The available capacity computation model based on artificial neural network for lead–acid batteries in electric vehicles

doi:10.1016/S0378-7753(99)00502-9

Journal of Power Sources

Volume 87, Issues 1–2, April 2000, Pages 201-204

https://doi.org/10.1016/S0378-7753(99)00502-9 Get rights and content

Abstract

The available capacity computation model based on the artificial neural network (ANN) for lead–acid batteries in an electric vehicle (EV) is presented. Comparing with the methods based on the Peukert equation, which is often used for the calculation of the available capacity for lead–acid batteries in EVs, this model is more accurate. The results of the experiment have proven the accuracy of the proposed model; the computation values are in good agreement with experimental data, the associated error has been considered acceptable from an engineering point of view.

Introduction

In response to the growing concern about energy conservation and environmental protection, the rekindling of interest in electronic vehicles (EVs) has been obvious. Due to its mature technology, lower cost and modest performance, lead–acid batteries are still widely used in the most commercially available EVs. Moreover, the present great improvement in lead–acid batteries has shown that, in the foreseeable future, it is almost impossible for other advanced batteries to replace lead–acid batteries completely in EVs.

However, the calculation of the available capacity of lead–acid batteries is always a tough task. Many empirical expressions have been presented, but only the Peukert equation [1], which describes the relationship between the available capacity (C_a) and discharge current (I_d), has found wide acceptance. It is expressed as: $C_{a} =K/I_{d}^{(n−1)}$ where the constants n and K depend on the temperature, the concentration of the electrolyte, and the structure of lead–acid batteries. Originally the Peukert constants were obtained only using two reference points, which are usually the maximum and minimum, within the whole range of discharge currents (Method-I). Later, multilevel Peukert equation (Method-II) was presented by using three reference points to obtain two sets of Peukert constants [2], and still the least-square method (Method-III) was proposed to estimate the Peukert constants using several reference points [3]. These two methods improve the accuracy of Peukert equation to some extent. In this paper, the available capacity computation model based on the artificial neural network (ANN, Method-IV) for lead–acid batteries in EVs is presented. The results of experiments have proven the further improvement of accuracy with the proposed model. The computation values are in good agreement with experimental data.

Section snippets

Available capacity computation model based on the ANN

With the comparison of the above-mentioned methods based on Peukert equation, here only the discharge current (I_d) need be considered in the available capacity computation model based on the ANN. So the simple configuration of the ANN model is presented and shown in Fig. 1.

In this configuration, the ANN has three layers, i.e., input, hidden, and output layers. The input layer has one node for the discharge current (I_d), the hidden layer has four nodes (this number was determined from studying

Experiment results and comparison

The selection of training pairs is essential to make the ANN achieve better performance. For the training of the proposed ANN model, discharge currents should cover the wide range of currents (e.g., from 0.2 C₅ to 1.0 C₅, here C₅ refers to the rated capacity of battery on a 5-h discharge rate) which are typically discharged in EVs. Fig. 2 shows the relationship between the available capacity and the discharge current of the CS-E105A traction battery installed in the tested EV.

From this figure,

Conclusion

The accurate estimation of the available capacity of battery is very important for EVs when EVs are on the road. An available capacity computation model based on the artificial neural network (ANN) has been proposed. The accuracy of this method has been verified by using the measured data. Comparing with the methods based on Peukert equation, the method based on the ANN gives the highly accurate estimation of the available capacity.

References (4)

W. Peukert
An equation for relating capacity to discharge rate
Electrotech, Z.
(1897)
S.-K. Song
State-of-charge measuring method using multilevel Peukert equation
Journal of Power Sources
(1998)

There are more references available in the full text version of this article.

Cited by (180)

