Model Reduction in Large Chemical Systems- an Alternative Method Applying Discrete Approximation

Dynamic model reduction techniques based on the decomposition of the stoichiometric matrix to find the chemical invariant, break down if axial diffusion is present in a tubular reactor.

Straightforward discretization of the partial differential operator does indeed show that the resulting discrete dynamic model cannot generally be partioned to obtain the reaction variant vector and the reaction invariant (asymptotic) vector. However, the paper demonstrate that, if the diffusional tubular reactor is discretely and approximatively represented by tanks-in-series, then matrix approaches to successfully find the chemical variant and invariant vectors of the resulting chemical process model is possible.