The Gregory Offer research group at Imperial College London, the Ouyang Minggao research group at Tsinghua University, and Billy Wu of the Faraday Institute and researchers from Shell Oil jointly published a review article on lithium-ion battery fast charging: A review in the international journal eTransportation. This paper comprehensively reviews the factors affecting the fast charging of lithium-ion batteries, the main problems and solutions of fast charging from the material level to the system level.
preface
Background
In recent years, in order to limit the effects of climate change and air pollution, the widespread use of lithium-ion batteries in pure electric vehicles is accelerating. However, compared with traditional fuel vehicles, problems such as range anxiety and long charging time have become the main problems hindering the development of electric vehicles. Therefore, the improvement of fast charging capabilities has become a common development goal for battery manufacturers and automakers. However, studies have shown that low temperature and large magnification charging will cause accelerated attenuation of the battery's capacity and output power; on the other hand, the large amount of heat generated by the battery during charging is difficult to dissipate uniformly and effectively, which will also cause attenuation acceleration and other safety issues. Figure 1 shows the factors that affect the rapid charging of lithium-ion batteries from the atomic level to the automotive system level. This paper focuses on a review and summary of the existing literature and analyzes the key technical constraints at each level.
Figure 1 Factors affecting the rapid charging of lithium-ion batteries at different levels
The classification of electric vehicle charging includes AC and DC, where DC charging is faster. Tesla was the first company to use 120kW fast charging; Bosch released a 350kW fast charging plan in 2017, which was implemented in the "Taycan" in 2019. Since the voltage of the current car battery Pack is about 400V, the high power charging of 350kW requires the voltage of the Pack to be higher to avoid the problems of excessive current and excessive heat production. Both the Bosch Taycan and Audi's e-tron GT concept cars (charging power up to 350 kW) are equipped with an 800V lithium-ion battery Pack. In December 2018, a joint research group of BMW, Bosch and Siemens achieved fast charging in 450kW CCS mode on two test vehicles in Germany.
Although research has made great progress in increasing the charging power of electric vehicles, these fast charging technologies are not adapted in all cases. According to the specific working conditions and charging environment of electric vehicles, the charging power will gradually decay during continuous charging. In addition, in fast charge mode, due to safety and other factors, the battery can usually only charge up to 80% of the charge; at higher levels, the charging rate will gradually decrease to avoid overcharge. In addition, the charging power is limited by the battery management system (BMS). Industry is becoming more and more interested in the field of battery fast charging, and it is necessary to understand the rapid steps of different charging methods and their impact on battery life. The purpose of this paper is to establish the relationship between microscopic processes, material properties, battery and Pack design and charging strategy optimization from the multi-scale and multidisciplinary characteristics of fast charging.
Background
The principle of fast battery charging
The ideal battery should exhibit long life, high energy density and high power density characteristics to be able to quickly charge and recharge at any temperature in any location to meet the requirements of electric vehicles driving long distances. However, there is a trade-off relationship between these physical properties, and the influence of the temperature of the material and the device determines the battery usage threshold. When the temperature drops, both the charge rate and the maximum voltage should be reduced to ensure safety, which makes temperature a key limiting factor for fast charging. Among them, as the temperature decreases, the risk of lithium evolution increases significantly. Although many researchers have pointed out that lithium evolution often occurs at temperatures below 25 °C, it is also prone to occur at high temperatures, especially when the charging rate is high and the energy density is high. In addition, the fast charging efficiency and temperature are also very closely related, the charging efficiency of the 50kW charging pile at 25 °C is 93%, but the charging efficiency at -25 °C is as low as 39%, mainly because the BMS will limit the rated power at low temperatures.
Common lithium-ion batteries are mainly composed of graphite anode, lithium metal oxide cathode, electrolyte, fluid collector, and porous diaphragm. As shown in Figure 2, when charging, Li+ is transmitted from the positive electrode through the electrolyte to the negative electrode, where the main transmission path is: 1) through the solid electrode; 2) through the positive and negative electrode/electrolyte interface; 3) through the electrolyte, including the solvation and debolytication of Li+. But improper use of batteries often causes a series of side reactions that affect performance and longevity. In addition, charge and discharge rate, battery internal resistance and battery polarization will affect the thermal characteristics of the battery, such as increasing heat production, reducing charging efficiency and safety.
