Rechargeable batteries are the primary components in EVs, mobile devices, and energy storage systems. The batteries’ working conditions, including state of health (SOH), state of charge (SOC), and temperature, are essential to reliably and efficiently operate devices or equipment. Predicting battery SOH and SOC is becoming a priority in order to increase their performance and safety.
Physically, you can represent the batteries as an electrical circuit model, as shown in Figure 1. The resistors (Rs) and capacitors (Cs) in the model have good correlations with battery states. Electrochemical impedance spectroscopy (EIS) technologies are crucial to characterize the elements of the model in order to obtain the batteries’ working conditions.
Figure 1 The equivalent circuit of a battery showing Rs and Cs that have a good correlation with battery states. Source: Texas Instruments
Rs and Cs change when the batteries are in different states, leading to impedance changes. With EIS techniques, applying AC signals to the batteries and measuring their voltage and current response enables calculations of the impedance data of the batteries in the frequency domains. By analyzing the impedance data, you can know the battery’s SOC, internal temperature, and battery life. EV manufacturers are now researching how to apply EIS techniques to a battery management system (BMS).
Nyquist tool
Applying an AC voltage to a circuit excites the AC current. Equation 1 calculates the impedance, which varies as frequencies change if the circuit is not a pure resistance load.
Figure 2 illustrates Ohm’s law for an AC voltage. You can plot the impedance by applying many frequencies. Typically, a battery is modeled as Rs and Cs in combination, as shown in Figure 1. Figure 3 illustrates the impedance plot using a Nyquist tool.
Figure 2 Ohm’s law in an AC circuit, impedance can be plotted by applying many frequencies. Source: Texas Instruments
Figure 3 The plot of impedance using the Nyquist tool. Source: Texas Instruments
Methods of excitation current generation
You can use the EIS technique for one cell, multiple cells, modules, or a pack. Performing an EIS measurement requires the application of AC current to the batteries. For different battery system voltages, there are four different methods to generate the excitation current. Let’s review them.
Method #1: Resistor dissipation at the cell level and module level
In Figure 4, the power switch (S1), power resistor (Rlimit), sense resistor (Rsense), and a controller produce the excitation source. The controller generates a sinusoidal pulse-width modulation (SPWM) signal for S1. One or several battery cells are connected in series with the excitation source. Turning on S1 draws the current from the batteries through Rlimit. The energy burns and dissipates. When the voltage is high, the power dissipation is significantly large.
Figure 4 EIS with a resistor load where S1, Rsense, and the controller source produce the excitation circuit. Source: Texas Instruments
You can use this method at the cell level and small module level with low voltage, but it is not a practical solution for high-voltage batteries in EVs or hybrid EVs (HEVs) because the power dissipation is too great.
Method #2: An isolated DC/DC converter at the pack level
In an EV powertrain, high-voltage batteries charge low-voltage batteries through an isolated DC/DC converter (as shown in Figure 5), which you can design to support bidirectional power flow. During EIS excitation, power transfers from high- to low-voltage batteries during the positive cycle; power is then reversed from the low- to high-voltage side during the negative cycle. This method uses existing hardware without adding extra costs. However, the excitation source is limited by the capacity of the low-voltage batteries. It is particularly challenging for 800V-12V battery systems.
Figure 5 EV power train with high-voltage batteries charging low-voltage batteries through an isolated DC/DC converter. Source: Texas Instruments
Method #3: A non-isolated DC/DC converter in stack mode for the pack
This method uses a non-isolated DC/DC converter to generate excitation current between two battery modules. During EIS excitation, the charge transfers from Vbat1 to Vbat2 during the positive cycle, and the charge transfers back to Vbat1 from Vbat2. In Figure 6, two battery modules are connected in stack mode. Two active half-bridges are connected in series, and their switching nodes are connected through an inductor and a capacitor.
There are several advantages to this method: one is the use of low-voltage rating switches in a high-voltage system; the other is that the switches operate under zero-voltage switching (ZVS) conditions. Additionally, this method enables the production of a larger excitation current without adding stress.
Figure 6 A non-isolated DC/DC converter connecting two battery modules in stack mode. Source: Texas Instruments
Method #4: A non-isolated DC/DC converter in parallel mode for the pack
This method connects two battery modules in parallel mode, as shown in Figure 7. Two modules share a common ground. The charges are transferred to the inductor and capacitor from VBat1; then the charges stored in the inductor and capacitor are transferred to VBat2. The parallel mode and stack mode are swappable by properly reconfiguring two modules to meet different charging stations or battery voltage systems ,such as 400 V or 800 V.
