Lithium-Ion Battery Condition Monitoring: A Frontier in Acoustic Sensing Technology

Abstract

Lithium-ion batteries (LIBs) are widely used in the fields of consumer electronics, new energy vehicles, and grid energy storage due to their high energy density and long cycle life. However, how to effectively evaluate the State of Charge (SOC), State of Health (SOH), and overcharging behavior of batteries has become a key issue in improving battery safety and lifespan. Acoustic sensing technology, as an advanced non-destructive monitoring method, achieves real-time monitoring of the internal state of batteries and accurate evaluation of key parameters through ultrasonic testing technology and acoustic emission technology. This article systematically reviews the research progress of acoustic sensing technology in SOC, SOH, and overcharge behavior evaluation of LIBs, analyzes its working principle and application advantages, and explores future optimization directions and industrialization prospects. Acoustic sensing technology provides important support for building efficient and safe battery management systems.

1. Introduction

As an efficient and stable electrochemical energy storage device, LIBs are widely used in consumer electronics, new energy vehicles, and grid energy storage due to their advantages of high energy density, long cycle life, low self-discharge [1,2,3], and significant progress in the past few decades [4,5,6,7]. Over the past two decades, researchers have been working on the development of LIB systems with high energy density and long cycle life and have achieved remarkable results [8,9,10,11]. From the first invention of LIBs using TiS2 cathode and lithium metal anode in 1976 by Whittingham et al. [12] to the doubling of energy density by Goodenough et al. [13] in 1980 using lithium cobalt oxide cathode to the successful commercialization of graphite anode by Sony in 1991 [14], LIB technology has been continuously innovated. However, with the improvement of people’s requirements for energy density, cycle life, and safety performance [15,16,17], traditional carbon-based anode materials have exposed bottlenecks due to capacity limitations, which has promoted extensive research on new electrode materials and technologies [18,19,20].

The internal structure of LIBs is complex, including multiple layers of electrodes and porous membranes, which directly affect their electrochemical performance and thermal management behavior. The actual operating conditions of LIBs are often complex and variable, which may lead to serious safety issues such as rapid capacity decay and even thermal runaway in extreme environments [21,22]. In the context of large-scale applications of renewable energy, alternative battery technologies such as redox flow batteries and lead-acid batteries have also received attention. The lifespan of redox flow batteries can reach 20 years, providing a promising solution for large-scale energy storage. Similarly, lead-acid batteries remain a viable option due to their cost-effectiveness and reliability.

Table 1. Comparison of battery technologies.

Battery Type	Energy Density	Cycle Life	Safety	Cost	Applications
Lithium-ion	High	Long	Moderate Risk	Moderate	Consumer electronics, EVs, Grid storage
Redox Flow	Moderate	Very Long	Low Risk	High	Large-scale energy storage
Lead-acid	Low	Moderate	Low Risk	Low	Backup power, grid storage

provides a brief comparison of these battery technologies.

Despite the advantages of redox flow and lead-acid batteries, LIBs remain the primary choice for many applications due to their superior energy density and cycle life. However, the safety and lifespan of LIBs are key factors that need to be addressed to ensure their reliable use in various applications. Therefore, researchers are committed to exploring the reaction mechanism, aging mechanism, and safety hazards inside lithium batteries, as well as achieving efficient characterization through the development of advanced characterization techniques. In 2020, the “Battery 2030+” R&D roadmap released by Europe clearly pointed out that the existing Battery Management Systems (BMS) have insufficient monitoring capabilities at the individual battery level, such as the inability to directly measure the internal temperature of the battery [23].

Therefore, the development of advanced non-destructive sensing technology has become a key way to improve the performance and safety of LIBs by converting physical changes and chemical reactions during battery operation into readable electrical signals for timely warning of potential risks [24,25]. In recent years, the introduction of new sensing technologies has made significant progress in the field of in situ battery monitoring [26,27], laying a solid foundation for the construction of efficient and safe LIB systems.

2. Acoustic Sensing

As a non-destructive testing method, acoustic sensing technology has the characteristics of strong penetration, non-destructiveness, and high sensitivity and has been widely used in medical, industrial, and architectural fields [28]. The essence of sound waves is mechanical waves, which are a form of periodic mechanical vibration propagation in a medium. When the sound wave passes through the object and interacts with it, the characteristic information such as sound velocity, attenuation, and frequency reflected by the object after passing through the object can be observed and measured, and the changes in the elastic modulus, internal stress and other parameters of the object material can be obtained, and then the material properties and internal structure can be accurately evaluated. According to these characteristics of acoustic sensing, only a probe needs to be placed outside the battery to detect the internal structure of the battery and obtain the internal information of the battery, which fundamentally solves the difficulties encountered by implantable sensors and is an ideal non-destructive monitoring method of batteries [29]. The principle of acoustic sensing can be divided into two categories: acoustic emission (AE) testing and ultrasonic testing (UT) (Figure 1); the differences are shown in

Table 2. Comparison between AE technology and UT technology.

