飞机硬着陆多因素特性判断仿真研究

Simulation study on multi-factor characteristic judgment of aircraft hard landing

针对国内航空公司飞机硬着陆识别精度不够理想,存在较高漏判率和误判率的问题,提出一种改进的AdaBoostSVM算法建立硬着陆判断模型。该模型在充分考虑产生硬着陆相关因素的基础上,增加主起落架缓冲支柱套筒压缩行程特征量作为飞机硬着陆判断因素,并调整AdaBoostSVM算法弱分类器的评价系数,优化对硬着陆识别能力强的弱分类器的权重。以国内某机队飞机为例,采用实际样本数据进行仿真验证,对比支持向量机和传统AdaBoostSVM算法分析飞机硬着陆判断的准确性和可靠性。仿真结果表明,改进的AdaBoostSVM算法在具有充分样本的情况下能够显著提高硬着陆事件的识别精度,有效降低飞机硬着陆的漏判率和误判率,减少飞机飞行的安全隐患和节约航空公司的维修成本,具有较强的工程实际意义。

Aiming at the low accuracy of aircraft hard landing recognition rate of domestic airlines, and the high false rate and miss rate of hard landing recognition, an improved AdaBoostSVM algorithm hard landing recognition model is proposed. On the basis of fully considering the relevant factors of hard landing, in order to improve the recognition rate of aircraft hard landing, the compressive stroke characteristic of main landing gear buffer pillar is used as the hard landing diagnostic index, adjusting the evaluation coefficients of AdaBoostSVM algorithm weak classifier and increasing the weight of weak classifier with high ability for recognizing hard landing. At the same time, the actual sample data of airlines fleet is simulated and verified, and the accuracy and reliability of hard landing recognition of three algorithms are compared. Simulation results show that in the case of sufficient sample, the improved AdaBoostSVM algorithm could significantly improve the accuracy of hard landing events recognition, reducing the false rate and miss rate of aircraft hard landing as well as security risks of aircraft flight and cutting down airlines＇ maintenance cost, having strong practical engineering value.