This paper focuses on a method to simulate dynamic models of chemical processes. Such models are usually sparse and stiff and the method in question splits the system into fast and slow modes before using an implicit integration method for the fast modes in the system and an explicit method for the slow modes. The splitting is achieved by analyzing the system and includes finding a suitable dimension for the fast modes and doing transformations on the system to obtain the best performance for the integrator when integrating the total system. Numerical experiments on a simple process model are done to illustrate the methods principles.