Recycle streams within a PSR cluster can be solved simultaneously with no extra solver iteration. However, recycle connections between clusters require an iterative convergence technique. In a simple two-cluster network in which a PSR initializes an IC HCCI Engine and part of the final product in the IC HCCI Engine recycles back to the PSR (such as in exhaust gas recirculation), the final product of the IC HCCI Engine between consecutive iterations must be monitored and compared to previous iteration results to determine whether we have achieved convergence in the solution of the IC HCCI Engine.

In a closed circular loop like this, either stream can be "torn" to become a tear stream. In more complex networks, the tear streams are selected to:

Then, the clusters connected by the tear streams are run iteratively with adjustments to the estimated properties of the torn stream until the properties of every reactor connected to the tear streams between consecutive iterations are within user-defined tolerances. There is a Tear Stream Control node in the project tree to control how this convergence is approached and assessed.