Research on process monitoring of DMTO based on SVR-PCA
Due to the nonlinear nature of DMTO reaction temperature and modeling data,conventional linear model could not accurately capture its characteristics.Therefore,the process monitoring model based on SVR-PCA was constructed to mitigate the impact of nonlinear features on the model.The results showed that the SVR-PCA process monitoring model could provide an alarm for significant and long-term fluctuations in DMTO reaction temperature 63 minutes earlier than the PCA process monitoring model,and 163 minutes earlier than the DCS system.It could also reduce the false alarm rate for short-term small fluctuations and long-term small fluctuations in reaction temperature.
NETL
NSTL
万方数据