Name of the Technology	AE	UT
The principle of detection	Detect acoustic radiation signals generated when the internal structure of a material changes	Detecting the sound wave signal after the interaction between ultrasonic waves and materials
Sound wave frequency	Full frequency band (usually 20 kHz~1 MHz)	Usually ranging from 0.1 to 15 MHz
Signal parameter indicators	Peak frequency and intensity	Flight time and peak intensity
Testing equipment	1 acoustic probe	2 acoustic probes
The way of detection	Passive detection	Active detection

.

2.1. Acoustic Emission Technology

When the internal structure of a material undergoes irreversible changes (deformation, fracture, etc.), the material will autonomously undergo acoustic radiation, and the technology that detects, records, and analyzes this acoustic signal using an acoustic detection probe is called AE technology (Figure 2). However, the battery will continue to undergo periodic volume changes during the cycle, and its mechanical evolution process will usually lead to irreversible structural changes in the electrode material, resulting in battery failure. Therefore, AE technology is very suitable for detecting the phase transformation process of electrode materials in cycling and their mechanical failure mechanisms.

In addition to the degradation mechanism of the electrolyte, the mechanical evolution of the electrode often leads to a fracture event and the corresponding energy release. Therefore, the sound waves emitted by these events depend on the intrinsic properties of the material (crystal structure, grain size, phase transition, etc.), as well as the stresses generated during cycling [30,31], as shown in Figure 3. Therefore, AE characterization, being a non-destructive and passive technique, is particularly suitable for monitoring electrode materials that undergo phase transitions and repeatedly expand and contract during cycling. This results in various types of cracks and locations within the material. AE is extensively used in civil engineering for the early detection of damage, such as cracks and fatigue, in concrete structures. Additionally, it is applied in the aerospace sector to facilitate predictive maintenance by mapping strains in aerospace components.

Thanks to AE technology, researchers used it to study the mechanical failure mechanism of alumina electrolyte in high-temperature sodium-sulfur batteries as early as the 70s of the 20th century and found that the failure of the electrolyte is closely related to the crack propagation caused by sodium dendrite growth, which laid the foundation for subsequent research on battery acoustic detection [32]. Subsequently, AE technology has been gradually applied to a variety of battery systems to detect the deformation and gas-producing behavior of electrode materials in batteries so as to explore their failure mechanisms. For example, by monitoring the graphite anode [33], the silicon anode [34], and the lithium cobalt oxide cathode [35], the researchers captured typical ultrasonic signals and demonstrated their potential to reveal the internal processes of the battery.

Despite the advantages of non-destructive testing of AE technology, its application in the battery field still faces challenges. The internal structure of the battery is complex, and the change in the acoustic signal can be caused by a combination of factors, such as electrode particle breakage, SEI or CEI growth, etc. This diversity makes it necessary to accurately distinguish between different types of AE events in combination with characterization methods such as scanning electron microscopy (SEM) and electrochemical impedance spectroscopy (EIS) [35]. For example, in LiCoO₂/C cells, AE signals can be attributed to LiCoO₂ fragmentation and SEI growth, respectively, by combining SEM measurements [35]. Similarly, in the study of LiNiO2/Li batteries, three different types of acoustic events have been identified by AE technology, but the precise connection between these signals and their physical origins still relies on cell voltage guidance and complementary characterization methods [36].

In addition, AE technology still has limitations in the breadth and depth of practical applications. For example, in the 90s of the 20th century, Ohzuku et al. installed an acoustic transducer on a Li/MnO₂ coin battery and, for the first time, correlated it with the mechanical grinding, phase transition, and gas evolution of the electrode by analyzing the AE signal generated during cycling, demonstrating the potential of AE technology in predicting battery failure [30]. This achievement lays a foundation for subsequent research on the formation of SEI, the volume change in graphite and silicon electrodes, and the chemical and mechanical effects of layered oxide electrodes. However, in order to widely apply AE technology in more complex battery systems, it is still necessary to solve the problems of signal interpretation and multi-factor interference further.

2.2. Ultrasonic Testing Technology

Compared with AE technology, UT technology is an active detection technology and is more widely used. Figure 4 is a schematic diagram of LIB ultrasonic transmission testing [37], in which the stress wave excited by the ultrasonic emitter propagates in the material, undergoes reflection, attenuation, transmission, and other behaviors, is received by the ultrasonic receiver, and then the transmitted or reflected ultrasonic signal (such as the time of flight (TOF) and the peak of the first echo detected by the ultrasonic wave) is analyzed to accurately evaluate the structural changes and defects inside the material. Therefore, UT technology can reflect the changing structure and possible defects during battery cycling in real-time so that important parameters related to the battery state can be obtained [38], which can be used to detect internal damage in the battery [39], interlayer gas accumulation [40], and electrolyte wetting [41].

