Data compression and recovery method for transient operation parameters of nuclear power plants

Abstract

Currently, because of many types of transient operation parameters of nuclear power plants, some problems are encountered in network transmission, data storage, and algorithm training processes, such as slow transmission speed, large storage space, slow algorithm training speed, and complex structure. In this study, a method for data compression and restoration of transient operation parameters of nuclear power plants is established by improving the characteristic engineering algorithm. The method uses principal component analysis to extract the feature vector of transient operation parameters, and combined with the Gaussian process regression method, the reduced dimension transient operation data can be restored with high precision. The results show that, after the transient operation data are compressed and restored, the maximum error between the restored and actual values does not exceed 0.000 002. The compressed data characteristics can also provide a basis for the judgment of fault diagnosis of nuclear power plants and further improve the safety and reliability of nuclear power plants.