The reliable operation of modern infrastructure, particularly in critical fields like seismic monitoring, is fundamentally dependent on continuous and stable power supply. The battery energy storage system (BESS) serves as the core component for energy autonomy in such remote or sensitive installations. Ensuring the green, economical, safe, and stable operation of these stations necessitates advanced management strategies. This article provides a detailed literature review on online monitoring technologies for battery energy storage systems, with a focus on applications supporting critical infrastructure. We synthesize existing research on monitoring systems, mature modeling and analysis methods for energy storage batteries, and the application status of battery management systems (BMS). By examining domestic and international literature, we systematically compare existing online state monitoring devices and deeply analyze the methodologies they employ.
The evolution of battery technology has been pivotal. While early energy storage relied on lead-acid batteries, the advent of lithium-ion chemistry marked a significant leap. Lithium-ion batteries offer superior energy density, higher nominal voltage, lower self-discharge, longer cycle life, and a lack of memory effect. A typical battery energy storage system is constructed by connecting individual cells in series, parallel, or a combination thereof to meet specific voltage and capacity requirements. The configuration fundamentally impacts system monitoring and management complexity.
The performance and longevity of a battery energy storage system are primarily governed by the accurate estimation of its State of Charge (SOC) and the implementation of effective cell balancing control. SOC, defined as the ratio of remaining capacity to maximum available capacity, is a critical but unmeasurable internal state.
$$ SOC(t) = SOC(t_0) – \frac{1}{C_n} \int_{t_0}^{t} \eta i(\tau) d\tau $$
where \( C_n \) is the nominal capacity, \( i \) is the current (positive for discharge), and \( \eta \) is the coulombic efficiency.
Accurate SOC estimation is challenging due to nonlinear battery dynamics, varying operating conditions, and aging. Common methods include:
| Method | Principle | Advantages | Disadvantages |
|---|---|---|---|
| Open-Circuit Voltage (OCV) | Uses the known relationship between OCV and SOC. | Simple, relatively accurate at rest. | Requires long rest periods, impractical for online use. |
| Coulomb Counting (Ampere-hour Integration) | Integrates current over time. | Simple to implement, works online. | Accumulates sensor error, requires accurate initial SOC. |
| Model-Based Filters (e.g., EKF, UKF) | Combines a battery model with real-time measurements using recursive estimators. | Online, accurate, can handle noise. | Computationally complex, depends on model accuracy. |
| Data-Driven Methods (e.g., Neural Networks) | Learns the mapping from measurements (V, I, T) to SOC. | Can model complex nonlinearities, no need for explicit physics. | Requires extensive training data, risk of overfitting. |
Advanced research focuses on hybrid and improved algorithms. For instance, an Adaptive Unscented Kalman Filter (AUKF) combined with a BP neural network uses the neural network’s output as an observation update for the filter, enhancing robustness. Other approaches employ optimization algorithms like the Sparrow Search Algorithm (SSA) to optimize the initial weights and thresholds of an Extreme Learning Machine (ELM) network, improving SOC prediction accuracy and convergence speed under dynamic loads.
$$ \text{SSA-ELM Objective: } \min_{\mathbf{w}, \mathbf{b}} ||\mathbf{H}\mathbf{w} – \mathbf{T}||^2 $$
where \( \mathbf{H} \) is the hidden layer output matrix, \( \mathbf{w} \) are the output weights, \( \mathbf{b} \) are the biases, and \( \mathbf{T} \) is the target SOC matrix.
Cell imbalance is another major challenge in a battery energy storage system. Due to manufacturing tolerances and differing operating temperatures, cells within a series string charge and discharge at slightly different rates. Over time, this leads to divergence in individual cell SOC, reducing the usable capacity of the entire pack and accelerating degradation. Balancing control is essential.
| Balancing Topology | Mechanism | Efficiency | Complexity & Cost |
|---|---|---|---|
| Passive (Dissipative) | Bleeds excess energy from higher-SOC cells as heat via resistors. | Low (energy wasted). | Low. |
| Active – Capacitive/Inductive | Shuttles energy from higher-SOC cells to lower-SOC cells or the entire pack using switches and passive elements. | Medium to High. | Medium. |
| Active – Converter-Based | Uses dedicated DC-DC converters to transfer energy between cells, modules, or pack-level buses. | High. | High. |
| Distributed Control | Modules autonomously communicate SOC and coordinate balancing via a sparse network. | High (system-level). | High (control logic). |
Recent innovations include multi-layer distributed management systems and novel switched-capacitor circuits that maintain performance with increasing cell count. The core equation for balancing current in a shuttling capacitor method between two adjacent cells can be simplified as:
$$ I_{bal} \approx \frac{C_{sh} \cdot (V_{cell\_high} – V_{cell\_low})}{T_{sw}} $$
where \( C_{sh} \) is the shuttle capacitance and \( T_{sw} \) is the switching period.
