We present a comprehensive evaluation framework for energy storage battery systems deployed in electric vehicle photovoltaic charging stations. As renewable energy sources expand rapidly and electric vehicles become ubiquitous, the role of energy storage battery systems in stabilizing grid operations, enhancing charging efficiency, and achieving energy equilibrium has become indispensable. Our work focuses on developing a robust methodology to assess the performance, health, and remaining lifetime of these energy storage battery systems. We propose a series-connected solar cell architecture to model the photovoltaic generation, followed by a detailed performance testing procedure for energy storage battery systems. Subsequently, we introduce a novel future life prediction method tailored for energy storage battery systems in charging stations. Finally, we employ a fuzzy evaluation technique to synthesize multiple risk factors and deliver a holistic assessment. Our findings underscore that battery aging risk requires heightened attention for energy storage battery systems in such environments.
1. Introduction
The rapid development of renewable energy sources coupled with the widespread adoption of electric vehicles has created an urgent need for efficient energy conversion and storage solutions. Energy storage battery systems serve as the core component in electric vehicle photovoltaic charging stations, enabling energy balance, reducing grid stress, and improving charging reliability. However, evaluating the condition and longevity of these energy storage battery systems remains challenging due to the complex interplay between charging cycles, temperature variations, and aging mechanisms. We aim to address this gap by proposing an integrated assessment methodology that combines solar cell modeling, battery state estimation, life prediction, and fuzzy logic-based risk evaluation.
Our approach begins with the establishment of a series solar cell model to optimize photovoltaic generation for the energy storage battery system. We then develop precise methods to estimate the state of charge (SOC) and state of health (SOH) of the energy storage battery, along with a data-driven remaining useful life (RUL) prediction technique. By leveraging real-time monitoring and big data analytics, we can capture critical parameters that influence the performance of energy storage battery systems. Finally, a fuzzy evaluation model aggregates multiple risk factors—such as control algorithm defects, communication failures, and data acquisition anomalies—into a comprehensive risk score. This methodology enables decision-makers to implement proactive maintenance and safety measures for energy storage battery systems.
The remainder of this paper is organized as follows. Section 2 details the series solar cell model used for photovoltaic energy conversion. Section 3 presents the performance testing methods for energy storage battery systems, including SOC estimation, SOH estimation, and RUL prediction. Section 4 describes the fuzzy evaluation framework for risk assessment. Section 5 concludes with key findings and practical implications.
2. Series Solar Cell Model for Photovoltaic Generation
To improve the efficiency of photovoltaic generation that charges the energy storage battery system, we adopt a series (or multi-junction) solar cell architecture. By stacking p-n junctions with different bandgaps, the top cell absorbs short-wavelength photons while the bottom cell captures the remaining longer-wavelength portion, thereby enhancing overall conversion efficiency. The critical challenge in such series-connected cells is achieving current matching between sub-cells to minimize energy losses due to current mismatch.
2.1 Single Solar Cell Structure
A single solar cell in our model comprises several functional layers:
- Transparent Conducting Oxide Layer (TCOL) – material: fluorine-doped tin oxide (FTO).
- Electron Transport Layer (ETL) – material: cadmium sulfide (CdS), which facilitates electron movement toward the absorber.
- Hole Transport Layer (HTL) – directs holes toward the metal electrode.
- Absorber Layer (AL) – material: Sb₂(S,Se)₃, which absorbs photons and generates charge carriers.
- Metal Electrode – material: gold (Au).
The absorber material Sb₂(S,Se)₃ exhibits varying optical and electrical properties depending on the Se/(S+Se) composition ratio. Table 1 summarizes the key characteristics of each layer in our single cell model.
| Layer | Material | Thickness (nm) | Role |
|---|---|---|---|
| TCOL | FTO | 500 | Transparent conductor |
| ETL | CdS | 50 | Electron transport |
| AL | Sb₂(S,Se)₃ | 300–1000 | Photon absorption |
| HTL | Spiro-OMeTAD | 200 | Hole transport |
| Metal | Au | 100 | Back electrode |
The absorption coefficient α(λ) of the absorber layer varies with wavelength and Se content. We model it using the following equation:
$$ \alpha(E) = A_\alpha \sqrt{h\nu – E_g} $$
where \(E_g\) is the bandgap energy in eV, \(h\) is Planck’s constant, and \(\nu\) is the frequency. For Sb₂(S₁₋ₓ,Seₓ)₃, the bandgap decreases linearly from 1.7 eV (x=0) to 1.2 eV (x=1) as the selenium fraction increases.