Figure 2 Schematic diagram of lithium ion transport a) charging, b) discharging
A large number of studies have shown that the decay of the positive electrode and the growth of the positive CEI membrane have no effect on the fast charging speed of traditional lithium-ion systems, so the negative electrode has become the main focus of the charging process. Under certain circumstances, lithium metal may continue to precipitate into lithium dendrites, and may even puncture the diaphragm to cause an internal short circuit. Factors affecting lithium deposition and deposition structure include the diffusion rate of lithium ions at the negative electrode, the electrolyte concentration gradient at the negative electrode interface, metal salt deposition of the collector fluid, and the side reaction of the electrode/electrolyte interface. Studies have shown that the performance of the negative electrode during lithium evolution can be attributed to the effect of the current at the beginning of lithium evolution on the internal resistance of the negative electrode isocular density. Reducing the internal resistance of the negative electrode through the battery design is very important to improve the fast charging ability of the battery. In addition, the temperature effect is also very important, too low or too high temperature will be considered unfavorable to the battery, but the battery temperature is higher when fast charging will be conducive to its own balance, especially for high specific energy batteries. The effect of electrode thickness on charging performance also needs to be paid attention to. Thin electrodes are often thought to be ideal for lithium ion transport, and when the electrode thickens, it becomes important to ensure sufficient lithium ion concentration at the electrode/electrolyte interface to maintain overpotential stability and reduce the potential for lithium evolution. Thick electrode batteries during fast charging, lithium salts may be deposited at the collector, resulting in an imbalance in electrode utilization and an increase in the current density of the negative electrode of the diaphragm.
The principle of fast battery charging
Attenuation effects
1) Temperature effect
The heat production of lithium-ion batteries can be divided into reversible and irreversible processes. The expression of non-thermogenerating Qirr is as follows:
U is the open circuit voltage, Vbat is the battery voltage, and I is the current (when charging). Most of the irreversible heat comes from internal resistance to heat production:
where R is the internal resistance of the battery. Joule heating is proportional to the square of the current, so when the current increases at fast charge, the irreversible heat increases significantly. Reversible heat Qrev is derived from entropy changes in electrochemical reactions, also known as entropy heat, and its expression is:
In lithium-ion batteries, the heat distribution and dissipation of soft pack, cylindrical and square shell batteries are unevenly distributed: for example, some battery materials have poor thermal conductivity, so their heat will accumulate more in the core position relative to the surface. In addition, the current density and heat production rate are different in different positions of the battery. These inconsistencies are further amplified on large-size batteries. As shown in Figure 3, the temperature in the inner center of the cylindrical battery is significantly higher than the surface. For soft packs or square batteries, as shown in Figures 4 and 5, the temperature at the polar ear is significantly higher than elsewhere. In addition, since the positive aluminum collector is more resistant than the negative copper collector, the temperature of the positive pole ear is often higher than that of the negative pole ear.
Figure 3 Simulation results of the distribution of temperature and current density inside a cylindrical battery
Figure 4 Surface temperature change of soft pack battery at 5C constant current discharge: a) simulation result and b) measurement result of t=250s; c) simulation result and d) measurement result of t=667s; e) 3D distribution of internal temperature
Figure 5 Temperature distribution of LFP batteries discharged at a)1C, b)2C, and c)5C
The uneven distribution of heat generation not only exists in the battery cell, but also requires more attention to the design of the thermal management system at the battery pack level, because it has a significant impact on the distribution of temperature within the Pack. Over time, the different aging paths of the battery cells will also have a great impact on the heat uniformity of the Pack, which is due to the different increases in the internal resistance of different batteries. To solve this problem, the design of the thermal management system will be described in Part 6.
Many of the aging mechanisms in lithium-ion batteries are temperature-dependent. At high temperatures, the SEI membrane grows at an accelerated rate at the negative electrode and becomes more loose and unstable. At low temperatures, the ion diffusion and reaction rate slow down, and the possibility of lithium evolution and lithium dendrite growth increases. Almost all aging reactions are accelerated at high temperatures; low temperatures can reduce the rate of side reactions but also reduce the diffusion of the active substance, and if lithium metal precipitates, it will accelerate attenuation. In addition, increased low-temperature polarization leads to increased heat production and reduced energy efficiency. In most cases, SEI membrane growth at the negative/electrolyte interface is the main attenuation mechanism. The SEI film will increase the internal resistance of the battery, reduce the power, and then lead to capacity attenuation. At high temperatures (60 C or more), SEI components dissolve and decompose, destroying the integrity of the negative protective film. In extreme cases, when the battery temperature exceeds a safe threshold, thermal runaway may occur.
