摘要
振动烈度是剩余寿命的一个评价指标。为提高剩余寿命预测精度,解决时序模型预测延时问题,文中提出了一种时间序列分析理论,对振动烈度数据进行平稳建模,得到符合其变化规律的模型方程;通过得到的模型方程推导出卡尔曼滤波算法的状态方程和观测方程;然后依靠卡尔曼预测递推方程进行预测,再对振动烈度进行预测,从而预测剩余寿命。实例分析表明,采用混合算法可以提高预测精度,且较好地解决了预测延时问题。
The vibration severity is an evaluation index of remaining life. To improve the vibration severity fore- casting accuracy and solve the problem of time delay of forecasting by time series model, the author proposes a hybrid algorithm integrating time series analysis with Kalman filter. The basic concept of this algorithm is as following： firstly, by use of time series analysis theory, the stationary modeling for vibration severity is proceeded to obtain the model equation conforming to its variation law ; secondly, by means of the obtained model equation the state equation and observational equation for Kalman filter are deduced; thirdly, the vibration severity is forecasted by Kalman forecasting recurrence equation; and finally, the forecasting for vibration severity is conducted to validate the pro-posed hybrid algorithm. Case study results show that by using this hybrid algorithm the forecasting accuracy of vibra-tion severity is improved and the time delay in the forecasting is well solved.
出处
《电子科技》
2013年第9期28-30,共3页
Electronic Science and Technology
