基于改进灰色波形预测模型的装备故障预测  

Equipment Fault Prediction Based on Improved Grey Waveform Prediction Model

作　　者：张佶 洪亮 杨利斌 ZHANG Ji;HONG Liang;YANG Libin(Naval Aviation University,Yantai 264001,China)

机构地区：[1]海军航空大学,山东烟台264001

出　　处：《火力与指挥控制》2025年第1期80-86,共7页Fire Control & Command Control

摘　　要：装备故障预测存在小样本、波动性强的特性,针对传统灰色波形预测模型对波动序列进行拟合时误差较大问题,探索一种基于神经网络优化改进灰色波形预测模型的装备故障预测方法,该方法主要采用BP神经网络与灰色波形预测模型中GM(1,1)模型组进行并联,优化等高点出现时间预测精度,从而得出中短期时间内预测波形,最终计算得出未来某一时间点的装备故障率,装备故障预测实例表明,改进后的波形预测模型较传统波形预测模型有更高的精度。Equipment fault prediction has the characteristics of small sample size and strong fluctuation.Aiming at the problem of large errors when fitting wave series with traditional gray waveform prediction model,an equipment fault prediction method based on neural network optimization and improvement of grey waveform prediction model is proposed.The method mainly uses parallel connection of BP neural network with GM(1,1)model group in grey waveform predication model.By optimizing the prediction accuracy of the occurrence time of the contour point,the prediction waveform in the short and medium time is obtained,and the failure rate of the equipment at a certain time point in the future is finally calculated.The example of the equipment fault prediction shows that the improved waveform prediction model has higher accuracy than that of the traditional waveform prediction model.

关 键 词：故障预测 灰色波形预测模型 灰色神经网络 GM(1 1) 

分 类 号：TJ760[兵器科学与技术—武器系统与运用工程]