New Fault-Tolerant Strategy Boosts Reliability for EV Charging Converters
The relentless push towards electrification in the automotive industry has placed unprecedented demands on the power electronics that form the backbone of charging infrastructure. A critical component in this ecosystem is the DC-DC converter, responsible for efficiently and safely managing the flow of power from the grid or a charging station to an electric vehicle’s battery pack. Any failure in this system can lead to downtime, costly repairs, and, in the worst-case scenario, safety hazards. Recognizing this, a team of researchers from Shanghai University of Electric Power and Shanghai Jiao Tong University has developed a groundbreaking, low-cost fault-tolerant strategy specifically designed for Semi-Dual Active Bridge (S-DAB) converters, a topology increasingly favored for its efficiency in unidirectional power applications like EV fast-charging. Their work, published in the prestigious Transactions of China Electrotechnical Society, offers a practical solution that could significantly enhance the resilience and uptime of future charging networks.
The S-DAB converter has emerged as a compelling alternative to the more complex Dual Active Bridge (DAB) design, particularly for applications where power only needs to flow in one direction—from the charger to the car. Its key advantage lies in its simplicity. By replacing two active switches on the secondary side (the side connected to the EV battery) with diodes, the S-DAB reduces component count, lowers cost, and can achieve Zero Voltage Switching (ZVS) over a wider operating range, which translates to higher efficiency and less heat generation. This makes it an ideal candidate for the next generation of compact, high-power DC fast chargers. However, this streamlined design also introduces a potential vulnerability: the failure of one of the remaining active switches on the secondary side. In the high-stakes world of power electronics, switch failures are among the most common points of system breakdown. While a short-circuit failure is catastrophic and usually triggers immediate protective shutdowns, an open-circuit failure is more insidious. The switch simply stops conducting, but the system may continue to operate, albeit in a degraded and potentially dangerous state.
When an open-circuit fault occurs in an S-DAB converter, the carefully choreographed dance of current flow is thrown into chaos. The current, seeking an alternative path, begins to flow through the body diodes of other components. This unintended rerouting has two major, detrimental consequences. First, it introduces a direct current (DC) bias into the transformer’s leakage inductance. Transformers are designed to handle alternating current (AC); a DC bias can drive the transformer core into saturation. A saturated core loses its ability to store energy effectively, causing the current to spike uncontrollably. This surge can quickly exceed the maximum current ratings of the remaining semiconductor switches, leading to their thermal destruction in a cascading failure. Second, the fault distorts the voltage waveform on the secondary side of the converter. Under normal operation, the average voltage over a switching cycle at this point is zero. An open-circuit fault causes this average voltage to shift dramatically, either becoming strongly positive or negative, depending on which specific switch has failed. Left unaddressed, this fault condition doesn’t just reduce the converter’s power output; it actively works to destroy it from the inside out.
The conventional approach to enhancing reliability in power converters often involves redundancy—adding extra, identical modules that can take over if one fails. While effective, this strategy is expensive and increases the system’s size and weight, making it impractical for cost-sensitive and space-constrained applications like consumer EV chargers. The research team, led by Guan Shuo and Ma Jianjun, took a radically different and elegantly simple approach. They asked: instead of adding more hardware, can we reconfigure the existing circuit to safely bypass the fault? The answer they arrived at is the “Fault-Tolerant Single Active Bridge” (SAB) mode. The brilliance of this method lies in its minimalism. When a fault is detected in one of the two secondary-side switches, the system doesn’t try to repair or replace it. Instead, it deliberately shuts off the gate drive signal to the other, perfectly healthy switch on the same side. This action effectively transforms the entire secondary-side bridge from a semi-active circuit into a completely passive, diode-only rectifier bridge.
This reconfiguration is transformative. By forcing both secondary-side switches into an “off” state, the circuit regains its electrical symmetry. The current, no longer forced down an asymmetrical and biased path, flows naturally through the diodes in a balanced manner. This immediately eliminates the dangerous DC bias in the transformer, preventing core saturation and the resulting current spikes. The system stabilizes, operating in a new, albeit less powerful, mode. While the maximum power transfer capability in this fault-tolerant SAB mode is reduced to approximately 48% (12/25) of the original S-DAB’s maximum capacity, this is a far better outcome than a complete system failure. It allows the charging session to either continue at a reduced rate or to shut down gracefully, protecting the vehicle’s battery and the charger’s internal components. For an EV driver, this could mean the difference between a minor delay and being stranded with a dead charger.
