Figure 1 - uploaded by Kai Goebel
Content may be subject to copyright.
Source publication
This paper presents an empirical model to describe battery behavior during individual discharge cycles as well as over its cycle life. The basis for the form of the model has been linked to the internal processes of the battery and validated using experimental data. Subsequently, the model has been used in a Particle Filtering framework to make pre...
Contexts in source publication
Context 1
... process by which ions are transported across the electrolyte from one electrode to another. Figure 1 depicts the typical polarization curve of a battery with the contributions of all three of the above factors shown as a function of the current drawn from the cell. Since, these factors are current-dependent, i.e. they come into play only when some current is drawn from the battery, the voltage drop caused by them usually increases with increasing output current. ...
Context 2
... into the discharge we can predict to within ±2 minutes 45 seconds, and so on. The performance of the PF algorithm for EOL prediction problem is shown in Figure 10. The measured capacity values are shown by the red solid line, the PF tracking by the green patch and the prediction points by the blue asterisks. ...
Similar publications
Accurately predicting the remaining useful cycle life of a lithium-ion battery is essential for health management of battery systems. Aiming at the time-varying and nonlinear problems of lithium-ion batteries, a remaining useful cycle life estimation method based on Takagi-Sugeno fuzzy model is proposed, which not only reduces the amount of data ca...
Citations
... The analysis of predicting the battery system behavior is critical, and the development of a robust model is highly anticipated. The UAV battery system was first studied in [10], and then in various studies such as [11], where remaining flying time was predicted for small electric aircraft; [12], where State of Health (SOH) was analyzed with a Bayesian framework; [13], where comparative Remaining Useful Life (RUL) was estimated; [14], where transient battery behavior was analyzed; [15], where battery impedance was analyzed; [16], where battery capacity was estimated using the Support Vector Regression (SVR) method; [17], where SOH was analyzed using an Artificial Neural Network (ANN) based regression process; [18], where Simulink model was used to analyze SOH; [19], where SOH was predicted using Gaussian Process Regression (GPR) with current and voltage data; [20], where RUL was estimated using the Auto Regression (AR) model; and [21], where an electrochemistry-based Li-Ion battery model was studied. However, there are a few modeling approach available to predict the exact behavior of battery SOH based on capacitive data. ...
Long-distance space systems generate enormous amounts of bigdata. These bigdata can be used to generate intelligent that can help us better understand the behavior of space systems. There is currently no such tool for precisely understanding and predicting the behavior of aerospace systems. In this study, three different aerospace systems are analyzed to build the respective artificial intelligence (AI) models to understand and predict their space behavior using the deep learning (DL) ecosystem. We studied the pulsed plasma thruster (PPT), an electric space propulsion system; the ARTEMIS-P1 spacecraft sensor array; and the UAV battery system. Three sets of comparative analyses are carried out to assess the model accuracy. A number of tests are utilized to assess and predict the exact physical behavior. The comparison and test results show that DL-based artificial models are capable enough (> 99%) to mimic the exact system behaviors. This DL-based approach provides a novel means of understanding and predicting the real behavior of the aerospace systems.
... In addition, empirical models incorporate the battery degradation characteristics and relax the strict constraints on the data requirements as assumed in data-driven models. The filtering methods, such as the extended Kalman filter (EKF) [18] and the particle filter (PF) [10,19] are utilized to update the model parameters. The model-based filtering method incorporates the characteristics of both the battery degradation process and the Bayesian inference methods to deal with uncertainties, which is more efficient and accurate for the real-time prognostics of lithiumion battery systems [20]. ...
... It would be an enormous waste to replace the battery early due to the failure consideration of the regeneration phenomenon. The regeneration phenomenon has been considered to be a self-recharge process and approximated with an exponential function [19,22]. Combined with the Coulombic efficiency as an adaptive parameter, the next charge/discharge cycle of capacity can be denoted by Equation (1). ...
... DH(k) denotes the difference of the SOH between cycle k + 1 and cycle k; Th is a specified threshold to determine whether regeneration happens or not; r(k) = −1 means the regeneration happens, and r(k) = 1 means the opposite. The detection methods avoid modeling the regeneration phenomenon at every cycle and capture the regeneration effects when they are detected, which is sensible compared to Saha's model [19]. However, these detection methods cannot detect the regeneration phenomena in a multi-step ahead manner and are unsuitable for the prediction of the RUL. ...
