Deep Extreme Learning Machine Research Articles

Improved Deep Extreme Learning Machine for State of Health Estimation of Lithium-Ion Battery

Prediction of state of health (SOH), a crucial aspect of battery management systems, necessitates accurate and reliable estimations for lithium-ion batteries. However, achieving high-precision SOH estimation using the deep extreme learning machine (DELM) in complex environments is challenging due to the instability caused by its random key parameters. To address this, we propose a novel approach that combines the improved bald eagle search (IBES) algorithm with DELM. By utilizing the IBES algorithm, we can extract highly relevant health indicators from the battery’s parameter curve during charging and discharging, as well as the incremental capacity curve. These indicators serve as inputs to the constructed estimation model, which predicts SOH under different working conditions. The proposed method has been validated using publicly available experimental data, with an overall error in SOH estimation of less than 1%. In comparison to the existing model, the proposed method achieves RMSE, MAE, and MAPE values lower than 0.35%, 0.26%, and 0.21%, respectively. These findings demonstrate that the proposed method excels in terms of both the speed and accuracy of SOH estimation, showcasing its enhanced robustness and reliability.

Accurate prediction of hyaluronic acid concentration under temperature perturbations using near-infrared spectroscopy and deep learning

Accurate prediction of the concentration of a large number of hyaluronic acid (HA) samples under temperature perturbations can facilitate the rapid determination of HA’s appropriate applications. Near-infrared (NIR) spectroscopy analysis combined with deep learning presents an effective solution to this challenge, with current research in this area being scarce. Initially, we introduced a novel feature fusion method based on an intersection strategy and used two-dimensional correlation spectroscopy (2DCOS) and Aquaphotomics to interpret the interaction information in HA solutions reflected by the fused features. Subsequently, we created an innovative, multi-strategy improved Walrus Optimization Algorithm (MIWaOA) for parameter optimization of the deep extreme learning machine (DELM). The final constructed MIWaOA-DELM model demonstrated superior performance compared to partial least squares (PLS), extreme learning machine (ELM), DELM, and WaOA-DELM models. The results of this study can provide a reference for the quantitative analysis of biomacromolecules in complex systems.

A State-of-Health Estimation Method for Lithium Batteries under Multi-Dimensional Features

In recent years, the number of new energy vehicles has increased rapidly. The online state-of-health (SOH) prediction of lithium-ion batteries, which are core components of new energy vehicles, is crucial for maintaining vehicle safety. However, existing data-driven methods encounter challenges such as the difficult application of health feature extraction methods in practice, single feature dimensions, and complex algorithm models. This study extracted the peak height of the incremental capacity (IC) curve, constant-current charging time, and time when the battery surface temperature reaches its maximum value as health features from multiple dimensions. Furthermore, by randomly generating prey, the Pelican Optimization Algorithm (POA) fundamentally overcomes the shortcomings of traditional swarm intelligence optimization algorithms which are prone to falling into local optimal solutions. POA was introduced to optimize the Deep Extreme Learning Machine (DELM), which maximally simplified the algorithm model while ensuring accuracy. The experimental results demonstrate that this method does not require extensive historical data support. Whether applied to batteries under the same or different working conditions, all four battery groups exhibit excellent prediction results, with Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) values below 1.2%.

Forecasting model for short-term wind speed using robust local mean decomposition, deep neural networks, intelligent algorithm, and error correction

Wind power generation has aroused widespread concern worldwide. Accurate prediction of wind speed is very important for the safe and economic operation of the power grid. This paper presents a short-term wind speed prediction model which includes data decomposition, deep learning, intelligent algorithm optimization, and error correction modules. First, the robust local mean decomposition (RLMD) is applied to the original wind speed data to reduce the non-stationarity of the data. Then, the salp swarm algorithm (SSA) is used to determine the optimal parameter combination of the bidirectional gated recurrent unit (BiGRU) to ensure prediction quality. In order to eliminate the predictable components of the error further, a correction module based on the improved salp swarm algorithm (ISSA) and deep extreme learning machine (DELM) is constructed. The exploration and exploitation capability of the original SSA is enhanced by introducing a crazy operator and dynamic learning strategy, and the input weights and thresholds in the DELM are optimized by the ISSA to improve the generalization ability of the model. The actual data of wind farms are used to verify the advancement of the proposed model. Compared with other models, the results show that the proposed model has the best prediction performance. As a powerful tool, the developed forecasting system is expected to be further used in the energy system.

