Biomass prediction method of nuclear power cold source disaster based on deep learning

Given the insufficient early warning capacity of nuclear cold source biological disasters, this paper explores prediction methods for biomass caused by disasters based on deep learning. This also uses correlation analysis method to determine main environmental factors. The adaptive particle swarm optimization was used optimize depth confidence network model Gaussian continuous constrained Boltzmann machine (APSO-CRBM-DBN). To train model, marine factors were as input and after a period time output training. Optimal results obtained, thus, disaster established. provides an accurate scientific basis in power plants has important practical significance solving problem blockage at inlet water plants.