With the rapid advancement of renewable energy technologies and the widespread adoption of electric vehicles (EVs), the integration of battery energy storage systems (BESS) in photovoltaic (PV) charging stations has become increasingly critical. These systems play a pivotal role in enhancing charging efficiency, stabilizing the power grid, and achieving energy balance by mitigating the intermittency of solar power. In this study, we focus on developing a comprehensive evaluation method for energy storage cell performance within EV PV charging stations. We propose a fuzzy logic-based assessment approach that incorporates key parameters such as state of charge (SOC), state of health (SOH), and remaining useful life (RUL) of energy storage cells. By establishing mathematical models and leveraging data-driven techniques, our method aims to provide accurate insights into the operational status and risks associated with energy storage cells, thereby supporting informed decision-making for system optimization and maintenance.
The foundation of our evaluation lies in the modeling of series-connected solar cells, which are essential for efficient energy harvesting in PV charging stations. Series or multi-junction solar cells improve overall efficiency by stacking multiple p-n junction cells with varying bandgaps. In such configurations, the top cell absorbs shorter wavelengths of the solar spectrum compatible with its absorber bandgap, while the bottom cell captures the residual spectrum based on its respective bandgap. The open-circuit voltage (VOC) of a series device is the sum of the voltages of its sub-cells. However, a significant challenge in series configurations is current density mismatch, which can lead to energy losses. To address this, we model a series solar cell structure that ensures tight current matching between sub-cells. The individual solar cell consists of several layers: a transparent conducting oxide layer (TCOL) made of fluorine-doped tin oxide (FTO), an electron transport layer (ETL) of cadmium sulfide (CdS), a hole transport layer (HTL), an absorber layer (AL) of Sb2(S,Se)3, and a metal electrode layer of gold (Au). The absorber material, Sb2(S,Se)3, exhibits a dielectric constant and charge carrier mobility that depend on the selenium content, influencing the absorption coefficient. The absorption coefficient α(λ) for different Se/(S+Se) composition ratios is calculated using the formula: $$ \alpha(E) = A_{\alpha} \sqrt{hv – E_g} $$ where E_g is the energy gap of the layer, h is Planck’s constant, and v is the frequency. This relationship allows us to optimize the absorption characteristics for enhanced performance. The transmission of light through the top sub-cell to the bottom sub-cell is determined by: $$ T(\lambda) = T_0(\lambda) \exp\left(-\sum_{k=1}^{n} \alpha_k(\lambda) t_k\right) $$ where n is the number of layers in the top unit, t_k is the thickness of each layer, and T_0 is the incident global AM 1.5 G spectrum. By varying the selenium content in the absorber layer, we can adjust the bandgap linearly from 1.7 eV for Sb2S3 to 1.2 eV for Sb2Se3, as shown in the energy band diagram where the valence band maximum (VBM) and conduction band minimum (CBM) shift with composition. This modeling enables the design of series solar cells that minimize current mismatch and maximize energy conversion efficiency in PV charging stations.
To evaluate the performance of energy storage cells in BESS, we focus on three key indicators: state of charge (SOC), state of health (SOH), and remaining useful life (RUL). SOC represents the transient amount of remaining battery capacity or operational time and is defined as: $$ \text{SOC} = \frac{Q_{\text{remain}}}{Q_{\text{rated}}} \times 100\% $$ where Q_remain is the remaining capacity and Q_rated is the rated capacity. Traditional methods, such as current integration, often lead to accumulating errors due to neglect of internal factors like resistance and aging. Therefore, we employ a Kalman filter algorithm for SOC estimation, which uses feedback from the error between actual and estimated SOC values to achieve closed-loop control and real-time computation. The capacity loss L_cal of energy storage cells under different temperatures exhibits a consistent increase in SOC with rising temperature, as illustrated in experimental data where higher temperatures accelerate reaction rates. SOH, on the other hand, quantifies the battery’s performance state relative to a new cell and is defined as: $$ \text{SOH} = \frac{C – C_{\text{EOL}}}{C_{\text{BOL}} – C_{\text{EOL}}} \times 100\% $$ where C is the actual capacity, C_EOL is the end-of-life capacity, and C_BOL is the beginning-of-life capacity. We use voltage curve fitting for SOH estimation due to its simplicity and independence from inherent battery parameters. Over time, SOH declines due to cycle degradation and calendar aging, with faster degradation in summer due to higher temperatures. For RUL prediction, we adopt a data-driven approach that utilizes extensive operational data, including cycle count, charge-discharge capacity, time, and temperature. This method trains predictive models to forecast battery degradation without requiring deep insights into internal mechanisms, though it demands high-quality and voluminous data. To monitor real-time battery data, we implement a battery monitoring module in vehicle networking systems that uses median expectation-based prediction to detect anomalies like temperature thresholds being exceeded, which could lead to electrolyte failure or overcharging. This module adjusts the battery management system controller to regulate charging current, enhancing energy storage cell longevity and grid service effectiveness. Additionally, big data technology is employed for online monitoring, where battery parameters are imported into driven models to estimate cycle life and trigger warnings when critical values are reached. The integration of these techniques ensures comprehensive assessment and protection for energy storage cells in BESS.
