What you'll learn:
- Details on a BMS based on electrochemical impedance spectroscopy (EIS).
- BMS advances to keep pace with EV rising complexity.
- Building better battery models based on EIS characterization.
The battery-management system (BMS) is the brains of the electric vehicle (EV). The BMS is all about monitoring the lithium-ion (Li-ion) cells in the battery pack and regulating the voltages, currents, and temperatures of the highly combustible assembly. By better understanding the internal state of the battery and adapting to different driving conditions, it ensures the EV battery’s peak performance along with its longevity, reliability, and safety.
Every commercial BMS collects lots of data. Typically, it measures the voltage of every battery cell and the voltage and current of the pack. A limited number of temperature sensors—strategically placed—map out the temperature distribution of the pack.
The algorithms at the heart of the BMS run through these signals to give a relatively coarse view of the battery’s condition, said Davide Cavaliere, a director of product management at Marelli, which is bringing one of the most widely used tools for lab-testing batteries into its latest BMS.
Due out in 2025, the Marelli-designed BMS will be based on electrochemical impedance spectroscopy (EIS), which is widely used as a diagnostic tool in battery labs. By tracking changes in impedance, Cavaliere said the next-gen BMS can deliver deeper insight into the state-of-charge (SOC) and state-of-health (SOH) of the battery pack. Then, it can use these insights to increase the EV’s range and bolster the battery’s lifespan.
The Tier-1 supplier said the BMS uses EIS to evaluate the degradation rate as the battery ages or operates in harsh conditions. Impedance is one of the main indicators of the battery’s internal state. Resistance tends to increase as the battery discharges, and the battery’s degradation over time also causes resistance to rise. These revelations into battery aging are crucial for calculating a battery’s remaining useful life, which is what determines its residual economic value.
EIS is typically done during battery R&D or as part of the manufacturing process of EV battery cells. But engineers have been striving for more than a decade to integrate EIS directly into the cell-level BMS. The problem? The cost of the complex hardware, software, and system design made it impractical for use in EVs. For its part, Marelli said its EIS-based BMS is finally cost-effective and scalable enough for mass production.
The BMS: The Difference Maker in EV Battery Packs
The BMS is playing a more pivotal role in the EV, especially due to the growing complexity of high-voltage battery packs under the hood, which can account for as much as 30% to 40% of the average EV’s net cost. Modern EV battery packs consist of hundreds to thousands of electrochemical cells, which can deliver high power density and last years before they must be replaced. The tradeoff is that they’re very sensitive.
Each battery is based on a different chemistry at the cell level and construction at the pack level. On top of that, the performance of the battery cells can be impaired by impurities or slight deviations in the production process.
These EV batteries also react to real-world conditions under the hood—electrical variations during use can cause premature charge termination, physical degradation, and catastrophic failures. Moreover, they’re vulnerable to vibrations and pressure as well as heat, cold, and any other harsh environmental conditions.
One of the main responsibilities of the BMS is also its most challenging: state estimation. The BMS uses parameters such as voltage, current, and internal resistance to indirectly estimate the SOC, which reflects the battery’s remaining charge, and SOH, which quantifies the battery’s internal condition by comparing its present capacity to its original capacity. The BMS is in charge of uniform charging and discharging of the battery’s cells as well. Cell balancing helps to maximize the EV’s battery performance and its useful lifespan.
Furthermore, the BMS is responsible for safety monitoring. The unit constantly monitors temperature, voltage, and current at the cell and pack level to prevent damage that can occur due to overcharging or undercharging. And it protects against dangers caused by short circuits or other faults, including thermal runaway.
The BMS also corrects for temperature fluctuations inside the battery, making sure it stays at an optimal temperature. Exposure to extreme heat and cold can lead to large variances in the battery’s performance and longevity.
The other role of the BMS comes down to data communication. It sends information about the battery to the driver and other systems, including the traction inverter, onboard charger (OBC), and DC-DC converter.
At the heart of the BMS is a microcontroller (MCU) that uses everything it learns at both the cell and pack levels to calculate the internal state of the battery pack. By keeping close tabs on the situation, it can adjust the charging and discharging to bring the best out of the battery. Battery-management ICs are tasked with cell-level voltage monitoring, while current sensors, isolation monitors, high-voltage contactors, and pyro fuses are used to physically disconnect the battery pack from the rest of the EV when in danger.
