Dynamic Load Profile Research Articles

Physics-guided fault diagnosis method for proton exchange membrane fuel cells based on LSTM neural network

In this paper, we propose a novel hybrid method for fault diagnosis of Proton Exchange Membrane Fuel Cells (PEMFCs) based on the combination of a physics-based model and a long short-term memory (LSTM) neural network. By incorporating the physics-based model in the fault diagnosis algorithm, we can access to several process variables not directly measured through sensors but related to the state of the PEMFC stack. The model estimates are subsequently combined with signals measured on the PEMFC stack and inputted to a LSTM neural network. The performance of the physics-guided LSTM is evaluated on an extensive dataset comprising a thousand hours of operation of a PEMFC stack under dynamic load profiles, proving the enhanced capability of the proposed fault diagnosis method in capturing complex fault patterns. Furthermore, the effectiveness of the proposed method in dealing with PEMFC aging is tested by using data from the initial phase of stack operation for algorithm training and reserving the data from the aged stack operation for the testing phase. The experimental results reveal that the proposed physics-guided LSTM method allows for a significant amelioration over purely data-driven LSTM.

Dynamic management of ground thermal response uncertainty through temporal analysis of parameter sensitivity

Accurately estimating the dimensionless ground thermal response factor (g-function) is critical for the design and operation of geothermal energy systems. This process, however, is fraught with uncertainties stemming from the difficulty in controlling on-site geometric parameters and estimating thermal properties through field experiments. The relative contributions of these parameters to response uncertainty change temporally as the response progresses. Thus, understanding this temporal sensitivity variation is crucial for effective response uncertainty management. To address this, we conducted transient global sensitivity analyses. Regardless of borefield setups, ground thermal properties were key factors in thermal response uncertainty. However, their influence decreased from the late-mid term as borehole radius and borehole distance became more significant in single-borehole and multi-borehole setups, respectively. This time-varying relative sensitivity suggests that prioritizing parameters to reduce response uncertainty depends on the temporal characteristics of the ground load. To demonstrate this, we systematically halved the uncertainty for each parameter and examined the resultant change in the fluid temperature uncertainty for a dynamic ground load scenario. Notably, prioritizing parameters with high early-response sensitivity significantly reduced fluid temperature uncertainty. These findings highlight the need for dynamic uncertainty management strategies in geothermal system design and operation. For dynamic load profiles utilizing the ground as a heat source or sink, prioritizing parameters with high sensitivity in the early- to mid-term is crucial. Conversely, for long-term storage applications, focusing on parameters with high sensitivity in the mid- to long-term response becomes essential.

Optimal adaptive coordination of overcurrent relays in power systems protection using a new hybrid metaheuristic algorithm

AbstractAdvent of distributed generation and progression towards an intelligent grid infrastructure within the domain of contemporary electrical power systems have created dynamic load profiles. Accompanying these developments, protective relays are faced with an evolving electrical load landscape and variable fault current conditions, resulting in disparate operational timings throughout the diurnal cycle. In light of these challenges, this paper delineates the formulation and simulation of a novel adaptive protection strategy for overcurrent relays, meticulously tailored to accommodate the fluctuations in electrical load. To construct a robust framework for this adaptive mechanism, a series of hypothetical fault current scenarios are meticulously crafted to activate the relays within the briefest time interval feasible. Further innovating within this sphere, this paper introduces a new hybrid algorithm, deftly amalgamating the strengths of three preeminent metaheuristic models: Improved Harmony Search, Particle Swarm Optimization, and Differential Evolution. Simulations and analyses substantiate the efficacy of the algorithm in optimizing the coordination among overcurrent relays aiming to uphold the overarching protective imperatives of the grid. For the IEEE 6‐bus system, the mean value of the objective function during 24 h in Monte Carlo is 292.6607 and very close to 272.0758 in the simulation of eight stochastic scenarios, which contributes to the validity of the approach in practical settings. Also, in the IEEE 30‐bus system, the results of the mean relay operation time set for the hours with the lowest and highest consumption load are 17.1297 and 14.8049 s, which reveals the increase in the operation speed of the relays.

