基于ARMA模型的动力系统缓变数据故障趋势预测  

Prediction of Slowly Varying Data of Propulsion System Based On ARMA Model

作　　者：朱晓乐[1,2] 王华[1] 符菊梅[2] 陈景鹏[1] 徐腊萍[1] 

机构地区：[1]装备指挥技术学院 [2]中国西昌卫星发射中心

出　　处：《载人航天》2011年第2期54-58,共5页Manned Spaceflight

摘　　要：针对运载火箭飞行过程中动力系统缓变数据可能存在的故障趋势不容易判断的特点,提出了对实时飞行数据进行趋势预测的思想,并采用ARMA模型(自回归移动平均模型)进行趋势预测研究。通过仿真数据实例分析得出该预测模型的预测精度比较高,跟踪速度比较快,实时性比较好,能够实现对飞行数据的实时趋势预测,提前告知潜在的故障趋势,为指挥员决策提供理论支持。Sime the potential fault tendency of the slowly varying data of propulsion system is not easy to determine during flight, the concept of predicting the tendency of real time flight data is proposed, and ARMA model （Auto-Regressive and Moving Average Model） is used to predict data tendency. Simulation shows that the predicting model is highly precise, fast tracking and almost in real time, so that it can help to realize real-time flight data tendency prediction and forecast potential fault in advance to provide theoretical support for the commander＇s decision.

关 键 词：ARMA 缓变数据 趋势预测 动力系统 

分 类 号：V421.4[航空宇航科学与技术—飞行器设计]