Driving Habits Directly Influence Battery Degradation, New Study Finds
A groundbreaking study conducted by researchers at Beijing Jiaotong University has revealed a critical connection between driver behavior and the rate at which electric vehicle (EV) batteries degrade. The research, led by Fu Zhicheng, Sun Bingxiang, Jia Yiming, Gong Minming, Ma Shichang, and Pang Junfeng from the National Active Distribution Network Technology Research Center, uncovers how subtle differences in acceleration patterns significantly impact the lifespan of lithium-ion batteries. This discovery not only challenges conventional methods of battery health assessment but also paves the way for more personalized and accurate battery management systems in future EVs.
For years, the automotive industry has focused on external factors such as temperature, charging cycles, and overall mileage when evaluating battery longevity. While these elements are undeniably important, the new study argues that internal, driver-specific behaviors play an equally, if not more, crucial role. The research team set out to answer a fundamental question: Why do two identical EVs, driven under similar environmental conditions and covering the same distance, often exhibit vastly different battery degradation rates? The answer, they found, lies in the driver’s foot.
The investigation began with a critical observation about existing data. Most prior research on battery aging relies on laboratory tests or real-world data collected at intervals of 10 seconds or longer. The team, however, argued that this coarse-grained data is insufficient to capture the rapid, second-by-second fluctuations in current that occur during real driving. “A 10-second snapshot of data is like trying to understand a symphony by listening to one note every 10 seconds,” explained Dr. Sun Bingxiang, the study’s corresponding author. “You miss the melody, the rhythm, and the dynamics. The same is true for driving. The nuances of how a driver accelerates—whether they do it in one smooth motion or a series of short bursts—happen on a timescale of tenths of a second. We needed data with a resolution of 0.1 seconds to see the real picture.”
To achieve this, the researchers utilized two primary data sources. The first was the China Light-Duty Vehicle Test Cycle for Passenger Cars (CLTC-P), a standardized driving cycle designed to reflect typical Chinese urban and suburban driving conditions. This provided a controlled baseline. The second source was real-world driving data collected from a fleet of test vehicles equipped with high-frequency data loggers, capturing battery voltage, current, state of charge (SOC), and temperature at a rate of 10 times per second. By analyzing this high-resolution data, the team could dissect each driving event with unprecedented precision.
The analysis focused on the discharge current, the flow of electricity from the battery to the motor, which is directly influenced by the driver’s throttle input. The researchers developed a novel method to segment the continuous stream of current data into distinct “fragments” based on the vehicle’s speed and acceleration. They categorized these fragments into three types: acceleration fragments, transition fragments, and braking fragments. Braking fragments, which involve regenerative braking and result in negative current, were found to have a negligible impact on degradation, accounting for less than 5% of the energy cycle in the CLTC-P test, and were therefore excluded from the core analysis.
The key insight emerged from the segmentation of the acceleration process. The team observed that what might appear as a single “acceleration” from a distance is often composed of multiple short, sharp bursts of power. They termed this a “segmented acceleration process.” In contrast, a “continuous acceleration process” involves a single, sustained application of throttle. The proportion of segmented versus continuous acceleration became a central metric for quantifying driving habits. For instance, a cautious, smooth driver might exhibit a high percentage of continuous acceleration, while an aggressive, impatient driver would show a much higher percentage of segmented acceleration, characterized by frequent “pulses” of power.
To make sense of the vast amount of data, the researchers employed two powerful statistical tools: Principal Component Analysis (PCA) and BI-KMEANS clustering. PCA is a dimensionality-reduction technique that identifies the most important variables within a complex dataset. When applied to the eight initial current characteristics—maximum current, average current, duration, current rise and fall rates, and so on—the analysis revealed that the two most significant factors were the maximum current (Imax) and the fragment duration (T). These two parameters captured the essence of the driving behavior’s impact on the battery.
With these two key parameters identified, the BI-KMEANS clustering algorithm was used to group the thousands of acceleration fragments into distinct categories based on their similarity. This process revealed five distinct “typical” current patterns within the CLTC-P data. The first, labeled “Typical(1),” was a short-duration, low-to-moderate current pulse, resembling a trapezoid. This fragment was associated with gentle, incremental acceleration, such as when a driver makes small adjustments to maintain a gap in traffic. The second, “Typical(2),” was a triangular wave, indicating a linear increase in current, which occurs when the driver accelerates smoothly but does not push the motor to its maximum power limit. The remaining three categories, “Typical(3),” “Typical(4),” and “Typical(5),” were all high-current, trapezoidal pulses of varying durations and intensities, representing more aggressive driving behaviors like rapid merging or overtaking.
When the same analysis was applied to the real-world driving data from five different test drivers, the differences became stark. The drivers were not selected for their aggressive or conservative styles; they were ordinary individuals asked to drive the same route under the same conditions. Yet, their driving habits varied significantly. One driver had an average speed of 34.9 km/h, with segmented acceleration accounting for 51.6% of the driving time, while another drove at an average of 24.3 km/h, with segmented acceleration occurring in only 27.8% of the time. The clustering analysis of their data showed clear differences in the distribution of the typical current fragments. The driver with more segmented acceleration produced a higher proportion of “Typical(1)” pulses, while the smoother driver had a higher proportion of longer, more continuous fragments.
