Predicting simultaneously fields of soot temperature and volume fraction in laminar sooting flames from soot radiation measurements - a convolutional neural networks approach

An original convolutional neural network, i.e. U-net approach, has been designed to retrieve simultaneously local soot temperature and volume fraction fields from line-of-sight measurements of soot radiation fields. A five-stage U-net architecture is established and detailed. Based on a set of N2 diluted ethylene non-premixed flames, the minimum batch size requirement for U-net model training is discussed and the U-net model prediction ability is validated fbr the first time by fields provided by the modulated absorption emission (MAE) technique documenting the N-2 diluted flame. Additionally, the U-net model's flexibility and robustness to noise are also quantitatively studied by introducing 5% & 10% Gaussian random noises into training together with the testing data. Eventually, the U-net predictive results are directly contrasted with those of Bayesian optimized back propagation neural network (BPNN) in terms of testing score, prediction absolute error (AE), soot parameter field smoothness, and time cost. (C) 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

Automated summary

An original U-net convolutional neural network is proposed for retrieving local soot temperature and volume fraction fields from line-of-sight measurements. The U-net model's prediction ability and robustness to noise are validated and quantitatively studied. Comparative analysis with other algorithms demonstrates the effectiveness and accuracy of the U-net approach.