In the realm of new energy vehicles, whether pure electric or hybrid, the driving energy is supplied by power batteries. The performance of these batteries is critically influenced not only by their intrinsic parameters but also by the efficacy of the battery management system (BMS). The battery management system plays a pivotal role in extending the service life of power batteries and enhancing their usage efficiency. To conduct comprehensive experimental research and simulation analysis on the performance of the power battery management system, we have embarked on developing an experimental platform based on mechatronics product development principles. This platform aims to facilitate仿真, debugging, and testing of the battery management system, addressing gaps in existing research where experimental platforms often lack完善 functionality, exhibit high energy consumption, and suffer from low testing accuracy.
The core objective of this work is to design and implement a robust experimental platform that enables detailed investigation of battery management system functionalities, including state monitoring, state-of-charge (SOC) estimation, charge-discharge management, and cell balancing. Throughout this paper, we will delve into the selection of key components, the design of the platform’s topology, associated circuit designs, and the development of control software and programming methodologies. Emphasis is placed on多次体现 the importance of the battery management system (BMS) in ensuring optimal battery performance and safety.
| Component Type | Selected Model/IC | Key Parameters and Features | Rationale for Selection |
|---|---|---|---|
| Front-end Acquisition IC | BQ76930 (Texas Instruments) | 5-10 cell voltage channels (14-bit precision), 16-bit current sense, internal temperature sensing, I2C digital interface, integrated protection features (overvoltage, undervoltage, overcurrent), passive balance MOSFET drivers. | Balances high integration, sufficient channel count for modular design, and precise measurement capabilities suitable for research-oriented BMS experimentation. |
| Main Control MCU | STM32F103C8T6 (STMicroelectronics) | ARM Cortex-M3 32-bit core @72 MHz, 64 KB Flash, 20 KB SRAM, multiple communication interfaces (CAN, I2C, SPI, UART), 12-bit ADC channels, wide operating voltage (2.0-3.6V). | Offers high performance for logical operations and时序 control, rich peripheral set for system integration, and widespread use in嵌入式 systems facilitating development. |
| Isolation Communication IC | NSi8100 (Nuvoton) | 2-channel bidirectional I2C isolator, isolation voltage up to 5000 VRMS, data rate up to 2 MHz. | Ensures reliable signal isolation between high-voltage battery sections and low-voltage control circuits, meeting safety and noise immunity requirements. |
| Battery Type | Lithium Iron Phosphate (LiFePO4) | Nominal voltage 3.2V per cell, enhanced safety, long cycle life. | Commonly used in automotive applications, providing a realistic basis for BMS algorithm testing. |
The selection of the front-end acquisition solution is a fundamental step. While battery voltage acquisition can be achieved using discrete components, integrated circuits are predominant in mainstream automotive BMS designs. We evaluated solutions from TI and NXP, ultimately selecting the BQ76930 for our platform. This choice was driven by its configurability (5-10 series cells per IC), digital MCU interface (I2C), and integrated protection features, which align well with the needs of a modular, research-focused battery management system. The functional block diagram of the BQ769x0 family illustrates its comprehensive integration.
For the main control unit, automotive-grade MCUs from vendors like NXP are common in production vehicles. However, for this experimental system focused on research flexibility and通用性, we opted for the STM32F103C8T6. Its ARM Cortex-M3 core, ample memory, and diverse communication peripherals (including CAN, crucial for BMS networks) make it suitable for handling the logic, data processing, and communication tasks of the主控 module in our battery management system.
The experimental platform’s topology is deliberately chosen to reflect common automotive architectures. Among the prevalent topologies—integrated, master-slave, and distributed—we adopted a master-slave structure. This topology is advantageous for systems with numerous battery cells, even though our platform uses a 16-cell pack for manageability. It allows for scalable investigation of BMS network通信 and management strategies. In this setup, one BMS master board communicates with multiple BMS slave boards via a CAN bus. Each slave board is responsible for monitoring a subset of battery cells (8 cells in our implementation), measuring voltages and temperatures, and relaying this data to the master. The master board consolidates information, performs system-level calculations (like total pack SOC), and interfaces with external components like relays, current sensors, charging equipment, and a host PC for data display and storage.
