ID 431

Is “DNS” Simulation of Spray Atomization Ready for Prime Time?

 

Abstract:

Liquid fuel atomization plays a critical role in the performance of combustion devices. Quantitative prediction of liquid atomization has remained a challenge, however, due to the complexity of the multi-scale processes involved and the significant limitations of experimental techniques in quantifying its characteristics. In the relatively recent past, first principles, high resolution computational simulation has arisen as a viable alternative for the analysis of this phenomenon. This type of simulation, often referred to as Direct Numerical Simulation (DNS), has been enabled by the development of advanced numerical methods and the availability of High Performance Computing (HPC). This lecture will describe our multiyear efforts to advance and leverage this type of simulation to atomization problems relevant to aerospace propulsion applications. The lecture will consist of three segments: In the first segment the background and development of the computational methodology will be described together with its detailed validation and demonstration across a variety of atomization problems ranging from canonical problems to realistic injectors. In the second segment, the power of the approach to yield detailed understanding of the physical phenomena involved, will be demonstrated in the context of analysis of the liquid jet in gaseous cross-flow atomization problem. A particular focus here will be the alteration of the physical processes driving atomization as the flow controlling parameters are varied. Finally, in the third segment, future directions will be discussed. In this, our current efforts to address the few remaining challenges in simulating traditional aerospace atomization problems will be elaborated upon and so will new frontiers of research in this area such as compressible and transcritical atomization.