Research on soft measurement model of flow in bends of primary circuits of the nuclear power plant

When measuring the coolant flow in a nuclear power plant using the elbow flowmeter, the complex fluid-heat coupling environment at the measurement location and other factors will affect the accuracy of the flow measurement, and the uncertainty of the influencing factors on the flow measurement needs to be considered to improve the measurement accuracy. To address this problem, this paper adopts the finite element simulation method to simulate and analyze the flow-heat field of the bend section of the primary circuit of a nuclear power plant and optimizes the Optimal Cross-section selection of the pipeline for flow measurement. Based on the pressure values measured using the traditional method, temperature information is added, and a BP neural network bend pipe flow soft measurement model based on the whale optimization algorithm is established to quantify the effects of temperature and pressure on flow measurement. The experimental results show that compared with the traditional engineering empirical method, the average absolute percentage error measured by the soft measurement method is reduced from 2.57 % to 0.21 %, which realizes the accurate measurement of the coolant flow rate of the elbow pipe. •CFD simulation for modeling curved section of primary circuit in nuclear power plant.•Uncertainty analysis of the elbow flowmeter and determination of the optimal section.•Propose the application of neural networks to conventional bend flow measurement in order to accurate measurement of coolant flow rate.