论文部分内容阅读
一些工业过程可以近似用一个传递函数描述,结合统计辨识方法和非线性优化策略提出传递函数参数辨识方法.该方法采用动态数据方案,使用系统观测数据获得系统更多的模态信息.基于动态观测数据,提出传递函数随机梯度参数辨识方法.为进一步提高辨识精度,利用动态窗数据将随机梯度参数辨识方法中的标量新息扩展为新息向量,提出传递函数多新息随机梯度参数估计方法.最后通过仿真例子对所提出的方法进行了性能分析和模型验证.
Some industrial processes can be approximated by a transfer function description, and combined with statistical identification method and nonlinear optimization strategy, a transfer function parameter identification method is proposed.The dynamic data scheme is adopted to obtain more modal information of the system by using the system observation data.Based on the dynamic observation Data, the transfer function stochastic gradient parameter identification method is proposed.In order to further improve the identification accuracy, the dynamic window data is used to extend the scalar new information in the stochastic gradient parameter identification method into the new interest vector, and a multi-innovation random gradient parameter estimation method is proposed. At last, the performance analysis and model verification of the proposed method are given through simulation examples.