Enhanced SOC estimation in Lithium-ion Batteries using GRU Neural Networks with GSA

State of charge (SOC) is one of the most important metrics to evaluate the performance of the battery management system (BMS) in electric vehicles (EVs). Lithium-ion batteries have been utilized often for SOC estimates in EV applications because of its attractive characteristics such as extended lifespans, high voltages, energy, and capacities. However, dynamic charging and discharging profiles, material degradation, battery aging, chemical reaction, and temperature fluctuations would diverge the accuracy of SOC estimation. Deep learning has proven to become an efficient method for SOC estimation in EVs due to its strong computational capabilities, enhanced generalization performance, and excellent accuracy under dynamic loading profiles. Therefore, this study proposes the use of the gated recurrent unit neural network (GRUNN) for estimating SOC in lithium-ion batteries. The best hyperparameters of GRUNN that enable improving SOC accuracy are found using gravitational search optimization (GSA). To evaluate the functional sustainability of the suggested method, a full battery test bench model is created and accordingly, robustness of the proposed approach is verified under diverse EV drive cycles and aging impacts. The effectiveness of the proposed approach is evaluated by comparing with several existing approaches. The results demonstrate that the suggested method archives SOC error and RMSE in EV driving cycles and ageing effects below 5% and 2%, respectively. The excellent outcomes obtained by the proposed method would certainly enhance the battery charging and discharging profile and EV performance. GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 9(1), 2022 P 1-12

A deep dive into enhancing frequency stability in integrated photovoltaic power grids

Voltage control strategies (VCS) and frequency stability analysis (FSA) are essential for power system reliability, particularly during high-load periods. Stable voltage and frequency levels prevent malfunction, power quality deterioration, and supply interruptions. Grid operators must skillfully manage VCS and FSA control to ensure system stability. Nonlinear loads, especially under transient conditions, significantly affect voltage stability (VS), introducing harmonics, waveform distortion, and stability complexities. Accurate modeling of these nonlinear loads is vital when traditional static load models fall short. Frequency fluctuations from power generation-demand imbalances require vigilant monitoring and regulation. Effective frequency control mechanisms are indispensable for preserving desired frequencies. Using a Western Algeria case study, this paper underscores FSA's significance in integrating photovoltaic (PV) systems into power grids. It addresses challenges from frequency fluctuations due to dynamic ZIP load profiles, emphasizing the importance of FSA for reliable grid operation. The study offers insights and practical approaches to enhance VS, FSA control, and energy management (EM), improving grid reliability and ensuring uninterrupted power supply. We must look into FSA's benefits in integrating PV systems to improve performance and lower grid interruptions. This includes looking into its control mechanisms and feedback systems.

Five-Port Isolated Bidirectional DC-DC Converter for Interfacing a Hybrid Photovoltaic–Fuel Cell–Battery System with Bipolar DC Microgrids

This paper introduces a novel five-port, three-input, dual-output isolated bidirectional dc-dc converter (FPIBC) topology with an effective controller for power-sharing and voltage-balancing in bipolar dc microgrids (BPDCMGs). The proposed converter acts as the interface for the integration of a hybrid generation system comprising a solid oxide fuel cell (SOFC), a photovoltaic (PV) system, and a battery into BPDCMGs. It employs a reduced number of circuit elements compared with similar multiport converter topologies suggested for BPDCMG applications. Symmetrical bipolar output voltages are ensured by a voltage-balancing circuit composed of a fully controlled switch and four diodes. The FPIBC is equipped with different controllers for output voltage regulation and balancing, power sharing, maximum power point tracking of the PV, the optimum operating region of the SOFC, and constant-current, constant-voltage charging of the battery. To verify the viability and effectiveness of the proposed system, a simulation model was developed with a 4.2 kW SOFC, a 3.7 kW PV, and a 140 V 10.8 Ah battery in MATLAB/Simulink. The performance of the FPIBC was evaluated through extensive case studies with different operational modes, including battery charge/discharge states and SOFC and PV parameter changes under varying load conditions. In addition, the proposed system was examined using a daily dynamic load profile. According to the simulation results, a peak efficiency of 97.28% is achieved and the voltage imbalance between the output ports is maintained below 0.5%. It is shown that the FPIBC has advantages over previous converters in terms of the number of ports, number of circuit elements, bipolar output voltage, bidirectional power flow, and efficiency.

