摘要
在大型旋转机组运行过程中,机组振动烈度值常常表现为具有趋向性的非平稳时间序列。建立灰色预测模型对烈度值进行预测。方法是用一阶灰色模型(GM(1,1))提取趋势项,然后用线性回归模型(AR模型)对残差进行预测。对于残差预测,采用Marple算法估计AR模型参数,用FPE准则对AR模型适用性进行检验。全部程序在LabVIEW平台下并结合VC编程实现。经工业现场数据验证,该预测方法具有较好的预报效果。
During the running course of large rotary machines, the data of intensity of oscillation tend to behave in non linear time series. Setting up a GM(1,1)combined with AR model to predict their future data. Using AR model to predict their remains and using Marple Algorithm to estimate their model parameters. And using FPE rule to test the applicability of the model.The whole process is realized in LabVIEW combination with VC. Using this new method has obtained more satisfactory predicting results to large rotary sets in the industrial application.
出处
《北京机械工业学院学报》
2003年第1期5-9,15,共6页
Journal of Beijing Institute of Machinery
基金
国家自然科学基金资助项目[59775002]
