Abstract

Nuclear data includes all of the nuclear interaction information required for computational modeling. These data are of fundamental importance in nuclear science and engineering. How trustworthy is our current knowledge of nuclear interaction probabilities? For materials used in light water reactor technology, some level of predictive capability has been reached by modeling and simulation codes using the current nuclear data base. However, there are large gaps in our knowledge of nuclear data for materials and characteristic particle energies used in areas such as next generation nuclear reactor designs, criticality safety, homeland security and nonproliferation, medical applications, space reactors, and structural materials in fusion designs. To understand the root of the problem of nuclear data uncertainty, we return to the birthplace of nuclear data where quantum theory intersects with large scattering experiments. In this retrospective, we uncover the sources of uncertainty in fundamental nuclear data and consider methods to eliminate or quantify them. With an improved understanding of the quantum nature of nuclear interaction probabilities and the human element in their evaluation, modern machine learning algorithms emerge as the ideal method to collect the vast information from merging quantum mechanics and experimental results to aid the eort to develop and understand nuclear data. A more thorough understanding, aided by machine learning algorithms, will lead to more accurate nuclear data and a quantiable understanding of the uncertainty in our knowledge. A more comprehensive understanding of nuclear data uncertainty will also help dene the consequences of this uncertainty as they aect the projects we implement.

 

Biography

Dr. Vladimir Sobes is currently a Research and Development Sta Member in the Nuclear Data and Criticality Safety group at Oak Ridge National Labora-tory (ORNL). He received a Bachelor of Science degree in Nuclear Science and Engineering from the Massachusetts Institute of Technology (MIT) in 2011. After receiving his Doctoral degree from the same department two and half years later, Sobes, began working at ORNL.

As a nuclear data specialist, Dr. Sobes focuses his research on evaluation of nuclear data and propagation of nuclear data uncertainty for a wide range of end-use applications. The motivating factor for his research is to optimize the benets of nuclear energy by increasing the safety of nuclear installations by providing better nuclear data and the understanding of how the uncertainty in nuclear data drives our ability to perform predictive modeling and simulation of nuclear systems. As a member of the nuclear data team at ORNL, Dr. Sobes shares responsibilities for the measurement, evaluation and validation of nuclear data for the international nuclear community, methods and code development for data applications, nuclear criticality safety analysis and reactor physics methods development. He is also the ORNL oversight technical lead for the design and construction of the University of Tennessee, Knoxville (UTK) sub-critical experiment facility.

Dr. Sobes has demonstrated his exibility and breadth of knowledge by taking on research outside the eld of nuclear data. He has contributed to the United States (US) Department of Energy eort of evaluating the feasibility of direct disposal of spent nuclear fuel in a geologic repository. Early on in his career, Dr. Sobes conducted reactor analysis research and in recent years he took on a research project to develop new statistical inference methods for criticality calculations for the US Nuclear Regulatory Commission.

Dr. Sobes has over 30 peer-reviewed publications in international journals, technical reports and con-ference proceedings. He is an active member of the American Nuclear Society serving on the Education Committee and the Technical Program Committee for the Nuclear Criticality Safety Division. He is also an active participant in two OECD/NEA1 Working Party on International Nuclear Data Evaluation Cooperation (WPEC) expert subgroups and group co-ordinator and founding member of a new subgroup on "Investigation of Covariance Data in General Purpose Nuclear Data Libraries."