Application of Particle Swarm Optimization In Phase Equilibrium Calculations

Application of Particle Swarm Optimization in Phase Equilibrium Calculations

Sameer Punnapala¹, Ali Elkamel^{1, 2}, Hatem Zeineldin³ and Francisco M Vargas¹

1.      Department of Chemical Engineering, Petroleum Institute, Abu Dhabi, UAE

2.      Department of Chemical Engineering, University of Waterloo, Ontario, Canada

3.      Department of Electrical Engineering, Masdar Institute, Abu Dhabi, UAE

AUTHORS EMAIL ADDRESS: spunnapala@pi.ac.ae, aelkamel@pi.ac.ae, aelkamel@uwaterloo.ca, hzainaldin@masdar.ac.ae, fvargas@pi.ac.ae

CORRESPONDING AUTHOR FOOTNOTE: email- fvargas@pi.ac.ae, Phone- 00971-561266149

Phase equilibrium calculations play a vital role in the design, development, operation, optimization and control of chemical processes. The performance of Particle Swarm Optimization (PSO), a novel evolutionary stochastic global optimization method that is recently gaining importance among the chemical engineering community [1], is investigated on typical thermodynamic applications. Initially, PSO is tested on different benchmark problems involving several local minima. This study focuses on using PSO for parameter estimation in Vapor-Liquid Equilibrium (VLE) modeling for different systems using both Equations of State and Activity Coefficient models, accurate prediction of which are of prime importance in industrial operations. 

The objective functions in nonlinear parameter estimation problems are mostly non-convex and therefore have potentially multiple local optima. Conventional solution methods may not be reliable since they do not guarantee convergence to the global optimum for the estimated parameters [2]. This leads to a discrepancy in the parameter value for the same model reported by different authors in literature. Our results show that PSO is very promising, reliable and offers the best performance for global minimization problems compared to other evolutionary techniques like Genetic Algorithms or Simulated Annealing [3].

Keywords: Particle Swarm Optimization, Phase Equilibrium, Global Optimization, benchmark problems, Vapor-Liquid Equilibrium, non-convex, Evolutionary techniques.

References

 ADDIN EN.REFLIST 1.         A. Bonilla-Petriciolet, J.G. Segovia-Hernandez, Fluid Phase Equilib. 289 (2010) 110–121.

2.         C.Y. Gau et al., Fluid Phase Equilib. 168 (2000) 1–18.

3.         A. Bonilla-Petriciolet et al., Fluid Phase Equilib. 287 (2010) 111–125