Ryan Ahmed

Reduced-Order Electrochemical Model Parameters Identification and SOC Estimation for Healthy and Aged Li-Ion Batteries Part I: Parameterization Model Development for Healthy Batteries

The current phase in our transportation system represents a paradigm shift from conventional, fossil-fuel-based vehicles into the second-generation electric and hybrid vehicles. Electric vehicles (EVs) provide numerous advantages compared with conventional vehicles because they are more efficient, sustainable, greener, and cleaner. The commercial market penetration and success of EVs depend on the efficiency, safety, cost, and lifetime of the traction battery pack. One of the current key electrification challenges is to accurately estimate the battery pack state of charge (SOC) and state of health (SOH), and therefore provide an estimate of the remaining driving range at various battery states of life. To estimate the battery SOC, a high-fidelity battery model along with a robust, accurate estimation strategy is necessary. This paper provides three main contributions: 1) introducing a new SOC parameterization strategy and employing it in setting up optimizer constraints to estimate battery parameters; 2) identification of the full-set of the reduced-order electrochemical battery model parameters by using noninvasive genetic algorithm optimization on a fresh battery; and 3) model validation by using real-world driving cycles. Extensive tests have been conducted on lithium iron phosphate-based cells widely used in high-power automotive applications. Models can be effectively used onboard of battery management system.

Automotive Internal-Combustion-Engine Fault Detection and Classification Using Artificial Neural Network Techniques

In this paper, an engine fault detection and classification technique using vibration data in the crank angle domain is presented. These data are used in conjunction with artificial neural networks (ANNs), which are applied to detect faults in a four-stroke gasoline engine built for experimentation. A comparative study is provided between the popular backpropagation (BP) method, the Levenberg-Marquardt (LM) method, the quasi-Newton (QN) method, the extended Kalman filter (EKF), and the smooth variable structure filter (SVSF). The SVSF is a relatively new estimation strategy, based on the sliding mode concept. It has been formulated to efficiently train ANNs and is consequently referred to as the SVSF-ANN. The accuracy of the proposed method is compared with the standard accuracy of the Kalman-based filters and the popular BP algorithms in an effort to validate the SVSF-ANN performance and application to engine fault detection and classification. The customizable fault diagnostic system is able to detect known engine faults with various degrees of severity, such as defective lash adjuster, piston chirp (PC), and chain tensioner (CT) problems. The technique can be used at any dealership or assembly plant to considerably reduce warranty costs for the company and manufacturer.

Reduced-Order Electrochemical Model Parameters Identification and State of Charge Estimation for Healthy and Aged Li-Ion Batteries—Part II: Aged Battery Model and State of Charge Estimation

Recently, extensive research has been conducted in the field of battery management systems due to increased interest in vehicles electrification. Parameters, such as battery state of charge (SOC) and state of health, are of critical importance to ensure safety, reliability, and prolong battery life. This paper includes the following contributions: 1) tracking reduced-order electrochemical battery model parameters variations as battery ages, using noninvasive genetic algorithm optimization technique; 2) the development of a battery aging model capable of capturing battery degradation by varying the effective electrode volume; and 3) estimation of the battery critical SOC using a new estimation strategy known as the smooth variable structure filter based on reduced-order electrochemical model. The proposed filter is used for SOC estimation and demonstrates strong robustness to modeling uncertainties, which is relatively high in case of reduced-order electrochemical models. Batteries used in this research are lithium-iron phosphate cells widely used in automotive applications. Extensive testing using real-world driving cycles is used for estimation strategy application and for conducting the aging test. Limitations of the proposed strategy are also highlighted.

Adaptive temperature monitoring for battery thermal management

Battery thermal management is crucial for avoiding disastrous consequences due to short circuits and thermal runaway. The temperature inside a battery (core temperature) is higher than the temperature outside (skin temperature) under high discharge/charge rates. Although the skin temperature is measurable, the core temperature is not. In this paper, a lumped thermal model is considered to estimate the core temperature from skin temperature readings. To take into account uncertainties in thermal model parameters, which are bound to occur as the battery ages, an adaptive closed-loop estimation algorithm called the adaptive Potter filter is derived. Finally, computer simulations are performed to validate the adaptive Potter filters ability to track the skin and core temperatures under high charge/discharge current pulses and model mismatches.