In the past decade, many researchers have used UT technology to monitor various battery systems and developed a series of methods to explore the battery state using ultrasonic technology. Huang et al. [41] used a focused ultrasonic beam to accurately scan the battery and detect electrolyte wetting in square and pouch cells using the inconsistencies of the ultrasonic signal attenuation in solids, liquids, and gases (Figure 5). While analyzing the changes in transmittance of different batteries, it can also quickly determine the minimum electrolyte injection amount and wetting time, which can accurately image the electrolyte distribution and gas production of the battery, which is of great significance for the process optimization in the actual battery production process. At present, this technology has been widely used in the study of battery gas production [42,43] and electrode-electrolyte interface [44].

Furthermore, Steingart et al. [45] carried out ultrasonic monitoring of pouch cells and cylindrical batteries during charging and discharging, took the changes in the amplitude of acoustic TOF and acoustic wave signals as the key indicators of critical phenomena in the batteries, and proposed to correlate the SOC and SOH of the batteries with the subtle changes in the materials and between layers, which provided a feasible operation for ultrasonic monitoring of the working state of the batteries. A simpler and practical approach to the use of ultrasound in the field of batteries was recently elaborated by J. Dahn et al. [41,46], who developed an ultrasonic imaging device consisting of a pair of movable ultrasonic transducers located on each side of a bag-shaped cell, and the entire cell was immersed in silicone oil for ultrasonic coupling. Due to the significant variation in signal attenuation between dry, wet, or gas-containing electrodes, the authors successfully visualized the electrolyte distribution within the battery by monitoring the amplitude of the transmitted wave, depending on immersion time and electrolyte content. Furthermore, by cycling lithium-ion NMC/C cells under different temperatures, cut-off voltages, and electrolytes, they were able to detect the early stages of gas formation and determine the number of cycles before the battery ran dry. This approach offers the potential for using ultrasound imaging to aid in the electrolyte-filling process of lithium-ion pouch batteries and to monitor battery performance over time. This method could greatly benefit in-line inspections at LIB manufacturing facilities, provided that proper management of silicone oil inclusions is ensured.

During the long-term cycling of the battery monitored by ultrasonic, Huang et al. [43] found that when the battery was aging, the overall waveform of the ultrasonic signal showed poor consistency due to the large fluctuation and poor reversibility of the physical structure characteristics in the battery, and its intensity decreased with the increase in the number of cycles of the battery. At the same time, this change in the ultrasonic signal is more drastic than the electrochemical reaction, which can be used as a reference index for battery life and provides a possibility for battery life prediction.

New research is using acoustic ultrasound characterization, an active but non-destructive technique that relies on the use of two piezoelectric sensors [39,40,45,47]: one sensor injects sound waves of a specified frequency into the battery as it cycles, while the other records sound waves after they have passed through the battery medium, see Figure 6. Its propagation speed and amplitude attenuation depend not only on the properties (modulus, density, thickness, and porosity) of the anode, cathode, separator, and collector through which the wave passes (remember, at each interface, the wave will be partially reflected and transmitted) but also on the electrolyte load or residual bubbles. Although complex, these variables can be simulated using elastic wave propagation theory, while the TOF and amplitude spectra of the received pulses are temporarily recorded as they cycle. From this information, a link between ultrasonic spectroscopy and SOC can be deduced. Through the operando study of commercial LiCoO2/C pouch batteries, a clear statement was conveyed by interrogating their various SOCs with a pulse of 200 kHz [48]. The amplitude and TOF of a peak of the rectified received pulse are linearly related to SOC [40], which can be used to construct a model for predicting SOH [49].

3. LIB-Based Acoustic Sensing Technology

3.1. The Application of Ultrasonic Testing Technology in the Field of Battery Inspection

UT technology can non-destructively, quickly, and accurately detect the state of materials inside the battery and analyze the electrochemical reaction, electrolyte decomposition, gas generation, lithium dendrite growth, and other processes in the battery in real-time. In recent years, many scholars have carried out applied research on SOC, overcharge behavior, and SOH estimation of LIBS in combination with UT technology [45,48,50], aiming to explore the response mechanism of ultrasonic signals to the electrochemical process of batteries and realize accurate monitoring of the internal state of LIBs under different working conditions.

