Reactor Network

You can reduce the computational cost of solving complex chemistry for each cell in the computational domain of a steady combustor by simulating with an economical combustion model—such as a flamelet model, or the Eddy Break-Up (EBU) model—then clustering cells into a specified number of reactors and solving detailed chemistry on this network of reactors.

The process of clustering CFD cells into reactors consists of three operations. In the first, the clustering algorithm is used to group cells with similar states—temperature and equivalence ratio by default. These cells are typically non-contiguous, and the second operation is to split non-contiguous clusters into separate clusters. The final step is to group clusters with the smallest number of cells to their closest neighbor in composition space until approximately the specified number of reactors are attained.