Thermal Error Modeling of Plenum by Improved ANFIS
To reduce the influence of the uneven temperature of the vent hole of the static plenum on the film stretching process,a thermal error model was established to analyze the temperature field of the vent hole of the static plenum under different input parameters.The optimal temperature measurement points are obtained by a comprehensive algorithm combining SOM-GRA(Self-organizing map-grey relational analysis)to ensure that the data input into the model is representative,and the number of temperature measurement points is reduced from 20 to 3.The thermal error model of the static plenum was established with the ANFIS(Adaptive-network-based fuzzy inference system)model,and the number parameters of the membership function in ANFIS were optimized by the RF(Random forest)algorithm,and the numerical simulation data verified by the experiment was used as the input training data.The prediction results show that compared with the original ANFIS model,the BP(Back propagation)model and the RBF(Radial basis function)model,the MAE(Mean absolute error)values have decreased by 22.43%,59.97%and 49.87%,respectively,and the optimized prediction model has higher accuracy.