Stacked encoded cascade error feedback deep extreme learning machine network for manufacturing order completion time

Stacked encoded cascade error feedback deep extreme learning machine network for manufacturing order completion time

Prediction tool wear using improved deep extreme learning machines based on the sparrow search algorithm

Accurate tool wear monitoring is crucial for increasing tool life and machining productivity. Although many prediction models can achieve high prediction accuracy, there are problems such as poor stability in the face of different working conditions or tool signals. A tool wear prediction method based on improved deep extreme learning machines (DELMs) was proposed as a solution to this issue; it uses the sparrow search algorithm (SSA) to upgrade the input weight of DELM to improve the model, and then extracts the time-domain, frequency-domain, and time-frequency domain characteristics from multi-sensor signals to construct and test the improved model SSA-DELM. The verification results show that the proposed model accurately reflects the tool wear. Meanwhile, the RMSE of the proposed model decreased by 53.39%, 19.95%, 43.86%, 23.80%, 24.80%, and 3.72%, respectively, and the MAE decreased by 67.81%, 17.87%, 32.70%, 29.90%, 30.30%, and 6.78%, respectively, compared to the with unimproved DELM, particle swarm optimization-least squares support vector machine, long short-term memory, stacked sparse autoencoder, recurrent neural network, and dung beetle optimizer-DELM.

Retrieval of purification ability of urban forest to SO2 stress based on the coupling of radiative transfer and AO-DELM models

Currently, hyperspectral remote sensing technology used for vegetation monitoring mainly uses empirical and semi-empirical statistical methods to calculate heavy metal content. Combining physical models and machine learning algorithms is an effective method for estimating biochemical vegetation parameters without much ground measurement data. However, a deep extreme learning machine (DELM) has a faster training speed and better generalization ability. By introducing the Aquila Optimizer (AO) algorithm, the training process of the DELM method can be accelerated. This study combined the PROSAIL model, chlorophyll concentration information, and vegetation SO2 purification ability, and comprehensively applied physical and empirical models to analyze the optical characteristics of the urban forest SO2 purification rate and other factors. A coupled model (PROSAIL + AO + DELM) was constructed to simulate urban forests canopy SO2 purification ability and was then applied to remote sensing inversion. These results indicated that the Syringa oblate Lindl. (S. oblate) and Ulmus pumila “Jinye” (U. pumila) had moderate SO2 purification capacities, whereas that of Prunus cerasifera var. atropurpurea Jack. (P. cerasifera) were low. The SO2 purification rate was highly sensitive in the green, red, and red-edge spectral ranges. In the SO2 purification rate estimation model constructed by the subset (Corresponding to Sentinel-2 band), NDI, DI, and RVI indices, the PROSAIL + AO + DELM model had the best performance, with R2, root mean square error (RMSE), and relative percent deviation (RPD) of 0.73, 0.056, 1.61, and 0.68, 0.096, and 1.06 for T2 (low concentration) and T3 (high concentration) treatments, respectively. The PROSAIL + AO + DELM model was extended to multispectral images (Sentinel-2), where the results of the NDI model inversion were the closest to those of field monitoring. These results indicate that the urban forest SO2 purification rate model constructed in this study has the potential to be applied on a large scale. This study addresses the research gap regarding rapid, non-destructive, and low-input evaluation of plant purification capacity and achieves the non-destructive detection of plant purification capacity from point to point and from static to dynamic. This enables efficient, non-destructive, and cost-effective evaluation of air purification levels, and provides a basis for predicting air purification in large regions in later stages.

Multimodal Feature Fusion Using Optimal Transfer Learning Approach for Lung Cancer Detection and Classification on CT Images

