Research on Fault Diagnosis and Control Strategies for Power Batteries

Electric vehicles have transformed the way we think about transportation, replacing the roar of an engine with the silent hum of electricity. At the heart of this revolution lies the power battery, a massive, complex energy pack that must be managed with extreme precision. If this battery fails, the car stops; if it fails dangerously, it can pose a serious risk to the driver and passengers. To prevent this, every electric vehicle is equipped with a brain called a battery management system. This system constantly watches the battery, looking for signs of trouble like electrical leaks or broken communication lines between its internal parts. The challenge for engineers is that these systems operate in a chaotic environment filled with noise and unpredictable conditions. A tiny fluctuation in voltage or a momentary glitch in a signal can look like a major failure, leading to false alarms that inconvenience drivers, or worse, missing a real danger that could cause a fire. The goal is to find the true signal in the noise, distinguishing a harmless blip from a genuine crisis with absolute certainty.

Researchers from Guangdong Polytechnic Normal University and GAC AION New Energy Automobile Co., Ltd. tackled this problem by developing a new way for the battery's brain to think. Instead of relying on rigid, black-and-white rules that often struggle with real-world messiness, they introduced a method based on fuzzy theory. In simple terms, this approach allows the system to handle uncertainty much like a human expert does. Rather than asking, "Is the voltage exactly at the danger line?" the system asks, "How close is it to the danger line, and how does that compare to other signs of trouble?" By weighing multiple factors together—such as temperature, current, and signal quality—the system can make a more nuanced judgment about whether a fault is actually happening. This method was specifically designed to catch two of the most critical types of failures: insulation faults, where electricity leaks out of the battery toward the car's metal frame, and communication faults, where the battery's internal parts stop talking to each other correctly.

To test their idea, the team built a sophisticated simulation model and then put it to the test on actual electric vehicles. They created scenarios where they deliberately introduced faults, such as connecting a resistor to simulate an electrical leak or cutting off communication signals between the battery modules. The results showed that their fuzzy logic system was remarkably effective at identifying these problems. When the researchers compared the new system to older, standard methods, they found a clear improvement in accuracy. For insulation faults, the ability to correctly identify a problem jumped from about 81 percent to over 92 percent. For communication faults, the accuracy rose from roughly 91 percent to nearly 98 percent. This means the new system is far less likely to miss a real danger or falsely accuse a healthy battery of being broken.

However, this increased safety comes with a small trade-off. Because the fuzzy system takes the time to weigh all the different factors and run multiple checks to be sure, it takes slightly longer to make a decision. The researchers found that the time it took to confirm a fault increased by about 4.6 seconds compared to faster, less careful methods. In the context of a car driving down the highway, this extra time is negligible, but it is a necessary cost for the higher level of safety. The system is designed to wait just long enough to be certain, ensuring that when it finally sounds the alarm, the problem is real. Once a fault is confirmed, the system immediately takes action, such as limiting the power the car can use or shutting down the battery completely to prevent a fire, all while keeping the driver informed.

The study also mapped out exactly how these faults should be handled once detected. If the system finds an insulation leak, it checks whether the problem is inside the battery pack or in the external wiring. If it is inside, the battery pack must be opened and repaired; if it is outside, the external components are fixed. Similarly, if the communication lines are broken, the system checks if the issue is a temporary glitch or a permanent failure of a sensor. In every case, the fuzzy logic system helps the car decide the right course of action, balancing safety with the need to keep the vehicle moving. The researchers concluded that while this approach requires a bit more processing time, the dramatic improvement in accuracy makes it a vital tool for the future of electric vehicles. As the industry moves forward, combining this kind of intelligent, flexible diagnosis with massive amounts of real-world data promises to make electric cars safer and more reliable for everyone.