The functional decomposition of the battery management system hardware is critical. The BMS master module comprises several key circuits: a power management circuit, the MCU circuit, relay control circuits (for main, pre-charge, and charging contactors), CAN and RS-485 communication interfaces, total current and voltage measurement circuits, insulation resistance measurement circuits, and opto-isolation circuits for signal integrity. Its primary responsibilities within the broader battery management system include aggregating data from slaves, executing high-level control logic based on battery state, managing contactors for safe power distribution, calculating pack SOC, and communicating with external vehicle controllers or test equipment.
| Module | Core Functions | Key Hardware Components |
|---|---|---|
| BMS Master Module |
|
STM32F103C8T6 MCU, CAN transceiver, relay drivers, isolation amplifiers, ADC circuits for pack measurements. |
| BMS Slave Module |
BQ76930 AFE IC, STM32F103C8T6 MCU, NSi8100 isolator, CAN transceiver, balancing circuitry. |
The BMS slave module is centered around the BQ76930 analog front-end (AFE) chip. It handles the direct connection to the high-voltage battery stack. To ensure safety and noise immunity, complete galvanic isolation is maintained between the high-voltage domain (powered by the battery stack itself) and the low-voltage domain (powered by a 12V system) of the slave board. This isolation is implemented for both power and communication signals. The communication between the slave’s MCU (low-voltage side) and the BQ76930 (high-voltage side) occurs through the NSi8100隔离芯片, specifically designed for bidirectional I2C isolation, a critical aspect of a reliable battery management system.
Cell voltage acquisition and balancing are intimately linked in BMS design. The BQ76930 uses shared pins for voltage sensing and passive balancing control. To achieve stable voltage readings, the input filter requires relatively large resistor values (e.g., Rc5, Cc5 in the schematic). While the IC has internal balancing MOSFETs, the achievable balance current is limited by these filter resistors. For more effective balancing in our experimental platform, we designed an external balancing circuit. This approach decouples the sensing stability from the balancing current magnitude. The external balance circuit works by utilizing the control signal from the BQ76930’s balance driver pin to turn on an external MOSFET (Q5). The balance current then flows through an external resistor (Rbal5), whose value can be adjusted independently to set the desired balance current (Ibal). The balance current for a given cell can be approximated by:
$$I_{bal} \approx \frac{V_{cell}}{R_{bal} + R_{DS(on)}}$$
where \(V_{cell}\) is the cell voltage, \(R_{bal}\) is the external balancing resistor, and \(R_{DS(on)}\) is the on-resistance of the external MOSFET. This design ensures that voltage measurement accuracy, governed by the filter time constant \(\tau = R_{filter} \times C_{filter}\), remains uncompromised while allowing for higher, adjustable平衡 currents to improve the efficiency of the battery management system’s均衡 function.
The isolation communication circuit design is paramount for safe operation. The NSi8100 chip creates a robust barrier. The I2C signals (SDA, SCL) from the low-voltage-side MCU are translated across the isolation boundary to the high-voltage-side BQ76930. The isolation voltage rating (5000 VRMS) far exceeds the potential differences in our platform, ensuring safety. The data rate capability (2 MHz) is more than sufficient for the periodic data polling required by the battery management system algorithms.
The embedded software for both master and slave modules was developed within the Keil uVision5 IDE. The software architecture directly supports the functional requirements of the battery management system experimental platform.
The BMS master module software follows a structured flow. After initialization of system clocks, peripherals (CAN, UART, ADC, GPIO), and communication protocols, it enters a main loop. Key tasks in this loop include:
- Periodically polling slave modules via CAN for cell voltages, temperatures, and fault status.
- Reading local ADC channels for total pack current (Ipack) and voltage (Vpack).
- Executing SOC estimation algorithms. A common coulomb counting method with compensation can be represented as:
$$SOC(t) = SOC(t_0) – \frac{1}{C_{nominal}} \int_{t_0}^{t} \eta(I) \cdot I(\tau) \, d\tau$$
where \(C_{nominal}\) is the battery’s nominal capacity, \(I\) is the measured current, and \(\eta(I)\) is a current-dependent efficiency factor. More advanced methods like Extended Kalman Filters (EKF) can also be implemented on this platform for research:
$$x_{k} = f(x_{k-1}, u_{k-1}) + w_{k-1}$$
$$z_{k} = h(x_{k}) + v_{k}$$
where \(x\) represents the state vector (e.g., SOC, internal resistance), \(u\) is the input (current), \(z\) is the measurement (voltage), and \(w\), \(v\) are process and measurement noise.
- Evaluating protection conditions based on predefined thresholds. For example, an overcurrent condition is triggered if:
$$|I_{pack}| > I_{overcurrent\_threshold}$$
- Processing commands from the host PC (e.g., start/stop discharge, set balancing parameters).
- Controlling contactors based on system state and faults.
- Calculating cell voltage differences and initiating balancing commands when the maximum difference \(\Delta V_{max} = \max(V_{cell,i}) – \min(V_{cell,j})\) exceeds a set limit.
- Periodically sending system status data (voltages, currents, temperatures, SOC, faults) to the host PC for display and logging.
The BMS slave module software is comparatively focused. Its primary cycle involves:
- Initializing the I2C peripheral and configuring the BQ76930 AFE (setting cell count, protection thresholds, ADC modes).