Lung cancer detection is the process of detecting the presence of lung tumor or abnormalities in the lungs. Early diagnosis is crucial for increasing the chances of patient survival and successful treatment. When compared to X-rays, Computed Tomography (CT) images are more sensitive and are increasingly being used for the diagnosis and screening of lung tumors. They provide complete cross-sectional images of the lungs and it will even detect small lesions. AI and Machine learning (ML) approaches are most commonly employed to analyse medical images (e.g. CT scans) and detect lung cancer. This algorithm can help radiologists identify patterns indicative or subtle abnormalities of cancer. Medical diagnosis, particularly in complex diseases such as lung cancer, frequently involves ambiguity. The diagnostic system can alleviate ambiguity via cross-verifying findings from various sources by fusing multimodal features. Multimodal feature fusion using deep learning (DL) algorithm is an advanced technology that leverages the abilities of deep neural networks to combine data from three different modalities or sources for better robustness in several applications, namely natural language processing, image, and data analysis, etc. This study introduces a Multimodal Feature Fusion using an Optimal Transfer Learning Method for Lung Cancer Detection and Classification (MFFOTL-LCDC) methodology on CT images. The chief objective of the MFFOTL-LCDC methodology is to exploit the feature fusion process for the identification and classification of lung tumor. To attain this, the MFFOTL-LCDC model undergoes a multimodal feature fusion approach to derive feature vectors using 3 DL approaches such as SqueezeNet, CapsNet, and Inception v3 models. Besides, the MFFOTL-LCDC technique applies the remora optimization algorithm (ROA) for the hyperparameter choice of 3 DL models. For lung cancer recognition, the MFFOTL-LCDC algorithm exploits the deep extreme learning machine (DELM) algorithm. A series of simulations were conducted to ensure the greater lung cancer recognition outcomes of the MFFOTL-LCDC methodology. The extensive outcomes determine the improved results of the MFFOTL-LCDC technique over recent DL approaches.

Power Increase due to Marine Biofouling: a Grey-box Model Approach

This paper proposes a grey-box modelling approach to predict marine biofouling growth and its effects on ship performance. The approach combines empirical or experimental-based white-box models with data-driven black-box models. First, a white-box model is built to predict ship resistance considering a bare hull. This prediction is based on calm water resistance, wind, waves, and temperature differences. Subsequently, marine biofouling growth is predicted using an experimental model that estimates the level of roughness on the ship hull. Finally, a deep extreme learning machine is used as a black-box model, employing a feedforward neural network technique. To test the approach, a superyacht case study was selected as a category of vessel heavily exposed to fouling. The study used a 2-year dataset obtained through a collaboration with Feadship. Results showed that the black-box approach outperforms the white-box approach in predictive capabilities. However, when the knowledge encapsulated in the white-box model is included in the grey-box approach, the model shows the highest prediction accuracy achieved by leveraging less historical data. This study demonstrates the potential of the proposed grey-box approach to accurately predict marine biofouling growth and its effects on ship performance, which can benefit ship operators and designers in improving operational efficiency and reducing maintenance costs.

HCDP-DELM: Heterogeneous chronic disease prediction with temporal perspective enabled deep extreme learning machine

The strain on healthcare systems is growing dramatically due to the significant rise in heterogeneous chronic disease incidence. Even though frequent patient monitoring is complex and expensive, analyzing chronic illness transitions and predicting patients at risk of developing ailments is crucial to healthcare. This paper proposes an innovative prediction framework for heterogeneous chronic disease prediction (HCDP), including hypertension (HT), heart disease (HD), and chronic kidney disease (CKD), and develops a patient symptom-weighted network (PSWN) by projecting the bipartite graph. Hence, PSWN represents the relationship between health conditions for patients with similar illnesses. LSTM was used to capture the temporal relationship between features in the patient network using a complex network theory. Finally, a deep extreme learning machine (DELM) was applied as a classifier to predict heterogeneous chronic disease. Subsequently, the proposed model was evaluated against previous studies, and the result revealed that the proposed model was more accurate. Remarkably, the proposed architecture achieves an accuracy of 95.02% for hypertension, 96.15% for heart disease, and 98.39% for chronic kidney disease prediction. The relevance of the model is also evaluated with the use of the Wilcoxon signed-rank test. As evidenced in this study, the suggested model outperformed previous research and state-of-the-art approaches, suggesting its potential for a pivotal role in the medical decision-making of heterogeneous chronic diagnoses. Furthermore, the HCDP-DELM model contributes to healthcare authorities by establishing a system of smart surveillance and precise forecasting that allows for delayed disease detection and provides insightful potential strategies for developing effective healthcare.

