摘要
针对一类缓变退化设备的样本平均在短时间内退化不明显的特点,将设备检测数据的样本标准差作为一个失效指标进行退化建模,从样本标准差的变化趋势来估计设备的性能状态,并且鉴于样本标准差数据非线性强波动幅度大的特点,提出了一种随机阈值的剩余寿命预测方法。通过惯性平台陀螺仪漂移数据实例验证证明,该方法能够对这类非线性强、波动幅度大的退化数据建模,并得到可信的预测结果。
Considering that the slowly degradating devices have the characteristic that the sample degenerates indistinctly when the monitoring data are used for prognostics,we used the sample standard deviation as an index to establish a degradation model .The trend of the sample standard deviation was used to evaluate the performance of the monitored device .Since the sample standard deviation was volatile and has high nonlinearity,a random threshold based forecasting method was proposed to predict the remaining useful life （RUL）.Finally,a practical case study for gyros in an inertial navigation system was provided to demonstrate the presented method.The experimental results indicate that the proposed method can effectively model the non-linear and volatility data and can achieve credible estimates .
出处
《电光与控制》
北大核心
2014年第3期80-83,共4页
Electronics Optics & Control
基金
国家杰出青年基金(61025014)
国家自然科学基金(610 04069
61174030)
关键词
剩余寿命预测
随机阈值
样本标准差
维纳过程
remaining useful life prediction
random threshold
sample standard deviation
Wiener process
