Research on Optimization Technology of Sensing Data for Nuclear Reactor Thermal Hydraulic Experiments

Abstract

The sensor signals collected by the nuclear reactor thermal hydraulic experimental system are mixed with complex noise information. The uncertainty of sensor data directly affects the mining and application effect of artificial intelligence on experimental data, which in turn affects the analysis and evaluation of the system's operating status. Traditional denoising methods have poor adaptability to different noisy data, are highly dependent on researchers, and take a long time. This paper proposes a data optimization technology based on a hybrid adaptive real-time (HART) denoising model, which can realize data distribution self-adaptation and algorithm hyperparameter self-optimization according to the characteristics of noisy source data. Using the joint denoising algorithm with the Variational Mode Decomposition (VMD) algorithm as the core to realize the partition denoising processing of the source data. In addition, the practical application of the source data of thermal hydraulic experiments in the nuclear field and the verification of simulation and noise addition experiments have been completed. The results show that the denoising model proposed in this article has good adaptability, self-optimization, real-time processing and other characteristics for different sensing signals. It can effectively suppress data uncertainty problems, realize the optimization of experimental data, and promote the application of artificial intelligence technology in nuclear technology. Engineering applications in the field of thermal hydraulic experimental systems. At the same time, related technologies can be further promoted and applied to the optimization of sensor data in nuclear power plants (NPP).

Keywords: Artificial Intelligence, Thermal hydraulic sensing signal, Self-adaptation, Data optimization, VMD