2) Analyze the effects of lithium
Lithium evolution refers to the process by which lithium ions in the electrolyte are deposited as faradays of lithium metal on the negative electrode, rather than embedded in the negative electrode particles. Lithium evolution may occur when the negative potential drops below Li/Li+. During the lithium evolution process, the lithium metal will first form a droplet shape to reduce the surface energy, and the surface metal and electrolyte react quickly to form a SEI film. As more lithium is deposited under the SEI film until the SEI film ruptures, a new SEI film is formed on the surface of the lithium, the lithium salt concentration gradually decreases, and the lithium metal begins to grow perpendicular to the surface of the pole sheet, forming lithium dendrites. Lithium dendrite growth is considered one of the worst side effects, and if the dendrite punctures the diaphragm to reach the positive electrode, the internal short circuit will cause the battery to quickly heat up. Lithium metal is more reactive than the negative electrode, further bringing internal side reactions, leading to problems such as SEI growth, gas production and electrolyte dissolution.
The researchers have proposed some models of lithium evolution observations. These include Fuller, Doyle, and Newman's P2D-based lithium evolution models, as well as the embedding process of reversible lithium proposed by Aorra, Doyle, and White. On this basis, Perkins proposed a control-oriented reduced order model; Hein and Latz proposed a three-dimensional microstructure analytical model. Ren considers both the re-embedding of reversible lithium and the reaction of irreversible lithium (dead lithium) during battery charging.
Non-destructive lithium evolution observation techniques are important for practical battery applications. Tests that can generally be used for lithium evolution characterization include SEM, TEM, NMR, and XRD, but these methods require the destruction of the battery or the use of a special battery configuration. Commonly used non-destructive lithium analysis observations use the external characteristics of the battery, including aging rate, lithium re-embedding voltage platform, model prediction and other methods. As shown in Figure 6, lithium evolution detection methods based on aging characteristics include (a) Arrhenius equation, (b) capacity and impedance change analysis of attenuation processes, (c) nonlinear frequency domain response analysis and (d) Coulomb efficiency analysis.
Figure 6 Lithium evolution detection method based on aging characteristics
Partially analyzed lithium is re-embedded in the negative electrode or dissolved during discharge. The relaxation process after the end of the charge or the immediate discharge process produces a new voltage platform, as shown in Figure 7. Voltage differentiation (DVA) and capacity differentiation (ICA) help find voltage platforms, but these methods require small-rate discharge, which increases polarization and covers the lithium-separated signal on the voltage curve. The process of lithium precipitation and re-embedding may also cause abnormal peaks in thermogenesis as one of the signals for lithium evolution.
The increase in battery thickness may also lead to lithium evolution, but the related mechanism needs to be further studied. Electrochemical models predict lithium evolution usually depend on charging conditions. However, these models are too complex and require a lot of computation, and they need to be further simplified to enable in-line detection. A few methods can achieve quantitative lithium-in-situ detection after abnormal battery charging. The authors believe that the detection method based on the abnormal voltage platform is the most promising to achieve application, but it is still highly knowledgeable and technical barriers to actual application.
Figure 7 Lithium evolution detection based on lithium re-embedding
a) Simulation of overpotential changes during CC-CV charging and standing. Stage I, where no lithium deposition is present on the negative electrode particles; Stage II, where lithium deposition begins to occur; Stage III, where partially reversible lithium is re-embedded in the negative electrode or dissolved, the rest becomes dead lithium; Stage IV, equilibrium, where dead lithium is no longer involved in subsequent cycles; b) Voltage differential analysis (DVA); c) Differential Capacity Analysis (ICA)
3) Mechanical influence
Mechanical chalking is another important aging phenomenon caused by fast charging, and has been confirmed in a variety of electrode materials (graphite, NMC, LCO, NCA, Si, etc.). According to the scale, mechanical attenuation can be divided into the following parts: rupture inside the electrode particles, separation of the electrode particles from conductive carbon and adhesives, separation of the active material from the collector, and electrode stratification. The main reason for these phenomena is the gradient distribution of lithium concentrations during fast charging, resulting in stress mismatch between components. When the energy release rate or stress exceeds a certain value, cracks appear in the particles, accompanied by the rupture of the SEI/CEI membrane. When the strain between the particles caused by fast charging cannot be matched to each other, it will cause contact between the electrode particles or between the particles and the conductive carbon and adhesive. Strain mismatch between the electrode material and the collector can also cause the active substance to fall off. High magnification can cause severe uneven distribution of current density between electrode plates, and if there is no external pressure, delamination may occur between electrode plates.