The implementation of these algorithms relies on a robust Battery Management System (BMS). The evolution of BMS parallels the advancement of the battery energy storage system itself. Early systems in the late 1980s, like the PBWC system, introduced multi-sensor modules for basic voltage, current, and temperature monitoring. The 21st century saw a shift towards higher integration and intelligence. Modern BMS architectures are multi-layered:
- Data Acquisition Layer: Comprises sensor hardware (for cell voltage, pack current, temperature) and isolation circuits (e.g., using high-voltage analog switches like MAX14752 and differential amplifiers like INA148 for voltage sensing).
- Control Layer: A microcontroller (e.g., STM32 series) runs core algorithms for SOC/SOH estimation, balancing control, and protection logic (over-voltage, under-voltage, over-temperature).
- Application/Communication Layer: Provides external interfaces (CAN, Ethernet, RS-485) for data reporting and system integration. The trend is towards cloud-connected BMS.
Cloud-based BMS platforms represent a significant leap forward. They leverage IoT devices for data acquisition and wireless communication, transmitting operational data to cloud servers. The cloud provides virtually unlimited storage and powerful computational resources for running advanced digital twin models, performing big data analytics on cell degradation, and enabling predictive maintenance across fleets of battery energy storage systems. This architecture separates the resource-constrained embedded tasks from the computationally intensive analytics.
| Communication Protocol | Typical Use in BESS | Key Characteristics |
|---|---|---|
| CAN Bus | Internal communication between BMS master, slaves, and system controller. | Robust, real-time, multi-master, automotive standard. |
| Modbus RTU/TCP | Communication between BMS and external SCADA or data loggers. | Simple, widely adopted in industrial settings. |
| Ethernet/Wi-Fi/4G/5G | Remote monitoring and cloud connectivity for data upload and management. | High bandwidth, enables remote access and cloud analytics. |
The application of these technologies in seismic monitoring stations is a critical use case. These stations, often remote and unmanned, require absolute power continuity to ensure the integrity of continuous seismic data records. Traditionally, stations use AC power with Uninterruptible Power Supply (UPS) systems based on valve-regulated lead-acid (VRLA) batteries. However, managing these battery energy storage systems has been reactive and manual, leading to potential failures.
Intelligent power controllers based on microprocessors like the STM32F107 have been developed to monitor battery voltage, capacity, and environmental conditions, automating control logic to prevent damage from power anomalies. Remote power monitoring and control systems use TCP/IP protocols to reboot frozen instruments or diagnose issues from a central location, significantly improving maintenance efficiency for unmanned stations. Furthermore, optimizing power architectures—such as using direct DC power for communication equipment instead of inefficient AC inverters—has proven to enhance system stability and reduce power consumption. The implementation of a sophisticated online monitoring battery energy storage system directly addresses these needs by providing real-time state awareness, enabling proactive maintenance, and maximizing the availability of seismic monitoring infrastructure.
In conclusion, research into online monitoring technologies for battery energy storage systems is of paramount importance for ensuring safety, reliability, and efficiency. For critical applications like seismic stations, it guarantees operational continuity, prevents catastrophic failures, optimizes energy utilization, and extends asset life. The core technical challenges remain high-accuracy SOC estimation under dynamic conditions and efficient, reliable cell balancing. The future direction points towards greater intelligence, integration, and connectivity:
- High-Precision & Intelligent Diagnostics: Development of more advanced sensors and algorithms, including AI-driven models for State of Health (SOH) and Remaining Useful Life (RUL) prediction, and multi-physics models coupling electrochemistry with thermal dynamics for internal state observation.
- Integration and Miniaturization: Creating highly integrated system-on-chip (SoC) BMS solutions that reduce size, weight, and cost while improving reliability.
- Wireless and Networked Systems: Widespread adoption of wireless BMS to reduce cabling complexity and enable flexible modular designs, integrated with cloud platforms for fleet-level intelligence and management.
- Deep Integration with Energy Management Systems (EMS): The battery energy storage system BMS will evolve from a standalone protector to an intelligent node within a broader EMS, providing real-time state data to optimize energy dispatch, peak shaving, and grid support functions based on actual battery capability.
The transition from traditional periodic maintenance to condition-based and predictive maintenance for battery energy storage systems is essential for modern, resilient infrastructure. This is not merely a technical improvement but a fundamental shift in asset management philosophy, maximizing return on investment and ensuring system readiness when it is needed most.
| Aspect | Traditional Approach | Future Intelligent BESS Monitoring |
|---|---|---|
| State Estimation | Periodic manual testing, voltage-based SOC guess. | Real-time, model- and AI-based SOC/SOH/RUL estimation. |
| Fault Detection | Reactive (after failure occurs). | Proactive and predictive (identifying trends leading to failure). |
| Data Management | Local logging, manual retrieval. | Cloud storage, centralized analytics, digital twin simulation. |
| Maintenance | Scheduled, time-based. | Condition-based and predictive. |
| System Integration | Standalone BMS with limited external communication. | Fully integrated node within IoT and smart grid ecosystems. |