Your opinion on a class of ODE problems

[Thomas Ritschel]

It's quite likely that one runs into conservation law violations when trying to (numerically) simulate complex chained reactions with simple ODE integrators (like the Newton model). Especially for long computations or when it comes to oscillating reactions one will notice inconsistencies in total mass (but also charge, energy, etc.).

One possible (quick-and-dirty) solution would be some kind of intermediate normalization. But better use a predictor-corrector integrator (e.g. Adams-Bashforth-Moulton) instead.

Here are two further links regarding chemical kinetics simulations:
http://chemistry.illinoisstate.edu/stand...12-S06.pdf
http://www.springer.com/cda/content/docu...306-c1.pdf