Optimisation based 3-dimensional polynomial regression to represent lithium-ion battery's open circuit voltage as function of state of charge and temperature
2022, Journal of Energy Storage
An accurate representation of Open Circuit Voltage (OCV) as a function of State of Charge (SoC) and temperature is a vital component of Battery Management System (BMS). The conventional representation of OCV through either look up table or temperature dependent equations requires enormous memory. In this article a novel approach of representing the OCV with a three dimensional polynomial equation through optimization based regression analysis is presented. The OCV of the cell is obtained from the incremental OCV test data performed at three different temperatures 0 °C, 25 °C and 45 °C. These experimental data of OCV is subjected to optimization based regression algorithms. The present study compares various optimization algorithms for three dimensional regressions through statistical parameters such as RMSE, MAE and Variance. Differential Evolution provides the least RMSE of 1.48 × 10⁻², 1.34 × 10⁻² and 0.37 × 10⁻² for Charging, Discharging and Average OCV with a reduced MAE and variance which is in order of 10⁻⁹ and 10⁻⁵. It also provides optimised output in only 178 iterations which is lower than other algorithms considered in the article. The memory requirement of the proposed methodology of representation is of 30 bytes irrespective of number of data points. The computational time is reduced by ‘3P + 1’ cycles where P is the time taken for computing a power term in an ECU. Therefore, the comparison of the proposed methodology for three dimensional OCV representation with aforementioned methods reveals reduction of memory requirement and computational complexity in BMS.
Differential evolution based regression algorithm for mathematical representation of electrical parameters in lithium-ion battery model
2022, Journal of Energy Storage
Accurate equivalent electric circuit models of a lithium-ion battery, generally consisting of lumped electrical parameters, are of utmost importance for implementing various algorithms in Battery Management System (BMS). The low-cost Electronic Control Unit for BMS application prompts for an accurate representation of these parameters with a reduced memory footprint is significant. This prompted the representation of SoC dependant parameters as polynomial, Gaussian, sum of sine and exponential equations through regression algorithms instead of a lookup table. This article proposes Differential Evolution (DE) based Regression algorithm to represent the experimentally pre-identified parameters as a function of SoC through the aforementioned equations. Based on the experimental test, data points representing the SoC dependency of battery parameters such as average open-circuit voltage, hysteresis, internal resistances, and capacitance have been obtained. These data points are subjected to a proposed algorithm to establish the mathematical relationship between the parameters and SoC. The flexibility of the algorithm has been verified by considering the aforesaid equations. The DE based regression algorithm's performance was compared with various other algorithms in terms of statistical parameters such as RMSE, Mean Average, and Variance. Based on the analysis of the obtained results for various algorithms, the equations developed by DE based regression algorithm have the best fit with experimental data for all the aforementioned equations. The analysis depicts that the eighth order polynomial equation fitted by the proposed algorithm is highly accurate. Various battery models have been validated with the experimental data based on the developed equations for all parameters.
Effects analysis on active equalization control of lithium-ion batteries based on intelligent estimation of the state-of-charge
2022, Energy
In this paper, the characteristics analysis results of lithium-ion battery show that the essence of the inconsistency of lithium-ion battery is State-Of-Charge (SOC) inconsistency. Therefore, the disparity of SOC can be used to describe the battery inconsistency degree and investigate the equalization control strategy of lithium-ion battery. Firstly, Extended Kalman Filter (EKF) algorithm is proposed to estimate the SOC on the basis of the Ampere-Hour integral and open circuit voltage. Then, the Simscape battery model is established to estimate battery parameters, and the HPPC experiment is employed to verify the SOC estimation precision. Moreover, based on the battery SOC inconsistency laws, the battery equalization control strategy is presented. Finally, the experimental bench is set up to validate the effectiveness of the equalization strategies. Simulation results depict that with the active equalization, the dispersion of SOC drops from 9.4% to 0.84% during charging phase, while the dispersion of SOC drops from 7.4% to 4.12% and the dispersion of voltage drops from 8.54% to 0.71% during discharging phase. The experimental results show that the maximum SOC estimation error is about 4%, so this SOC estimation method meets the accuracy requirement; the 6 batteries are fully charged at the same time with charging equalization, the voltage error is within 15 mV and the voltage variance is within 2%; discharging processes of the 6 batteries are over at the same time with discharging equalization, voltage error is below 15 mV and the voltage variance is below 2%. These results indicate that active equalization control in this paper not only can improve cell inconsistency but also can improve the energy utilization of the battery pack in the process of charging and discharging.
Lithium-Ion Battery State of Charge (SoC) Estimation with Non-Electrical parameter using Uniform Fiber Bragg Grating (FBG)
2021, Journal of Energy Storage
Conventional SoC estimation methods mainly rely on electrical parameters such as the current and voltage of the battery. However, recent studies have shown that the non-electrical parameters such as strain and temperature that have non-linear relationships with the battery SoC can be adopted for SoC estimation. In this work, the use of non-electrical parameters for SoC estimation using Deep Neural Network (DNN) was proposed. Fiber Bragg Grating (FBG) sensors are employed for the simultaneous measurement of strain and temperature of the battery. Besides, Pseudohigh-Resolution interrogation (PHRI) method is adopted for demodulating the output spectra to improve the detection accuracy of the small wavelength signals from the FBG sensors. Our findings have shown a great improvement in the FBG signal quality based on a high up-sampling rate, k in the spectral processing using PHRI. This has a significant impact on the performance of the SoC estimation using DNN. In the comparison with SoC estimation based on electrical parameters, the proposed model based on non-electrical parameters has a better estimation performance.
Combining physicochemical model with the equivalent circuit model for performance prediction and optimization of lead-acid batteries
2020, Electrochimica Acta
Lead-acid batteries with the largest markets among all kinds of batteries have advantages of mature technology, low price, good high/low temperature stability, high reliability, high safety and convenient maintenance. Improving their performance such as energy density and cyclability is still active research topic. In this paper, a new modeling method for lead-acid batteries through combining the physicochemical model with the equivalent circuit model has been developed, which can broaden the application of the two models. The physicochemical model is used to replace the parameter identification experiment of the equivalent circuit model, which reduces the experimental cost. The physicochemical model is used to replace the calculation module of the electromotive force and the state of charge of the circuit model, which simplifies the equivalent circuit model and reduces its calculation time. A new empirical formula of equivalent circuit parameter is proposed based on the physicochemical model, the average relative error of the simulation results is reduced significantly. The results suggest that this paper’s approach can be used for performance prediction and optimization of lead-acid batteries.
Fuzzy Logic Control of Maximum Power Point Tracking Controller in an Autonomous Hybrid Power Generation System by Extended Kalman Filter for Battery State of Charge Estimation
2023, International Journal of Engineering, Transactions A: Basics

View all citing articles on Scopus

View full text