Currently, some fault prognosis technology occasionally has relatively unsatisfied performance especially for in- cipient faults in nonlinear processes duo to their large time delay and complex internal connection. To...Currently, some fault prognosis technology occasionally has relatively unsatisfied performance especially for in- cipient faults in nonlinear processes duo to their large time delay and complex internal connection. To overcome this deficiency, multivariate time delay analysis is incorporated into the high sensitive local kernel principal component analysis. In this approach, mutual information estimation and Bayesian information criterion （BIC） are separately used to acquire the correlation degree and time delay of the process variables. Moreover, in order to achieve prediction, time series prediction by back propagation （BP） network is applied whose input is multivar- iate correlated time series other than the original time series. Then the multivariate time delayed series and future values obtained by time series prediction are combined to construct the input of local kernel principal component analysis （LKPCA） model for incipient fault prognosis. The new method has been exemplified in a sim- ple nonlinear process and the complicated Tennessee Eastman （TE） benchmark process. The results indicate that the new method has suoerioritv in the fault prognosis sensitivity over other traditional fault prognosis methods.展开更多
A new empirical approach for the seasonal prediction of annual Atlantic tropical storm number (ATSN) was developed using precipitation and 500 hPa geopotential height data from the preceding January February and April...A new empirical approach for the seasonal prediction of annual Atlantic tropical storm number (ATSN) was developed using precipitation and 500 hPa geopotential height data from the preceding January February and April May.The 2.5°×2.5° resolution reanalysis data from both the US National Center for Environmental Prediction/the National Center for Atmospheric Research (NCEP/NCAR) and the European Center for Medium-Range Weather Forecasting (ECMWF) were applied.The model was cross-validated using data from 1979 2002.The ATSN predictions from the two reanalysis models were correlated with the observations with the anomaly correlation coefficients (ACC) of 0.79 (NCEP/NCAR) and 0.78 (ECMWF) and the multi-year mean absolute prediction errors (MAE) of 1.85 and 1.76,respectively.When the predictions of the two models were averaged,the ACC increased to 0.90 and the MAE decreased to 1.18,an exceptionally high score.Therefore,this new empirical approach has the potential to improve the operational prediction of the annual tropical Atlantic storm frequency.展开更多
A higher correlation tends to yield a more accurate prediction,so that a correlation as high as possible has been searched for and employed in the prediction of solar activity.Instead of using geomagnetic activity dur...A higher correlation tends to yield a more accurate prediction,so that a correlation as high as possible has been searched for and employed in the prediction of solar activity.Instead of using geomagnetic activity during the descending phase of the solar cycle,the minimum annual aa index (aa min) is used as an indicator for the ensuing maximum amplitude (R m) of the sunspot cycle.A four-cycle periodicity is roughly shown in the correlation between R m and aa min.The widely accepted Ohl's precursor prediction method often fails due to the prediction error relative to its estimated uncertainty.An accurate prediction depends on the positive variation of the correlation rather than a higher correlation.Previous experiences by using this method indicate that a prediction for the next cycle,R m (24)=80 ± 17,is likely to fail,implying that the sunspot maximum of Cycle 24 may be either smaller than 63 or greater than 97.展开更多
基金Supported by the National Natural Science Foundation of China(61573051,61472021)the Natural Science Foundation of Beijing(4142039)+1 种基金Open Fund of the State Key Laboratory of Software Development Environment(SKLSDE-2015KF-01)Fundamental Research Funds for the Central Universities(PT1613-05)
文摘Currently, some fault prognosis technology occasionally has relatively unsatisfied performance especially for in- cipient faults in nonlinear processes duo to their large time delay and complex internal connection. To overcome this deficiency, multivariate time delay analysis is incorporated into the high sensitive local kernel principal component analysis. In this approach, mutual information estimation and Bayesian information criterion （BIC） are separately used to acquire the correlation degree and time delay of the process variables. Moreover, in order to achieve prediction, time series prediction by back propagation （BP） network is applied whose input is multivar- iate correlated time series other than the original time series. Then the multivariate time delayed series and future values obtained by time series prediction are combined to construct the input of local kernel principal component analysis （LKPCA） model for incipient fault prognosis. The new method has been exemplified in a sim- ple nonlinear process and the complicated Tennessee Eastman （TE） benchmark process. The results indicate that the new method has suoerioritv in the fault prognosis sensitivity over other traditional fault prognosis methods.
基金supported by the Major State Basic Research Development Program of China (Grant No.2009CB421406)the National Natural Science Foundation of China (Grant Nos. 40631005 and 40875048)
文摘A new empirical approach for the seasonal prediction of annual Atlantic tropical storm number (ATSN) was developed using precipitation and 500 hPa geopotential height data from the preceding January February and April May.The 2.5°×2.5° resolution reanalysis data from both the US National Center for Environmental Prediction/the National Center for Atmospheric Research (NCEP/NCAR) and the European Center for Medium-Range Weather Forecasting (ECMWF) were applied.The model was cross-validated using data from 1979 2002.The ATSN predictions from the two reanalysis models were correlated with the observations with the anomaly correlation coefficients (ACC) of 0.79 (NCEP/NCAR) and 0.78 (ECMWF) and the multi-year mean absolute prediction errors (MAE) of 1.85 and 1.76,respectively.When the predictions of the two models were averaged,the ACC increased to 0.90 and the MAE decreased to 1.18,an exceptionally high score.Therefore,this new empirical approach has the potential to improve the operational prediction of the annual tropical Atlantic storm frequency.
基金supported by the Chinese Academy of Sciences (Grant No.KGCX3-SYW-403-10)the National Natural Science Foundation of China (Grant Nos.10973020,10673017 and 40890161)
文摘A higher correlation tends to yield a more accurate prediction,so that a correlation as high as possible has been searched for and employed in the prediction of solar activity.Instead of using geomagnetic activity during the descending phase of the solar cycle,the minimum annual aa index (aa min) is used as an indicator for the ensuing maximum amplitude (R m) of the sunspot cycle.A four-cycle periodicity is roughly shown in the correlation between R m and aa min.The widely accepted Ohl's precursor prediction method often fails due to the prediction error relative to its estimated uncertainty.An accurate prediction depends on the positive variation of the correlation rather than a higher correlation.Previous experiences by using this method indicate that a prediction for the next cycle,R m (24)=80 ± 17,is likely to fail,implying that the sunspot maximum of Cycle 24 may be either smaller than 63 or greater than 97.
