1. Introduction
Lithium-ion batteries (LIBs) have been extensively implemented in the industry of electric vehicles and energy storage due to their high energy/power density, long lifetime and low self-discharge rate. The impedance of LIBs with graphite negative electrodes increases drastically in cold environments, sharply decreasing the discharge power and available energy [
1]. In addition, low-temperature operation may induce the Li-ion plating in-between the graphite particles and electrolytes, resulting in battery degradation and even safety hazards [
2]. Therefore, heating up LIBs to a suitable temperature range (such as 0 °C or 5 °C) before usage is pragmatically required at low temperatures.
To strengthen the low-temperature performance of LIBs, external and internal preheating methods are usually adopted. For external preheating methods, various heat actuators are generally used to heat the batteries through convection or conduction [
3,
4,
5]. However, related methods are considered to be inefficient due to the long heat transfer path and high thermal resistance in between components. Contrarily, internal preheating methods, generally warming up the batteries themselves using the heat produced by their internal resistances, removes the process of heat conduction, exhibiting higher rates and better temperature uniformity.
Alternating current preheating (ACP) has drawn a lot of attention among internal preheating methods due to its high preheating rate and capability to prevent lithium plating. Zhang et al. [
6] developed an internal preheating method for LIBs using sinusoidal alternating current (SAC). Based on an equivalent electrical circuit (EEC) model, a heat generation model in the frequency domain was proposed so as to predict the battery temperature evolution during the ACP. Utilizing the method, an 18,650 battery was warmed up from −20 °C to 5 °C within 15 min. Zhu et al. [
7] examined the temperature distribution on a LiFePO
4/C battery with a large capacity by adjusting the amplitude and frequency of alternating current. The maximum temperature difference on the battery surface is no higher than 2 °C in the test using the external alternating current excitation with an amplitude of 1.25 C and at a frequency of 1 Hz. Regarding the implementation of the ACP method, hardware circuits, including a soft-switching circuit [
8], buck–boost conversion [
9], and drive circuity of an electric vehicle [
10], have been designed to preheat the signal battery or battery module.
A key concern about the ACP method is the risk of lithium plating when the battery is forcefully charged at low temperatures. According to the Li plating mechanism, the anode potential criterion that the sum of the negative electrode equilibrium potential (
Ue) at a given state of charge (SOC) and over-potential of the negative electrode (
η) shall be above 0, i.e.,
Ue +
η > 0, is widely introduced in the previous studies to develop the preheating strategies of lithium-plating-prevention [
8,
11,
12].
Based on the anode potential control and negative electrode impedance, Ge et al. [
11] proposed a lithium-plating-free ACP method in the scenario of using SAC. The current limiting curves, capable of inhabiting lithium plating, are also provided at different temperatures. Integrating the current limiting curves with their proposed heat generation rate model, a multistep ACP method has been developed, successfully warming up a laminated Li-ion battery from −20 °C to 5 °C in 800 s. However, the method demands the anode potential and negative electrode impedance, which rely on the implementation of a three-electrode battery.
To simplify the implementation of ACP, two full-battery-based strategies based on full battery impedance control and terminal voltage control have been extensively adopted by limiting the full battery reaction overpotential [
8,
12] or keeping the maximum polarization voltage of a full battery constant [
13,
14,
15].
As for the first strategy, replacing negative electrode impedance using full battery impedance is generally adopted according to some electrochemical impedance spectroscopy (EIS) results showing that the charge transfer impedance derived from the negative electrode (NR
ct) is much larger than that from the positive electrode (PR
ct) [
11]. However, the result lacks general applicability for different batteries and at different temperatures. As for the cases with the PR
ct close to the NR
ct, such a strategy can tighten up the maximum permissible amplitudes of the lithium plating prevention and reserve excessive parameter redundancy, resulting in the severe depression of the preheating rate [
16]. In addition, the strategy evades the issue of obtaining the negative electrode equilibrium potential by referring to existing results [
12] or neglecting
Ue via directly making
η > 0 [
8].
As for the second strategy, Ruan et al. [
13] proposed an ACP method with overvoltage ≤0.5 V or ≥0.5 V to achieve a good balance between the heating time and the battery lifespan. Zhu et al. [
7,
14] developed an alternating current pulse heating approach by adjusting the current parameters to warrant the terminal voltage of the battery within the working voltage range. Despite the fact that the related works achieved a fast and lithium-plating-free preheating, the principle of setting the voltage threshold lacks a solid theoretical basis. Fundamentally, this strategy cannot distinguish the anode potential and cathode potential. The anode potential criterion for preventing lithium plating may not be precisely guaranteed, e.g., the strategy is found to induce lithium plating at low frequencies in work [
16], or may be overly limited by controlling the terminal voltage, e.g., when open circuit voltage (OCV) is close to the upper or lower voltage limitation of the battery.