Accurate prediction of remaining useful life (RUL) is crucial to the safety and reliability of the lithium-ion battery management system (BMS). However, the performance of a lithium-ion battery deteriorates nonlinearly and is heavily affected by capacity-regeneration phenomena during practical usage, which makes battery RUL prediction challenging. In this paper, a rest-time-based regeneration-phenomena-detection module is proposed and incorporated into the Coulombic efficiency-based degradation model. The model is estimated with the particle filter method to deal with the nonlinear uncertainty during the degradation and regeneration process. The discrete regeneration-detection results should be reflected by the model state instead of the model parameters during the particle filter-estimation process. To decouple the model state and model parameters during the estimation process, a dual-particle filtering estimation framework is proposed to update the model parameters and model state, respectively. A kernel smoothing method is adopted to further smooth the evolution of the model parameters, and the regeneration effects are imposed on the model states during the updating. Our proposed model and the dual-estimation framework were verified with the NASA battery datasets. The experimental results demonstrate that our proposed method is capable of modeling capacity-regeneration phenomena and provides a good RUL-prediction performance for lithium-ion batteries.
... The decreasing phase coincides with the capacity "knee" point [43]. Thus, CE is an indicator of the side reactions taking place within the battery [44]. In addition, it can give an indication of the capacity fade except when applied in the early stages; however, this is not a concern when considering second-life batteries that already have an aging history. ...
... On the other hand, the CCDT is lower for the electrically aged cells (A1, A2) compared to the non-electrically aged cells (F1, F2). The CCCT and CCDT decrease with aging [20,21,45,46], which is due to both IRI, which leads to an increase in the polarization of the cell upon charging [44], reaching higher voltages faster, and LAM, which contributes to reach the cell cut-off voltage faster [20,21]. From the results obtained, therefore, it can be assumed that A1 and A2 are characterized by higher LAM and IRI. ...
Upon reaching certain limits, electric vehicle batteries are replaced and may find a second life in various applications. However, the state of such batteries in terms of aging and safety remains uncertain when they enter the second-life market. The aging mechanisms within these batteries involve a combination of processes, impacting their safety and performance. Presently, direct health indicators (HIs) like state of health (SOH) and internal resistance increase are utilized to assess battery aging, but they do not always provide accurate indications of the battery's health state. This study focuses on analyzing various HIs obtained through a basic charging-discharging cycle and assessing their sensitivity to aging. Commercial 50 Ah pouch cells with different aging histories were tested, and the HIs were evaluated. Thirteen HIs out of 31 proved to be highly aging-sensitive, and thus good indicators. Namely, SOH upon charging and discharging, Coulombic efficiency, constant current discharge time, voltage relaxation profile trend, voltage-charge area upon discharging, hysteresis open circuit voltage HIs, and temperature difference between the tabs upon charging. The findings offer valuable insights for developing robust qualification algorithms and reliable battery health monitoring systems for second-life batteries, ensuring safe and efficient battery operation in diverse second-life applications.
... The proposed methodology is validated using the B5, B6, and B7 batteries from the NASA dataset [16]. The operating conditions of the batteries are detailed in Table 2. Figure 3 illustrates the capacity degradation curves for each battery. ...
Predicting the remaining useful life (RUL) of lithium-ion batteries is of paramount significance for the lifespan management of electric vehicles. This study delves into a feature extraction-based approach for indirectly predicting RUL in lithium-ion batteries, circumventing the limitations posed by the challenge of measuring capacity during battery operation. Initially, the health feature (HF) exhibiting strong correlations with capacity is extracted from localized charging voltage intervals, and capacity estimates are obtained through polynomial fitting. Subsequently, a dynamic long short-term memory (D-LSTM) neural network is established, iteratively incorporating prediction outcomes into training data to continuously enhance the model and ultimately accomplish RUL prediction. Experimental results conducted on batteries with varying initial states from the NASA dataset highlight the substantial accuracy and robustness of the proposed D-LSTM method. It outperforms other commonly utilized models, offering the capability for online RUL prediction.
... Although it ranks poorly in Table 5, and is rarely used in other space applications, LCO is often used in NASA missions. The possible reasons for this may include a customized battery cell, maturity of the technology, optimized anode resisting lithium-plating at low temperature, better stability, etc. [45]. For satellite missions, both LiPo and Lithium-ion cells are being considered depending on the features of interest [46,47]. ...