Path‐Dependent Ageing of Lithium‐ion Batteries and Implications on the Ageing Assessment of Accelerated Ageing Tests

AbstractPath dependency in ageing of Lithium‐ion batteries (LIBs) still needs to be fully understood, and gaps remain. For realistic operational scenarios that involve dynamic load profiles, understanding this path dependency is essential for effective monitoring and accurate modelling of LIBs‐ageing. Our research investigates the impact of different ageing sequences on capacity reduction and resistance increase, key metrics for determining the state of health (SOH). Moreover, we argue that relying solely on SOH‐based monitoring is insufficient for predicting the ongoing ageing trajectory. Our findings underscore that recent operational history influences subsequent degradation. This degradation is attributed to the emergence of uneven lithium distribution, which can both induce capacity recovery and amplify degradation during cycling phases. Such insights are particularly interesting for ageing studies where accelerated battery degradation is achieved through continuous cycling, a common approach in most cyclic ageing investigations. We demonstrate that capacity difference analysis (CDA) holds promise in tracking this unevenness and the potential for capacity restoration through re‐homogenisation. In conclusion, our work highlights the importance of utilising advanced tools, such as CDA and degradation mode (DM) analysis, to ensure accurate conclusions are drawn in accelerated LIB‐ageing experiments.

Analysis of voltage control using V2G technology to support low voltage distribution networks

AbstractThe decarbonization of the power generation and transport sector encourage the analysis of connection of distributed energy resources (DER), such as electric vehicles (EVs), to the electrical system, as well as the evaluation of their impact on smart cities. A better understanding of the negative impacts on the power systems will lead to propose mitigation measures and eventually revolutionize the way distributed generation works. This paper aims at modelling and evaluating the impact of EVs on a real distribution network. The energy system chosen operates at 60 Hz, 34.5 kV (medium voltage) and 0.208 kV (low voltage) and it is simulated using PSCAD/EMTDC. To reproduce realistic user consumption profiles, dynamic load profiles based on EV owners behaviour have been simulated. The vehicle‐to‐grid (V2G) technology is modelled to mitigate the impacts of high penetration of EVs by supporting the network from undervoltage. The results show the importance of active management in modern power systems, especially considering the increase in DER penetration expected for the coming years. This work shows the benefits of implementing V2G technology while highlighting the challenges involved in a real case.

Dynamic ultrasonic response modeling and accurate state of charge estimation for lithium ion batteries under various load profiles and temperatures

Accurate estimation of the State of Charge (SoC) plays a vital role in ensuring the efficient and safe operation of lithium iron phosphate (LFP) batteries. However, the flat open circuit voltage (OCV) curve of the LFP battery implies a low sensitivity to SoC, which results in large SoC estimation errors in the presence of noisy terminal voltage measurements. To address this challenge, an SoC estimation methodology utilizing an ultrasonic reflection response model is proposed, which is the first methodology regarding highly accurate and robust ultrasonic model-based SoC estimation under dynamic load profiles. Since ultrasound waves enable non-destructive acquisition of battery internal physical property changes directly associated with SoC, the ultrasonic battery near-surface reflection feature is extracted and demonstrated to exhibit a highly linear correlation with and higher sensitivity to SoC. We pioneeringly construct an empirical differential ultrasonic model to describe how the ultrasonic feature depends on the SoC and dynamic current. The advantage of such an ultrasonic model is demonstrated by theoretical and experimental results of an Adaptive Extend Kalman Filter (AEKF) and an Adaptive H-infinity Filter (AHIF) under various dynamic load profiles and temperatures. The Root Mean Square Error (RMSE) for ultrasonic model-based SoC estimation remains at approximately 1% across all tests, reducing by 36.7% compared to the voltage model, which shows its great potential in accurate SoC estimation.