While many companies are bringing AI into the fold, the algorithms that run inside the MCU are primarily based on “battery models” that stem from rigorous testing in research and development. These depend on highly accurate cell voltage measurements to estimate the battery’s internal state. The accuracy is a bigger deal in lithium-iron-phosphate (LFP) batteries, which can present challenges for state estimation due to flat voltage profiles. Even small deviations in cell-level monitoring can lead to large prediction errors.
EV Battery Testing: The Hidden Secret to the Modern BMS
As noted by Veronica Wright, a leading consultant on EV technology, no “one-size-fits-all” solution exists for battery management. Every battery pack is a very complex interplay of electrical, chemical, and mechanical properties.
Ensuring the optimal performance and safety of the battery pack for years under different driving conditions and unpredictable use patterns is not trivial, nor is building a BMS that can adapt to these and all of the other variables out on the road. The BMS must also be calibrated to the battery’s architecture.
To understand the unique profile of the individual cells and the pack design, batteries are subjected to years of rigorous testing under various conditions. This tends to involve repeated charging and discharging of the cells—also known as cycling. By analyzing the fluctuations in current and voltage, researchers can uncover patterns that predict how long a battery will endure before it can no longer hold enough charge for EV use.
These vast amounts of data are the foundation for battery models, which act as the backbone of traditional BMS algorithms, said Wright. These models are also backed up by huge databases that correlate different parameters of the battery. For instance, the open circuit voltage (OCV) is connected to the SOC of the battery over a wide range of temperatures, typically −20 to 60°C. Other characteristics such as predictions about degradation rate and the self-discharge rate can be incorporated, too.
One of the most widely used diagnostic tools is EIS. It measures the impedance—the AC resistance—of the battery cell. By applying a small amount of AC to the cells and measuring the response over a wide frequency range, EIS maps the impedance spectrum. This reveals minute details about the electrochemical processes at work inside the battery pack that are invisible to other types of evaluation, specifically any negative influences on battery life and performance.
The important metric with EIS is the frequency range, which is critical for capturing all of the electrochemical processes inside the battery. These occur at different speeds, with slow reactions such as ion diffusion—in which the lithium ions move between the anode and cathode inside the battery—typically observed at lower frequencies. The physical movement of these specks from one side of the battery’s structure to the other is what creates electrical current, and it is impedance in the battery that impedes this movement.
In contrast, rapid reactions such as charge transfer—a fundamental process that occurs at the insulating interface between the electrode and the electrolyte that separates the two sides of the battery—require high-frequency measurements. Most commercial test equipment can cover frequencies from under 1 Hz up to 1 MHz. Higher-frequency measurements are used to highlight what happens in the metal conductors inside the battery and at the surfaces of the electrodes, while lower frequencies capture other phenomena.
The impedance value at any given frequency serves as a unique signature for the electrochemical reactions taking place in the battery without cutting it open or damaging it. EIS also helps predict the degradation rate of the battery and pinpoint defects and other imperfections that impact quality and safety. For instance, the underlying process that leads to lithium metal plating is highly undesirable. But EIS can predict the risk of it.
EIS and other electrical tests are supplemented by more strenuous physical tests, where companies pound the batteries, heat them up, and punch holes through them to see what it takes to get them to malfunction.
Bringing the Benefits of the Battery Lab Inside the EV
Using EIS to measure impedance inside the EV is no easy task. But Marelli and other companies are trying to tackle the difficulties by building battery models based on EIS characterization and bringing EIS directly into the cell-level BMS. The main hurdle is the cost, said Cavaliere. While companies building battery labs can invest in high-quality test equipment for EIS, the issue is bringing the same performance to the EV and keeping it within the constraints of a commercially viable BMS.
Marelli called its current generation of battery-management systems “EIS-ready,” and the company said it made several significant improvements in its ability to measure impedance in the lower-frequency ranges.
To do EIS, the Marelli-designed BMS solution uses controllers already present in the EV, such as the OBC or DC-DC converter, to apply the AC voltage to the battery pack, enabling pack-level stimulation. As Cavaliere explained, the BMS first measures the time-domain signals of the cell voltages, pack voltage, and pack current. Then it runs fast Fourier transform (FFT) analysis in the frequency domain to figure out the impedance of the cells and the entire pack at varying frequencies.