2.2 Series-Connected Solar Cell Model
In the series configuration, the top sub-cell is illuminated by the standard AM 1.5 G spectrum, while the bottom sub-cell receives only the filtered spectrum transmitted through the top layers. The transmitted spectrum is given by:
$$ T(\lambda) = T_0(\lambda) \exp\left[ -\sum_{k=1}^{n} \alpha_k(\lambda) t_k \right] $$
where \(T_0\) is the incident AM 1.5 G spectrum, \(t_k\) is the thickness of layer \(k\), and \(\alpha_k\) is the absorption coefficient of that layer. The index \(k=1,2,3,4\) corresponds to TCO, ETL, AL, and HTL respectively. To avoid significant current mismatch, we tune the absorber thickness and Se composition in the top and bottom sub-cells so that their current densities are closely matched. Table 2 provides an example of optimized parameters for a tandem device.
| Sub-cell | Se/(S+Se) ratio | Bandgap (eV) | Absorber thickness (nm) | Expected Jₛ (mA/cm²) |
|---|---|---|---|---|
| Top | 0.3 | 1.55 | 400 | 16.2 |
| Bottom | 0.7 | 1.33 | 600 | 16.0 |
The series-connected solar cells supply power to the energy storage battery system in the charging station. Accurate modeling of the photovoltaic generation is essential for predicting the input energy to the energy storage battery and for designing efficient charging strategies.
3. Performance Testing of Energy Storage Battery Systems
To evaluate the condition of energy storage battery systems, we focus on three key metrics: state of charge (SOC), state of health (SOH), and remaining useful life (RUL). These parameters are non-directly measurable and must be estimated using advanced algorithms combined with sensor data.
3.1 State of Charge (SOC) Estimation
The SOC of an energy storage battery is defined as the ratio of remaining capacity to rated capacity:
$$ \text{SOC} = \frac{Q_{\text{remain}}}{Q_{\text{rated}}} \times 100\% $$
Conventional current integration methods suffer from accumulated errors due to sensor drift and ignore internal factors like resistance and aging. We adopt a Kalman filter approach that uses the error between the measured SOC and the estimated SOC as feedback to correct the estimate in real time. This closed-loop method significantly improves accuracy. The relationship between daily capacity loss and temperature at various SOC levels is critical for understanding calendar aging. Table 3 presents simulated data showing how capacity loss increases with temperature and SOC for a typical lithium-ion energy storage battery.
| Temperature (°C) | SOC=20% | SOC=50% | SOC=80% | SOC=100% |
|---|---|---|---|---|
| 10 | 0.02 | 0.04 | 0.07 | 0.10 |
| 25 | 0.08 | 0.15 | 0.25 | 0.35 |
| 40 | 0.20 | 0.38 | 0.60 | 0.80 |
| 60 | 0.50 | 0.90 | 1.40 | 1.90 |
Kalman filtering effectively handles the nonlinearities present in energy storage battery dynamics. We define the state vector including SOC and polarization voltages, and update it iteratively using measurements of voltage, current, and temperature.
3.2 State of Health (SOH) Estimation
The SOH quantifies the ability of an energy storage battery to store energy compared to a fresh cell. It is defined as:
$$ \text{SOH} = \frac{C – C_{\text{EOL}}}{C_{\text{BOL}} – C_{\text{EOL}}} \times 100\% $$
where \(C_{\text{BOL}}\) is the capacity at beginning of life, \(C_{\text{EOL}}\) is the capacity at end of life (typically 80% of nominal), and \(C\) is the current actual capacity. We employ a voltage curve fitting method that does not require intrinsic battery parameters. By analyzing the incremental capacity or differential voltage curves during charging, we can infer the capacity fade. Figure 1 (not shown) illustrates the typical SOH degradation over time due to cycle degradation and calendar aging. Table 4 provides example SOH values over a hypothetical 5-year operation period for an energy storage battery system.
| Time (years) | Cycle degradation factor (%) | Calendar aging factor (%) | Total SOH (%) |
|---|---|---|---|
| 0 | 0 | 0 | 100 |
| 1 | 5 | 3 | 92 |
| 2 | 9 | 6 | 85 |
| 3 | 12 | 8 | 80 |
| 4 | 15 | 10 | 75 |
| 5 | 18 | 12 | 70 |
The degradation trends show that calendar aging is slower but persistent, while cycle aging accelerates with usage. Ambient temperature variations cause seasonal fluctuations—summer months typically induce faster SOH decline.
