基于原子搜索Kriging模型方法的民用飞机液压系统温度监测

温度是民用飞机液压系统性能的重要表征参数之一，为有效实现液压系统温度监测，基于Kriging 模型和原子搜索算法（ASO），提出了一种基于原子搜索Kriging 模型（ASOKM）方法。首先，针对液压系统温度故障发生原因进行分析，建立其故障逻辑图，明确影响液压系统温度的特征参数；然后，结合快速存取记录器（QAR）数据，论述了ASOKM 方法建模原理；最后，通过某型国产民用飞机液压系统温度监测分析，对所提出的ASOKM 方法进行有效性验证。研究结果表明：ASOKM 方法的训练平均绝对误差、监测平均绝对误差低于响应面（RSM）、Kriging、BP 神经网络（BP-ANN）模型，具有较高的精度、效率和鲁棒性。所提方法可以为液压系统可靠性分析、故障诊断、预测维修提供借鉴。

Civil aircraft hydraulic system temperature monitoring based on atom search Kriging model method

Temperature is one of the important characteristic parameters of civil aircraft hydraulic system performance. To effectively realize the temperature monitoring of hydraulic system, the method of Kriging model based on atomic search (ASOKM) is proposed based on Kriging model and atomic search algorithm (ASO). Firstly, the temperature fault causes of hydraulic system are analyzed, and fault logic diagram is established to clarify the characteristic parameters affecting the hydraulic system temperature. Then, combined with the quick access recorder (QAR) data, the modeling principle of ASOKM method is discussed. Finally, the effectiveness of the proposed ASOKM method is verified by the temperature monitoring analysis of domestic civil aircraft hydraulic systems. The results show that the training and monitoring average absolute errors of the ASOKM method are lower than the response surface method (RSM), Kriging, and back-propagation-artificial neural network (BP-ANN) models, and have higher accuracy, efficiency, and robustness. The proposed ASOKM method can provide a reference for reliability analysis, fault diagnosis, and predictive maintenance of the hydraulic system.