The effect of mechanical decay on battery performance can be divided into active material loss (LAM), active lithium loss (LLI) and impedance increase. First, cracks can cause the electrical contact to deteriorate; second, cracks will expose more fresh surfaces to react with the electrolyte, and the high temperatures brought by fast charging will accelerate the above side reactions. These reactions in turn accelerate the growth of SEI, exacerbating the increase in impedance, LAM and LLI, etc. Finally, the consumption of the electrolyte reduces the wettability of the electrode surface and hinders ion transport. The relevant positive feedback mechanism can be described as follows: large magnification current leads to crack formation; crack aggravates the difference between electron and ion transfer rates, because ions can be transported through the electrolyte to the crack and electrons cannot, which in turn leads to uneven charging states, further aggravating crack generation. In addition, the authors also briefly introduced the effect of particle size on the mechanical attenuation of the fast charging process, the effect of high magnification on the rupture of secondary particles, and the optimization of fast charging strategy according to mechanical attenuation limitations.
In general, there are many problems to be studied in mechanical decline under fast charging conditions. Different experiments on this problem have produced different conclusions, and there is still controversy on some important issues, such as the relationship between the charging rate and the rate of crack generation. Mechanical decay is also often difficult to understand with other aging machines. Compared with aging mechanisms such as SEI growth or lithium evolution, few models have studied mechanical effects under large currents, and very few of them have been experimentally verified. The determination of model parameters and boundary conditions has become a major problem hindering the development of mechanical models.
4) Multi-scale fast charging performance design
Fast charge-induced aging and aging modes are influenced by a variety of factors, such as battery material composition (intrinsic characteristics of electrode materials and electrolytes), operating conditions (high-rate charge and discharge, extreme voltage and temperature), battery production process, and Pack design. Multi-scale design and composite means will help develop high-performance fast-charging batteries.
Selecting the right electrolyte and electrode material for the electrode material to achieve high specific capacity and high magnification performance has always been a challenging challenge in battery design. There are many studies dedicated to the development of dendrite-free fast-charging anode materials, such as carbon-based materials, metal oxide composites and alloys, which have achieved a certain degree of success. The traditional graphite anode potential is very close to the redox potential of lithium, which can make the battery perform higher energy density, but at the same time increase the possibility of lithium evolution. Therefore, improving the anode material has become one of the important ways to improve the performance of lithium-ion batteries. In addition, the LTO, because it does not dissolve lithium and does not form a SEI film, is believed to be used to develop long-life super fast-charging batteries. On the other hand, the potential of the LTO is high, and acting as a negative electrode material will reduce the voltage of the whole battery, limiting the energy density of the battery. Some metal oxides and alloy materials also have good energy and power characteristics, but are limited by severe volume changes, chalking and agglomeration, and their cycle stability is usually poor.
Other graphene-like two-dimensional materials have a high surface area/mass ratio and unique physico-chemical properties, shorten the ion transport path, accelerate electron transport and increase lithium ion activity sites, and are considered potential negative electrode materials. These materials mainly include transition metal oxides, transition metal sulfides, metal carbides and nitrides. Among them, the oxide electrochemical window of titanium and niobium-based is usually between 1.0-1.6V, which matches the current commercial electrolyte and is very suitable for negative electrode materials. Recently, goodenough research group proposed that TiNb2O7, a high-magnification anode material, has a theoretical specific capacity comparable to graphite, and can achieve rapid lithium ion de-embedding and long cycle life, which is expected to replace LTO as a new negative electrode material.
Designing a suitable electrode structure at the nanoscale can also achieve high power and energy density, such as 2D hollow structure, core-shell structure, yolk-core structure, etc. Integrating 2D materials into macroscopic 3D structures can also enhance the transport of electrons and ions of the material at the electrodes.
Lithium metal is one of the negative electrode materials that can improve the energy density of batteries, but its power performance is poor due to the low specific surface area of pure metal lithium foil. The introduction of lithium metal into a 3D structural framework to speed up the rate of ion diffusion can significantly improve its magnification performance. In addition to the selection, modification and nanostructure design of the anode material, the electrode/electrolyte interface can also greatly affect the performance of the anode material. By optimizing the negative/electrolyte interface such as amorphous carbon-coated graphite to form a uniform SEI film, selecting suitable lithium salts and co-solvents can also inhibit the growth of lithium dendrites.
