Li-Ion Battery Active–Passive Hybrid Equalization Topology for Low-Earth Orbit Power Systems

Abstract

The lithium-ion battery equalization system is a critical component in Low-Earth Orbit (LEO) satellite power supply systems, ensuring the consistency of battery cells, maximizing the utilization of battery pack capacity, and enhancing battery reliability and cycle life. In DC bus satellite power systems, passive equalization technology is widely adopted due to its simple structure and ease of control. However, passive equalization suffers from drawbacks such as complex thermal design and limited operation primarily during battery charging. These limitations can lead to inconsistent control over the depth of discharge of individual battery cells, ultimately affecting the overall lifespan of the battery pack. In contrast, active equalization technology offers higher efficiency, faster equalization speeds, and the ability to utilize digital control methods, making it the mainstream direction for the development of lithium-ion battery equalization technology. Nevertheless, active equalization often requires a large number of switches and energy storage components, involves complex control algorithms, and faces challenges such as large size and reduced reliability. Most existing active equalization techniques are not directly applicable to DC bus satellite power systems. In this study, based on the operational characteristics of LEO satellite power storage batteries, an active–passive hybrid equalization topology utilizing a switching matrix is proposed. This topology combines the advantages of a simple structure, ease of control, and high reliability. Its feasibility has been validated through experimental results.

1. Introduction

Because the single-cell voltage of a lithium-ion battery is low and its capacity is small, it is necessary to combine multiple cells in series and parallel to improve their open-circuit voltage and discharge power when they are used as energy storage devices in satellite systems [1,2]. In the manufacturing process of battery cells, there will be some differences in electrical characteristics due to technological reasons, and these differences will gradually increase during use. Even though the battery cells in the lithium-ion battery pack are carefully selected before leaving the factory so that their performances (such as capacity, internal resistance, etc.) are as consistent as possible, in the process of use, the differences will gradually expand with the aging of the battery cells due to the slight differences in working temperature and circuit parameters, such as the ability to withstand high voltage and large current, internal resistance, and self-discharge rate [3]. Without effective balanced management, the consistency between lithium-ion batteries will be divergent in the process of recycling, which will accelerate the attenuation or even failure of individual batteries with poor performance, thus deteriorating the overall characteristics of the battery pack and affecting the performance and life of the whole lithium-ion battery pack [4].

In the energy storage system, the lithium battery equalizer occupies a very important position and shoulders the important responsibility of improving the reliability and prolonging the life of the battery pack. In essence, the balancing process of the battery is the process of “peak-cutting and valley-filling” of the single battery, which can be realized by energy dissipation or energy transfer. At present, there are many kinds of lithium-ion battery equalizers [5,6,7,8,9,10,11], and there are many classification methods according to different topological structures [12], different energy-carrying devices [4], or different energy transfer paths [1,13].

Passive equalization technology has the advantages of simple structure and easy control [14,15] and occupies an absolute main position in DC bus satellite power supply systems [16,17]. However, passive equalization has the disadvantage of complex thermal design, and it is often only balanced when the battery is charged, which will lead to inconsistent control of the discharge depth of the battery, thus affecting the life of the battery pack.

Active balancing technology has high efficiency and balancing speed and can adopt intelligent control methods, which is the mainstream development direction of lithium-ion battery balancing technology [1,3,4,18,19]. However, active equalization technology often requires a large number of switches and energy storage devices and complex control algorithms and has the disadvantages of large volume and low reliability [20,21]. Most existing active equalization technologies cannot be directly applied to DC bus satellite power supply systems [22,23].

The DC bus satellite power supply system is divided into one side and the second side with the intermediate supply bus line. One side is responsible for power generation and control, and the second side is responsible for power distribution. The block diagram of the DC bus power supply system for the LEO satellite is shown in Figure 1. It is mainly composed of solar arrays, the Power Conditioning Unit (PCU), battery packs, capacitor banks, and various types of loads. The PCU combines MPPT (optimizing solar output) and Direct Energy Transfer (DET) to route excess solar energy directly to loads, reducing battery strain. A bidirectional DC/DC converter manages battery charge/discharge and stabilizes bus voltage. Integrated with the BMS, it employs balancing circuits to maintain cell uniformity, enhancing efficiency and extending battery lifespan under extreme orbital cycles.

