CLChuanfeng Li

Chuanfeng Li

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Chuanfeng Li has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Cross-cell SOH prediction of lithium-ion batteries based on SPMe parameter identification

    by Yun Wang, Yang Yu, Xingyu He, Huhe Dai, Chuanfeng Li

    Cell-to-cell variations in degradation trajectories limit the cross-cell generalization of state of health (SOH) prediction models trained on one lithium-ion cell. Training with labeled data from multiple cells can improve generalization, but SOH labels can only be obtained through full-capacity tests that interrupt operation, making such data difficult to acquire. To solve this problem, this paper proposes a cross-cell SOH prediction framework based on the single particle model with electrolyte dynamics (SPMe). Automatic matching of baseline parameter sets and two-stage particle swarm optimization identify electrochemical parameters from the discharge voltage and current of each cycle, producing parameter sequences for each cell. For each dataset group, only one cell with SOH labels is used for feature selection and model training. Electrochemical features are selected by considering their correlations with SOH in the training cell, parameter redundancy, and physical meaning. A long short-term memory network learns the relationship between the selected feature sequences and SOH to predict future SOH values. The trained model is directly applied to target cells without adaptation. Target-cell SOH labels are used only for evaluation. The framework is validated on five dataset groups covering four cathode materials. With equal weighting across groups, the root mean square error and mean absolute error for 10-step-ahead prediction are 0.0269 and 0.0212, respectively. These values are 12.52% and 15.06% lower than those obtained using voltage statistical features. The results show that electrochemical parameters identified using the SPMe provide useful degradation information for cross-cell SOH prediction without target-cell SOH labels during model development.

  • Open Access

    Robust non-autoregressive Transformer for short-term wind power forecasting via structured NWP perturbation training

    by Chuanfeng Li, Zhixiang Wang, Liang Jia, Ziwen Xiao, Han Ding

    Numerical weather prediction (NWP) uncertainty is a key factor limiting the accuracy and robustness of short-term multi-step wind power forecasting. Under NWP uncertainty, existing methods face three main problems: (i) a training–inference distribution shift between the historical observations used for training and the uncertain future NWP inputs used at inference; (ii) error accumulation over long horizons caused by recursive prediction; and (iii) limited characterization of abrupt power-change dynamics by static physical constraints. To address these problems, this paper proposes PUDirectFormer, a robust non-autoregressive Transformer for wind power forecasting based on structured NWP perturbation training. First, before physical features are computed, synchronized amplitude and temporal phase-shift perturbations are injected into the multi-height wind speed series, simulating how NWP uncertainty propagates into the derived features and reducing the training–inference distribution shift. Second, the Transformer encoder is configured for non-autoregressive forecasting through a last-observed-power placeholder, allowing the future horizon to be predicted directly and reducing error accumulation associated with recursive decoding. Third, a lightweight Ridge-regression residual calibrator, built on explicit physical features (the nonlinear wind-speed–power energy relationship, vertical wind shear, and wind speed dynamics), corrects residual biases under complex wind conditions and abrupt power-change events. On two operational wind farms in China (66 and 36 MW), under a controlled stress test with a 30% maximum perturbation intensity, PU-DirectFormer reduces the RMSE and MAE by 30.19% and 26.21%, respectively, relative to an LSTM baseline; in validation on the independent 36 MW farm, it lowers the RMSE by 4.24% and the MAE by 13.29%. These results indicate that the proposed method improves the accuracy and robustness of short-term wind power forecasting and remains stable across wind farms, providing more reliable forecasts for day-ahead scheduling and wind power integration.

  • Open Access

    Non‐Intrusive Online ESR Estimation for Interleaved Buck Converters Based on Virtual Signal Construction and DWT

    by Bao Hou, Chuanfeng Li, Cuiyu Liu, Wangwang Huang, Peng Qi

    Interleaved DC–DC converters are widely applied in modern power conversion systems for their high power density, low input current ripple, high reliability and fast dynamic response. However, the reliability of these converters is critically affected by aluminium electrolytic capacitors, which are prone to degradation over time. The equivalent series resistance (ESR) of capacitors serves as a key indicator of their health. This paper proposes an ESR estimation method for the aluminum electrolytic capacitors in a three‐phase interleaved Buck converter, based on virtual‐signal construction and discrete wavelet transform. Virtual voltage and current signals are reconstructed from the sampled output voltage and inductor current using an algorithm based on rising and falling slope trends. To capture ESR‐related frequency components, the discrete wavelet transform is applied to extract high‐frequency features, and the ESR is determined as the ratio of the dominant wavelet coefficients of the virtual voltage and current signals. Operating in the time–frequency domain, the proposed method exhibits superior robustness against sampling noise and dynamic load variations compared with the conventional time‐domain technique. Furthermore, it eliminates the need for any extra current sensors for measuring the capacitor current. In addition, simulation analysis and hardware experimental results validate the accuracy and robustness of the proposed approach under various operating conditions.