Figure 7 A non-isolated DC/DC converter in parallel mode for the pack. Source: Texas Instruments
EIS measurement
Figure 8 divides the battery pack into two modules. S2 and S3 are battery-management ICs. The BQ79826 measures the voltage of every cell through an analog front end. Applying the AC current to battery modules builds up the AC voltage of each cell, which the BMICs then measure. A current measurement IC is used to measure the excitation current sensed by a current shunt. A communication bridge IC connects all BMICs through a daisy-chain communication bus. The BQ79826 uses the EIS engine to calculate the impedance, which is transmitted to a microcontroller for the Nyquist plot. The Controller Area Network (CAN) protocol provides communication while MCU1 controls the generation of excitation current.
Figure 8 Block diagram of an EIS measurement that divides the battery pack into two modules, each monitored by BMICs. Source: Texas Instruments
In a simulation of a non-isolated stacked active bridge (SAB) based excitation circuit, the conditions were VBat1 = VBat2 = 400 V, Fs =100 kHz, and current amplitude = 5 A. Figure 9 shows the excitation current waveform from the simulation. The blue trace is VBat1 current, while the green trace is VBat2 current.
Figure 9 Excitation current generated by stacked active bridge, the blue trace is the VBat1 current, the green trace is the VBat2 current. Source: Texas Instruments
The frequency synchronization between the controller of the excitation source and the BQ79826 is essential to minimize measurement errors. One solution is to take the SPWM signal generated by the BQ79826 as the reference of the excitation source (Figure 10). The excitation source and EIS engine of BQ79826 are automatically synchronized.
Figure 10 Block diagram for system timing synchronization of the excitation source and the BQ79826 in order to minimize measurement errors. Source: Texas Instruments
When building hardware to evaluate an EIS measurement, the excitation source should have high efficiency to minimize charge losses in the batteries. The total current harmonics should also be small in order to increase the signal-to-noise ratio (SNR). Figure 11 shows the efficiency measurement of the converter using a DC voltage and a DC load. With a higher excitation current, the power efficiency is higher because of the larger ZVS range. Above a 1-A amplitude of excitation current, efficiency is >95%. All the power dissipates in the traditional method of using a load resistor.
Figure 11 Efficiency measurement of a stacked active bridge-based power stage. Source: Texas Instruments
A fast Fourier transform (FFT) is a tool to evaluate the SNR of the excitation current. Placing six 18650 batteries in series for one module, with two modules connected to the stacked active bridges, demonstrates the quality of excitation current. In Figure 12, two tones of 10 Hz and 100 Hz are generated simultaneously to reduce the excitation time.
Figure 12 FFT of excitation current where two 10 Hz and 100 Hz tones are generated simultaneously. Source: Texas Instruments
Figure 13 is a Nyquist plot showing the impedances of different cells using two 200-V battery modules. At lower excitation frequencies, the difference between the measurements is small. There are more discrepancies at higher excitation frequencies (shown on the left side of the graphs), but the impedances within this range are not important.
Figure 13 Nyquist plot showing the impedances of different cells using two 200-V battery modules. Source: Texas Instruments
EIS technique
EIS is an evolutionary technique with applications for EV and HEV batteries. EIS techniques enable users to obtain real-time information about the SOC, SOH, and temperature during battery system operation.
Achieving good EIS results still requires resolving challenges such as developing accurate algorithms, utilizing reliable excitation systems, and minimizing noise sensitivity.
Sean Xu currently works as a system engineer in Texas Instruments’ Power Design Services team to develop power solutions using advanced technologies for automotive applications. Previously, he was a system and application engineer working on digital control solutions for enterprise, data center, and telecom power. He earned a Ph.D. degree from North Dakota State University and a Master’s degree from Beijing University of Technology, respectively.
Related Content
- Power Tips #144: Designing an efficient, cost-effective micro DC/DC converter with high output accuracy for automotive applications
- Power Tips #143: Tips for keeping the power converter cool in automotive USB PD applications
- Power Tips #75: USB Power Delivery for automotive systems
- Power Tips #101: Use a thermal camera to assess temperatures in automotive environments
The post Power Tips #145: EIS applications for EV batteries appeared first on EDN.
