Database for Properties of Nuclear Reactor Materials Based on the Ontology and NoSQL Data Format

In the paper, the development of a new generation database for the properties of nuclear reactor materials is discussed. The goal of this project is to accumulate experimental and theoretical data on the mechanical, thermophysical, and service properties of different materials, their microstructure, and transformations due to radiation-induced swelling, corrosion, and other degradation processes for future use in data analysis and computer simulations that can predict unknown material properties and therefore accelerate the development of new materials for contemporary and planned nuclear reactors. Due to the large variety of data standards, types, and formats, as well as the various physical quantities and their dimensions used in different data sources, the use of typical relational databases requires a huge amount of handwork for restructuring and fitting the data into the common tables. Moreover, adding new data types may require a redesign of the database architecture. In this case, NoSQL databases where the information is stored as a set of JSON documents might greatly improve the flexibility of the data storage, although searching, extracting, and merging the data from different documents becomes more complicated. As a solution for this problem, we propose using an ontology for data annotation and integration that allows one to store the datasets in their original forms and convert them on-the-fly when processing a particular user query.