ANN approach in the creation of thin gas-sensitive films based on modified polyacrylonitrile

Abstract

Abstract In this study, the dependence gas sensitivity coefficient of silver-containing polyacrylonitrile (PAN) films on the technological parameters using artificial neural networks (ANN) is studied. Modelling is performed based on 38 datasets at different conditions. Initially, the experimental data on the fabrication of the films with gas-sensing properties have been collected, which constitutes the simulation space. At the next stage, the analysis of the measurement results has been carried out, self-tuning and training of the neural network have been done to provide the required output signal. Using the neural network model, the optimal technological parameters for creating efficient gas sensors for nitrogen dioxide have been established. Verification of the model was performed: the data convergence was 8-13%.