Operando impedance measurements are required for monitoring batteries in the field. In this work, we present pseudorandom sequences (PRSs) for low-cost operando battery impedance measurements. The quadratic-residue ternary (QRT) sequence is known to possess special properties related to eigenvectors of the discrete Fourier transform (DFT) matrix; it is proven in this article that these properties extend to direct-synthesis ternary (DST) sequences derived from the former sequence. A method is proposed to employ these properties to efficiently compensate for drifts and transients while detecting nonlinearities in operando impedance measurements. Practical considerations, such as the computational load, memory requirements, and choice of measurement parameters, are discussed. An experiment is performed on a commercial Li-ion battery cell during fast charging from 20% to 80% state-of-charge (SOC) to illustrate the feasibility of the proposed technique. The impedance is successfully measured at 20 different SOC levels across a charging time of 35 min. Low-cost hardware requirements, fast measurements, and simple data processing make the method practical for embedding in battery management systems (BMSs).

Single-crystal LiNi1-y-zCozMnyO2 (SC-NCM) materials are emerging as promising alternatives to polycrystalline NCMs by addressing intergranular cracking and suppressing structural degradation associated with phase transitions. However, their inherently large particle size and anisotropic morphology lead to sluggish solid-state lithium-ion transport, resulting in diffusion limitations during electrochemical cycling. To optimise SC-NCM electrodes for enhanced energy and power performance, a comprehensive understanding of how electrode level parameters, such as thickness, porosity, and active material volume fraction, influence transport properties is necessary. In this study, commercial-grade SC-NCM electrodes with areal capacities of 2 and 4 mAh cm-2 were examined using a suite of electrochemical and physicochemical techniques. Electronic, kinetic, and ionic transport limitations were identified in the 4 mAhcm-2 electrode designed for high-energy applications. Experimental data were used to parameterise and validate a physics-based half-cell model in PyBaMM, which was subsequently employed in PyBOP for numerical optimisation. By tuning key design parameters, including electrode thickness and active material volume fraction, the optimised design is predicted to deliver a 23% increase in areal 1C-discharge capacity (reaching 4.07 mAh cm-2). This integrated approach combining modelling and experimental validation provides a pathway for optimisation of SC-NCM electrode architectures to improve energy and power metrics in lithium-ion batteries.

Solar home systems provide low-cost electricity for rural off-grid communities. As access to them increases, more long-term data becomes available on how these systems are used throughout their lifetime. This work analyses a dataset of 1,000 systems across sub-Saharan Africa. Dynamic time warping clustering was applied to the load demand data from the systems, identifying five distinct archetypal daily load profiles and their occurrence across the dataset. Temporal analysis reveals a general decline in daily energy consumption over time, with 77% of households reducing their usage compared to the start of ownership. On average, there is a 33% decrease in daily consumption by the end of the second year compared to the peak demand, which occurs on the 96th day. Combining the load demand analysis with payment data shows that this decrease in energy consumption is observed even in households that are not experiencing economic hardship, indicating there are reasons beyond financial constraints for decreasing energy use once energy access is obtained.

A key challenge with large battery systems is heterogeneous currents and temperatures in modules with parallel-connected cells. Although extreme currents and temperatures are detrimental to the performance and lifetime of battery cells, there is not a consensus on the scale of typical imbalances within grid storage modules. Here, we quantify these imbalances through simulations and experiments on an industrially representative grid storage battery module consisting of prismatic lithium iron phosphate cells, elucidating the evolution of current and temperature imbalances and their dependence on individual cell and module parameter variations. Using a sensitivity analysis, we find that varying contact resistances and cell resistances contribute strongly to temperature differences between cells, from which we define safety thresholds on cell-to-cell variability. Finally, we investigate how these thresholds change for different applications, to outline a set of robustness metrics that show how cycling at lower C-rates and narrower SOC ranges can mitigate failures.

Grid-scale battery energy storage can generate revenue by stacking services across electricity and frequency response markets, yet identifying the lifetime profit-maximising stacking strategy remains challenging. Decisions across services are coupled through shared battery system capacity, constrained by system operator energy management rules, and further shaped by product-specific technical requirements that govern system operation, degradation, and lifetime profitability. This paper presents an ageing-aware receding-horizon framework for co-optimising multi-service stacking that explicitly captures product-specific characteristics and state-of-energy compliance rules. The framework is applied to the Great Britain market, where storage operators can stack electricity trading with multiple dynamic frequency response services procured through the newly introduced 'Enduring Auction Capability' platform under energy management requirements imposed by the National Energy System Operator. Using real market data, we demonstrate that degradation modelling, discount rate, and battery ageing jointly govern both lifetime value and optimal stacking strategy. Accounting for ageing increases lifetime revenue by up to 32

