In this paper,an efficient model structure composed of a second-order resistance-capacitance network and a simply analytical open circuit voltage versus state of charge(SOC) map is applied to characterize the voltage ...In this paper,an efficient model structure composed of a second-order resistance-capacitance network and a simply analytical open circuit voltage versus state of charge(SOC) map is applied to characterize the voltage behavior of a lithium iron phosphate battery for electric vehicles(EVs).As a result,the overpotentials of the battery can be depicted using a second-order circuit network and the model parameterization can be realized under any battery loading profile,without a special characterization experiment.In order to ensure good robustness,extended Kalman filtering is adopted to recursively implement the calibration process.The linearization involved in the calibration algorithm is realized through recurrent derivatives in a recursive form.Validation results show that the recursively calibrated battery model can accurately delineate the battery voltage behavior under two different transient power operating conditions.A comparison with a first-order model indicates that the recursively calibrated second-order model has a comparable accuracy in a major part of the battery SOC range and a better performance when the SOC is relatively low.展开更多
Recurrent events data with a terminal event （e.g. death） often arise in clinical and observational studies. Most of existing models assume multiplicative covariate effects and model the conditional recurrent event r...Recurrent events data with a terminal event （e.g. death） often arise in clinical and observational studies. Most of existing models assume multiplicative covariate effects and model the conditional recurrent event rate given survival. In this article, we propose a general mSditive-multiplicative rates model for recurrent event data in the presence of a terminal event, where the terminal event stop the further occurrence of recurrent events. Based on the estimating equation approach and the inverse probability weighting technique, we propose two procedures for estimating the regression parameters and the baseline mean function. The asymptotic properties of the resulting estimators are established. In addition, some graphical and numerical procedures are presented for model checking. The finite-sample behavior of the proposed methods is examined through simulation studies, and an application to a bladder cancer study is also illustrated.展开更多
Semiparametric transformation models provide a class of flexible models for regression analysis of failure time data. Several authors have discussed them under different situations when covariates are time- independe...Semiparametric transformation models provide a class of flexible models for regression analysis of failure time data. Several authors have discussed them under different situations when covariates are time- independent （Chen et al., 2002; Cheng et al., 1995; Fine et al., 1998）. In this paper, we consider fitting these models to right-censored data when covariates are time-dependent longitudinal variables and, furthermore, may suffer measurement errors. For estimation, we investigate the maximum likelihood approach, and an EM algorithm is developed. Simulation results show that the proposed method is appropriate for practical application, and an illustrative example is provided.展开更多
基金Project (No. 61004092) supported by the National Natural ScienceFoundation of China
文摘In this paper,an efficient model structure composed of a second-order resistance-capacitance network and a simply analytical open circuit voltage versus state of charge(SOC) map is applied to characterize the voltage behavior of a lithium iron phosphate battery for electric vehicles(EVs).As a result,the overpotentials of the battery can be depicted using a second-order circuit network and the model parameterization can be realized under any battery loading profile,without a special characterization experiment.In order to ensure good robustness,extended Kalman filtering is adopted to recursively implement the calibration process.The linearization involved in the calibration algorithm is realized through recurrent derivatives in a recursive form.Validation results show that the recursively calibrated battery model can accurately delineate the battery voltage behavior under two different transient power operating conditions.A comparison with a first-order model indicates that the recursively calibrated second-order model has a comparable accuracy in a major part of the battery SOC range and a better performance when the SOC is relatively low.
基金Supported by the National Natural Science Foundation of China(No.11371354)Young Elite Program of Beijing(YETP150)Science and Technology Project of Beijing Municipal Education Commission(KM201411232019)
文摘Recurrent events data with a terminal event （e.g. death） often arise in clinical and observational studies. Most of existing models assume multiplicative covariate effects and model the conditional recurrent event rate given survival. In this article, we propose a general mSditive-multiplicative rates model for recurrent event data in the presence of a terminal event, where the terminal event stop the further occurrence of recurrent events. Based on the estimating equation approach and the inverse probability weighting technique, we propose two procedures for estimating the regression parameters and the baseline mean function. The asymptotic properties of the resulting estimators are established. In addition, some graphical and numerical procedures are presented for model checking. The finite-sample behavior of the proposed methods is examined through simulation studies, and an application to a bladder cancer study is also illustrated.
文摘Semiparametric transformation models provide a class of flexible models for regression analysis of failure time data. Several authors have discussed them under different situations when covariates are time- independent （Chen et al., 2002; Cheng et al., 1995; Fine et al., 1998）. In this paper, we consider fitting these models to right-censored data when covariates are time-dependent longitudinal variables and, furthermore, may suffer measurement errors. For estimation, we investigate the maximum likelihood approach, and an EM algorithm is developed. Simulation results show that the proposed method is appropriate for practical application, and an illustrative example is provided.