3.1.1. Ultrasonic Testing of SOC

SOC indicates the remaining capacity of a battery during a specific charging or discharging process and is one of the most important characteristics that reflect the operating state of a battery [51]. It is defined as follows [52]:

$$SOC={\displaystyle \frac{C_r}{C_a}}\times 100\%,$$

(1)

Among them, C_r and C_a are the remaining capacity and the actual maximum capacity of the battery, respectively. At the same time, the non-destructive and efficient measurement of the SOC of the LIBs can provide a reference for battery design and process parameter optimization.

At present, numerous researchers have conducted extensive studies on evaluating and analyzing the SOC. Based on the advantages of LIB multi-physics properties, it remains valuable to explore new feature parameters related to SOC. As an alternative detection technique, ultrasonic testing proves to be an effective tool for analyzing different structures due to its rapid detection capability and low attenuation [53]. Recently, ultrasonic techniques have been increasingly used to detect the internal features of LIBs. First, Sood et al. [40] successfully identified the folding and expansion of the electrodes inside LIBs using ultrasonic transmission experiments. Following this, Hsieh et al. [45] examined the correlation between the acoustic properties of pouch and cylindrical cells and their SOC, revealing that the mechanical properties of the electrodes change the acoustic behavior during battery cycling. Gold et al. [48] applied ultrasonic transmission techniques to battery monitoring and discovered the presence of both fast and slow waves in the ultrasonic transmission data of batteries. These findings were explained using the Biot theory of ultrasound in fluid-saturated porous media.

Furthermore, Ladpli et al. [50] employed guided wave detection techniques to analyze the ultrasonic guided wave properties at different SOC. Their findings revealed a nearly linear correlation between the TOF and SOC during the discharge process. Despite this, the study did not provide comprehensive details regarding the specific guided wave modes or patterns identified during the experimental phase.

Li et al. [54,55] explored the intrinsic connection between the multi-characteristic indices of ultrasonic-guided waves and the SOC in LIBs. They introduced an adaptive FNN-XGBoost model to estimate the SOC of LIBs. Additionally, they developed a SOC estimation approach utilizing a simplified physics-based battery model (PBM) combined with an adaptive cubic Kalman filter (ACKF), which is particularly effective for high-precision SOC estimation in batteries with long-cycle lives [56]. Concurrently, to address the non-positive definite issue encountered with the CKF in SOC evaluation, Tian et al. [57] proposed three matrix factorization strategies to enhance the convergence rate of the CKF. Hao et al. [58,59] utilized piezoelectric transducer excitation and laser Doppler vibrometer reception to capture guided wave information from LIBs during cycling, enabling them to derive a consistent distribution curve correlating acoustic parameters with SOC. This groundbreaking research demonstrates the potential of acoustic detection technology in identifying lithium-related processes within LIBs. However, earlier investigations have not provided an in-depth theoretical examination of the guided wave propagation mechanism. Additionally, the choice of suitable mode types and frequency bands is essential for deriving precise and comprehensive characteristic indicators, which play a pivotal role in enhancing the accuracy of SOC estimation in battery systems.

In addition, Zhang et al. [60] proposed a joint estimation method for SOC and temperature of lithium iron phosphate batteries based on ultrasonic reflection waves. The method utilizes piezoelectric transducers attached to the surface of the battery to generate and receive ultrasonic signals. By analyzing the characteristic parameters of the ultrasonic signals, such as time-domain peak value, time-domain envelope peak value, energy integration, waveform index, kurtosis coefficient, and shape coefficient, high-precision estimation of SOC and temperature of lithium-ion batteries is achieved. The root mean square errors (RMSE) of SOC and temperature are 7.42% and 0.40 °C, respectively. This study highlights the potential of ultrasound reflection waves in non-destructive and real-time monitoring of battery status, which is crucial for preventing thermal runaway and improving battery management systems.

Since many characteristics of the acoustic signal are related to the SOC of the battery, Huang et al. [61] argue that it is unreasonable to estimate the SOC based on a single feature. Therefore, with the help of deep learning algorithms, the fitting features of the whole waveform are used as indicators for estimation, which can improve the accuracy of SOC estimation, and this method is suitable for a variety of different batteries.

In addition, scholars have proposed a model of the law of conservation of acoustics to describe the SOC of a standard battery, linked the change in sound speed with the SOC and SOH of the battery, and designed a low-cost online measurement system [62] to verify the applicability and effectiveness of ultrasonic technology in determining battery performance. Kirchev et al. [63] applied low-cost transducers to the safe monitoring of battery SOC, which can provide long-term stable and accurate acoustic data. Ladpli et al. [64] developed an efficient feature extraction strategy based on match-tracing technology, which combined with voltage data to improve the prediction accuracy of battery SOC to a certain extent. Gaul et al. [65] measured the high correlation of battery SOC with signal amplitude (SA) and phase using elastic-guided waves and evaluated the SOC of the battery more accurately than the battery voltage. Sun et al. [66] fused and analyzed the sonic ringing count, SA, and TOF indicators, which effectively improved the characterization accuracy of battery SOC and provided reliable precursor information for battery failure. Meng et al. [67] proposed a new method for quantifying battery SOC based on frequency-domain ultrasonic damping analysis, which used a time-harmonic continuous wave with a wide frequency range as the incident wave to detect the LIBs of different SOCs and described the propagation behavior of ultrasonic waves in the LIBs based on the multilayer model, and the model simulation results were in good agreement with the experimental data.