Lithium-ion batteries (LIBs) have an important role in the shift required to achieve a global net-zero carbon target of 2050. Electrode manufacture is amongst the most expensive steps of the LIB manufacturing process and, despite its apparent maturity, optimised manufacturing conditions are arrived at by largely trial and error. Currently, LIB manufacturing plants are controlled to follow the fixed "recipe" obtained by trial and error, which may nonetheless be suboptimal. Moreover, regulating the process as a whole to conform to the set conditions is not widespread. Inspired by control approaches used in other film and sheet processes, we discuss opportunities for implementing real-time process control of electrode-related products, which has the potential to reduce the electrode manufacturing cost, CO2 emissions, usage of resources by increases in process yield, and throughput. We highlight the challenges and significant opportunities of implementing real-time process control in LIB electrode production lines.

Estimating state of health is a critical function of a battery management system but remains challenging due to the variability of operating conditions and usage requirements of real applications. As a result, techniques based on fitting equivalent circuit models may exhibit inaccuracy at extremes of performance and over long-term ageing, or instability of parameter estimates. Pure data-driven techniques, on the other hand, suffer from lack of generality beyond their training dataset. In this paper, we propose a hybrid approach combining data- and model-driven techniques for battery health estimation. Specifically, we demonstrate a Bayesian data-driven method, Gaussian process regression, to estimate model parameters as functions of states, operating conditions, and lifetime. Computational efficiency is ensured through a recursive approach yielding a unified joint state-parameter estimator that learns parameter dynamics from data and is robust to gaps and varying operating conditions. Results show the efficacy of the method, on both simulated and measured data, including accurate estimates and forecasts of battery capacity and internal resistance. This opens up new opportunities to understand battery ageing in real applications.

Non-invasive estimation of Li-ion battery state-of-health from operational data is valuable for battery applications, but remains challenging. Pure model-based methods may suffer from inaccuracy and long-term instability of parameter estimates, whereas pure data-driven methods rely heavily on training data quality and quantity, causing lack of generality when extrapolating to unseen cases. We apply an aging-aware equivalent circuit model for health estimation, combining the flexibility of data-driven techniques within a model-based approach. A simplified electrical model with voltage source and resistor incorporates Gaussian process regression to learn capacity fade over time and also the dependence of resistance on operating conditions and time. The approach was validated against two datasets and shown to give accurate performance with less than 1 % relative root mean square error (RMSE) in capacity and less than 2 % mean absolute percentage error (MAPE). Critically, we show that changes from the open circuit voltage versus state-of-charge function will strongly influence the learnt resistance. We use this feature to further estimate in operando differential voltage curves from operational data.

