Temperature field and curing degree prediction of large composite blades based on coupled finite element analysis and machine learning

Abstract Since large composite blades are variable curvature and thickness components with large dimensions, the temperature field analysis will produce a large inhomogeneous temperature field, which requires a lot of time. In this paper, a new method combining finite element analysis and machine learning is proposed. By constructing a numerical model of blade temperature field curing and using zoned heating to optimize the temperature and curing gradient in the tongue and groove region, the maximum temperature gradient is reduced by 74.18% and the maximum curing degree gradient is reduced by 21.987% compared with the conventional temperature profile curing. A long short‐term memory(LSTM) neural network was used to predict the temperature field and curing gradient variations, and the Grey Wolf algorithm was used to optimize the parameters with high model prediction accuracy. The method is instructive for online monitoring and control of the curing process in customized hot press tanks. Highlights The zoned heating method improves the temperature‐field balance. A tandem LSTM neural network is constructed as an agent model. Enabling the temperature field to be connected to the degree of cure. Optimizing tandem LSTM model parameters using the grey wolf algorithm.