Prediction model of temperature field in dual-mode combustors based on wall pressure

Abstract

Effective acquisition of the complex flow fields in dual-mode combustors is valuable in improving dual-mode combustor performance. Introducing deep learning methods for fast flow field prediction is an effective way to provide high-quality temperature fields in a dual-mode combustor. A data-driven model to obtain temperature fields based on wall pressure of a dual-mode combustor is proposed in this paper. The temperature fields in dual-mode combustors depend on the combustor geometry, incoming flow condition, and equivalent ratio. The prediction model takes convolutional neural network as the main framework and includes multiple branches. The trained temperature field prediction model can effectively output the temperature distribution of the combustor with the input of wall pressure. The temperature field prediction accuracy has been discussed in depth. From the analysis results, the effectiveness and accuracy of the approach have been demonstrated under the current dataset.