It is well known that lithium-iron-phosphate (LFP) cathodes exhibit significant voltage hysteresis when open-circuit voltage (OCV) is accessed via charge or discharge. This work focuses on the combined role of graphite and LFP in determining the hysteresis in LFP/graphite Li-ion battery cells. Since many control systems infer state-of-charge (SOC) from cell voltage, voltage hysteresis, which can range in the tens of millivolts for LFP/graphite, can incur large estimation errors. Battery management systems for LFP/graphite cells—already made more complicated by the large voltage plateau of the cathode—need to incorporate effective hysteresis models. Previous work by our group [1] focused on exploring the empirical single-state hysteresis model proposed by Plett [2][3], by comparing predictions from Plett’s model to pseudo-OCV measurements of LFP/graphite cells. It was found that inaccuracies in the model predictions correlated strongly with graphite stage transitions. This work focuses on the alternative hysteresis model of Dreyer [4], which assumes that LFP voltage hysteresis arises from a single spinodal decomposition among the particles that make up the cathode. Dreyer’s mechanistic model is compared to pseudo-OCV hysteresis loops obtained by partially charging LFP-graphite cells across different windows of total cell SOC. These experiments clearly show that the magnitude of hysteresis depends on the SOC range spanned. The largest hysteresis occurs in experiments that span the entire SOC range. Comparing model predictions to the experimental data in Figure 1a-c) shows that Dreyer’s original model does not account for ‘bumps’ in the hysteresis loops and cannot match the lower voltage hysteresis when SOC spans narrow ranges around 50% SOC. Both observations may be attributable to graphite hysteresis, with the bumps coinciding with graphite stage transitions [5][6] and a decreased magnitude of hysteresis occurring when the cell is cycled between graphite stage transitions. These observations highlight the importance of augmenting the Dreyer model to include graphite hysteresis. The Gibbs free energy surface used by the Dreyer model can be modified to account for additional spinodal decompositions. Here, we probe the impact of graphite hysteresis by modifying the Gibbs free energy surface from the symmetric-well for LFP used by Dreyer to a double-well system centered around 30% SOC (Figure 1d,e) to account for one level of graphite staging. Multiple stage transitions in graphite [5][6] are modelled through a multiple well Gibbs free energy functional (Figure 1f). The LFP and graphite phases are modelled collectively by creating a Gibbs free energy functional with the stage transitions in LFP and graphite. The findings further emphasise the importance of accounting for graphite hysteresis in physics-based battery models. Further work will focus on incorporating dynamics into the hysteresis model. [1] Hagopian, Emmanuelle, Charles W. Monroe, and David Howey. “Improving the Relationship between State of Charge, Charge History and Voltage Hysteresis Evolution in First Order Differential Equation Voltage Hysteresis Models.” 245 th ECS Meeting (May 26-30, 2024) . ECS, 2024. [2] Plett, Gregory L. Battery management systems, Volume II: Equivalent-circuit methods. Artech House, 2015. [3] Plett, Gregory L. “Advances in EKF SOC estimation for LiPB HEV battery packs.” Consultant to Compact Power, Inc (2003). [4] Dreyer, Wolfgang, et al. “The thermodynamic origin of hysteresis in insertion batteries.” Nature materials 9.5 (2010): 448-453. [5] Mercer, Michael Peter, et al. “Voltage hysteresis during lithiation/delithiation of graphite associated with meta-stable carbon stackings.” Journal of Materials Chemistry A 9.1 (2021): 492-50 [6] Ovejas, V. J., and A. Cuadras. “Effects of cycling on lithium-ion battery hysteresis and overvoltage.” Scientific reports 9.1 (2019): 14875. Figure 1: a-b) Data from hysteresis experiments using 1/20 C rate for pseudo-OCV measurements on NX energies LFP/graphite 18650 cells, following a clockwise shrinking-loop charge/discharge protocol. a) Data plotted as normalized hysteresis (with 1 and –1 corresponding to the hysteretic voltages obtained during complete charge and discharge, respectively) versus charge. Red bands highlight bumps in the cycle loops, and blue bands highlight the lack of convergence to full charge and discharge curves when the charge-state window is narrower. b) Raw data from panel (a) plotted as voltage versus charge, without renormalization. c) Dreyer model prediction for LFP hysteresis. d-f) intercalated-lithium chemical potentials obtained from different free-energy surfaces: d) Dreyer LFP symmetric two-phase chemical potential [4], e) asymmetric two-phase chemical potential and f) three phase chemical potential. The chemical potentials are modified to account for graphite stage transitions in the Dreyer model. Figure 1

Degradation mechanisms affecting the long-term performance of lithium-ion batteries should be monitored and characterized. Such mechanisms, such as loss of lithium inventory (LLI) or active material, can be translated into parameter variations in electrochemical battery models. Here, a reduced-order model (the equivalent hydraulic model) is considered as it provides a good tradeoff between physical interpretability and complexity. The aim is to detect and characterize degradation, namely, to indicate the parameters subject to change, from standard (dis)charge data. To this end, change indicators (or residuals) are computed by combining a state observer and a local statistical approach. Model parameter changes induce changes in the mean of the residual vector which is asymptotically normally distributed with a specified variance. Degradation detection and characterization is achieved by processing the latter residual by statistical tests relying on log-likelihood ratios between multiple simple hypotheses. Results indicate the long-term changes in the main degradation modes affect battery performance. Most degradation modes considered are active at the 0.1% relative parametric change level, but active material loss reaches the 1% parametric change level over the battery lifetime, and 10% parametric change levels are obtained for sluggish diffusion and impedance rise. We show how the proposed methodology could be a useful alternative to methods based only on parameter identification.

The Python Battery Optimisation and Parameterisation (PyBOP) package provides methods for estimating and optimising battery model parameters, offering both deterministic and stochastic approaches with example workflows to assist users. PyBOP enables parameter identification from data for various battery models, including the electrochemical and equivalent circuit models provided by the popular open-source PyBaMM package. Using the same approaches, PyBOP can also be used for design optimisation under user-defined operating conditions across a variety of model structures and design goals. PyBOP facilitates optimisation with a range of methods, with diagnostics for examining optimiser performance and convergence of the cost and corresponding parameters. Identified parameters can be used for prediction, on-line estimation and control, and design optimisation, accelerating battery research and development.

