The invention discloses a rapid prediction method for a deposition state in a
nuclear power plant condenser
pipe, and the method comprises the following steps: 1, carrying out random sampling on deposition characteristic parameters in the
pipe in a condenser cooling
water pipe bundle enveloping region range, including a
foreign matter deposition position and an equivalent blocking
radius, and obtaining a plurality of groups of deposition
simulation working conditions in the condenser
pipe; 2, three-dimensional steady-state
simulation is carried out based on each group of working conditions; 3, performing post-
processing to obtain
temperature monitoring point data at the
secondary side M of the condenser under each group of working conditions; 4, building a
back propagation neural network; 5, taking the post-processed data of the M
temperature monitoring points as input, taking each group of in-pipe deposition
simulation working condition characteristic parameters as target output, and training a neural network; 6, inputting new data measured by
temperature monitoring points at all positions of the real condenser in real time into the trained neural network, and quickly predicting the deposition characteristic parameters in the current condenser pipe; the method can rapidly predict the deposition state in the
nuclear power plant condenser pipe in the operation state, provides reference for dredging and cleaning the condenser blockage, greatly saves the operation and maintenance cost, and is of great significance to the safety operation monitoring of a
nuclear power plant.