基于优化支持向量机的飞机重着陆智能诊断

Airplane’s Hard Landing Diagnosis Based on Optimized Support Vector Machine

为提高飞机重着陆判断的准确性,研究了将最小二乘支持向量机应用于飞机重着陆诊断的方法;首先,根据飞机着陆阶段运动方程确定5类关键的重着陆诊断指标,将传统的单一指标诊断扩展到多指标诊断;接着,对支持向量机的分类算法进行扩展,实现了支持向量机的多类分类,建立了飞机重着陆诊断模型;然后,分别利用遗传算法和粒子群算法优化了模型参数,并对优化结果进行了分析比较;最后,利用飞行品质监控数据库中的样本数据对某航空公司B737型飞机进行了重着陆诊断实验,结果表明:支持向量机模型具有较高的诊断精度,适用于飞机重着陆诊断。

In order to enhance the diagnosis accuracy, least square support vector machine （LS--SVM） was used to diagnose airplane＇ s hard landing. Firstly, according to airplane＇ s motion equation of landing phase, five major diagnosis indexes were determined and extended airplane~ s hard landing diagnosis from one index to several. Next, classification algorithm of SVM was expanded, multi--classes classifica- tion was realized and airplane＇ s hard landing diagnosis model was established. Then, genetic algorithm and particle swarm optimization al- gorithm were used to optimize the model parameters of SVM. The optimization result was analyzed and compared. Last, using the data of flight quality monitoring database, B737 airplanes of some airline were carried on the hard landing diagnosis experiment. The result shows that SVM model produces accurate diagnosis results and is suitable for airplane＇ s hard landing diagnosis.