Abstract
Particle Swarm Optimization with shrinkage factor (XARPSO) has the capacity of reaching good global optimization and fast convergence speed. It aims to avoid Particle Swarm Optimization getting global optimal solution prematurely. With shrinkage factor, Particle Swarm Optimization can not only particle’s vitality but obtain global optimal solution quickly. With this algorithm we can make the parameter selection of A Library For Support Vector Machines (LIBSVM) on wind power turbine bearing fault diagnosis model. Thus we can transform the LIBSVM model parameter selection into optimization with this approach. It can conquers the shortage that PSO falls easily into local optimum in the optimization and can also make the optimization performance improved. Using UCI database to do a series of classification-experiments. Comparing with CV — LIBSVM and GA LIBSVM, we make the classification more accurate. Then we use the XARPSO —LIBSVM model in the wind turbine bearing diagnosis and get a good outcome.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call