3.3 Remaining Useful Life (RUL) Prediction for Energy Storage Battery Systems
To predict the future performance of energy storage battery systems, we adopt a data-driven approach that utilizes historical records of cycle counts, charge/discharge capacities, temperatures, and SOC/SOH trajectories. Machine learning models (e.g., support vector regression, Gaussian process regression) are trained on these data to forecast the remaining cycles until the SOH drops below 80%. Equation (3) represents a typical degradation model used in RUL prediction:
$$ C(k) = C_0 \cdot \exp(-\beta \cdot k) + \epsilon $$
where \(C(k)\) is the capacity at cycle \(k\), \(C_0\) is the initial capacity, \(\beta\) is the aging coefficient, and \(\epsilon\) represents noise. The RUL is then determined by solving for \(k\) when \(C(k) \approx 0.8 \cdot C_0\). Table 5 lists example RUL predictions under different operating conditions for an energy storage battery.
| Average temperature (°C) | Average depth of discharge (%) | RUL (cycles) |
|---|---|---|
| 25 | 50 | 3200 |
| 25 | 80 | 2200 |
| 40 | 50 | 1800 |
| 40 | 80 | 1100 |
These predictions allow operators to schedule proactive replacement or maintenance of energy storage battery systems before catastrophic failure occurs.
3.4 Data Monitoring and Early Warning for Energy Storage Battery Systems
In a vehicle-grid integration system, energy storage battery systems undergo frequent charge/discharge cycles, making them susceptible to anomalies such as temperature excursions, SOC calculation errors, and internal short circuits. We implement a dedicated battery monitoring module that tracks parameters like voltage, current, temperature, and electrolyte leakage. A median-expectation-based prediction algorithm continuously estimates changes in battery internal resistance and bottom-plate current to adjust the battery management system (BMS) controller in real time. Figure 2 (conceptual) shows the architecture of this online monitoring system.
Big data technology plays a crucial role in safety protection. The monitoring terminal collects raw data from the energy storage battery system and feeds it into a data-driven model that estimates cycle life and SOH. When critical thresholds are approached, an alarm is triggered. Table 6 summarizes the key monitored parameters and their alarm thresholds for energy storage battery systems.
| Parameter | Normal range | Warning threshold | Action |
|---|---|---|---|
| Temperature | 15–45 °C | >55 °C | Reduce charge current |
| Voltage deviation | ±0.05 V/cell | >0.15 V/cell | Balance cells |
| SOC jump | <5% per minute | >10% per minute | Check sensor fault |
| Internal resistance increase | <20% from baseline | >50% from baseline | Schedule replacement |
4. Fuzzy Evaluation Based Comprehensive Assessment of Energy Storage Battery Systems
Fuzzy evaluation provides a powerful framework to handle the uncertainty and subjectivity inherent in assessing energy storage battery systems. We convert linguistic terms into fuzzy rules and aggregate multiple risk factors to produce an overall risk score.
4.1 Translating Linguistic Terms into Fuzzy Rules
We define linguistic variables for inputs such as “battery remaining capacity” (low, medium, high) and “charging demand” (low, high). Membership functions (e.g., triangular) are designed for each term. For example, the membership function for “high remaining capacity” might be:
$$ \mu_{\text{high}}(x) = \begin{cases} 0 & x \le 70 \\ \frac{x-70}{30} & 70 < x < 100 \\ 1 & x \ge 100 \end{cases} $$
Based on expert experience, we create fuzzy rules such as: “If remaining capacity is high and charging demand is high, then charging power is high.” Table 7 lists a sample rule base for an energy storage battery system.