The selection and modification of materials is undoubtedly the focus of future research. Many new materials exhibit better fast charging performance than current commercial materials. However, these materials are in the early stages of development, and there is still a long way to go before large-scale commercialization, and in many cases, problems such as developing new production processes and equipment and reducing costs may also hinder the application of new materials. In addition, the evaluation of many new materials and new structural designs remains only at the laboratory level, and when applied to commercial batteries or Packs, the actual effect may be greatly reduced. Materials science will undoubtedly play an important role in the development of batteries in the future, but engineering also requires a lot of effort to truly solve the problem of fast charging.
In addition to material selection and microstructure design, the geometric parameters of electrode design also have an important impact on battery performance. Increasing porosity and negative electrode thickness inhibits lithium evolution, but at the same time reduces energy density.
The capacity ratio (N/P) of the negative to positive electrode materials can significantly affect lithium deposition, and the N/P in commercial lithium-ion batteries is often greater than 1, and higher N/P helps to reduce the mechanical stress of the negative electrode, reduce SEI formation and loss of active lithium. In NMC811/ graphite batteries, the N/ P ratio will gradually decrease with the increase of the charging rate, which is due to the fact that the surface capacity of graphite decreases more sharply than the surface capacity of NCM811 with the increase of charging magnification, and the N/P ratio is 1.15 at 0.1C, 1.0 at 3C, and 0.5 at 4C. Studies have shown that during the standing process after charging, the lithium metal precipitated in the main area of the negative electrode will diffuse to the convex part of the negative electrode under the drive of the concentration gradient. During subsequent discharges, more lithium is received at the edge of the cathode. Continue charging, transferring the excess lithium to the negative and negative protruding areas corresponding to the positive edge. This leads to a local increase in lithium concentration and a decrease in potential, increasing the likelihood of lithium evolution. Therefore, the negative protruding area should be designed to be as small as possible to avoid lithium evolution.
The geometric parameters of the battery are also important factors affecting the ability to charge fast. The shape of the battery affects the distribution of current density and temperature, and large batteries are more likely to cause uneven distribution of temperature and current. The position, material, structure and welding process of the polar ear are important for the uniform distribution of current density, limiting local heat production and delaying aging.
In addition, the relationship between battery Pack performance and monomer performance is not very clear. Although there have been many fast-charging models of battery cells, few studies have attempted to extend them to The Pack design, due to the need to consider more parameters when designing the Pack. There are still many problems in the design of fast-charging battery Pack: first, fast-charging Pack requires the high performance of the battery cell and the low inconsistency between the cells; second, the monitoring and balance of the battery requires more sensors and circuit control of the advanced BMS; third, the need to design an advanced thermal management system to maintain safe temperature and reduce the temperature difference between the battery and the Pack.
Attenuation effects
Fast charging strategy
While many solutions at the material level work well, commercialization is difficult in the near future. The researchers moved the fast-charging solution to the battery and Pack tiers to enable applications in the short term. The design of the charging strategy is the key to solving this problem.
1) Types of charging strategies
Standard charging
CCCV is currently the most common charging protocol, that is, first constant current charging to the cutoff voltage (CC stage), and then constant voltage charging to a small current close to 0 (CV stage). The constant voltage process can make the ion concentration distribution in the electrode material more uniform, which is crucial for the material to exert a high specific capacity; but the current at constant voltage gradually decreases, making the charging time of CV significantly longer than that of CC. The simple operability of the CC-CV charging mode makes it the most widely used standard charging protocol. But there are many other key strategies that can reduce charging time, improve charging efficiency, and improve capacity/power retention. Figure 8 shows several common fast charging strategy curves.
Figure 8 Common fast charging strategy curve a) constant current-constant voltage (CC-CV); b) constant power-constant voltage (CP-CV); c) multi-stage constant current-constant voltage (MCC-CV) ;d) pulse charging; e) CC-CV-CC-CV mode continuous charging (Boostcharging); f) variable current charging (VCP)
Multi-stage constant current charging
Many studies have proposed that adjusting the current of the charging process can slow down the aging of the battery while reducing the charging time. The aim of these studies is often to reduce heat production, avoid lithium evolution or reduce mechanical stress. MCC was one of the earliest strategies used for fast charging, consisting of a two- or multi-step constant-current phase, followed by a constant-voltage phase. Since the negative potential at the beginning of charging does not easily drop to the lithium evolution potential, the current in the early CC stage is larger. However, some researchers have adopted the opposite, that is, the charging strategy of gradually increasing the current in the CC segment, which is because the internal resistance of the battery will gradually decrease.
