Application of PSO-LSSVM in Bias Correction of Shipborne Anemometer Measurement

EasyChair Preprint no. 629

Abstract

Wind measurement from shipborne anemometer is susceptible to the airflow distortion due to ship hull and superstructure. The measurement bias needs to be minimized with regard to various meteorological and navigation applications. To address this problem, this study illustrates the feasibility to correct the measurement bias due to airflow distortion by applying Least Squares Support Vector Machine with Particle Swarm Optimization (PSO-LSSVM) method. The airflow field around hull and superstructure of an experimental ship is simulated by computational fluid dynamics (CFD) techniques. And then the nonlinear relationship between the airflow through conventional anemometer mounting sites on the main mast and the airflow through the reference point above bridge is implicitly obtained using the PSO-LSSVM regression. The dataset of relative wind observation taken during a sea trial is used to validate the effectiveness of this method. The results show that the established model efficiently eliminates most of the speed bias and reduces half of the direction bias of the mean relative wind, which indicates this method could be extended to estimate the undisturbed freestream on the open sea surface.

Keyphrases: CFD, LSSVM, PSO, Ship Airflow Field, wind measurement

@Booklet{EasyChair:629,
  author = {Tong Hu and Suiping Qi and Zhijin Qiu and Jing Zou and Dongming Wang},
  title = {Application of PSO-LSSVM in Bias Correction of Shipborne Anemometer Measurement},
  howpublished = {EasyChair Preprint no. 629},
  doi = {10.29007/nkdz},
  year = {EasyChair, 2018}}