Parallel connected battery modules are a necessary component of large battery packs to increase the overall pack capacity. Unlike series connected cell systems, conventional parallel battery modules have a sparse sensing arrangement such that the battery management system (BMS) does not measure individual cell currents and temperatures. In this arrangement, single cell failures and current and temperature gradients are liable to go undetected by the BMS. As a result of this, it is important to understand the current and temperature dynamics within a parallel connected cell module, so that packs can be designed to mitigate these failures, and BMS diagnostic tools can overcome the aforementioned sensing challenges. In this work, we combine parallel module cycling experiments with a nonlinear affine model of the system to investigate which parameters contribute the most to thermal gradients, and to derive safety limits on cell-to-cell variability for each parameter. Cycles were performed on a 1S4P module of prismatic LFP cells, recording individual cell currents and temperatures, while adjusting the contact resistance arrangement between each cycle to represent different failure and degradation scenarios. The module was discharged at 0.45C (504 A) to match the power capability and current draw of a 2-hour grid storage module in a large scale storage system. We then built an empirical model of parallel cells that consists of a temperature dependent equivalent circuit model (ECM), and a thermal circuit for each cell. The differential algebraic equation which describes the parallel coupling of the system is resolved using Schur decomposition in order to calculate the individual currents for each cell, so that total module current can be used as the model input, and cell surface temperatures are the outputs. Furthermore, we account for the Arrhenius behaviour of the charge transfer kinetics using a temperature dependent resistance in each cell ECM. The model was then fit to the experimental data from cycling to instil confidence that it accurately reproduces the behaviour of a true grid storage system. Finally, a global sensitivity analysis has been completed to analyse which system parameters have the greatest effect on the thermal gradients, and we derive safety limits on cell-to-cell variability for each parameter intended to serve as design constraints for the system. The sensitivity analysis shows that for prismatic LFP cells, the thermal gradient is primarily dependent on inconsistencies amongst contact resistance and internal cell resistance, although there are a myriad of possible parameter variation scenarios, and so capacity imbalances should also be minimized according to the defined safety limits.

Reaching net zero requires substantial large-scale energy storage systems (LESS) deployment. This strategy poses key challenges, including understanding how different LESS technologies compare in terms of both economic benefits and environmental impact, as well as analysing the complex interactions within and between markets when storage is deployed. To help shed light on these aspects, we investigate how LESS location, rated power, duration, and technology can affect welfare and carbon emissions in the Italian electricity system by modelling the day-ahead and the ancillary services markets. We considered lithium-ion batteries, pumped-storage hydro, and vanadium redox flow batteries. The results show that deploying LESS is always beneficial in the day-ahead market, but ancillary services costs can increase due to spillover effects because these markets run sequentially. Lithium-ion is the technology that yields the best social welfare increase. Location, rated power, and duration significantly impact carbon emissions, with changes ranging from-260 kgCO2 to 190 kgCO2 per MWh traded. These results suggest that LESS can help increase welfare and induce unintended consequences, such as spillovers across markets with a mixed effect on emissions.

Non-invasive parametrisation of physics-based battery models can be performed by fitting the model to electrochemical impedance spectroscopy (EIS) data containing features related to the different physical processes. However, this requires an impedance model to be derived, which may be complex to obtain analytically. We have developed the open-source software PyBaMM-EIS that provides a fast method to compute the impedance of any PyBaMM model at any operating point using automatic differentiation. Using PyBaMM-EIS, we investigate the impedance of the single particle model, single particle model with electrolyte (SPMe), and Doyle-Fuller-Newman model, and identify the SPMe as a parsimonious option that shows the typical features of measured lithium-ion cell impedance data. We provide a grouped-parameter SPMe and analyse the features in the impedance related to each parameter. Using the open-source software PyBOP, we estimate 18 grouped parameters both from simulated impedance data and from measured impedance data from a LG M50LT lithium-ion battery. The parameters that directly affect the response of the SPMe can be accurately determined and assigned to the correct electrode. Crucially, parameter fitting must be done simultaneously to measurements across a wide range of states-of-charge. Overall, this work presents a practical way to find the parameters of physics-based models.