On the PMU Placement Optimization for the Detection of False Data Injection Attacks

The graph signal processing (GSP) approaches could facilitate the design of detectors for false data injection attacks (FDIAs) on smart grids. However, enhancing the performance of FDIA detection by optimizing the placement of phasor measurement units (PMUs) on smart grids remains an open problem. We formulate herein the detection of FDIAs as a binary hypothesis testing problem within the Grid-GSP framework and develop a subspace-based projection detection (SPD) algorithm, incorporating a quasi-optimal PMU placement strategy and two double thresholds. The proposed quasi-optimal PMU placement strategy enhances the performance of the SPD in FDIA detection. We further propose to enhance the SPD with the dynamic double threshold for high dynamic load profiles beyond a simple yet efficient static double threshold. Illustrative simulations based on the IEEE 118 bus test grid and the ERCOT dataset validate the superior performance of the proposed quasi-optimal PMU placement strategy along with the two double thresholds.

A Combined DNN-NBEATS Architecture for State of Charge Estimation of Lithium-Ion Batteries in Electric Vehicles

In this paper, a new hybrid architecture combining Deep Neural Network (DNN) and Neural Basis Expansion Analysis for Time Series (N-BEATS) is proposed first time for estimating the State of Charge (SoC) of Lithium-ion batteries. The input and target vectors used for training and testing the proposed hybrid architecture is experimentally determined using a low-priced microcontroller. The proposed hybrid architecture is trained under different operating conditions, tested with untrained data, and is shown capable of accurately estimating the SoC. The architecture is also trained and tested with four different dynamic loading profiles. Further, existing and well-established neural networks, namely Long short-term memory (LSTM), Bidirectional long short-term memory (Bi-LSTM), and Gated recurrent unit (GRU), are employed and tested under identical conditions, and results are compared. Extensive computation results are presented for various temperatures and battery chemistries, and comparison highlights the superior performance of the newly proposed hybrid architecture.

A quantitative method for early-stage detection of the internal-short-circuit in Lithium-ion battery pack under float-charging conditions

Detecting the internal short circuit (ISC) of Lithium-ion batteries is critically important for preventing thermal runaway. Conventional approaches mainly focus on ISC detection for dynamic load profiles, while the commonly seen float-charging scenarios with a high risk of ISC are rarely considered. Technical challenges arise from not only the lack of models describing battery dynamics at fully charged conditions but also the computational burden caused by the high number of series-connected cells in a battery pack. To address these issues, we here propose a simple and accurate method to quantitatively identify the leakage current of the battery with ISC, by checking the behaviors of the battery equalization system. Battery-in-the-loop experiments are carried out to verify the proposed method under a wide range of the short circuit resistances (∼20 to ∼500Ω), using both LiFePO4 and Li(NiCoMn)0.33O2 batteries with different aging degrees. The typical errors of the identified leakage currents can be well-bounded within ±1 mA, including the cases of using both passive and active balancing. The proposed method is completely model-free and, therefore, can be transplanted into low-cost embedded systems to support broader applications.

Uncertainty Management in Lebesgue-Sampling-Based Li-Ion Battery SFP Model for SOC Estimation and RDT Prediction

The state-of-charge (SOC) estimation and remaining-dischargeable-time (RDT) prediction based on the simplified first principle (SFP) model under Lebesgue sampling (LS) has advantages of less computation and higher performance. The accuracy, precision, and convergence speed of SOC estimation and RDT prediction are significantly influenced by the model. This article introduces an online dichotomy method for the parameter adaptation of the LS-based SFP model to improve the accuracy and convergence speed. Besides, the uncertainty of RDT prediction is managed by adjusting model noises through a short-term correction loop to improve the precision of RDT prediction. The developed SFP is then integrated with an extended Kalman filter (EKF) in the LS framework for lithium-ion battery SOC estimation and RDT prediction. Experimental results show that the proposed approach can greatly improve the performance in terms of convergence speed, accuracy, and computing efficiency. Compared with the LS-EKF, the convergence speed and RDT prediction accuracy of the proposed method are improved by 98.6% and 47.4% under dynamic loading profiles, respectively. The absolute average error of SOC estimation keeps in the range of 2%.