While ultrasonic testing methods have been applied to assess the internal conditions of LIBs, there has been limited theoretical exploration dedicated to this purpose. Addressing this gap, Gao Jie et al. [68] developed an analytical acoustic model to investigate the relationship between the SOC of LIBs and the propagation features of ultrasonic-guided waves. Their research identified the frequency range as most sensitive to SOC variations and provided a detailed analysis of how SOC impacts the dispersive properties of acoustic waves in LIBs. This work further substantiates the reliability of acoustic wave detection as a method for identifying operational states in LIBs.

3.1.2. Ultrasonic Testing of Overcharge Behavior

SOC is the ratio of the remaining charge of a battery to its rated capacity under the same conditions at a certain discharge rate [69] and is defined as an overcharge state when it exceeds 100% SOC [70]. Overcharging behavior can cause a series of complex physical and chemical changes inside the battery, which can affect the safety and lifespan of the battery. Ultrasonic testing technology can monitor the internal state changes in batteries in real time during overcharging, providing an important basis for the detection and early warning of overcharging behavior.

The formation of lithium plating is an important sign of battery overcharge behavior. Li et al. [71] found through research that ultrasonic non-destructive testing technology can be used to detect lithium plating during battery charging. When the lithium plating is formed, gas will be generated inside the battery, resulting in significant changes in the ultrasonic signal. Specifically, ultrasound waves undergo reflection and scattering at the interface of different acoustic impedance media during propagation, resulting in significant signal attenuation [72]. The formation of lithium plating will increase the attenuation of ultrasonic waves, so the occurrence time of lithium plating can be diagnosed by ultrasonic detection. In the experiment, Li et al. used a phased array ultrasound probe (Figure 7), placing the probe directly above the battery and placing a coupling layer between the probe and the battery. Ultrasonic waves are emitted by the array elements in the probe and propagate inside the battery. The experimental results showed that batteries without lithium plating exhibited strong reflection signals in ultrasonic detection images, while batteries treated with lithium plating showed significantly reduced reflection signal intensity. This indicates that the formation of lithium plating leads to more gas generation, which prevents the ultrasonic wave from propagating to the back of the battery.

Oca et al. [73] measured the SA and TOF of the ultrasonic signal and found that during light overcharging, the SA began to decrease, and the TOF began to increase, while the battery voltage and temperature did not change significantly. Wu et al. [74] studied the voltage, current, temperature, and ultrasonic signals of the pouch LIB in the overcharged state and found that the overcharging of the battery led to a change in the internal electrode interface and an increase in the macroscopic thickness of the battery (Figure 8), when the TOF and the first echo peak increased rapidly, indicating the overcharge state of the battery before the battery expanded, thereby assessing the degradation of the battery, and demonstrating the effectiveness of ultrasound in monitoring the risk of electrical abuse. Copley et al. [75] found that the correlation between ultrasonic response and battery capacity is stronger than that between voltage and capacity and that the change in battery capacity during charging can be tracked by selecting the peak sound wave, effectively preventing the occurrence of battery overcharging.

Appleberry et al. [76] used ultrasonic transmission signals to warn of LIB attenuation during overcharging, demonstrating that UT can warn of battery thermal runaway earlier and more sensitively.

In order to improve the accuracy and reliability of LIB overcharge detection, Dou et al. [77] designed an online ultrasonic ringing counting measurement device (Figure 9) to introduce ringing count into the battery overcharge state evaluation to realize real-time detection and early warning of LIB overcharge state.

3.1.3. SOH Estimation Based on Ultrasonic Testing

During electrical cycling and aging, the modulus and density distribution of the battery change. Therefore, different methods (i.e., coulombs or impedance methods) have been proposed to monitor the physical material changes associated with electrochemical cell processes. However, none of this can be applied in real time. Therefore, ultrasonic measurement has recently been proposed as a non-destructive technique to assess SOH and identify changes in the materials inside the battery [50,78,79].

UT is a battery test that actively emits ultrasonic, which has the advantages of controllability and high ultrasonic intensity. On the other hand, AE mainly uses filters, amplifiers, and signal converters to process the acoustic signal caused by the transient deformation of the internal material of the battery captured by the real-time acoustic sensor into effective information that can reflect the battery state [21]. Therefore, the current research on lithium-ion SOH estimation based on UT is more extensive.