Diagnosing lithium-ion battery degradation is a crucial part of managing energy storage systems. Recent research has explored ultrasonic testing for non-invasive health assessment as an alternative to traditional, time-consuming, electrical-only methods. Assessing the state of health is vital for determining quality at end of 'first' life, with retired batteries at 70-80 % health still holding value for secondlife applications. Over the coming years, tens of GWh of salvaged batteries will hit the market, requiring rapid noninvasive methods to classify retired batteries according to their state of health. This study uses a 64 - element ultrasonic array to obtain mid-band quantitative ultrasound spectroscopy parameters - including mid-band fit, spectral slope, and intercept - from circumferential waves around cylindrical batteries. Thirteen cylindrical cells were used to evaluate the methodology: three pristine and ten retired from the same source. The mid-band fit showed the ability to track the state of charge and discriminate between the state of health levels in accelerated degradation experiments both with pristine batteries, and also with recovered secondlife batteries with unknown historical use. Linear-array ultrasonic transducers, coupled with quantitative spectral parameters, show promise for future non-destructive battery health screening methods, offering valuable insights for the emerging used battery market. An ultrasonic array efficiently evaluates cylindrical battery health using quantitative ultrasound spectroscopy parameters, such as the mid-band fit. This non-invasive approach distinguishes state of charge and state of health levels, showcasing promise for rapid screening of second-life batteries. Valuable insights for the evolving used battery market emerge, making this method crucial for energy storage systems. image

Dispatch of a grid energy storage system for arbitrage is typically formulated into a rolling-horizon optimization problem that includes a battery aging model within the cost function. Quantifying degradation as a depreciation cost in the objective can increase overall profits by extending lifetime. However, depreciation is just a proxy metric for battery aging; it is used because simulating the entire system life is challenging due to computational complexity and the absence of decades of future data. In cases where the depreciation cost does not match the loss of possible future revenue, different optimal usage profiles result and this reduces overall profit significantly compared to the best case (e.g., by 30-50%). Representing battery degradation perfectly within the rolling-horizon optimization does not resolve this - in addition, the economic cost of degradation throughout life should be carefully considered. For energy arbitrage, optimal economic dispatch requires a trade-off between overuse, leading to high return rate but short lifetime, vs. underuse, leading to a long but not profitable life. We reveal the intuition behind selecting representative costs for the objective function, and propose a simple moving average filter method to estimate degradation cost. Results show that this better captures peak revenue, assuming reliable price forecasts are available.

Like many optimizers, Bayesian optimization often falls short of gaining user trust due to opacity. While attempts have been made to develop human-centric optimizers, they typically assume user knowledge is well-specified and error-free, employing users mainly as supervisors of the optimization process. We relax these assumptions and propose a more balanced human-AI partnership with our Collaborative and Explainable Bayesian Optimization (CoExBO) framework. Instead of explicitly requiring a user to provide a knowledge model, CoExBO employs preference learning to seamlessly integrate human insights into the optimization, resulting in algorithmic suggestions that resonate with user preference. CoExBO explains its candidate selection every iteration to foster trust, empowering users with a clearer grasp of the optimization. Furthermore, CoExBO offers a no-harm guarantee, allowing users to make mistakes; even with extreme adversarial interventions, the algorithm converges asymptotically to a vanilla Bayesian optimization. We validate CoExBO's efficacy through human-AI teaming experiments in lithium-ion battery design, highlighting substantial improvements over conventional methods. Code is available https://github.com/ma921/CoExBO.

Grid energy storage can help to balance supply and demand, but its financial viability and operational carbon emissions impact is poorly understood because of the complexity of grid constraints and market outcomes. We analyse the impact of several technologies (Li-ion and flow batteries, pumped hydro, hydrogen) on Great Britain balancing mechanism, the main market for supply-demand balancing and congestion management. We find that, for many locations and technologies, financially optimal operation of storage for balancing can result in higher carbon emissions. For example, the extra emissions associated with a 1 MW 2-hour duration Li-ion battery in winter vary between +230 to -71 kgCO2/h. Although storage enable higher usage of renewables, it can also unlock additional demand leading to greater use of gas. In addition, balancing services alone are presently insufficient for financial viability of storage projects. This work highlights the need for market reform aligning financial incentives with environmental impacts.

Batteries are pivotal for transitioning to a climate-friendly future, leading to a surge in battery research. Scopus (Elsevier) lists 14,388 papers that mention "lithium-ion battery" in 2023 alone, making it infeasible for individuals to keep up. This paper discusses strategies based on structured, semantic, and linked data to manage this information overload. Structured data follows a predefined, machine-readable format; semantic data includes metadata for context; linked data references other semantic data, forming a web of interconnected information. We use a battery-related ontology, BattINFO to standardise terms and enable automated data extraction and analysis. Our methodology integrates full-text search and machine-readable data, enhancing data retrieval and battery testing. We aim to unify commercial cell information and develop tools for the battery community such as manufacturer-independent cycling procedure descriptions and external memory for Large Language Models. Although only a first step, this approach significantly accelerates battery research and digitalizes battery testing, inviting community participation for continuous improvement. We provide the structured data and the tools to access them as open source.