Robust lithium-ion state-of-charge and battery parameters joint estimation based on an enhanced adaptive unscented Kalman filter

Accurate modeling and state of charge (SOC) estimation of lithium-ion battery against the model uncertainty and data uncertainty are difficult tasks nowadays. In this paper, a model and data uncertainties-robust method is proposed simultaneous estimation of the model parameters and the SOC using an enhanced adaptive unscented Kalman filter (AUKF). An extended state observer is established to integrate all unknown variables including parameters and SOC into a vector. An covariance matching technique with adaptive forgetting factor is proposed to obtain uncertain model and data statistics, in combination with a singular value decomposition based unscented transform to guarantee the positive definiteness of the error covariance matrix. Furthermore, establishing new protocols to handle missing input and missing output separately, the battery SOC and parameters can be estimated from missing measurements. Benefits from above procedures, the proposed method is more robust to model uncertainties and the data uncertainties compared to the conventional SOC estimation method. The robustness of the proposed method is verified at different operation temperatures and dynamic load profiles. The results shows that the proposed method possesses high accuracy and excellent robustness.

Automotive Fuel Cell Systems: Testing Highly Dynamic Scenarios

PEM fuel cell systems face highly dynamic load profiles in automotive application. This work showcases the impact of media supply adaption, system architecture and test rig restrictions on the transient voltage response of an automotive fuel cell stack. Current step and load profile experiments were conducted on a system test rig, featuring automotive balance of plant components, and a short stack test bench. A time scale analysis allowed us to identify the predominant effect for the voltage response in each test case. The voltage response measured in the test cases was dominated either by air supply, membrane humidification or coolant temperature dynamics. This systematic comparison of different types of test setups highlights the importance of application-like system level testing as, in contrast to common experiments, different phenomena shape the electrical stack behavior.

Comprehensive co-estimation of lithium-ion battery state of charge, state of energy, state of power, maximum available capacity, and maximum available energy

In developing an efficient battery management system (BMS), accurate battery state estimation is always required. However, the trade-off between computational efficiency and accuracy of state estimation is hard to maintain. This work proposes the comprehensive co-estimation method for battery states, maximum available capacity, and maximum available energy estimation. The existing correlation between different battery states is effectively utilized to achieve high accuracy and reduce the computational burden. A combined state of charge (SOC) and state of energy (SOE) estimation using the dual forgetting factor adaptive extended Kalman filter (DFFAEKF) algorithm and experimental quantitative relations between SOC and SOE are utilized to estimate the SOC and SOE. Due to low computational cost and simplicity, the multiple constraints model-based SOP estimation using the Rint model is employed. The maximum available capacity and maximum available energy estimation are performed using a new sliding window-approximate weighted total least square (SW-AWTLS) algorithm. The performance of the proposed co-estimation method is validated by two different chemistry battery cells under dynamic load profiles at different operating temperatures. Moreover, the comparison with other existing co-estimation methods is also conducted, whose results indicate the superior accuracy of the proposed comprehensive co-estimation method.

Comparative Study of Parameter Identification with Frequency and Time Domain Fitting Using a Physics-Based Battery Model

Parameter identification with the pseudo-two-dimensional (p2D) model has been an important research topic in battery engineering because some of the physicochemical parameters used in the model can be measured, while some can only be estimated or calculated based on the measurement data. Various methods, either in the time domain or frequency domain, have been proposed to identify the parameters of the p2D model. While the methods in each domain bring their advantages and disadvantages, a comprehensive comparison regarding parameter identifiability and accuracy is still missing. In this present work, some selected physicochemical parameters of the p2D model are identified in four different cases and with different methods, either only in the time domain or with a combined model. Which parameters are identified in the frequency domain is decided by a comprehensive analysis of the analytical expression for the DRT spectrum. Finally, the parameter identifiability results are analyzed and the validation results with two highly dynamic load profiles are shown and compared. The results indicate that the model with ohmic resistance and the combined method achieves the best performance and the average voltage error is at the level of 12 mV.

