Dynamic simulation of a model is a very intensive process in computer time. Wastewater treatment plant models are a class of model particularly long to solve because of the great variety of dynamics in the model. A method that could be used to reduce significantly the calculus time is to consider very fast state variables as being at steady state at each integration step. However, the performance of the algorithms depends a lot on external knowledge about the model like initial estimates of the state variables. Since WWTP models are well understood, it is possible to include knowledge such as initial estimates or boundary limits of state variables to help the convergence of the algorithms.

The first step of my project is to develop a strategy to include a maximum of available knowledge in the existing algorithms. The second step will be to develop a sorting method to extract the state equations showing very fast dynamics. With this project, it is hoped that simulation time can be reduced significantly.