SOH indicates the state of aging of a battery over a long period of time and is defined as follows [52]:

$$SOH={\displaystyle \frac{C_a}{C_{rated}}}\times 100\%,$$

(2)

C_a and C_rated are the actual maximum capacity and rated capacity, respectively.

An important parameter of ultrasonic characteristics is the ultrasonic wave velocity, as shown in Equation (3) [80].

$$v=\sqrt{{\displaystyle \frac{K+4G/3}{\rho }}},$$

(3)

In this equation, v is the ultrasonic wave velocity; K and G are bulk moduli and shear modulus, respectively. ρ is the mass density of the material inside the battery.

When ultrasonic waves propagate into the material, the reflection coefficient R is different due to the difference in acoustic impedance Z, as shown in Equation (4).

$$\left\{\begin{array}{c}R=\sqrt{{\displaystyle \frac{Z_2-Z_1}{Z_2+Z_1}}}\\ Z=\rho v\end{array}\right.,$$

(4)

In this equation, Z₁ and Z₂ are the acoustic impedances of different materials, respectively. Ultrasonic waves are transmitted and reflected multiple times between different materials, and the signal is captured by the receiving device. By analyzing the captured signals, TOF and SA can generally be obtained, as shown in Equation (5) and Equation (6), respectively.

$$T_{TOF}={\displaystyle \frac{d}{v}},$$

(5)

$$D_{SA}={\displaystyle {\int }_{t_1}^{t_2}\left|h(t)\right|dt}$$

(6)

In this equation, T_TOF is the time of flight; d is the flight distance; D_SA is the sound wave amplitude; h(t) is the ultrasonic waveform; t1 and t2 are the start and end times of waveform observation, respectively.

Davies et al. [49] conducted ultrasound testing on LIBs with different cycles and used elastic wave propagation theory to demonstrate that TOF and SA are closely related to physical quantities such as acoustic impedance and elastic modulus inside the battery that change with battery aging. This directly indicates that ultrasound perception can be an effective means of estimating the SOH of lithium-ion batteries, and TOF is used as the input of SVR to estimate SOH, with a final error of only 1%. LADPLI et al. [50] distinguished between reflection and projection modes by using guided wave signals based on one transmitter and one receiver mode and demonstrated that TOF decreases with battery aging, while SA is the opposite. Based on the statistical method of generalized additive model (GAM) to estimate SOH, the adjusted R2 coefficient and bias interpretation coefficient of the final result are close to 1, indicating that the accuracy of the results is high. In view of the different trends of TOF and SA with battery aging, Kim et al. [81] suggested that the battery material will be elastically softened, the wave velocity will decrease, and the ultrasonic attenuation will increase after long-term charge–discharge cycles, which will cause this changing trend. Wu et al. [74] estimated SOH by ultrasonic testing of LiCO₂ cells of two capacities and concluded that the correlation between TOF and SOH was higher than that of SA.

Contrary to the above-mentioned rules, Zhao et al. [59] found that both TOF and SA decreased with the aging of the battery based on guided wave detection, which may be due to the influence of different ultrasonic excitation frequencies and SOC. Scholar Yin [82] explored the relationship between the characteristic parameters of the AE signal and the cycle time and verified that it can be used as a basis for judging the SOH of lithium-ion batteries, and the stress wave signal can reflect the SOH of the battery.

3.1.4. In Situ Detection

In spite of this, technology has become a powerful tool in the field of battery research, which can monitor the internal status of batteries in real time during operation. Compared with non-in situ characterization techniques, in situ techniques can better establish the relationship between the obtained signals and electrochemical behavior, thereby allowing a deeper understanding of the dynamic processes occurring inside the battery.

The latest advances in in situ ultrasound technology have demonstrated their potential in real-time monitoring of battery behavior. Shen et al. [83] used A-scan and 2D/3D fully focused (TFM) ultrasound technology to monitor lithium-ion batteries under overcharging (Figure 10). They demonstrated that ultrasound signals can detect side reactions with high accuracy (0.4% SOC) at 102% SOC. This study emphasizes the sensitivity of ultrasound signals to physical changes in electrodes, such as density and elastic modulus, and their ability to visualize the internal battery state. This study comprehensively analyzed the ultrasonic behavior of batteries during normal and abnormal operations. During normal charging/discharging processes, it was found that the amplitude of the first echo in A-scan technology is closely related to the acoustic impedance difference between the anode and cathode, which varies with the SOC of the battery. Specifically, the amplitude of the first echo increases with battery charging and decreases with battery discharging, showing an approximately linear relationship with SOC. This relationship is consistent at different current densities (1C, 2C, 3C, and 4C), although the amplitude variation range and certain stages may vary due to concentration polarization.