This was the crucial link: different driving habits create different patterns of electrical stress on the battery. But how does this translate into actual physical degradation? To answer this, the research team moved from data analysis to physical experimentation. They designed a series of controlled laboratory tests to isolate the effects of the different current patterns they had identified.
The experiment used brand-new ternary lithium-ion battery cells, identical to those used in the real-world test vehicles. The goal was to simulate the effects of different driving habits over an extended period. The researchers created five distinct “discharge conditions” that mirrored the key characteristics of the real-world fragments. Condition 1 simulated a driving style dominated by short, frequent pulses. Condition 2 and Condition 3 simulated longer, more sustained acceleration events of different durations, representing different average speeds. Condition 4 was a high-intensity version of Condition 1, with much higher peak current. Finally, Condition 5 was designed to have the same total energy output as Condition 4 but with fewer, longer pulses, thus reducing the number of times the battery experienced a sharp current change.
Each battery cell was subjected to 100 full charge-discharge cycles under one of these five conditions. The charging process was kept identical for all cells—constant current followed by constant voltage—to ensure that only the discharge profile varied. After every 50 cycles, the researchers measured the battery’s remaining capacity, the most direct indicator of its state of health (SOH).
The results were both dramatic and definitive. After 100 cycles, the batteries subjected to Condition 4, the high-intensity, high-frequency pulsing, showed the most severe capacity loss. This was followed closely by the cells in Condition 1 and Condition 5, which also involved frequent pulsing, albeit at lower intensity. The batteries in Condition 2 and Condition 3, which experienced longer, more stable discharge periods, showed the least degradation. This finding directly contradicted the common assumption that higher average speeds (represented by longer discharge durations) are the primary driver of battery wear.
The study concluded that the dominant factor in battery degradation is not the total energy drawn or the average speed, but rather the number and intensity of the “polarization events” caused by rapid changes in current. Every time a driver rapidly presses the accelerator, the battery experiences a sudden surge in current. This forces the electrochemical reactions inside the battery to accelerate, creating a phenomenon known as “polarization.” This is a temporary imbalance in the voltage across the electrodes. When the driver releases the pedal, the current drops, and the battery “relaxes.” This constant cycle of stress and relaxation causes irreversible damage over time.
The damage occurs at the microscopic level. The repeated expansion and contraction of the electrode materials during charge and discharge cycles can cause them to crack and lose contact with the conductive matrix. The solid electrolyte interphase (SEI) layer, a protective film on the anode, can grow thicker and more unstable under the stress of frequent pulses, consuming active lithium ions and increasing internal resistance. Furthermore, high-current pulses generate more heat, which can accelerate all of these degradation mechanisms.
The research provides a clear explanation for the differences observed between drivers. A driver who frequently “dabs” the accelerator, making small adjustments, subjects their battery to hundreds of polarization events during a single trip. A smoother driver, who accelerates in longer, more deliberate bursts, creates far fewer of these damaging events, even if they cover the same distance at a similar average speed. The study’s data showed that the number of polarization events and their intensity (linked to the peak current) are the primary drivers of the observed differences in battery degradation.
This research has profound implications for the future of electric mobility. For consumers, it suggests that adopting a smoother, more predictable driving style can significantly extend the life of their EV’s battery, potentially saving thousands of dollars in replacement costs. For automakers, it means that future battery management systems (BMS) could be made much smarter. Instead of using a one-size-fits-all model for SOH estimation, a BMS could analyze a driver’s historical current patterns and provide a personalized, more accurate prediction of battery lifespan. This could be used to offer tailored driving tips, optimize charging strategies, or even inform warranty policies.
Moreover, the findings are crucial for fleet operators and car-sharing companies. By training drivers in smoother driving techniques, these organizations could dramatically reduce their maintenance costs and extend the usable life of their vehicles. It also provides valuable data for urban planners and policymakers. If a city’s traffic flow is designed to minimize stop-and-go driving and encourage smoother traffic patterns, it could have a direct, positive impact on the longevity of the EVs operating within it.
The study also highlights the importance of high-resolution data in understanding complex systems. The shift from 10-second to 0.1-second data was not just a technical upgrade; it was the key that unlocked the entire discovery. It demonstrates that to truly understand the real-world performance of EVs, researchers must look beyond standardized tests and delve into the granular details of how these vehicles are actually used.
In conclusion, this research from Beijing Jiaotong University provides a fundamental shift in our understanding of EV battery degradation. It moves the conversation from external, environmental factors to the internal, behavioral ones. It proves that the driver is not just a passenger in the vehicle’s lifecycle but an active participant in the battery’s health. By understanding the physics of how driving habits create electrical stress, we can now take concrete steps to mitigate that stress and build a more sustainable, cost-effective future for electric transportation. The simple act of pressing the accelerator pedal, it turns out, carries a much heavier consequence than previously thought.
Fu Zhicheng, Sun Bingxiang, Jia Yiming, Gong Minming, Ma Shichang, Pang Junfeng, National Active Distribution Network Technology Research Center, Beijing Jiaotong University, Power System Protection and Control, DOI: 10.19783/j.cnki.pspc.246019