Untethered robots carry their own power supply in the form of a battery pack, which has a crucial impact on the robot’s performance. Although battery technologies are richly studied and optimized for applications such as electric vehicles, computers and smartphones, they are often a mere afterthought in the design process of a robot system. This tutorial paper proposes criteria to evaluate the suitability of different battery technologies for robotic applications. Taking into consideration the requirements of different applications, the capabilities of relevant battery technologies are evaluated and compared. The tutorial also discusses current limitations and new technological developments, pointing out opportunities for interdisciplinary research between the battery technology and robotics communities.
... The authors used the finite element method to establish the mechanical characteristics of the wing of the firefighting aircraft in a specific environment and uses the machine learning method to find the optimal flight strategy. Saha and Goebel (2009) introduced a method based on particle filter to estimate the degradation of key battery parameters in real time. ...
Dynamic evolution is the most typical feature of a digital twin, making it different from a traditional digital model. Dynamic evolution is also the core technology for building equipment digital twins because it ensures consistency between physical space and virtual space. This paper proposes a dynamic evolution framework for black box equipment digital twins. The framework consists of three main parts: data acquisition and processing, an evolution triggering mechanism and an evolution algorithm. A formal description of the dynamic evolution of a black box digital twin is also given. Furthermore, by synthetically considering the computational accuracy and efficiency, we design an incremental external attention temporal convolution network (IExATCN) model to instantiate the proposed framework. Finally, the significance of digital twin dynamic evolution and the effectiveness of the IExATCN is verified by 3D equipment attitude estimation datasets.
... Meru et al. [17] employed the support vector machine (SVM) [18,19] method to predict the battery RUL and achieved an accurate estimation for multiple batteries simultaneously. Other available methods, such as the probabilistic neural network (PNN) [20], recurrent neural network (RNN) [21], particle filtering (PF) [22][23][24], extended Kalman filter (EKF) [25,26], adaptive EKF (AEKF) [27,28], support vector regression (SVR) [11], relevance vector machine (RVM) [29], principal component analysis (FCA), etc., have been widely used recently to prognosticate battery health state and investigate battery degradation behavior. Long short-term memory (LSTM) [30] has recently become a hot data-driven research method, as battery SOHs and RULs are truly time-series research problems. ...
The development of a machine-learning method with high accuracy, high generalization, and strong robustness for evaluating battery health states is essential in the field of battery health management. In this work, the data-driven stacking regressor (SR) method with a two-layer diagnostic framework was proposed to estimate the state of health (SOH) and predict the remaining useful life (RUL). Five individual estimators were merged in the first layer, including bagging, gradient boosting regression (GBR), support vector regression (SVR), Hist-GBR, and AdaBoost, and linear regression (LR) was used in the second layer to construct the SR model. The SR model produces highly accurate results without the requirement of excessive parameter adjustment. Fifteen batteries from the NASA dataset were used for our experiments, resulting in rather low values of average root mean square error (ARMSE) and relative error (RE) for the SOH estimation and RUL predictions of the different batteries, demonstrating the superiority of the SR model.
... Furthermore, the characterization cycles were performed regularly to obtain the battery capacity values. 27 As it can be observed from Figure 2(a), the true EOL is difficult to define due to the capacity regeneration phenomenon during relaxation periods. 27 In order to execute the prognostics experiments, the minimum capacity value C min of each battery cell is defined as its EOL. ...
... 27 As it can be observed from Figure 2(a), the true EOL is difficult to define due to the capacity regeneration phenomenon during relaxation periods. 27 In order to execute the prognostics experiments, the minimum capacity value C min of each battery cell is defined as its EOL. ...
... 7,12,29,30 In this paper, considering the two selected battery capacity degradation datasets, the empirical ageing model shown in equation (9) is employed to describe the capacity degradation trend. 27 Hence, the rest is around how to take advantage of this degradation model to execute degradation extrapolation and its uncertainty assessment for the predictive maintenance. ...