Of course, a fault-tolerant strategy is only as good as its ability to detect the fault in the first place. A slow or inaccurate diagnosis renders the entire system useless. The researchers’ diagnostic method is as clever as their fault-tolerant approach. They leverage the very symptom that makes the fault so dangerous—the shift in the average secondary-side voltage (Us_avg)—as the primary diagnostic signal. Their system requires just one additional voltage sensor to monitor this specific point. When an open-circuit fault occurs, the Us_avg doesn’t just change slightly; it undergoes a significant and unambiguous shift. If the upper switch (S6) fails, Us_avg jumps to a large positive value. If the lower switch (S8) fails, it plunges to a large negative value. A simple logic circuit, comparing this measured average against pre-set positive and negative thresholds, can instantly and accurately pinpoint which of the two switches has failed. The team meticulously designed these thresholds to be larger than any voltage fluctuation caused by normal dynamic operation (like adjusting the charging current) but smaller than the shift caused by an actual fault, ensuring the system is sensitive to real problems while being immune to false alarms. In their experimental validation, this diagnostic system was able to identify and locate the faulty switch within just four switching cycles—a response time measured in mere microseconds, which is critical for preventing damage.
The implications of this research extend far beyond the laboratory. For manufacturers of EV charging equipment, this strategy offers a direct path to building more robust and reliable products without a significant increase in the bill of materials. The cost of adding a single voltage sensor and implementing the relatively simple control logic for the fault-tolerant SAB mode is negligible compared to the cost of adding redundant power modules or dealing with field failures and warranty claims. For charging network operators, increased converter reliability translates directly into higher station uptime and lower maintenance costs. A charger that can “limp home” after a component failure, rather than dying completely, is a charger that causes less disruption for drivers and generates more revenue over its lifetime. From a safety perspective, the ability to prevent catastrophic current surges and component explosions is paramount. This technology acts as a vital safety net, protecting not just the expensive power electronics but also the connected EV and, by extension, its owner.
Furthermore, the transition to the fault-tolerant SAB mode brings an unexpected benefit: an expanded range for Zero Voltage Switching (ZVS). ZVS is a highly desirable switching technique where a power switch turns on only when the voltage across it has naturally fallen to zero. This eliminates the energy loss that occurs when a switch turns on while still blocking a high voltage, known as “switching loss.” Reducing switching loss is crucial for improving overall efficiency and reducing heat, which in turn allows for smaller heatsinks and more compact designs. The analysis in the paper shows that the fault-tolerant SAB mode can maintain ZVS over a wider range of input and output voltages compared to the standard S-DAB operating mode under certain conditions. This means that even in its degraded state, the converter can operate more efficiently than one might expect, further mitigating the impact of the reduced power capacity.
The research team didn’t just theorize; they built a physical prototype to validate their concepts. Their experimental setup meticulously recreated the conditions of normal S-DAB operation, the chaotic state of an open-circuit fault, and the stabilized fault-tolerant SAB mode. The results were compelling. Oscilloscope traces clearly showed the dangerous DC bias current appearing immediately after a simulated switch failure (achieved by deliberately blocking its gate signal). They then demonstrated how, upon activating the fault-tolerant strategy, this bias current was rapidly and completely eliminated, restoring a clean, symmetrical current waveform. The diagnostic system successfully identified which switch was “faulted” within the promised four cycles. Power measurements confirmed that while the maximum power in SAB mode was lower, it was still a substantial and usable amount, proving the converter’s ability to continue functioning. The team also verified that ZVS was maintained for the primary-side switches in the SAB mode under the tested conditions, confirming the efficiency benefits of their approach.
In an industry where reliability and cost are locked in a constant tug-of-war, this research offers a rare and valuable win-win solution. It demonstrates that sophisticated fault tolerance doesn’t always require complex, expensive hardware. Sometimes, it can be achieved through intelligent software control and a deep understanding of the underlying circuit physics. By turning a potential point of failure into a trigger for a safe, reconfigured mode of operation, the team has created a system that is not just resilient, but also smarter. This “graceful degradation” philosophy is crucial for the future of EV infrastructure. As charging power levels continue to climb—moving from 50 kW to 150 kW, 350 kW, and beyond—the stresses on power electronics will only increase. Having converters that can withstand component failures without catastrophic consequences will be non-negotiable for ensuring the safety and reliability of these high-power systems.
The work by Guan Shuo, Ma Jianjun, Zhu Miao, and Zhang Dezhen represents a significant step forward in power converter design for electric mobility. Their strategy is not a niche academic exercise; it is a practical, implementable, and cost-effective engineering solution to a real-world problem. It addresses a critical gap in the literature, as previous fault-tolerance research had largely focused on the more complex DAB or other converter topologies, leaving the increasingly popular S-DAB without a dedicated, low-cost solution. By publishing their findings in the Transactions of China Electrotechnical Society, they have made this knowledge accessible to the global engineering community, accelerating its potential adoption in commercial products. As the world accelerates its transition to electric vehicles, innovations like this fault-tolerant S-DAB strategy will be instrumental in building the robust, reliable, and safe charging infrastructure that drivers can depend on.
By Guan Shuo, Ma Jianjun, Zhu Miao, Zhang Dezhen. Published in Transactions of China Electrotechnical Society, Vol.39, No.6, Mar. 2024. DOI: 10.19595/j.cnki.1000-6753.tces.230073.