Minimizing battery wear in a hybrid energy storage system using a linear quadratic regulator

A battery-ultracapacitor Hybrid Energy Storage System (HESS) combines the advantages of both Li-ion batteries and ultracapacitors. Li-ion batteries sustain a relatively long electric only driving range but degrade if exposed to high C-rates and large number of cycles. Ultracapacitors are robust, have a quasi infinite cycle life and can sustain highly dynamic power profiles. This paper proposes a HESS Linear Quadratic Regulator (LQR) design to mitigate issues related to battery wear and peak power demands for electric and hybrid electric vehicles. The LQR controller imposes the battery current with a bidirectional power electronic converter that interfaces the battery to the ultracapacitor. The HESS is accurately modeled using experimental battery and ultracapacitor data in conjunction with equivalent circuit models. Simulations are carried out to validate the LQR controller on a UDDS drive cycle. Reduced battery wear is quantified using a spectral analysis of the battery current, which identifies microcycles.

Li-Ion Battery SoC Estimation Using a Bayesian Tracker

Hybrid, plug-in hybrid, and electric vehicles have enthusiastically embraced rechargeable Li-ion batteries as their primary/supplemental power source of choice. Because the state of charge (SoC) of a battery indicates available remaining energy, the battery management system of these vehicles must estimate the SoC accurately. To estimate the SoC of Li-ion batteries, we derive a normalized state-space model based on Li-ion electrochemistry and apply a Bayesian algorithm. The Bayesian algorithm is obtained by modifying Potter’s squareroot filter and named the Potter SoC tracker (PST) in this paper. We test the PST in challenging test cases including high-rate charge/discharge cycles with outlier cell voltage measurements. The simulation results reveal that the PST can estimate the SoC with accuracy above 95% without experiencing divergence. INTRODUCTION Recently, hybrid and electric vehicles have received substantial attention due to their high fuel efficiency, low cost of operation and reduced greenhouse gas emission. At the heart of these vehicles lies rechargeable (secondary) batteries as a source of energy. Specifically, Li-ion batteries have become a more popular choice than NiMH batteries in newer generation hybrid and electric vehicles due to their high energy density, slow selfdischarge, and zero memory effect. A battery management system (BMS) is required to keep the Li-ion cells within their specified operating range by sensing voltage, current, temperature and internal pressure signals. This ensures the availability of reliable electrical power, improves overall energy efficiency, protects cells from damage and prolongs battery lifespan. The battery SoC is one of the most important parameters that BMS estimates in real time to accomplish its goals. The SoC needs to be estimated accurately in a broad range of environmental and operating conditions, including environmental temperature (from freezing cold to scorching hot) and battery age (from new to old). To this end, the following two key elements must be readily available: • An accurate battery model and • A robust SoC estimation strategy. A Brief Review of Battery Modelling Modeling refers to the process of analysis and synthesis to determine a suitable mathematical description that characterizes the relevant dynamics of a component under test. To be useful, the model must be scalable and easy to be simulated. A battery model is required to capture battery physics accurately. The two broad approaches to battery modelling are • Equivalent electrical-circuit modeling and • Electrochemical modeling Equivalent electrical-circuit models use RC (ResistorCapacitor) circuits to model the charge and discharge behavior of Li-ion batteries [15, 24, 10]. They may consist of the first-order, second-order or the third-order RC models coupled with a hysteresis effect [24, 10]. Equivalent circuit models are conceptually simple to understand and use a few parameters to be identified. However, they provide little insight into underlying physical battery limitations. On the other hand, electrochemical modeling uses the first principles– they use partial differential equations to capture the diffusion dynamics of Li-ions in the solid phase composite electrodes and the electrolyte. Although electrochemical modeling provides a more accurate SoC estimate, it comes with a price; it requires many unknown parameters to be identified. However, once an electrochemical model is developed, it can be easily manipulated to meet different battery specifications. For these reasons, electrochemical modeling is often preferred to equivalent electrical-circuit modeling [17, 7]. A Brief Review of SoC Estimation methods Estimating the SoC in an electric vehicle is analogous to gauging fuel in a conventional vehicle. In general, SoC estimation techniques can be broadly categorized into two types:

