航空器孔探人员脑力负荷预测  

Workload Prediction of Aircraft Borescope Inspection

作　　者：钱锋 贺强[2] QIAN Feng;HE Qiang(Engineering Techniques Training Center,Civil Aviation Flight University of China,Sichuan Guanghan 618307,China;Aviation Engineering Institute,Civil Aviation Flight University of China,Sichuan Guanghan 618307,China)

机构地区：[1]中国民用航空飞行学院工程训练中心,四川广汉618307 [2]中国民用航空飞行学院航空工程学院,四川广汉618307

出　　处：《机械设计与制造》2024年第5期35-40,共6页Machinery Design & Manufacture

基　　金：四川省科技计划项目(2021YJ0537);中国民用航空局人才教育类项目(MHJY2022028)。

摘　　要：孔探检测是航空器维修中最重要的检测项目,在孔探检测过程中人员脑力状态与检测质量好坏直接相关,在高脑力负荷状态下常常会诱发检测人员出现错检、忘检、漏检等安全风险,从而造成维修差错。为解决NASA-TLX等主观量表测量脑力负荷的非即时问题,因此通过眼动仪无接触实时的测量孔探人员在不同脑力负荷状态下的眼动数据,再利用单因素方差分析来寻找到构建脑力负荷状态预测模型的关键眼动指标。此外由于眼动指标与脑力负荷存在非线性关系,因此选用支持向量回归机SVR来构建脑力负荷预测模型,并使用GASA算法来对SVR参数进行寻优,获得了具有足够精度和泛化能力的孔探人员脑力负荷预测模型,从而帮助飞机维修单位实时把握孔探人员状态,降低孔探检测中人为差错导致的风险,同时为民航局制定相应的孔探人员管理规定提供依据。Borescope inspection is the most important detection item in aircraft maintenance.In the process of borescope inspec-tion,the workload is directly related to the quality of detection.Under the condition of high load,it often induces safety risks such as wrong detection,forgetting detection and missing detection,resulting in maintenance errors.In order to solve the non real-time measurement of load such as NASA-TLX,the eye tracking data of borescope inspection under different workload mea-sured by Tobii Glasses 2 without contact and then the key eye tracking indexes of constructing workload prediction model are found by ANOVA.In addition,because there is a nonlinear relationship between eye tracking index and workload,SVR is se-lected to build the workload prediction model,and GASA algorithm is used to optimize the SVR parameters to obtain the work-load prediction model of borescope inspection with sufficient accuracy and generalization ability.So as to help the maintenance and repair organization supervise borescope inspection personnel in real-time,reduce the risk caused,and provide the basis for the CAAC to formulate corresponding regulations on the management of borescope inspection.

关 键 词：脑力负荷 孔探 NASA-TLX量表 单因素方差分析 支持向量回归机SVR GASA算法 

分 类 号：TH16[机械工程—机械制造及自动化] V219[航空宇航科学与技术—航空宇航推进理论与工程]