- Regularly reading registers from the BQ76930 to obtain raw ADC values for the 8 cell voltages and 2 temperatures.
- Converting raw ADC values to physical quantities. For voltage, the conversion for cell \(i\) is:
$$V_{cell,i} = \frac{ADC_{raw,i} \cdot V_{ref}}{2^{n} \cdot G_{ain}}$$
where \(n\) is the ADC resolution (14-bit for BQ76930), \(V_{ref}\) is the internal reference voltage, and \(G_{ain}\) is any internal gain factor.
- Checking the BQ76930’s status flags for any cell-level faults (OV, UV).
- Listening for CAN messages from the master, which may contain balancing enable/disable commands for specific cells.
- Applying balancing commands by writing to the BQ76930’s control registers, which activate the external平衡 circuit described earlier.
- Packetizing the acquired data and transmitting it to the master via the isolated CAN interface at a fixed interval.
The hardware design, encompassing schematic capture and PCB layout, was finalized alongside the software development. Components were selected not only for functionality but also for accessibility and ease of prototyping. The final experimental platform was constructed by mounting all subsystems onto a sturdy baseplate (900mm x 700mm), creating an integrated test bed. This platform includes the 16-cell LiFePO4 battery pack, the custom-designed BMS master and slave boards, contactors and pre-charge circuitry, a programmable DC load for模拟 discharge, an AC/DC charger模拟 unit, Hall-effect current sensors, and necessary wiring and fusing for safety. A host PC runs a dedicated graphical user interface (GUI) developed in a environment like LabVIEW or Python, which communicates with the BMS master via RS-485 or CAN. This GUI provides real-time visualization of all battery management system parameters (cell voltages, temperatures, pack current/voltage, SOC, balance status), allows for manual control of platform operations (charge/discharge cycles), and enables data logging for post-test analysis.
The constructed platform, as illustrated, serves as a comprehensive tool for电池管理系统 research. It validates the core functionalities of a typical automotive battery management system. We can conduct experiments to characterize battery charge and discharge profiles under controlled loads, assess the accuracy of voltage and current measurement circuits against calibrated instruments, evaluate the performance of different passive balancing strategies in equalizing cell voltages, test the robustness of SOC estimation algorithms under dynamic current profiles, and verify the correctness of protection logic (overvoltage, undervoltage, overcurrent). The modular nature of the design, especially the master-slave topology and the use of generic MCUs, allows for easy modification and expansion. For instance, one could replace the SOC estimation algorithm in the master software, change the balancing current by modifying a resistor on the slave board, or even integrate different battery chemistry models into the test routines.
| Experimental Category | Measurable/Testable Parameters | Relevant BMS Function |
|---|---|---|
| Battery Characterization | Charge/discharge curves, internal resistance, capacity fade over cycles. | Provides foundational data for BMS algorithm calibration and validation. |
| Measurement Accuracy | Cell voltage error (< 10 mV), pack current error (< 1%), temperature sensing accuracy. | Core competency of the battery management system’s sensing subsystem. |
| Cell Balancing | Balancing current magnitude, time to achieve voltage convergence, energy dissipation. | Evaluates the effectiveness of the BMS’s passive balancing management strategy. |
| State Estimation | SOC estimation error under static and dynamic profiles, convergence time. | Tests the core intelligence algorithms of the battery management system. |
| System Protection | Response time to overvoltage, undervoltage, overcurrent, and short-circuit events. | Validates the safety-critical reactive functions of the BMS. |
| Communication & Network | CAN bus loading, data latency between slave and master, error handling. | Assesses the reliability of the distributed BMS architecture. |
In conclusion, the design and development process of this experimental platform for a new energy vehicle power battery management system has been详细 described. The platform is grounded in a systematic mechatronics development approach. Key decisions included the selection of the BQ76930 as the front-end acquisition IC and the STM32F103C8T6 as the main control MCU, balancing performance with research flexibility. The adoption of a master-slave topology provides a scalable framework representative of real-world automotive BMS designs. Innovative circuit designs, such as the external balancing circuit that decouples measurement stability from balance current, and the use of the NSi8100 for robust isolated I2C communication, enhance the platform’s capabilities. The embedded software for both master and slave modules implements the essential control, estimation, and communication逻辑 of a modern battery management system. The fully assembled platform offers a versatile environment for conducting a wide array of experiments, from basic parameter measurement to advanced algorithm validation. It effectively serves as a testbed for researching methods to extend battery life, improve efficiency, and enhance safety through sophisticated battery management system strategies. Future work may involve integrating the platform with hardware-in-the-loop (HIL) simulators for more complex scenario testing or adapting it to explore active balancing techniques and advanced battery state estimation methods like neural networks. The successful implementation of this platform underscores the critical role that dedicated experimental systems play in advancing the technology of battery management systems for sustainable transportation.