Accurate estimation of battery parameters such as resistance, capacitance, and open-circuit voltage (OCV) is absolutely crucial for optimizing the performance of lithium-ion batteries and ensuring their safe, reliable operation across numerous applications, ranging from portable electronics to electric vehicles. Here, we present a novel approach for estimating parameters that combine the two RC equivalent models with the variational and logistic map cuckoo search (VLCS) algorithm. To accurately estimate the parameters of a battery, an experimental setup is designed to carry out a range of tests under controlled laboratory operating conditions. These tests include the Hybrid Pulse Power Characterization (HPPC), OCV, and capacity tests. The OCV test helps to establish the relationship between the state of charge and the OCV, while the HPPC test provides a variable schedule of ‘C’-rates, which allows for a better understanding of the battery’s behavior under different load conditions. The result of the experiment shows that the proposed establishment is effective to accurately determining parameters under different C-rates. After performing a comparative analysis, it is found that the VLCS algorithm outperforms in contrast to standard algorithms such as genetic algorithm, particle swarm optimization, and cuckoo search algorithm. The algorithm mitigates voltage variation between experimental and simulation results, resulting in an approximate error percentage of 0.23%. The root mean square error is employed as a performance indicator, which demonstrates the superiority of the proposed approach. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

The state of charge (SoC) of a battery is a crucial monitoring indicator for battery management systems and it helps to assess how much further an electric vehicle can travel. This work proposes a novel approach for predicting battery SoC by developing a closed-loop system that integrates a bidirectional long short-term memory neural network with an innovative algorithm-extended Kalman filter. A second-order equivalent circuit model is selected, and its parameters are computed using the variational and logistic map cuckoo search approach. Further, an Extended Kalman filter is combined with an innovation algorithm to update process noise in real-time, and a bidirectional long short-term memory neural network takes the input from the Extended Kalman filter and gives the compensated error value for the final SoC estimation.

Lithium-ion batteries are an intuitive choice for electric vehicles and many other gadgets. Parameters play a critical role in addressing its performance characterization. Accurate parameter estimation and real-time monitoring of lithium-ion batteries are important in modeling equivalent circuits. The characteristics of lithium-ion batteries are dynamic due to energy storage. Dynamical behavior is characterized by RC equivalent models. This work presents the estimation of parameters associated with the n-RC equivalent circuit model in integration with the Improved Cuckoo Search Algorithm (ICSA). To get it, battery tests such as HPPC test, static capacity test, and open circuit voltage test in consideration of temperatures are carried out. The experiments are carried out under different temperature ranges to record the valid data sets. ICSA is advantageous over existing algorithms in estimating the battery parameters under temperature ranges. The performance of the proposed approach captures and estimates the parameters in the dynamic range of temperatures of the lithium-ion battery. The error profile is addressed with the root mean square error and it is found to be 0.23 % at 30 °C. It is observed that experimental data with ICSA accurately matches the simulated model data at different temperature ranges

Predicting battery capacity is essential for enhancing battery management systems (BMSs), ensuring safety, and extending battery life. However, lithium-ion battery faces the challenge of performance degradation over the period due to electrochemical phenomena. It can be addressed with data-driven techniques to estimate the battery capacity and remaining useful life (RUL). The machine learning (ML) algorithm efficacy directly impacted by the data types. NASA and CALCE datasets are used to validate the applicability of ML algorithms. The dataset are divided into training and testing sets based on charged-discharged cycles. Pretraining datasets are tested in time series with forward prediction judgment of the data size to predict RUL. There may be an overfitting or underfitting problems in estimating capacity of the battery. However, such problems can be addressed with proper tuning of hyperparameters in time series model with, number of trees, maximum depth of the tree and splitting the data points. As data may be noisy or nonlinear, in most cases, RF prevents overfitting or underfitting by building multiple decision trees which reduces the variance and increases accuracy. RF achieves comparable prediction accuracy, even when trained on limited data as compared to existing data-driven techniques in terms of error metrics RMSE, MSE, MAPE, and R2. The findings highlight RF as a preferred choice with an average RMSE error reduced to 5.66E-16 and predict the battery RUL maximum error in four cycles to lithium-ion battery. These techniques may provide robustness to BMS in real-time applications. This is a preview of subscription content, log in via an institution

It has been observed that the Electric vehicles production is increasing rapidly as it will be a substitute for the usage of fossil fuels. The objective is to minimize usage of fossil fuels to reduce pollution as well as the cost of refueling the vehicle with efficient and reliable chargers. The Charging of an Electric Vehicle is still an open problem regarding its performance. Hence, the charging technologies and their improvements can be tailored by increasing the efficiency of battery charging and adjustable charging voltages. A DSP controller-based off-board battery charger with a three-stage AC-DC charger is presented in the study. To serve it, a step-up AC-DC converter with power factor correction for line regulation and a full-bridge phase-shifted DC-DC converter is placed in the circuit with a transformer and a full bridge rectifier at the output end. The addressed scheme is having external feedback from the battery to a DSP controller that senses battery voltage and charges the battery accordingly. The presented simulated and hardware results are verified and compared to conventional chargers. The results are attaining the charging efficiency up to 94% at different voltage levels.