Prediction interaction responses between railway subgrade and shield tunnelling using machine learning with sparrow search algorithm

Tunnelling-induced uneven ground structure settlement is a hot research topic involving various interrelated factors. This paper employs hybrid algorithms to establish the predictive model for the interaction responses, including maximum settlements, the longitudinal settlement curve, and the shield operational parameters. We choose four machine learning (ML) models: back-propagation neural network (BPNN), long short-term memory neural network (LSTM), least squares support vector machine (LS-SVM), and deep extreme learning machine (DELM). The sparrow search algorithm (SSA) searches for optimal hyperparameter combinations to improve prediction performance. We comprehensively compare the above models' accuracy and generalization ability for different predicting objects. The database used in this study is collected from a subway project in Beijing, China, where the excavation of twin shield tunnels caused subgrade differential settlements on four national railway lines. The in-situ data from the right line of twin shield tunnels is used to train and test the models, while that from the left line is applied to verify the generalization ability of the models. The DELM-SSA model performs well in predicting maximum settlement, while the LSTM-SSA model excels at predicting shield operational parameters. The LS-SVM-SSA model accurately predicts the monitoring points' longitudinal settlement curve. According to the results, different models are recommended for predicting the interaction responses. The analysis of the Pearson correlation coefficient also reveals that shield operational parameters, such as shield driving speed (Sds) and cutterhead rotational speed (Crs), correlate relatively strongly with the settlement.

Application Deep Extreme Learning Machine in Multi-dimensional Smart Teaching Quality Evaluation System

INTRODUCTION: The construction of the wisdom teaching evaluation system, as the essential part of the institution's teaching reform, is conducive to developing the institution's disciplines, making the existing teaching more standardized, and making the means of teaching diversified, intelligent, and convenient.
 OBJECTIVES: Aiming at the current intelligent teaching evaluation design method, there are evaluation indexes that need to be more comprehensive, a single method, and system standard limitations.
 METHODS: Proposes an intelligent optimization algorithm for a multi-dimensional innovative teaching quality evaluation method. First of all, the multi-dimensional wisdom teaching evaluation system is constructed by analyzing the influencing factors of teaching quality evaluation; then, the parameters of the depth limit learning machine are optimized by the bird foraging search algorithm, and the multi-dimensional wisdom teaching evaluation model is constructed; finally, the validity and stability of the proposed method are verified by the analysis of simulation experiments.
 RESULTS: The results show that the proposed method improves the accuracy of the evaluation model.
 CONCLUSION: Solves the problem of low evaluation accuracy and incomplete system of teaching quality evaluation methods.

Study on Short-Term Load Combination Forecasting Model Considering Historical Data Interval Construction

In response to the insufficient accuracy of load forecasting in power system and the wide range of intervals, a combined short-term power load forecasting model considering the interval construction of historical data is proposed. First, the data are decomposed into relatively stable subsequences using extreme-point symmetric mode decomposition (ESMD), and the adaptive dispersion entropy (DE) of C–C algorithm is proposed to recombine similar subsequences. Then, periodicity and correlation analysis are used to determine the input set of each reconstructed component, and the hybrid strategy improved equilibrium optimizer (HSIEO) is proposed to optimize the output weights of the deep extreme learning machine (DELM) to obtain the prediction values of different components and the historical data errors, and the historical data intervals are constructed based on the errors of each component. Then, based on the upper and lower bound estimation method (LUBE), the proposed improved objective function is optimized using the HSIEO-DELM to obtain the component prediction intervals, and the optimal prediction intervals are obtained after superposition. Finally, the experimental comparison shows that the proposed algorithm has higher accuracy and better quality of model prediction intervals.

Prediction of compressive strength of nano-silica modified engineering cementitious composites exposed to high temperatures using hybrid deep learning models