In terms of the underlying hardware, the BMS relies on high-precision battery-monitoring ICs that can sample the voltage and current of the battery cells with high accuracy and high frequency, along with the synchronization of current-voltage acquisitions. The MCU has the high clock frequency and fast sampling rate to run the FFT, which occurs after the battery-monitoring IC sends the raw data on cell voltage, pack voltage, and pack current.
The "Full EIS" solution will be able to handle higher-frequency measurements when it launches later in 2025, offering a fuller depiction of the battery’s internal condition. This will improve real-time predictions of the SOC and SOH of the battery as well as its internal temperature, said Cavaliere.
In the EIS-ready solution, the MCU samples the voltage and current of the battery cell and the larger pack every 1 to 10 ms due to the constraints on the daisy-chain communication protocol between the MCU and the battery-monitoring IC. That limits the maximum frequency of the EIS to approximately 10 Hz, which is significantly less than the current standard in EV battery labs.
Marelli said plans are to upgrade the battery-monitoring IC to run FFT directly, increasing the frequency range above 1 kHz, which will allow for more accurate measurements of the SOC and SOH as well as temperature. "By performing the FFT on the BMIC directly, the daisy-chain communication channel is no more a bottleneck, so higher frequencies can be measured," Cavaliere said.
The EIS-based BMS can also keep close tabs on the thermal situation inside the battery pack, said Cavaliere. As the temperature inside the battery pack rises, the impedance drops due to lower resistance in the electrolyte, which influences the rate of charge transfer and other electrochemical reactions in the battery cell, he said. Thanks to that, the EIS-based BMS can potentially detect the onset of overheating or even thermal runaway.
"Measuring the impedance at high frequencies is important to evaluate the resistive components of the cells," Cavaliere said. "This resistance strongly depends on the internal temperature of the cell. Consequently, [by] measuring the resistance, it is possible to have an estimation of the internal temperature of the cell during its operation. This information is particularly important during fast charging to limit the damage due to overheat."
One of the challenges in bringing EIS inside the EV is creating the “stimulus” for the measurement. While it uses other controllers present in the EV to feed the AC into the battery before measuring the frequency response, the BMS must control the amount of AC entering it very tightly.
Cavaliere explained, “It is important that the current stimulus is reproduceable and not too high in amplitude to avoid non-linearity issues, yet not too small in amplitude in order to generate a sufficient signal that can be distinguished by noise.”
On a related note, he said it makes more sense to measure the impedance of the battery cells and the larger pack when the EV isn’t charging or discharging rapidly—for instance, when it’s parked or charging slowly.
The BMS: The Secret to Longer-Lasting EV Batteries?
Today, the world’s largest battery manufacturers and car companies are kicking the tires on new materials, new chemistries, or other tweaks to build batteries that have larger storage capacities, charge more rapidly, or are safer than those currently in use. But these technological leaps tend to have extensive research and development phases or require rigorous testing. There tend to be tradeoffs in performance or cost, too.
But innovations inside the BMS present a route to improving several performance metrics at the same time, said Alex Holland, EV battery technology analyst at IDTechEx, and that’s a tall order when it comes to Li-ion battery chemistries.
For instance, faster EV charging times can be accomplished by tracking the internal state of each cell in the battery, said Holland. If the EV detects the onset of lithium plating during charging, the BMS can reduce the charging current to reduce the risk before increasing the current once the risk is reduced. He said that by offering more accurate state estimation, the BMS can increase the charging rate without increasing degradation.
EIS is only of several new technologies entering the BMS. For instance, Marelli said it plans to bring wireless cell monitoring, active balancing, and contactless current sensors to the table within several years. Cavaliere added that it also plans to integrate AI algorithms locally into the BMS, as well as in the cloud, to capture the SOC and SOH more accurately and track these metrics over longer periods of time.
Giovanni Mastrangelo, CTO of Marelli’s Propulsion Solutions, said the integration of EIS and AI would give car companies a technological edge by “enhancing safety, reliability, and performance standards in [EV] battery management.”
