Volume 05: Proceedings of 11th International Conference on Applied Energy, Part 4, Sweden, 2019

Improved Predictions of Onset Temperature In Twin Thermoacoustic Heat Engine by Neural Network Based Calibrated Thermoacoustic Model Anas A. Rahman, Xiaoqing Zhang

Abstract

Thermoacoustic heat engines are promising devices converting thermal power into acoustic power with the distinct merits of simplicity, and possibility for utilizing low-grade heat sources. As well known, design parameters are usually determined based on weakly non-linear thermoacoustics theory which of course produces significant deviations due to non-capturing of nonlinear phenomena. In order to improve inaccurate predictions of onset temperature obtained by DeltaEC linear thermoacoustic model, artificial neural network is first proposed to be hybridized with DeltaEC model to provide a new synergistic approach. This synergistic approach was applied to a twin thermoacoustic heat engine for improving the computational efficiency of DeltaEC model itself through considering some nonlinearities existing in the whole thermoacoustic system. The onset temperature was predicted as the responses to both resonator length and charging pressure and the obtained results had been proven to be desirable in their accuracy compared to experimental ones and better than literature DeltaEC results under same given conditions.

Keywords thermoacoustics, artificial neural networkdeltaec hybrid model, twin thermoacoustic heat engine, onset temperature, resonator length, charging pressure

Copyright ©
Energy Proceedings