Pulse charging
During pulse charging, the current shows periodic changes to reduce the polarization of the concentration difference, avoid negative local potential or reduce the increase in mechanical stress caused by local lithium ion de-embedding.
Enhanced charging
CC-CV charging is performed with a larger average current during the initial charging phase and then a reduced current. The first stage of charging can be the CC phase (the entire charging strategy is equivalent to MCC-CV), the CV phase (CV-CC-CV) after the battery voltage reaches the maximum voltage set, or a full CC-CV stage (CC-CV-CC-CV). This strategy sets higher current and voltage than CC-CV to reduce the total charging time. However, under the same charging time, enhanced charging has faster capacity attenuation than CC-CV, and pulse charging is not significantly different from CC-CV. Some researchers have shown that CC-CV is suitable for high-power battery fast charging, and MCC is often used in charging scenarios where lithium is easily analyzed.
Variable current charging
In order to achieve the purpose of fast charging, the researchers proposed a series of more complex variable current charging curves, including VCD, UVP and so on. As the battery ages, the current curve needs to be adjusted for changes in internal resistance at the same voltage. In addition to the aging factor, the current is always low during the initial charging phase and then rises rapidly due to the maximum internal resistance of the 0% SOC, after which it decreases rapidly. The maximum current often occurs in the lower SOC region, and then the current gradually decreases due to the increase in the amount of lithium embedded in the particles and the limited transmission of Li. In addition, the distribution of temperature within the battery and Pack during charging is very important, but the charging control strategy often only considers the surface temperature as the main factor of attenuation.
Schindler combined the different charging strategies in Figure 9 to perform cycle experiments on the battery, and compared it with CC-CV to study the capacity decay of the battery under different cycles. The results show that combined with all charging strategies for cycle experiments, the battery maintained 80% of the capacity after 800 cycles, which performed best in all cycles; the battery with only CC-CV cycle attenuated to the same capacity was only used 400 times; while under the cycle combined with CC-CV and cold derating, the battery only cycled 330 times, the worst performance.
Figure 9 Current curve: a) AC pulse; b) cooling derating; c) polarization retention; d) pulse charging
Most fast charging strategies are only effective at standard temperatures and specific battery configurations. Since large currents can cause greater mechanical stress inside the electrode particles, accompanied by significant uneven current and temperature distribution, fast charging should be used with care when using different types of batteries. The universality of many current charging strategies lacks further experimental verification. With the promotion of electric vehicles in low temperature areas, more research on fast charging strategies at low temperatures is needed. In addition, the battery performance is determined by its own temperature, not the ambient temperature, and the change in battery temperature during charging also needs to be considered. Finally, the impact of different charging strategies at pack levels needs to be studied.
2) Model-based strategy optimization
Fast charging strategy based on ECM model
Some researchers optimize charging strategies based on equivalent circuit models, and they use formulas to embed these models into single-target or multi-objective optimization constraint problems. In these problems, first- or higher-order equivalent circuit models are used to describe battery behavior by setting multiple cost functions to achieve maximum charging efficiency or minimal charge loss.
Based on the equivalent circuit model, a coupled thermal-electrical-aging model can be established, which can describe the thermal effect or battery aging caused by charging, and can optimize the battery temperature and aging based on the model for fast charge optimization. In addition to the commonly used lumped models, some enhanced models can separate the temperature inside the battery and the surface, or improve the accuracy of simulation at large magnifications. Combining charge rate, activation energy, total discharge capacity, and temperature, the Arrhenius formula can be used to accurately simulate aging phenomena.
Once the framework for the optimization problem is established, a suitable algorithm can be developed for fast charge control according to the cost function and constraints. Common algorithms include dynamic programming, Pontiac minimum principle, genetic algorithms, LGR pseudo-spectroscopy, and minimum-maximum strategies.
The equivalent circuit can describe the external characteristics of the battery, but cannot provide information about its internal state, especially side reactions during charging, such as SEI film thickening, lithium deposition, etc. As a result, electrochemical models have received attention.
Fast charging strategy based on electrochemical model
Electrochemical models can estimate the internal state of the battery (solid/liquid phase potential, ion concentration, reaction flow, etc.) to predict side reactions during charging, and the most commonly used electrochemical model is the P2D model proposed by Doyle, Fuller, and Newman. However, in full-order models (FOMs), the computational value of solving partial differential equations (PDEs) is very large. As a result, the researchers did a lot of simplification work based on FOM to increase the computation rate. Some models also incorporate side reactions to more realistically simulate the inside of the battery. In recent years, some physically significant ECMs have also been used to describe electrochemical processes inside batteries, and their parameter identification is simpler than P2D.
