Artificial Neural Network-DeltaEC Hybrid Model for Thermoacoustic Systems: New Synergistic Approach

Abstract

Despite the wide-spread use of DeltaEC model in thermoacoustics community, intrinsic nonlinear phenomena are still hindering its applicability to real practical situations due to assumptions and limitations. In the present study, artificial neural network (ANN) as an intelligent technique is hybridized with conventional DeltaEC model to provide a new synergistic approach called (ANN-DeltaEC) hybrid model for thermoacoustic research field. The aim of this paper is to improve the prediction accuracy of single DeltaEC model by integrating it with distributed and synergistic neural networks. One application for this new approach has been conducted on one standing wave thermoacoustic heat engine based on published literature work to predict the acoustic wave parameters, namely, oscillating frequency and acoustic pressure amplitude under given design considerations of stack geometry and resonator length. The results from hybrid synergistic model had been proven to be desirable in its accuracy compared to experimental work and better than the results of DeltaEC model itself. The present work had shown the capability of the new synergistic approach in accurately predicting the outputs for any new given inputs within given range. Further applied research will be devoted in order to identify complex mappings between thermoacoustic system parameters and their corresponding responses.