Liquid ammonia is an ideal zero-carbon fuel for internal combustion engines. High-pressure injection is a key technology in organizing ammonia combustion. Characteristics of high-pressure liquid ammonia injection is lack of research. Spray behaviors are likely to change when a high-pressure diesel injector uses liquid ammonia as its fuel. This study uses high-speed imaging with a DBI method to investigate the liquid penetration, width, and spray tip velocity of high-pressure liquid ammonia injection up to 100 MPa. Non-flash and flash boiling conditions were included in the experimental conditions. Simulation was also used to evaluate the results. In non-flash boiling conditions, the Hiroyasu model provided better accuracy than the Siebers model. In flash boiling conditions, a phenomenon was found that liquid penetration and spray tip velocity were strongly suppressed in the initial stage of the injection process, this being the “spray resistance phenomenon”. The “spray resistance phenomenon” was observed when ambient pressure was below 0.7 MPa during 0–0.05 ms ASOI and was highly related to the superheated degree. The shape of near-nozzle sprays abruptly changed at 0.05 ms ASOI, indicating that strong cavitation inside the nozzle caused by needle lift effects is the key reason for the “spray resistance phenomenon”.

Ammonia is a highly promising zero-carbon fuel for engines. However, it exhibits high ignition energy, slow flame propagation, and severe pollutant emissions, so it is usually burned in combination with highly reactive fuels such as hydrogen. An accurate understanding and modeling of ammonia–hydrogen combustion is of fundamental and practical significance to its application. Deep Neural Networks (DNNs) demonstrate significant potential in autonomously learning the interactions between high-dimensional inputs. This study proposed a deep neural network-based method for optimizing chemical reaction mechanism parameters, producing an optimized mechanism file as the final output. The novelty lies in two aspects: first, it systematically compares three DNN structures (Multi-layer perceptron (MLP), Convolutional Neural Network, and Residual Regression Neural Network (ResNet)) with other machine learning models (generalized linear regression (GLR), support vector machine (SVM), random forest (RF)) to identify the most effective structure for mapping combustion-related variables; second, it develops a ResNet-based surrogate model for ammonia–hydrogen mechanism optimization. For the test set (20% of the total dataset), the ResNet outperformed all other ML models and empirical correlations, achieving a coefficient of determination (R²) of 0.9923 and root mean square error (RMSE) of 135. The surrogate model uses the trained ResNet to optimize mechanism parameters based on a Stagni mechanism by mapping the initial conditions to experimental IDT. The results show that the optimized mechanism improves the prediction accuracy on laminar flame speed (LFS) by approximately 36.6% compared to the original mechanism. This method, while initially applied to the optimization of an ammonia–hydrogen combustion mechanism, can potentially be adapted to optimize mechanisms for other fuels.