In summary, model-based optimization of charging optimization usually prefers ECM, SP, ROM, etc. to FOM, which is due to the small amount of computation of the former and is more suitable for real vehicle applications. But this is often at the expense of accuracy, so careful verification is required in certain abusive conditions such as fast charge simulations. Although there are many model-based optimization methods, few model results can fully match the experimental data, and these anastomosiss are only suitable for fresh battery scenarios, and the problem of establishing a long-term aging model for batteries needs to be solved urgently.
Fast charging strategy
Impact of thermal management
Fast charging is often accompanied by a large number of heat production and heat production uneven problems, and the large magnification charging at low temperatures has great damage to battery life and safety. Effective thermal management is therefore important for achieving lossless fast charging under all conditions. The design of battery thermal management systems at different temperatures can vary greatly. Cooling the Pack requires high thermal conductivity, while at low temperatures the Pack requires better thermal insulation to retain enough heat for itself. Adjusting thermal conductivity according to temperature is one way to solve the problem.
1) Cool down
Common cooling media for electric vehicle Pack are air, liquid and phase change materials (PCM). Air cooling systems are low cost and simple, but due to their low heat capacity and poor thermal conductivity, air cooling rate and temperature consistency are poor, and they are not suitable for fast charging systems. Liquids are 3500 times more efficient than air, but they are costly, complex and potentially leaky. To avoid short circuits, the cooling medium must be an insulator, and commonly used liquids include deionized water and mineral oil. PCM cooling is the use of the material phase change process to absorb battery heat, but its disadvantages are also obvious: when the room temperature is very high, even if the battery does not produce heat PCM will completely melt, low thermal conductivity liquid PCM will hinder the battery heat dissipation.
Since fast charging inevitably further worsens the unevenness of the temperature distribution, efficient and uniform cooling technology is more important than standard charging. The thermal conductivity inside the battery relative to the surface is worse, and the battery surface is usually connected to the cooling system, which exacerbates the uneven distribution of temperature inside and outside the battery, and there are similar problems in the battery module and pack.
Finally, some electric vehicle charging piles will be equipped with corresponding external cooling systems according to the charging conditions while increasing the fast charging rate. If achievable, this approach will reduce the cost of the on-board cooling system.
2) Preheating in low temperature environment
Low temperature and fast charging of lithium-ion batteries is very difficult. This section only describes methods for rapidly heating the entire battery, as fast heating is indispensable for fast charging. The internal heating method is favored for its high efficiency and high uniformity. The four common methods are: self-discharge heating. This method is less efficient; the battery drives the heating wire and heats it with a fan. This method heats relatively quickly but is not efficient enough and unevenly heated; two-way pulse heating. That is, a battery Pack is divided into two groups of equal capacity batteries, and the power is pulsed between the two groups of batteries, and the internal resistance is used for heating. This method is more efficient, mainly limited by DC/DC conversion, simulation results show that this method can heat 2.2Ah 18650 batteries from -20 °C to 20 °C within 120s; AC heating. This method of heating is faster, but its effect on battery aging and cycle stability is unclear. Designing lithium-ion battery configurations for rapid preheating is also one of the ways to solve low temperature fast charging. For example, an electrochemically separated nickel foil can be inserted in the middle of two single-sided negative electrode layers, and the DC current flowing through the nickel foil can be controlled by a switch for rapid heating.
Although the internal heating method is more efficient and enables a more uniform temperature distribution, the effect of coupling internal heating and fast charge on battery cycle life has been rarely studied. Since the current passes through the low resistance zone more easily, the corresponding zone temperature increases, so even the small temperature gradient caused by the warm-up will be amplified at fast charging. Since the internal temperature is difficult to measure experimentally, cycle testing or reliable modeling is required to evaluate preheating methods. Nickel foil preheating, while promising, requires a non-standard battery design and adds weight and other possible problems.
Impact of thermal management
security
1) The effect of fast charging on thermal runaway
Studies have shown that the thermal runaway behavior of batteries changes after fast charging. For example, arc testing of high-energy soft pack batteries after fast charging found that the thermal runaway temperature of the battery after fast charging is significantly reduced compared with fresh batteries, and these effects can be eliminated if there is enough standing time. As the standing time increases, the precipitated lithium is gradually re-embedded in the negative electrode, and some of the lithium and electrolyte react to form a new SEI membrane. As a result, the lithium involved in the thermal runaway process decreases, and the thermal runaway characteristics of the battery gradually return to the level of fresh batteries.