In summary, among the above three kinds of ACP, the method based on the anode potential control and negative electrode impedance is the most scientific and accurate for preventing lithium plating, despite the fact that it requires a reference electrode. Actually, as regards its application in real situations, calibrating the maximum permissible current parameters in the development phase and looking up a table in the usage phase is a feasible and commonly-used-in-engineering strategy. In addition, combined with the state of health estimation methods, this kind of ACP method can be further optimized when considering battery inconsistency and battery aging. Moreover, due to the advantages of in-situ monitoring and characterizing the electrochemical performance of electrodes, the reference electrode configuration has aroused many battery manufacturer’s interests to embed it in commercial LIB [
17,
18]. Although the present LIBs have not equipped reference electrodes, several patents have reported the implementation of the reference electrodes in practical rechargeable batteries [
19,
20,
21,
22,
23]. The researchers of [
24] also declared a remarkable breakthrough in improving the lifetime and accuracy of the three-electrode battery. The lifetime of their developed three-electrode battery has exceeded 500 h, already having the potential for application in real situations. Consequently, the ACP approach based on the anode potential criterion shall be an applicable and promising technology and herein is employed to acquire the operation boundaries of preventing lithium plating in terms of current amplitude and frequency.
The waveform of the alternating current is another crucial issue in the implementation of the ACP method. In previous papers, the alternating current with sinusoidal waves has been widely adopted. However, it is very difficult to generate in the case of electric vehicles or chargers. Furthermore, the sinusoidal wave current may suppress the heating efficiency due to its low energy flux density. Consequently, searching for an applicable waveform with high efficiency is urgent. Since the square wave alternating current (SWAC) can be easier to generate through switching devices and carries higher energy flux density, it shows higher potential to apply in real situations. Although several works have investigated the effect of a square wave current or likewise on preheating performance [
7,
12,
25,
26], most works adapt full-battery-based strategies, and few have provided the deterministic and intrinsic operation boundaries of lithium plating prevention and fully released the preheating capability of ACP using the SWAC excitation.
In this paper, we develop a rapid and lithium-plating-free ACP approach using the excitation of SWAC. A heat generation model in the frequency domain is established based on the EEC model. The impact of current parameters, including frequency, amplitude and waveform, on the preheating rate is experimentally investigated. Operation boundaries of lithium plating prevention are determined using the fitted EEC parameters of negative electrode EIS and the lithium plating potential. The effectiveness of operation boundaries, together with their effect on battery health, is explored through the aging tests. By combining the operation boundaries and the ACP method with SWAC, a temperature-adaptive procedure with different frequencies, temperature intervals and heat transfer coefficients is explored. The procedure is found to be able to significantly speed up the preheating rate with higher frequency, smaller temperature intervals, and better thermal insulation. In addition, the proposed method in this paper is evaluated in terms of preheating rate by comparing it with the preheating methods based on the two full-battery-based strategies. The results demonstrate that the proposed method has more robust strength than these using the full-battery-based strategies at high frequency.
The rest of this paper is organized as follows. The theoretical considerations about heat generation and Li-ion plating prevention are introduced in
Section 2. The experimental setups are presented in
Section 3. The model validation, operation boundaries of lithium plating prevention and temperature-adaptive preheating are discussed in
Section 4. Finally, conclusions are summarized in
Section 5.
5. Conclusions
In this paper, a rapid preheating method capable of preventing lithium plating was developed using a square wave alternating current. The method achieves a higher preheating capability, capable of warming up the battery faster than that with sinusoidal alternating current. The strategy of anode potential control is adopted to maximize the preheating rate. Operation boundaries of lithium plating prevention, in terms of frequency and maximum permissible current amplitude, are determined using the anode impedance and potential. The effectiveness in preventing lithium plating, together with its influence on battery health, is experimentally investigated. After repeated preheating of 800 cycles at −20 °C, no lithium plating is observed on the graphite electrode, and negligible effect on battery degradation is confirmed.
By integrating the operation boundaries and the square wave preheating method, a temperature-adaptive amplitude procedure with different control parameters is explored. The procedure is found to be able to significantly speed up the preheating rate with higher frequency, smaller temperature intervals, and better thermal insulation. Another two methods based on the full battery impedance control and the terminal voltage control are implemented. The results demonstrate that the proposed method in this paper has a larger preheating rate than these using the full-battery-based strategies at high frequency. When the battery is preheated at a frequency of 400 Hz, with a temperature interval of 5 °C and a heat transfer coefficient of 5 Wm−2 K−1, the preheating rate can reach 6.61 °C/min, exceeding the method based on the terminal voltage control by 5.4%, and larger than that based on the full battery impedance control strategy by 41.8%.
In extending the method to the battery module or pack, despite the fact that several excellent works have provided implementation schemes in terms of hardware circuits, the effect of the battery-to-battery variations on the uniformity in terms of temperature and operation boundaries of lithium plating prevention has not been considered yet. The related work will be carried out in a future study so as to revise the operation boundaries and improve the robustness of the proposed method for practical applications.