This article proposes the shunt compensation capability improvement using a deep belief learning network approach (DBLN) with anti-wind-up regulator-supported distributed static compensator (DSTATCOM). Six subnets make up this proposed DBLN controller. Three subnets for each active and reactive mass part are employed to isolate the basic component of the output current. Numerous issues such as past and normalising weight and learning rates are engaged in the DBLN-based weight-updating formula to have a superior dynamic presence, reduce the computational load and achievesfaster estimation, etc. This proposed DBLN is suggested for both proportional-integral (PI) and anti-wind-up regulator to showcase the better DC link voltage which further leads to providing better PQ improvement. This method offers excellent dynamics and resilience to outside disturbances. The suggested study is examined by simulation and experimental development using MATLAB/Simulink by a real-time interface based on a dSPACE 1104 for healthier potential regulation, potential balancing, input current harmonic distortion and PF correction under different load scenarios. © 2024 Informa UK Limited, trading as Taylor & Francis Group.

The Distributed Static Compensator (DSTATCOM) is being recognized as a shunt compensator in the power distribution networks (PDN). In this research study, the superconducting magnetic energy storage (SMES) is deployed with DSTATCOM to augment the assortment compensation capability with reduced DC link voltage. The proposed SMES is characterized by a DC-DC converter with different circuit elements like one inductor, two diodes and two insulated gate bipolar transistors. The Deep Bayesian Active Learning algorithm is suggested to operate SMES supported DSTATCOM for the elimination of harmonics under different loading scenarios. Apart from this, the other benefits like improvement in power factor, load balancing, potential regulation are attained. The simulation studies obtained from the proposed method demonstrates the correctness of the design and analysis compared to the DSTATCOM. To show the power quality effectiveness, balanced and unbalanced loading are considered for the shunt compensation as per the guidelines imposed by IEEE-519-2017 and IEC- 61000-1 grid code. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

Concerning the power quality issues in the power distribution network due to load uncertainties and improper impedance matching of the inductances, deep reinforced learning (DRL)-based inductively coupled DSTATCOM (IC-DSTATCOM) is proposed. First, by analyzing the impedance matching principle, the expression of source, load and filter current is derived with the help of inductive filtering transformer. And second, an individual DRL subnet structure is accumulated for each phase using mathematical equations to perform the better dynamic response. A 10-kVA, 230-V, 50-Hz prototype direct coupled distributed static compensator (DC-DSTATCOM) and IC-DSTATCOM experimental setup is buit to verify the experimental performance under uncertainties of loading. The IC-DSTATCOM is augmented better dynamic performance in terms of harmonics curtailment, improvement in power factor, load balancing, potential regulation, etc. The benchmark IEEE-519-2017, IEC-61727 and IEC-61000-1 grid code are used to evaluate the effectiveness of the simulation and experimental study. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

Estimating battery parameters is essential for comprehending and improving the performance of energy storage devices. The effectiveness of battery management systems, control algorithms, and the overall system depends on accurate assessment of battery metrics such as state of charge, state of health, internal resistance, and capacity. An accurate estimation of the battery parameters is a key challenge in the battery management system due to its nonlinear characteristics. The primary objective of this work is to provide a comprehensive, understandable overview of the existing key issues, methods, technical challenges, benefits, and emerging future trends of the battery parameter estimation. This work presents different parameter estimation approaches, including conventional and modern techniques, to characterize the battery. The comparative analysis has been carried out for techniques improvised by different methods. The benefits, drawbacks, and prospective features of each method are discussed in the work along with recent advancements and their future directions

This paper discusses the novel parameter estimation approach by incorporating the Cuckoo Search Algorithm (CSA) to estimate the internal parameters of the battery occurring inside it due to chemical reactions. Several tests such as an open circuit voltage test and hybrid pulse power characterization test are conducted at variable C rates under three different temperature conditions to validate the approach. The Simulink model is developed to obtain the equivalent circuit model parameters. The objective function is taken as experimentally measured voltage and estimated voltage, the RMSE voltage error magnitude is found to be 0.28%

This work presents the novel maximum power point tracking (MPPT) approach for a small 50 W photovoltaic (PV) system using the DC-DC converter. The method of modeling of PV module is discussed. The firefly (FFY) algorithm and the perturbation and observation (P&O) algorithm are combined to implement MPPT of the PV system connected to battery load. The operating principle is discussed in detail and steady-state analysis of the proposed system is implemented through simulation. The charging profile of the 7.5 Ah VRL battery is also studied using simulation. Furthermore, a low-cost microcontroller-based experimental setup rated at 50 W system connected to battery load was built to implement a hybrid FFY-P&O algorithm. The experimental results are in same as the simulation result. In contrast to the traditional P&O approach, it demonstrated the quick and efficient maximum power point operation triggered by a sudden transition in the environment.

The elemental parts of nonlinearities associated with MEMS devices which were modeled by distributed-parameter equations. The numerical computation of fidelity model derived from discretization scheme creates a large number of ODEs which demands the huge computational efforts. The dynamics of MEMS device reduced with a Galerkin—POD approach, which will suffer their own drawbacks in nonlinearity computation. Therefore, improved strategy over the POD for MOR intricate. This work addressed the drawbacks of POD for MEMS model specifically the dynamical behavior of center point deflection of a beam and their improvements for efficient simulation in data driven framework.