基于IBA-LSSVM强迫选择模型的民用飞机着陆滑行仪表灯常亮故障关联检测

Correlative Detection of the Constant Illumination of the Taxi Indicator Light of the Civil Aircraft Landing by Introducing IBA-LSSVM Forced Selection Model

摘要:

民用飞机着陆滑行仪表灯是故障检测系统的重要组成部分，因民用飞机在地面滑行过程中，电气接口易受到电磁干扰，引起着陆滑行仪表灯常亮故障，在飞行过程中增加了安全隐患。为此，提出引入IBA-LSSVM强迫选择模型的民用飞机着陆滑行仪表灯常亮故障关联检测方法。分析仪表灯常亮故障产生的原因，利用Python编程数据采集飞机起飞、近进、降落、地面滑行时的照明信号及电气交联数据，通过控制逻辑，建立数据关联检测逻辑关系数据库。利用自适应多普勒补偿方法改进蝙蝠算法（IBA），检测着陆滑行灯常亮故障特征，在强迫选择部分，构建IBA-LSSVM模型，将分布式博弈和线性输出调节理论相结合，抵消外部干扰，完成着陆滑行灯常亮故障检测。Matlab仿真测试结果表明：所提方法的故障识别率均在95%以上，故障识别准确率可以达到97.3%，故障识别时间低于2ms，可有效识别故障数据，降低飞行安全隐患。

Abstract:

The landing taxiing instrument light of civil aircraft is an important part of the fault detection system. Because the electrical interface of civil aircraft is vulnerable to electromagnetic interference during the ground taxiing process, the landing taxiing instrument light is always on, which increases the potential safety hazard during flight. For this reason, a correlation detection method for the constant light fault of the taxiing instrument lights of civil aircraft landing is proposed by introducing the IBA-LSSVM forced selection model. Analyze the cause of the instrument light constant lighting fault, use Python programming data to collect the lighting signals and electrical cross-linking data during aircraft takeoff, approach, landing and ground taxiing, and establish a data association detection logic relationship database through control logic. The bat algorithm (IBA) is improved by using adaptive Doppler compensation method to detect the fault characteristics of the landing taxiing light always on. In the forced selection part, the IBA-LSSVM model is constructed. The distributed game theory and linear output regulation theory are combined to eliminate external interference, and the landing taxiing light always on fault detection is completed. Matlab simulation test results show that the fault identification rate of the proposed method is above 95%, the fault identification accuracy can reach 97.3%, and the fault identification time is less than 2ms, which can effectively identify fault data and reduce flight safety hazards.