回归型支持向量机的机翼结冰冰型预测

Airfoil Ice Shape Predictions Based on Support Vector Regression

飞机结冰严重威胁着飞行安全。针对机翼结冰冰型预测时不确定性因素较多、难以准确预测的问题,提出了一种基于回归型支持向量机的机翼结冰冰型预测方法。在建立冰型模型的基础上,利用回归型支持向量机(Support Vector Regression,SVR)获得冰型多项式系数,从而预测出相关飞行条件和大气条件下的冰型。仿真结果表明,该方法具有较好的预测能力,可以及时提供可靠的结冰信息,为保证结冰条件下的飞行安全提供了保障。

Aircraft icing has been one of the fatal causes to aircraft accidents.Considering the uncertainty of factors and difficulty in predicting accurately in airfoil ice shape predictions,a new predicting method based on support vector regression is proposed.Coefficients of polynomial can be obtained by using support vector regression based on the establishment of ice shape model.As a result,ice shape predictions under certain circumstance can be accomplished.The simulation results show that the method has good ability of prediction,and reliable icing information can be presented in time,which will benefit flight safety greatly under icing conditions.