During overcharging, the characteristics of ultrasonic signals undergo significant changes. For example, when the battery is charged to the cut-off voltage of 5.0 V, the first echo completely disappears, indicating a large amount of electrolyte decomposition and gas generation. Two-dimensional TFM imaging technology can observe the uneven distribution of reflection patterns inside battery bags and accurately locate the location of side reactions. This warning capability allows for timely intervention to prevent further damage and extend battery life.

Despite the advantages of in situ ultrasound technology, it faces challenges such as sensor integration into existing systems, data interpretation complexity, and cost. Future research on in situ ultrasound technology should focus on innovative sensor design and advanced data analysis methods. For example, the development of miniaturized and high-sensitivity ultrasonic sensors can improve the integration of in situ technology with battery systems. In addition, machine learning algorithms can be used to analyze complex ultrasonic signals and extract meaningful information about the internal state of the battery.

3.2. Application of Acoustic Emission Technology in the Field of Battery Testing

3.2.1. Acoustic Emission Detection for SOH

When the LIB is working, the generation of bubbles, the expansion of the electrode, and the formation of electrode cracks all generate stress waves, which can be collected and analyzed by AE technology.

Kai Zhang et al. [84] set up an AE measurement platform for LIBS (Figure 11) and set up a LIB cyclic experiment to analyze the stress wave signal of LIBs, and obtained two stress wave signals that can characterize the SOH of lithium-ion batteries: continuous AE signal and pulsed AE signal. The experimental results show that the amplitude of the continuous AE signal decreases with the increase in the number of battery cycles during the discharge process, which can be used to characterize the degradation of battery performance. The pulsed AE signal is more at the first cycle of the battery, less at the time of a small number of cycles, and slowly increases at the time of a large number of cycles, which is in line with the bathtub curve and can be used for aging monitoring.

The study of lithium-ion battery SOH by AE technology provides a new idea and method for the detection of lithium-ion battery SOH. However, due to the research conditions, the current experiment does not consider the combination of AE data and electrochemical data.

3.2.2. Active Acoustic Emission Sensing for Fast Co-Estimation of SOC and SOH

Battery material properties can be seen as a function of SOC and SOH. Wang et al. [80] first proposed a new active AE sensing technique for rapid co-estimation of battery SOC and SOH. Leveraging the alterations in battery material characteristics during charge–discharge cycles, power ultrasound is utilized to induce AE events, enabling AE transducers to actively detect a broader spectrum of battery state information across a wide frequency range. To evaluate the efficacy of the proposed active AE sensing technique for swift battery status monitoring, a specialized test system was constructed, incorporating a narrowband power ultrasonic transducer paired with a wideband AE sensor. Figure 12 is the schematic diagram of the designed test bench. It is composed of a host computer, a battery tester, a temperature chamber, a data acquisition system, an AE acquisition system, a preamplifier, an AE sensor, a power ultrasonic generator, an ultrasonic vibrator, a DC power supply, a current clamp, a bag battery, a thermocouple, a G clamp, etc.

Power ultrasound, when applied at suitable intensities, serves to stimulate the battery, thereby initiating acoustic emission (AE) phenomena. This stimulation is attributed to the dual mechanical and chemical influences of ultrasound [85]. As ultrasonic waves traverse through solid or liquid mediums, they impart shear forces and generate shock waves, which are the principal mechanical actions. These actions facilitate the acceleration of ion transport, thereby promoting electrochemical processes [86]. By harnessing these effects, power ultrasound can induce minute structural alterations within the battery, leading to an increase in AE events that may reflect variations in the battery material’s performance.

Conversely, AE sensors are employed to detect the elastic waves produced by battery materials of varying characteristics, providing insights into the battery’s condition. AE technology has emerged as a proficient technique for monitoring rapid elastic or transient waves emanating from localized material changes, a task that conventional auditory and vibrational sensors are ill-equipped to perform due to their limitations in capturing high-frequency data. As power ultrasound permeates a battery, the cell’s stratified architecture leads to wave attenuation, and the ultrasonic waves induce stress waves within the electrodes. AE sensors are capable of capturing these ultrasonic and stress wave signals, which, upon analysis, can reveal alterations in battery material performance. The evolution of battery material properties during charge, discharge, or cycling processes alters the acoustic impedance and AE events. Moreover, AE transducers offer the capability to convert narrowband signals to wideband, enhancing the detection of material property changes. Consequently, the AE approach ensures a comprehensive perception of battery material property variations, leading to more precise estimations of SOC and SOH.

Based on the fundamental excitation frequency f₀, a wide range of frequency components can be obtained, which are described as follows [87,88]:

(1)

Fundamental frequency f₀: Generated by the excitation frequency.