Modeling, parameterization, and state of charge estimation of Li-Ion cells using a circuit model

This paper presents a general procedure applied for modeling, parameter identification, and state of charge (SOC) estimation of a Li-Ion battery cell. The paper explains a battery tester with a number of experiments conducted to investigate the cell physical properties. Dynamics of the Li-Ion cell are modeled using an equivalent circuit model, whereas parameters of the model are calculated using particle swarm optimization. This method minimizes the output error that is the difference between the simulated output from the model and the measured terminal voltage. The provided equivalent circuit model with optimized parameters was used for SOC estimation. Two different state estimation methods have been applied to estimate the cell SOC based on real-time measurements. The estimation methods include the extended Kalman filter (EKF), and the novel smooth variable structure filter (SVSF). The SVSF method was used as it can produce more accurate state estimates for dynamic systems with modeling and parametric uncertainties. This paper compares the performance of these two estimators for real-time SOC estimation using tester data.

Artificial neural network training utilizing the smooth variable structure filter estimation strategy

Abstract A multilayered neural network is a multi-input, multi-output nonlinear system in which network weights can be trained by using parameter estimation algorithms. In this paper, a novel training method is proposed. This method is based on the relatively new smooth variable structure filter (SVSF) and is formulated for feed-forward multilayer perceptron training. The SVSF is a state and parameter estimation that is based on the sliding mode concept and works in a predictor–corrector fashion. The SVSF training performance is tested on three benchmark pattern classification problems. Furthermore, a study is presented comparing the popular back-propagation method, the extended Kalman filter, and the SVSF.

Comparison of Kalman Filter-based state of charge estimation strategies for Li-Ion batteries

Currently, the automotive industry is experiencing a significant technology shift from internal combustion engine propelled vehicles to second generation battery electric vehicles (BEVs), hybrid electric vehicles (HEVs) and plug-in hybrid electric vehicles (PHEVs). The battery pack represents the core of the electric vehicle powertrain and its most expensive component and therefore requires continuous condition monitoring and control. As such, extensive research has been conducted to estimate the battery critical parameters such as state-of-charge (SOC) and state-of-health (SOH). In order to accurately estimate these parameters, a high fidelity battery model has to work collaboratively with a robust estimation strategy onboard of the battery management system (BMS). In this paper, three Kalman Filter-based estimation strategies are analyzed and compared, namely: The Extended Kalman Filter (EKF), Sigma-point Kalman filtering (SPKF) and Cubature Kalman filter (CKF). These estimation strategies have been compared based on the first-order equivalent circuit-based model. Estimation strategies have been compared based on their SOC estimation accuracy, robustness to initial SOC error and computation requirement.

Fault detection of an engine using a neural network trained by the smooth variable structure filter

A multilayered neural network is a multi-input, multi-output (MIMO) nonlinear system in which training can be regarded as a nonlinear parameter estimation problem by estimating the network weights. In this paper, the relatively new smooth variable structure filter (SVSF) is used for the training of a nonlinear multilayered feed forward network. The SVSF is a recursive sliding mode parameter and state estimator that has a predictor-corrector form. Using a switching gain, a corrective term is calculated to force the network weights to converge to within a neighbourhood of the optimal weight values. SVSF-based trained neural networks are used to classify engine faults on the basis of vibration data. Two faults are induced in a four-stroke, eight-cylinder engine. Furthermore, a comparative study between the popular back propagation method, the extended Kalman filter (EKF), and the SVSF is presented. Experimental results indicate that the SVSF is comparable with the EKF, and both methods outperform back propagation.