首页> 外文期刊>Journal of information and computational science >Sparse Improved Method Based on AR Model for Parameter Extraction of Vibration Signal
【24h】

Sparse Improved Method Based on AR Model for Parameter Extraction of Vibration Signal

机译:基于AR模型的振动信号参数提取的稀疏改进方法

获取原文
获取原文并翻译 | 示例
           

摘要

The parameter extraction method based on AR model holds high frequency resolution. It is sensitive to noise. It is also difficult to eliminate the false modal effectively when a high order AR model is used for improving the noise resistance performance. Stabilization diagram method can establish the order to solve it. But it needs to calculate too many times and the order established may be not accurate if the orders of AR model are not suitable. So a sparse improved method is put forward based on it. In the improved method, a least square algorithm is replaced by sparse optimization method to calculate the coefficient of vibration modal. According to the nonzero elements of coefficient, the modal could be filtered automatically from the calculated modal. Though the sparse solutions are calculated many times for an accurate solution, the number of repeatedly computing is greatly reduced than stabilization diagram method. The order can be established without changing the order of AR model. Simulating experiments verifies that improved method can achieve accurate and steady result when the order of AR model is changing within a large range. It ensures that improved method is easier to use and decreases the possibility of error.
机译:基于AR模型的参数提取方法具有较高的频率分辨率。它对噪音敏感。当使用高阶AR模型来改善抗噪声性能时,也难以有效地消除错误模态。稳定图法可以建立求解顺序。但是它需要计算太多次,如果AR模型的阶数不合适,则建立的阶数可能不准确。因此在此基础上提出了一种稀疏的改进方法。在改进方法中,将最小二乘算法替换为稀疏优化方法以计算振动模态系数。根据系数的非零元素,可以从计算出的模态中自动过滤出模态。尽管为精确的解计算了许多次稀疏解,但与稳定图方法相比,重复计算的次数大大减少了。可以建立顺序,而无需更改AR模型的顺序。仿真实验证明,当AR模型的阶数在较大范围内变化时,改进的方法可以获得准确,稳定的结果。它确保改进的方法更易于使用,并减少了出错的可能性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号