This study estimated the compressive strength of nano-silica-modified engineering cementitious composites subjected to high temperatures using innovative hybrid deep learning models. The innovative hybrid models in this study were designed using autoencoder (AE)-decision tree (DT) and autoencoder (AE)-extreme learning machine (ELM). Additionally, ELM, DT, and deep AE models in this study were designed to compare the results of innovative hybrid deep learning models. The sensitivity analysis was used for the statistical assessment of the experimental results. The input variables of the models were selected as the cement amount, fly ash amount, sand amount, water amount, high-range water reducer amount, PVA (polyvinylalcohol) fiber amount, nano-silica amount, and the degree of exposure to high temperatures. The compressive strength was used as the output variable of the models. The mixture ratio in the experimental study was 583 kg/m3 cement, 467 kg/m3 sand, 700 kg/m3 fly ash, 187 kg/ m3 water, PVA fiber (0.5 %, 1 %, 1.5 % and 2 %) and nano silica (0 %, 1 %, 2 %, 3 % and 4 %) were used. The ELM, DT, and deep AE models estimated the compressive strength of nano-silica-modified engineering cementitious composites subjected to high temperatures with 93.86 %, 77.35 %, and 86.5 % accuracy, respectively. Also, the same compressive strength was estimated with 94.28 % and 98 % accuracy using the hybrid deep AE-DT and AE-ELM models. This study found that the innovative hybrid deep AE-ELM model predicted compressive strength with higher accuracy than the deep AE-DT, DT, ELM, and deep AE models. Additionally, the deep AE-DT model predicted compressive strength with higher accuracy than non-hybrid models. Thus, it can be stated that innovative hybrid deep models are more advantageous than other models in estimating the compressive strength of ECC. The sensitivity analysis obtained that the PVA fiber was the most significant variable affecting the compressive strength results of nano-silica-modified engineering cementitious composites subjected to high temperatures.

Optimized deep extreme learning machine for traffic prediction and autonomous vehicle lane change decision-making

Deep learning occupies a prominent position in the field of machine learning, where researchers have introduced autoencoder into extreme learning machine (ELM) to establish deep extreme learning machine (DELM), with the aim of enabling autonomous feature learning from data. This approach has found widespread applications in various domains, including finance and electrochemistry. However, the network accuracy of DELM is sensitive to the initial weighting and bias configurations, which can be negatively impacted by random initialization. To mitigate the adverse effects of such random initialization on network performance, this study employs two swarm intelligence optimization algorithms, particle swarm optimization (PSO) and harris hawks optimization (HHO), to optimize the input weights and biases of DELM. In addition, we apply the variational mode decomposition (VMD) technique to decompose the original data, enabling a deeper exploration of latent information within the data. Subsequently, two integrated models are constructed: VMD-HHO-DELM and VMD-PSO-DELM. We apply these models to address road traffic condition prediction and autonomous lane change decision-making problems, with a specific focus on the transportation domain. Through extensive experimentation on two real-world datasets, we observe that deep learning models based on swarm intelligence optimization algorithms exhibit excellent performance across various scenarios. In the context of road traffic condition prediction, our experiments reveal superior results under different time intervals and traffic volumes compared to traditional models, all while maintaining higher operational efficiency. For the autonomous lane change decision experiment, we establish a driver dissatisfaction model that aids in formulating more reasonable lane change decisions, thus reducing the frequency of abrupt lane changes.

A novel deformation monitoring model for high arch dams using impulse response-based equivalent temperature and machine learning-aided separate modeling

The anatomy of service status through health monitoring models is essential to long-term structural safety. Since deformation is the intuitive representation of the dam's operating condition, it is crucial to investigate the deformation monitoring model with high accuracy and strong interpretability for the safety management of high arch dams. A novel deformation monitoring model is proposed by incorporating impulse response-based equivalent temperature (ET) and machine learning-aided separate modeling technique (SMT). This methodology has three main sources of novelty. First, the impulse response-based equivalent temperature and corresponding temperature component are derived from heat transfer and temperature convolution theories. Then, the model parameters are identified by an improved firefly algorithm. Second, the components in the monitored displacement are progressively stripped out by clustering analysis, separation under equal water level conditions, and a robust signal decomposition technique. Third, the separated deformation components and environmental factors are modeled by the multi-output deep extreme learning machine (DELM) with autoencoder, and the monitoring model ET-SMT-DELM is thus established. The world's highest arch dam is selected to illustrate the proposed model, and the prediction accuracy, early warning performance, component shares, and impulse response mechanism are comprehensively investigated by comparing with several typical baseline models. The results show that the overfitting of the proposed model is reduced, and the prediction and early warning performance is significantly improved. The resulting temperature impulse response function and component shares imply that the interpretability of the model is also enhanced.