Spark Ablation for the Fabrication of PEM Water Electrolysis Catalyst-Coated Membranes

Proton-exchange-membrane (PEM) electrolyzers represent a promising technology for sustainable hydrogen production, owing to their efficiency and load flexibility. However, the acidic nature of PEM demands the use of platinum-group metal-electrocatalysts. Apart from the associated high capital costs, the scarcity of Ir hinders the large-scale implementation of the technology. Since low-cost replacements for Ir are not available at present, there is an urgent need to engineer catalyst-coated membranes (CCMs) with homogeneous catalyst layers at low Ir loadings. Efforts to realize this mainly rely on the development of advanced Ir nanostructures with maximized dispersion via wet chemistry routes. This study demonstrates the potential of an alternative vapor-based process, based on spark ablation and impaction, to fabricate efficient and durable Ir- and Pt-coated membranes. Our results indicate that spark-ablation CCMs can reduce the Ir demand by up to five times compared to commercial CCMs, without a compromise in activity. The durability of spark-ablation CCMs has been investigated by applying constant and dynamic load profiles for 150 h, indicating different degradation mechanisms for each case without major pitfalls. At constant load, an initial degradation in performance was observed during the first 30 h, but a stable degradation rate of 0.05 mV h−1 was sustained during the rest of the test. The present results, together with manufacturing aspects related to simplicity, costs and environmental footprint, suggest the high potential of spark ablation having practical applications in CCM manufacturing.

A Model Fusion Method for Online State of Charge and State of Power Co-Estimation of Lithium-Ion Batteries in Electric Vehicles

In this paper, a model fusion method (MFM) is proposed for online state of charge (SOC) and state of power (SOP) co-estimation of lithium-ion batteries (LIBs) in electric vehicles (EVs). Firstly, a particle swarm optimization-genetic algorithm (PSO-GA) method is cooperated with a 2-RC <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CPE</sub> fractional-order model (FOM) to construct battery open-circuit voltage (OCV)-SOC curve, which only relies on a part of dynamic load profile without the prior knowledge of an initial SOC. Secondly, a dual extended Kalman filter (DEKF) algorithm based on a 1-RC model is employed to identify the model parameters and estimate battery SOC with the extracted OCV-SOC curve. Furthermore, battery polarization dynamics in a SOP prediction window is analyzed from two aspects: (1) self-recovery; and (2) current excitation. They are separately simulated using 2-RC <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CPE</sub> FOM and 1-RC model, and then integrated through a model fusion for online SOP estimation, which enables an analytical expression of battery peak charge/discharge current in a prediction window without weakening the nonlinear characteristic of FOM. Experimental results demonstrate the improved performance of the proposed MFM for online discharge SOP estimation, where the mean absolute error and root mean square error are only 0.288 W and 0.35 W, respectively, under the urban dynamometer driving schedule profile.

Online state of charge and state of power co-estimation of lithium-ion batteries based on fractional-order calculus and model predictive control theory

Accurate battery modelling is the cornerstone to state of charge (SOC) and state of power (SOP) co-estimation of lithium-ion batteries in electric vehicles. Due to strong battery nonlinearity over a broad frequency range, traditional integer-order models are incapable of capturing complex battery dynamics for SOC and SOP co-estimation. This paper proposes a fractional-order modified moving horizon estimation (FO-mMHE) algorithm and a fractional-order model predictive control (FO-MPC) algorithm. Firstly, a second-order FOM is constructed by performing a series of hybrid pulse tests at different SOC regions, and its model parameters are identified through a particle swarm optimization-genetic algorithm method. Secondly, online SOC estimation is converted into a constrained optimization problem in a past moving horizon and then solved by the FO-mMHE algorithm, which enables fast convergence speed and proactive smoothing of estimation outcomes. Thirdly, the FO-MPC algorithm is devised to manipulate the current sequence in a prediction horizon for maximizing discharge/charge power accumulation and determining battery SOP in real time. Moreover, different battery current–voltage behaviors are comprehensively researched in the prediction horizon over a whole battery operating range. The proposed co-estimation method is validated under different dynamic load profiles. The experimental results demonstrate a SOC estimation error reduction of up to 1.2 % compared with the commonly used fractional-order extended Kalman filter while the SOP estimation error could be limited below 0.35 W.