Nuclear big data processing and management consists of data pre-processing for the purpose of formalizing nuclear big data, and data post-processing that optimizes learning data for the purpose of using nuclear big data. The data pre-processing framework consists of data filtering, data type conversion, data purification, and data labeling. The data post-processing framework consists of data integration, data transformation, and data reduction. The effectiveness of the entire framework plan should be verified through the construction and testing of the nuclear power plant big data processing system. However, there is a difficulty in securing nuclear power plant big data, and due to realistic research restrictions, it is only necessary to test the effectiveness of the nuclear power plant big data processing system with limited functions and scope.
Therefore, this study first selects the scope and function of the pre-processing implementation to be piloted, secures the nuclear power plant big data pre-processing technology, analyzes the simulation results at a level that can be realistically input, and formalizes it according to the characteristics. The purpose is to implement the nuclear power plant big data pre-processing technology by determining the categorization and labeling method and determining the level of automation.
In addition, by selecting the scope and function of post-processing implementation to build a pilot, securing nuclear power plant big data post-processing technology, and determining the level of automation by determining the data correlation analysis and dimension reduction method according to the nature of nuclear power plant big data. Its purpose is to implement processing technology. And, based on the nuclear power plant big data quality management items, the pre-processing and post-processing results are checked to confirm the possibility of diagnosing transient situations and producing response support data.