Thermal runaway is caused by a series of chain reactions. Thermal runaway of fresh batteries is usually caused by a short circuit, followed by a reaction of the electrolyte and the battery temperature reaches its maximum. The thermal runaway process of the battery after fast charging can be divided into three stages, as shown in Figure 10. The first stage (60 °C
Fig. 10 Chain reaction of the thermal runaway process of the battery after fast charging
2) Thermal runaway caused by overcharging
Some fast-charging battery Packs may be overcharged due to inconsistencies between battery cells, which may lead to thermal runaway. This process can be divided into 4 stages:
Stage 1 (100%
Stage 2 (120%
Stage 3 (140%
Phase 4 (140%
Based on the internal material and reaction kinetics of the battery, two design methods have been proposed to protect the battery from overcharging:
1) Increase the oxidation potential of the electrolyte from 4.4V to 4.7V, which makes the electrolyte more stable and increases the SOC that thermal runaway occurs to 183%. This can be achieved by adding functional additives to the electrolyte or additives that can undergo reversible redox reactions.
2) Increase the thermal runaway temperature of the battery to 300 °C to delay the occurrence of short circuits in a large area, and increase the SOCs that occur thermal runaway to 180%. By optimizing the pressure design of the battery, or by using a separator with high heat exchange stability, the rupture of the battery pack can be delayed.
security
conclusion
The electrification of vehicles is undoubtedly one of the important means to solve climate change. In order to cope with mileage anxiety and meet customer needs, many manufacturers have taken Pack's fast charging capacity as an important indicator. Although there have been many studies on fast charging over the years, there are still many problems:
1. To this day, there is still no reliable on-board method to detect the aging of batteries (such as lithium evolution and mechanical cracking). The lithium evolution detection method based on the voltage platform is expected to achieve online application, but how to effectively distinguish between lithium dissolution platform and other phenomena, as well as detect lithium evolution without voltage platform, there has been no relevant research.
2. Many new electrode materials have good fast charging capability, but their stability, attenuation mechanism, large-scale production and cost are still debatable. Although graphite anodes are very easy to analyze lithium, considering the cost, wide range of applications and technological maturity, graphite will occupy the main market for lithium-ion battery anode materials in the foreseeable future.
3. Existing model methods have obvious limitations: ECM-based models cannot predict the internal state of the battery and can only be used within a limited range. On the other hand, high-precision FOM models cannot be applied in real time due to the large amount of calculations. Therefore, a reduced-order model is needed to accurately describe the internal core state of the battery for application in future fast-charging BMS systems.
4. Many fast charging strategies are based on experience or experiments, and the results are only applicable to specific batteries of specific configurations or specific operating conditions, and cannot be extended to other types of batteries. In addition, many model-based charging optimization studies are based on SP or ECM models, and the prediction accuracy of models at high currents is often insufficient.
5. There are very few studies on the strategy of fast charging optimization at low temperatures, and these studies are crucial for the promotion and application of electric vehicles in cold areas.
6. In order to further optimize the charging process of the single battery in the battery pack and avoid local aging or overcharging, it is also necessary to develop an advanced BMS system with balanced battery consistency.
7. Although there has been a lot of research devoted to the design of thermal management systems, the efficiency and uniformity of various preheating and cooling systems still need to be thoroughly evaluated. Few researchers have evaluated the effect of alternating current preheating coupled fast charging on battery life. Optimizing the ear design, position and geometry of the cooling system is also an important means of improving temperature and current uniformity. External cooling technology coupled charging pile is also an important way to reduce the cost and quality of on-board cooling systems, but its real effect remains to be seen.
8. Finally, the relationship between the battery cell and the attenuation rate of the Pack is not clear. Although many charging and preheating strategies are effective on battery cells, the effects, feasibility, and cost of their application on Pack have not been studied. The application of effective charging strategies for some monomer batteries to the Pack may cause uneven distribution of temperature and current density, so any non-traditional charging technology needs a lot of research before practical application. In addition, almost no model takes into account the effects of inconsistencies between cells inside the battery pack. Since fast charging amplifies inconsistencies, multi-size research needs to be carried out urgently. Multi-size studies are critical for cell integration and Pack design of batteries.
Erratum regarding previous published articles
eTransportation, Volume 6, November 2020, Pages 100088
conclusion