| Remaining capacity \ Demand | Low demand | Medium demand | High demand |
|---|---|---|---|
| Low | Low power | Low power | Medium power |
| Medium | Low power | Medium power | High power |
| High | Medium power | High power | High power |
4.2 Membership Computation and Fuzzy Aggregation
After obtaining normalized data for each metric, we compute the membership degree using the designed functions. The similarity between two fuzzy sets \( \tilde{A} \) and \( \tilde{B} \) is given by:
$$ S_{uv}(\tilde{A},\tilde{B}) = 1 – \frac{1}{n}\sum_{i=1}^{n} |a_i – b_i| $$
where \( S_{uv}(\tilde{A},\tilde{B}) \in [0,1] \). Larger values indicate better performance. The weighted average of rule \( E_u \) is calculated as:
$$ WA(E_u) = \frac{\sum_{v \neq u} W(E_v) \cdot S_{uw}(\tilde{R}_u,\tilde{R}_v)}{\sum_{v \neq u} W(E_v)} $$
where \(M\) is the number of linguistic terms, and \(W(E_v)\) is the importance weight of rule \(v\). If all rules are equally important, \(W(E_v)=1/M\). The relative agreement of rule \(E_u\) is:
$$ RA(E_u) = \frac{WA(E_u)}{\sum_{n=1}^{M} WA(E_n)} $$
The consistency coefficient is then:
$$ CC(E_u) = \beta \cdot W(E_u) + (1-\beta) \cdot RA(E_u) $$
The parameter \(\beta\) balances the importance of the rule weight versus the relative agreement. For \(\beta=0.5\), both factors contribute equally. Finally, the overall fuzzy evaluation result \(\tilde{R}\) is:
$$ \tilde{R} = CC(E_1) \cdot \tilde{R}_1 + \cdots + CC(E_M) \cdot \tilde{R}_M $$
The value of \(\tilde{R}\) indicates the overall state of the energy storage battery system and can be used to trigger protective actions.
4.3 Risk Factor Assessment for Energy Storage Battery Systems
We identified six key risk factors that affect the reliability of energy storage battery systems in electric vehicle photovoltaic charging stations, as listed in Table 8.
| ID | Risk category | Description |
|---|---|---|
| N1 | Control algorithm defect | Flawed algorithms may cause incorrect SOC/SOH estimation |
| N2 | Controller functional failure | Hardware or software failure in BMS |
| N3 | Communication fault | Loss of data between sensors and BMS |
| N4 | Control execution anomaly | Relay or switch malfunction while executing commands |
| N5 | Data acquisition anomaly | Sensor drift or noise causing inaccurate data |
| N6 | Command response anomaly | Delayed or missing response to BMS commands |
Using the fuzzy evaluation method described above, we quantified the relative occurrence probability of each risk factor based on expert input (with appropriate weighting for expertise levels). The results are presented in Table 9.
| Risk factor | N1 | N2 | N3 | N4 | N5 | N6 |
|---|---|---|---|---|---|---|
| Probability | 0.35 | 0.28 | 0.42 | 0.31 | 0.65 | 0.58 |
Data acquisition anomaly (N5) and command response anomaly (N6) exhibit the highest risk levels. This emphasizes that battery aging, which leads to increased internal resistance and unstable sensor readings, is a critical concern for energy storage battery systems. As the energy storage battery ages, the risk of thermal runaway and internal short circuits escalates, potentially jeopardizing the entire storage system.
5. Conclusion
We have developed a comprehensive evaluation method for energy storage battery systems in electric vehicle photovoltaic charging stations. Our approach integrates a series solar cell model for efficient photovoltaic generation, advanced state estimation techniques (SOC, SOH) based on Kalman filtering and voltage curve fitting, a data-driven remaining useful life prediction method, and a fuzzy evaluation framework that synthesizes multiple risk factors. The experimental results demonstrate that data acquisition and command response anomalies pose the highest risks, highlighting the need for special attention to battery aging in energy storage battery systems. The proposed method provides accurate and holistic assessments that enable informed decision-making for maintenance, replacement, and operational optimization of energy storage battery systems. Practical application of this methodology can reduce charging time and cost while improving charging efficiency and quality.