Modeling-and-simulation-based intelligent concrete vibration system for nuclear power plants.

Nuclear power plants have extremely high requirements for the compactness of concrete after forming, due to its unique operating environment. As a key link in concrete construction, vibration has a crucial impact on the forming quality of concrete. However, in practice, vibration mainly relies on empirical judgment and lacks objective quantitative standards, which may lead to uneven vibration, such as vibration leakage, undervibration and overvibration, which seriously affects the compactness of concrete. Therefore, this paper proposes an intelligent vibration simulation method for concrete in nuclear power plants. This method provides professionals with visual analysis tools by simulating the distribution of concrete energy during the vibration process, so that they can formulate scientific vibration schemes and optimize the vibration process based on the simulation results, so as to guide the on-site construction personnel to implement standardized vibration operations and effectively improve the vibration compactness of concrete. This paper proposed an optimized model as the simulation model, which combines the existing energy transfer model with the neural network (NN). Direct3D for simulation visualization. Compute unified device architecture (CUDA) is used for parallel computing to improve the simulation performance. The NN model is trained, verified and tested, and the relative root mean square error (RRMSE) is 5.8%, 3.9% and 4.9%, respectively, and the final model can generate reliable and accurate prediction results. In this paper, the energy transfer model is optimized based on the NN model, and a data acquisition platform and energy distribution visualization simulation system of concrete vibration process are constructed. [ABSTRACT FROM AUTHOR]