(2)

Harmonics of the fundamental frequency nf₀ (where n denotes a positive integer): they are caused by the transducer itself or possibly by nonlinear propagation in the cell.

(3)

Elementary f₀/n subharmonics: they are excited by bubbles twice the size of the resonance or non-spherical bubbles with surface oscillations.

Time-domain analysis reveals that the energy of the collected AE signals varies with different SOC or SOH levels. In addition, frequency domain analysis shows that a series of harmonics occur in the spectrum, such as 2f₀, 3f₀, 4f₀, 5f₀, 6f₀, and 7f₀. The statistical parameter RMS was chosen as the basic feature to characterize the SOC/SOH of batteries in different frequency bands. Experimental studies show that different frequency bands have different sensitivities to battery SOC/SOH. Bandpass RMS decreases as the battery’s SOC and SOH decline. In the frequency range of 270~300 kHz, 7f₀ can successfully realize the simultaneous estimation of battery SOC and SOH in any SOC region, which verifies the feasibility and performance of the proposed method.

4. Results and Discussion

Acoustics excels at visualizing mechanical effects and visualizing changes in internal structures. However, this technology has certain limitations in practical applications. For example, it is difficult to accurately distinguish between sound sources and to obtain direct observations related to temperature or thermodynamics. At the same time, compared with temperature sensors, strain gauges, and electrochemical sensors, acoustic technology can provide multi-dimensional data, but the accuracy and reliability of its measurement results still need to be improved. Nevertheless, acoustic technology has shown important research value in battery condition assessment, which is not available in other sensing methods.

UT technology directly characterizes the material properties inside batteries by capturing changes in the TOF and amplitude of ultrasonic waves. This offers greater analytical accuracy and intuitiveness than conventional indirect monitoring methods based on voltage and temperature. UT technology also includes in situ ultrasound techniques, which have demonstrated the ability to monitor battery behavior in real time, emphasizing the sensitivity of ultrasound signals to physical changes in electrodes, such as density and elastic modulus. These techniques can visualize the internal battery state and detect the side with high accuracy. However, UT technology still faces several bottlenecks in practical reaction applications. Battery thickness can limit the detection effect, and the detection speed is relatively slow. The variation range of signal amplitude is small, making acquisition challenging. Additionally, TOF measurement requires consideration of baseline offset, increasing computational complexity and hindering widespread adoption.

AE sensing technology has been successfully applied to a variety of battery chemistries, including Na-S, Ni-MH, Li-ion, Na-ion, and Li-S cells. It can be used to estimate battery SOC during short-term charging and discharging. In long-term use, it can also be used to evaluate battery SOH by analyzing the AE signal under different degradation states. However, the commercialization of this technology still faces multiple challenges, such as the difficulty of integration with existing battery management systems, the lack of standardized data interpretation protocols, and the guarantee of sensor reliability under variable conditions. Future research directions should focus on innovative sensor design, optimized data analysis methods, and multi-dimensional information processing fused with electrochemical data. This will not only improve the applicability and accuracy of AE technology but will also promote its widespread application in electric vehicles and grid energy storage systems, thereby improving battery performance, longevity, and safety [89]. With the growing demand for efficient energy storage, AE technology is expected to be an important pillar driving the development of battery technology and laying the foundation for a sustainable energy future.

5. Summary

As a new star in the field of battery monitoring, the laboratory research of acoustic sensing method has successfully realized real-time, in situ, and non-destructive monitoring of lithium-ion batteries and obtained a large amount of internal physical and chemical information, including electrolyte infiltration, gas generation, temperature distribution, electrode stress/strain, and interface side reactions. However, this information is currently used to help understand the evolution mechanism of battery operation and internal chemical composition, and challenges such as insufficient sensor reliability, complex data interpretation, and cost control still need to be overcome from laboratory research to actual industrial application.

In this paper, the application of AE and UT in lithium-ion batteries is systematically reviewed, focusing on the monitoring and evaluation of battery state-of-charge SOC, state-of-health SOH, and overcharge behavior. By capturing the stress wave signal during battery operation, AE technology reveals key processes such as electrode particle rupture and SEI/CEI growth. UT technology accurately evaluates SOC, SOH, and overcharge behavior through flight time and amplitude changes in sound waves. Both techniques provide new solutions for information that is difficult to obtain with traditional electrochemical methods.

Although acoustic sensing technology has shown unique advantages in battery condition monitoring, its application still faces problems such as complex signal analysis, insufficient cross-system applicability, and difficult device integration. In the future, it is expected to further improve the accuracy and reliability of acoustic sensing technology by optimizing sensor design, improving signal processing algorithms, and integrating multimodal sensing technology, thereby accelerating its industrialization process. With the growing demand for energy storage technology, acoustic sensing technology will provide important support for the safety, life improvement, and performance optimization of lithium-ion batteries.