Prediction model of the large commercial building cooling loads based on rough set and deep extreme learning machine

Efficient prediction of building air conditioning cooling loads contributes to optimizing and controlling heating, ventilation, and air conditioning (HVAC) systems. In this paper, the K-means++ algorithm is used to classify the air conditioning cooling loads to determine the number of discretization intervals and influencing factors of cooling loads. Then, the rough set (RS) theory is used to screen out the critical influencing factors of cooling loads to reduce the input dimensions of the prediction model. Finally, the particle swarm optimization (PSO) algorithm is used to optimize the weights and thresholds between input layers and hidden layers of the deep extreme learning machine (DELM) neural network. Based on the above methods, a K-means++-RS-PSO-DELM prediction model is established for building air conditioning cooling loads. In this paper, a total of 50 large commercial buildings are investigated to find the critical influencing factors of air conditioning cooling loads. Data from a large commercial building in Guilin is collected to train and test the proposed model. Meanwhile, the comparisons between the proposed model and the K-means++-RS-BP, the K-means++-RS-PSO-BP, the K-means++-RS-Elman, the K-means++-RS-PSO-Elman, the K-means++-RS-ELM, the K-means++-RS-PSO-ELM, and the K-means++-RS-DELM models are conducted to validate the applicability of the prediction model. Results show that the prediction deviation of the proposed model is 1.0016 and 0.9942 in short-term and medium-term predictions, which are smaller than that of other prediction models. The proposed model has a strong generalization ability to predict the air conditioning cooling load in large commercial buildings, which is beneficial for online optimal control of the HVAC system.

Gaussian mixture model sample selection strategy–based active semi-supervised soft sensor for industrial processes

Soft sensors have become reliable tools for estimating difficult-to-measure target variables in modern industrial processes. In order to make full use of labeled and unlabeled samples, an active semi-supervised soft sensor modeling method is proposed, which combines active learning and semi-supervised learning to maximize model performance and minimize the laboratory analysis cost of expanding the labeled sample data set. First, manifold regularization is introduced into the deep extreme learning machine (DELM) algorithm to form a semi-supervised DELM that improves the performance of a model trained with unlabeled samples. Then, considering non-Gaussian processes and the error information between the predicted and true values, an active sample selection strategy based on error Gaussian mixture model is developed. Using this strategy, the most uncertain and representative unlabeled samples are selected for labeling, and thereby expanding the labeled sample data set. Finally, the effectiveness of the proposed method is verified using industrial debutanizer process data.

Enhanced Whale Optimization Algorithm with Wavelet Decomposition for Lithium Battery Health Estimation in Deep Extreme Learning Machines

Lithium battery health state estimation can help optimize battery usage and management strategies. In response to the challenges faced by traditional battery management systems in accurately estimating the State of Health of lithium-ion batteries and addressing issues such as capacity recovery and noise interference, this paper proposes a method based on wavelet decomposition and an improved whale optimization algorithm optimized deep extreme learning machine for estimating the SOH of lithium-ion batteries. Firstly, the lithium-ion battery capacity degradation sequence is extracted, and the wavelet decomposition method is used to decompose the battery capacity into global and local degradation trends. Next, the non-linear convergence factor and the whale optimization algorithm with adaptive weights are employed to optimize the deep extreme learning machine for predicting each trend component. Finally, the prediction results are effectively integrated to obtain the lithium-ion battery SOH. This experimental method is validated using NASA and CALCE datasets, and the results indicate that the root mean square error and mean absolute percentage error are both below 0.95%, with relative accuracy and absolute correlation coefficients exceeding 98%. This demonstrates the method’s excellent accuracy and robustness.

State-of-health estimation for lithium-ion battery using model-based feature optimization and deep extreme learning machine

Accurate and reliable estimation of state of health (SOH) is essential for the safe and efficient operation of battery energy storages. However, random charging/discharging behaviors and ambient conditions complicate the online SOH estimation. Aiming at an accurate and robust online SOH estimation for lithium-ion batteries, this paper proposes a SOH estimation method using model-based feature optimization and improved machine learning. First, an empirical model-based voltage reconstruction method is proposed to reconstruct the voltage curve for solving disturbance-free different time (DT) curves under measurement noises and high C-rates and extract model-associated health features (HFs). Then, an optimal charging voltage window (CVW) determination method is proposed, which extracts the CVW-associated HFs by determining the optimal CVW from partial charging voltage ranges. Finally, a set of informative and multi-attribute HFs are extracted to train an improved deep extreme learning machine (DELM) mode for online SOH estimation. The proposed method is validated based on three datasets of batteries with different materials. The results demonstrate high accuracy and strong robustness to partial voltage range, noise corruption, and ambient temperature fluctuation.