基于新陈代谢GM(1,1)模型的导弹遥测故障预测算法

Missile Telemetry Fault Forecast Algorithm Based on Metabolic GM(1,1) Model

在人工监视遥测关键参数是否超出阈值范围以判断导弹故障的传统模式基础上,对GM(1,1)预测模型进行了改进,研究了基于新陈代谢的GM(1,1)预测算法对遥测参数实时预测效果。该算法能不断去掉最老的信息,充分利用最新的信息,避免了预测模型老化的现象;由于用于预测的数列维度较少,便于计算,能够保证算法的实时性。用MATLAB编程实现算法,比对实测数据的预测值与实际值,并进行误差分析。结果表明,该算法对遥测参数的预测精度较高,可为及时发现导弹故障提供科学的依据。

The traditional model of missile fault judgment is manual monitoring of whether key telemetry meters exceed thresholds.Improvement is made on the GM （Grey Model） （1,1） forecast model and real-time forecast of telemetry parameters based on metabolic GM （1,1） forecast algorithm is studied.The algorithm increases forecast accuracy and avoids the phenomenon of aging by continuously removing old information and making full use of the latest information.Real-time performance of the algorithm is ensured as there are fewer sequences during forecast computing.The algorithm is programmed with MATLAB.Comparison between the measured data and forecast values and error analysis show that the algorithm has a high forecast accuracy and provides confident foundation for timely detection of missile faults.