Towards 4D Emission Tomography of Reacting Waves

We present a method for time-resolved, volumetric chemiluminescence tomography developed to characterize detonation waves. The intricate 3D structure of self-sustaining gaseous detonations presents significant challenges for numerical prediction, complicating the design of detonation-based engines. High-fidelity experimental measurements that resolve the internal density and heat release structures of detonation waves in curved domains are essential for understanding their dynamics and validating computational models. Chemiluminescence tomography offers a promising solution, but ultra-high-speed imaging constraints limit the number of projections, resulting in an underdetermined reconstruction problem. To address this, we developed an algorithm based on the neural-implicit reconstruction technique (NIRT) that can leverage temporal information and incorporates a physics-based observation operator to account for depth-of-field and refraction effects. These innovations enhance the accuracy and applicability of chemiluminescence tomography for studying complex reacting flows, such as those in detonation waves. Synthetic demonstrations are reported for a turbulent premixed methane/air flame.