A Study on Reconstructing In-Cylinder Combustion Images Based on Local Images

The current leading experimental platform for engine visualization research is the optical engine, which features transparent window components classified into two types: partially visible windows and fully visible windows. Due to structural limitations, fully visible windows cannot be employed under certain complex or extreme operating conditions, leading to the acquisition of only local in-cylinder combustion images and resulting in information loss. This study introduces a method for reconstructing in-cylinder combustion images from local images using deep learning techniques. The experiments were conducted using an optical engine specifically designed for spark-ignition combustion modes, capturing in-cylinder flame images under various conditions with high-speed cameras. The primary focus was on reconstructing the flame edge, with in-cylinder combustion images categorized into three types: images where the flame edge is fully within the partially visible window, partly within the partially visible window, and completely outside the partially visible window. For images with the flame edge partly within the partially visible window, a flame edge completion model was developed using the Generative Adversarial Network (GAN), effectively completing incomplete flame edges and providing accurate information beyond the partially visible window. For images where the flame edge is entirely outside the partially visible window, lacking spatial information, a flame prediction model based on Convolutional Long Short-Term Memory (ConvLSTM) networks was constructed to predict the combustion process over time, with a quantitative standard established to assess prediction accuracy. By integrating the flame edge completion and prediction models, this study successfully reconstructed the complete combustion process inside the cylinder from local images. The reconstructed images were analyzed to extract the flame propagation speed, which was validated through correlation. This method offers a highly accurate reconstruction of partially visualized combustion processes under various conditions, providing more effective data for combustion analysis.