Phase stability analysis using a new objective function and a global optimization method.

In the last decade there has been considerable interest in developing (both theoretically and algorithmically) new methods for the solution of a numbe r of problems in chemical engineering process design, synthesis, optimizati on and control. The phase stability analysis problem is among the most chal lenging, because its formulation, either as a tangent plane distance functi on minimization or as a Gibbs energy minimization, requires robust and reli able numerical techniques to determine the global solution. In our paper we use a modified tangent plane distance function and we advoc ate a stochastic sampling and clustering method to optimize it. The efficie ncy and robustness of the method are demonstrated on the example of the hyd rogen sulfide-methane system, which is modeled by a Redlich-Kwong-Soave cub ic equation of state. The obtained results demonstrate the high level of re liability of the method in comparison with other stochastic techniques. The procedure is user-friendly and it can well be tuned by the user. The new m ethod significantly decreases the computational cost regarding the number o f function-evaluations, and CPU time requirements which compare very favour ably with other stochastic optimization methods, like plain Newton methods.