  • Reconstructed Phase Current-Based Fault Diagnosis Method for Interleaved Buck Converter

    by Yang Yu, Chuanfeng Li, Zhiming Yang, Xiyuan Peng

    The interleaved converter is widely used in clean energy systems due to inherent advantages, such as low current ripple, high reliability, and superior energy density. However, an open circuit fault (OCF) will increase the current ripple and reduce the reliability of an interleaved converter. Therefore, detecting and locating the OCF as soon as possible is essential. This paper proposes a robust method based on the reconstructed phase current to detect and locate the OCF. The phase current averages are reconstructed by establishing mathematical relationships between total input current and phase currents. The reconstructed phase currents can easily indicate the fault phase. Instead of requiring N sensors to sample the phase current, only a single current sensor is necessary for input current sampling. Additionally, the reconstructed phase current can also be used for current-sharing control, offering cost-effectiveness. Both simulation and hardware experiments confirm the effectiveness and robustness of the proposed method. The proposed method quickly diagnoses the OCF within a switching period in hardware experiments. Furthermore, the robustness of the proposed method against input voltage and load disturbances and its operation at higher switching frequencies are demonstrated. Notably, the proposed method is adaptable to multi-phase interleaved buck converter, regardless of the number of phases.


  • A Virtual Phase Current-Based Current Balance Method for the Interleaved Boost Converter

    by Chuanfeng Li, Yang Yu, Zhiming Yang, Xiyuan Peng

    This article presents a novel current-balancing control method using virtual phase current for the interleaved boost converter. By theoretical derivation, it can be found that the output voltage contains useful information used to calculate the ratio between phase currents. The proposed method uses the ratio and the total input current to generate the virtual phase current. Once the virtual phase current is achieved, the phase current can be autotuned to balance due to the same reference from the voltage regulation control loop. Despite the N number of the phase, only a current sensor is needed. Moreover, instead of the input current details, only effective values are required, which will reduce sampling costs. A limiter is added to reduce the impact of the current balance on the stability and transient response of the voltage regulation control loop. It is also beneficial to improve the robustness against noise and external disturbances. The proposed method is easy to implement in the existing controller, including the double closed-loop controllers. The simulation and experimental results demonstrate the effectiveness and practicability.


  • A Sensorless Open-Circuit Fault Identification Method for Interleaved Boost Converter

    by Chuanfeng Li, Yang Yu, Zhiming Yang, Wang Wang, Xiyuan Peng

    The interleaved DC-DC converter is widely used for its low current ripple and voltage ripple, alongside high energy conversion efficiency. However, open-circuit faults (OCFs) will increase current and voltage ripple while reduce energy conversion efficiency. Therefore, the OCF identification is very significant. Most current fault diagnosis methods typically depend on current sensors to obtain current information, elevating diagnostic costs. Consequently, this paper introduces a cost-effective, sensorless OCF identification method using the step change in output voltage at specific sampling moments in the interleaved boost converter. It is proved that the ratio of output voltage step value serves as an indicator of the phase current relationships. Therefore, the output voltage is only used to locate the open-circuit fault, avoiding any current sensor. Both simulation and hardware experiments verify the effectiveness and practicality of the proposed method.


  • A Robust Open-Circuit Fault Diagnosis Method for Three-Phase Interleaved Boost Converter

    by Chuanfeng Li, Yang Yu, Tingyan Tang, Qingxin Liu, Xiyuan Peng

    An open-circuit fault will increase input current ripple and reduce the energy transmission efficiency, which will deteriorate the inherent advantages of an interleaved boost converter. To guarantee the efficient and reliable operation of a system, it is necessary to diagnose the open-circuit fault by sensing the input current. In practical application, the sampling value of the input current will be polluted by noise and disturbance, which will weaken the robustness of fault diagnosis. To address the problem, this article proposes a robust open-circuit fault diagnosis method. By theoretically deriving the input current, it can be found that the input current contains helpful information, which can be used to identify the fault phase. The proposed method extracts this useful information by sampling the input current and holding it at specific moments. Moreover, this method integrates the sampling values to enhance robustness against noise. The fault phase can quickly be detected and located by observing and comparing the integrals during two switching periods. The method has robustness against the noise and the disturbance. Additionally, the method needs no extra sensors to sense the input current, which the control loop uses. Simulations and hardware experiments are carried to prove the practicability and the robustness of the proposed method.