Application of a hybrid PSO-GA optimization algorithm in determining pyrolysis kinetics of biomass

A hybrid optimization algorithm, combining both Particle Swarm Optimization (PSO) and Genetic Algorithm (GA), is proposed to gain the favorable features of each individual algorithm when determining the pyrolysis kinetics of biomass. High convergence efficiency and the capability of avoiding being trapped in local optimal solution are primarily associated with PSO and GA, respectively. Gene operations in GA, including selection, crossover and mutation, are partially incorporated into PSO to increase the population diversity. Pyrolysis of beech wood was experimentally studied at three heating rates, and a numerical solver was established to simulate the pyrolysis details. In order to demonstrate the improved performance of PSO-GA, two pyrolysis models with given reaction schemes and kinetic parameters were adopted to create the acritical thermogravimetric analysis (TGA) curves. Then the kinetics was estimated using PSO-GA and individual GA and PSO. Subsequently, the experimental data were analyzed with the same manner. The results show that PSO-GA has the highest possibility of obtaining desired outcomes followed by PSO and then GA. With fixed population size, PSOGA converges to a lower fitness function value, corresponding to higher accuracy. The attained kinetics of wood falls into the reported ranges in the literature. In some scenarios, the optimized results of hemicellulose and lignin contradict with the existing conclusions even though the global curves match the experimental measurements well. This implies the general concept of the pyrolysis process should also be given adequate consideration to avoid potential compensation effect when encountering complex issues.