The predictive maintenance is a major challenge for improving battery safety without compromising performance. Its main objective is to predict the end-of-life (EOL) and assess the associated prediction uncertainty. In this paper, we consider this issue and propose a hybrid method combining empirical mode decomposition (EMD) and particle filter (PF) to perform the battery early EOL prediction and its uncertainty assessment. The proposed approach is tested on the two most widely accessible public lithium-ion battery degradation datasets from the Prognostics Center of Excellence at NASA Ames and the Center for Advanced Life Cycle Engineering at the University of Maryland. To avoid the overfitting, a major constraint in the prediction phase, only the residual sequence obtained after EMD decomposition of the raw data is used as the measurement input for PF. Besides, the sum of standard deviation of all intrinsic mode functions (IMFs) is used to determine the noise characteristic for the PF algorithm. Extensive experiments are conducted to show the importance of uncertainty assessment for the early EOL prediction. Although the distance between the true EOL and the mean predicted EOL has no obvious decrease when more operation data is available, the results show a clear decreasing trend of EOL prediction uncertainty when the prediction starts from later operation cycles. Particularly, a strong experimental support is provided for filtering-based prediction methods in the early EOL prediction.
... However, the dependence on impedance measurement hinders the use of this prediction method in practical applications, which is mainly due to equipment costs, strict measurement requirements and space constraints. Saha and Goebel [69] constructed an empirical computing power model that addresses the Coulomb efficiency factor and relaxation effects without the usage of EIS equipment. The particle filter is used to extrapolate future capacity values for RUL estimates and to estimate the value of battery model components. ...
Lithium-ion battery remaining useful life (RUL) is an essential technology for battery management, safety assurance and predictive maintenance, which has attracted the attention of scientists worldwide and has developed into one of the hot issues in battery systems failure prediction and health management technology research. This paper focuses on developing a Lithium-ion battery remaining practical life prediction algorithm to improve its adaptability and accuracy. To achieve this goal, the fusion model methods based on data-driven, model-driven and the combination of the two are summarized, and the problems they face are discussed. Accurate estimation of the remaining life of lithium batteries not only allows users to obtain battery life information in time, replace batteries that are about to fail, and ensure the safe and efficient operation of the battery pack but also ensures that lithium-ion batteries are used as the primary energy supply and energy storage to a large extent. The safety and reliability of the equipment in its operation avoid accidents and reduce operating costs. It focuses on the methods and research status of lithium-ion battery remaining life prediction at home and abroad and the main factors affecting battery life and prediction accuracy. In this paper, the advantages and limitations of various prediction methods are summarized and compared, the current technical research difficulties are outlined, the urgent problems to be solved, and the development trend of battery life prediction technology research are given.
... Generally, data driven techniques are employed in the early phase, plausibly offline, to capture and learn the damage progression using machine learning techniques that lead to statistically derived mathematical models that fit the given degradation database. Some notable works include: non-linear least square regression [67], relevance vector machine regression [72], DM approximated by a linear part and logarithmic/exponential part [73], residual based statistical DM [74], end of life (EOL) in lithium-ion batteries [75], battery health management [72], estimation-prediction of crack growth [76], fuel cell prognostics [74], prognostics of pneumatic valves [77], mechatronic systems [78] and system level prognostics [79]. ...
Unpredicted failure of barrels is a common problem for cannons, artilleries, and other ballistic missiles. The major cause of the barrels’ failure is high pressure, which generates large residual stresses in the inner layers of the barrel. To reduce the stresses, autofrettage of the barrels is carried out in which the core material is allowed to deform plastically and the upper layers produce a continuous elastic force over the inner layers. Autofrettage is a common engineering practice which is carried out while manufacturing of barrels to increase the operational life significantly. The other reason for barrel failure is material degradation which occurs due to thermal fluctuations, fatigue, wear, erosion, corrosion, etc.In order to correlate them with the life of the gun barrels, various tools such as borescope, variety of gauges, optical bore-mapping, ultrasonic sensors, and diamond indenter-based instruments are used. The measured data is then used to formulate the empirical mathematical relation which is further used to calculate the remaining useful life of the barrels. The other mathematical models are the prognostics models which are generally Probabilistic models or Data-driven models (such as machine learning) or Hybrid models. By using these models, we estimate the remaining life of barrels by using any of the techniques such as heat emission method, dimension increment method, muzzle velocity method, and strain-based methods. The paper concisely summarize various techniques used for diagnostic and prognostic of barrels to estimate the degradation profile and to calculate the remaining useful life (RUL).
