Inverse estimation of wall heat flux by using particle swarm optimization algorithm with Gaussian mutation

An inverse analysis based on particle swarm optimization algorithm is developed to estimate the unknown wall heat flux from transient temperature measurements at the boundary of a plane wall. Two methods of modifications on the basic particle swarm optimization algorithm are presented. The first improvement is taken on the equation update for each dimension of the velocity of particles. The second procedure improves the inverse solutions by including a Gaussian mutation into the algorithm. The two modified algorithms are tested against the exact solutions and the solutions obtained from basic particle swarm optimization. A statistical analysis is performed to establish confidence bounds for the measurement errors involved in the determination of wall heat flux. Computational results show that both methods improve the quality of the inverse solutions.

► An inverse analysis based on particle swarm optimization algorithm is developed to estimate the unknown wall heat flux.
► Two modifications including MPSO and MPSOG on the basic particle swarm optimization algorithm are presented.
► The accuracy of the estimated results is examined using simulated temperatures with or without measurement errors.
► A statistical analysis is performed to establish confidence bounds for the measurement errors involved in the determination of wall heat flux.
► Computational results show that both methods improve the quality of the inverse solutions.