Our proposed evaluation method for BESS relies on fuzzy logic to handle uncertainties in energy storage cell performance. The fuzzy evaluation process begins by converting linguistic terms into fuzzy rules. For instance, we define fuzzy sets for input variables like “battery remaining capacity” with terms such as high, medium, and low, and output variables like “charging power” with terms like large. Membership functions, such as triangular functions, are designed to describe the degree of membership for each term. For example, the membership function for “high” battery capacity might peak at SOC values above 80%. Fuzzy rules are then formulated based on expert knowledge, e.g., “If battery remaining capacity is high and charging demand is high, then charging power is large.” These rules cover all possible combinations of input variables to ensure robust decision-making. Next, we compute membership degrees and perform fuzzy combination operations. The similarity between fuzzy sets A’ and B’ is calculated as: $$ S_{uv}(A’, B’) = 1 – \frac{1}{n} \sum_{i=1}^{n} |a_i – b_i| $$ where S_uv ranges from 0 to 1, with higher values indicating better performance. The weighted value for a fuzzy rule E_u is given by: $$ WA(E_u) = \frac{\sum_{u \neq v, v=1}^{M} W(E_v) \cdot S_{uw}(R’_u, R’_v)}{\sum_{u \neq v, v=1}^{M} W(E_v)} $$ where M is the number of linguistic terms, W(E_v) is the weight representing the relative importance of each rule, and S_uw is the similarity measure. If all rules are equally important, W(E_1) = W(E_2) = … = W(E_M) = 1/M. The relative agreement RA(E_u) and consistency coefficient CC(E_u) are then derived as: $$ RA(E_u) = \frac{WA(E_u)}{\sum_{n=1}^{M} WA(E_n)} $$ $$ CC(E_u) = \beta \cdot W(E_u) + (1 – \beta) \cdot RA(E_u) $$ where β is a factor balancing the importance of weights and relative agreement. By setting β appropriately, decision-makers can emphasize either aspect; for example, β=1 focuses solely on weights, while β=0 prioritizes agreement. Finally, the overall evaluation result R’ for the BESS is computed as: $$ R’ = CC(E_1) \cdot R’_1 + \cdots + CC(E_M) \cdot R’_M $$ This result quantifies the system’s state, guiding safety measures and warnings. To illustrate the application, we assess risk factors in BESS, such as control algorithm defects, controller failures, communication faults, control execution anomalies, data acquisition issues, and command response abnormalities. A table summarizing these risk factors is provided below, highlighting their relative occurrence probabilities based on fuzzy evaluation. Our analysis shows that data acquisition and command response anomalies pose the highest risks, underscoring the need to monitor energy storage cell aging closely. As energy storage cells age, their safety performance degrades, increasing the risk of internal short circuits and thermal runaway, which could compromise the entire BESS.
| Identifier | Type | Description |
|---|---|---|
| N1 | Control Algorithm Defects | Imperfections in control logic leading to inefficiencies or failures. |
| N2 | Controller Function Fault | Malfunctions in controller hardware or software. |
| N3 | Communication Fault | Disruptions in data exchange between system components. |
| N4 | Control Execution Anomaly | Errors in executing control commands, such as delayed responses. |
| N5 | Data Acquisition Anomaly | Inaccuracies or failures in collecting operational data from energy storage cells. |
| N6 | Command Response Anomaly | Irregularities in responding to control commands, potentially due to aging energy storage cells. |
In conclusion, our comprehensive evaluation method for battery energy storage systems in electric vehicle photovoltaic charging stations integrates series solar cell modeling, energy storage cell performance testing, and fuzzy logic-based assessment. By addressing key parameters like SOC, SOH, and RUL, and incorporating real-time monitoring and big data analytics, we provide a robust framework for evaluating energy storage cell health and risks. Experimental results validate the method’s effectiveness, showing that it accurately identifies aging-related risks and supports proactive maintenance. This approach not only enhances charging efficiency and grid stability but also contributes to the sustainable development of renewable energy integration. Future work could explore adaptive fuzzy systems and machine learning techniques to further refine the evaluation process for evolving energy storage cell technologies.