localized corrosion prediction of stainless steels and Ni-base alloys using a combination of experimental and modeling approaches for a diverse set of applications. Next, the talk will extend the approach to stress corrosion cracking of these alloys in oil and gas applications. The lessons learned from applying this approach will be presented. Finally, an approach to synthesize knowledge using Bayesian networks will be presented.

Bio

Dr. Sridhar is a Vice President of DNV and the Director of Materials Program in DNV Strategic Research & Innovation. He also serves as an Adjunct Professor at The Ohio State University, Materials Science & Engineering Department. His main technical interests involve risk management of corrodible systems, advanced materials applications in diverse industries, and carbon dioxide utilization technologies. Prior to joining DNV in 2007, he worked at Southwest Research Institute, San Antonio, Texas for 18 years as a Program Director, where he was involved in developing life-prediction methods and sensors for nuclear waste disposal, pipelines, DoD systems, and NASA. From 1981 through 1989, he worked at Haynes International, Kokomo, Indiana as a group leader, where he was involved in the development of advanced Ni and Co-base alloys. He obtained a Ph.D. from University of Notre Dame in 1980. 

Dr. Sridhar has published over 180 papers and book chapters. He is a Fellow of NACE International and the recipient of NACE Technical Achievement award. He is also the recipient of the Vaaler award, the Guy Bengough award, the R&D 100 award, the Edward C. Greco International Corrosion Council Award and the Tech Columbus award. He is currently serving as an Associate Editor of Corrosion Journal and Corrosion Engineering Science & Technology journal. He has received a number of patents in the areas of advanced alloys, sensors, fuel cells, and CO2 conversion process.