The power system of LEO satellites exhibits the following characteristics [16,23]: (1) significant load variations; (2) short orbital periods; and (3) high frequency of battery charge and discharge cycles. According to the working characteristics of satellite power storage batteries, the demand principle of lithium battery balancing systems for satellites is put forward, and the existing balancing technology is screened and evaluated, and the feasible topology is selected for improvement and integration with the existing Battery Management System (BMS) of satellite power supply so as to ensure and improve the reliability of lithium batteries to the greatest extent, which not only has theoretical significance but also has application value.

In this paper, an active–passive hybrid equilibrium topology based on a switch matrix is proposed, which has the advantages of simple structure, easy control, and high reliability and is suitable for the application of energy storage systems in low-orbit satellites.

2. Battery Equalization Methods for Low-Earth Orbit Satellites

Battery equalization methods are critical for ensuring the reliability and longevity of energy storage systems in Low-Earth Orbit (LEO) satellites. These methods are typically classified based on energy transfer paths, which can be broadly divided into the following six categories (as shown in Figure 2): (1) Dissipative Balancing (Cell-to-R, C2R), where excess energy is dissipated as heat through passive components such as shunt resistors; (2) Adjacent Cell-to-Cell (AC2C), which transfers energy between neighboring cells using switched capacitors or inductors, offering simplicity in control but facing limitations in balancing speed and electromagnetic interference in large-scale systems; (3) Direct Cell-to-Cell (DC2C/C2C), enabling targeted energy transfer between non-adjacent cells via switch matrices or isolated DC/DC converters, though constrained by complexity and scalability; (4) Cell-to-Pack (C2P) and (5) Pack-to-Cell (P2C), dual methods that either distribute energy from a single cell to the entire pack or vice versa, with applications in topologies such as flyback converters but limited by their single-cell focus; and (6) Any Cells-to-Cells (ACs2Cs), a flexible approach allowing multi-cell energy redistribution through advanced switching strategies, albeit with high computational demands.

Passive balancing methods, predominantly C2R, remain widely adopted in aerospace due to their low complexity and high reliability, despite suboptimal efficiency. In contrast, active methods like DC2C and ACs2Cs demonstrate superior energy efficiency but introduce challenges in control complexity, electromagnetic compatibility, and system stability. For instance, AC2C-based topologies (e.g., switched capacitor chains [24]) simplify hardware but struggle with balancing speed in large battery strings, while DC2C architectures (e.g., resonant LC converters [25]) require sophisticated switch matrices and isolation mechanisms. C2P and P2C methods, often implemented with bidirectional DC/DC converters [26], reduce high-frequency switching losses but are ineffective for applications requiring simultaneous multi-cell balancing.

The selection of an optimal balancing strategy for LEO satellites must prioritize reliability, compactness, and weight minimization. Passive methods, though inefficient, align with the stringent reliability requirements and limited computational resources of spacecraft power systems. Active methods, while promising for high-performance applications, face barriers due to their intricate algorithms and susceptibility to failure in radiation-intensive environments. Current aerospace BMS designs predominantly utilize shunt resistor-based passive balancing, underscoring the need for future research into hybrid or adaptive topologies that reconcile efficiency with robustness.

3. Topology Constructure

The active–passive hybrid equalization topology based on a switching matrix integrates multiple subset topologies designed to separate charging equalization from discharging equalization in alignment with the energy storage characteristics of LEO satellite power systems. The conceptual framework of this topology is outlined as follows:

• LEO satellites have a short orbital period [16]. The orbital period of LEO satellites is about 100 min, and the illumination period is also short, about 1 h. Lithium batteries can be charged for a short time, so it is necessary to keep a large current in a short time to store electric energy quickly. It is simpler and more efficient to use bypass resistance equalization under the condition of high current. If active equalization is used, it is necessary to isolate the converter with high power, which is inefficient and affects reliability. In addition, in the lighting stage, the designed power of the solar cells on the satellite is greater than the power required by the system, and the bus regulator will short-circuit the excess solar cells so that electric energy can be converted into heat energy and radiated into space. Therefore, charging equalization does not need to consider efficiency, and passive equalization is enough to meet the requirements.
• Lithium batteries have many charge and discharge cycles [23]. LEO satellites have more than 5000 shadow periods every year, and the charging and discharging cycles have a great influence on the battery life, and the cycle life of lithium batteries is only about 1000 times. Passive equalization can only achieve charge equalization. Although LEO satellites rarely completely discharge batteries, adding active equalization during discharge can effectively control the discharge depth of lithium batteries, keep the attenuation of each battery cell consistent, and optimize the life of lithium batteries. Especially at the end of the design life of the satellite, when both solar cells and lithium batteries decay to a certain extent, discharge equalization can maximize the energy stored in the battery pack and maintain the normal operation of the satellite system. Active equalization can be designed with low power, and the equalization current can be about C/5, which can achieve good results.

The topology structure of active/passive hybrid equalization based on a switch matrix is shown in Figure 3. The passive equalization adopts shunt resistor topology, and the active equalization part chooses flyback topology based on the switch matrix. Passive equalization works when charging, and active equalization works when on standby or discharging.

4. Working Principle

In Figure 3, the left side is a flyback equalization converter, the input of which is the battery pack voltage, and the output is connected to the switch matrix equalization port. On the right are the switching matrix and passive equalization, and R₄~R₉ are bypass resistors. Because passive equalization also includes switch structure, the structure in the frame in this figure is collectively referred to as the switch matrix in the following section.

4.1. Working Process

The hybrid active–passive balancing system proposed in this paper integrates two complementary operational modes to optimize efficiency and safety across the battery lifecycle. Charging-phase balancing leverages passive energy dissipation to prevent overvoltage risks, while standby/discharging-phase balancing employs active energy redistribution to maximize pack capacity utilization.

• Charging-Phase Balancing (Passive Mode): During charging, the system employs a passive balancing strategy to ensure uniform cell voltages and prevent overcharging. Each cell is continuously monitored, and when the nth cell reaches the predefined cutoff voltage (e.g., 4.2 V for Li-ion), its corresponding bypass switch S_na is activated. This connects the cell to a shunt resistor R_n, diverting excess charging current and stabilizing the voltage within safe limits. The process operates sequentially: once a cell is bypassed, charging continues for the remaining cells until all bypass switches are closed, indicating a full charge of the entire pack. This approach minimizes complexity and aligns with conventional passive balancing practices, as illustrated in Figure 4.

2.

Standby/Discharging-Phase Balancing (Active Mode): In standby or discharge states, the system transitions to an active balancing mechanism to address voltage imbalances dynamically. The monitoring circuit identifies the cell with the lowest voltage (e.g., Cell n), and a switch matrix reconfigures to connect this cell to the output of a flyback-based DC-DC converter. The converter, powered by the entire pack voltage, transfers energy from the pack to the low-voltage cell, compensating for its reduced state of charge (SoC). This process iterates until all cells achieve voltage consistency, maximizing usable capacity during discharge. The active balancing workflow is detailed in Figure 4, emphasizing energy-efficient redistribution without resistive losses.

4.2. Driving of Switch Matrix

Each battery cell is connected with a switch matrix basic unit, and each basic unit consists of five MOSFETs (metal-oxide semiconductor field-effect transistors). Figure 5a shows an N-channel basic cell in which back-to-back MOSFETs (Q_{n_a}~Q_{n_d}) form a bidirectional switch, and two bidirectional switches corresponding to each battery cell share a common drive and are simultaneously turned on in the circuit so that the battery cell can be connected to an equalization channel, and then the battery voltage can be sampled or equalized, and this equalization only occurs when the battery pack is on standby or discharging. When the switch Q_{n_e} on the right side of the battery is turned on, the bypass shunt resistor R_N will be merged into the battery circuit to charge the battery monomer evenly. Because the battery pack is composed of battery cells in series, the relative potentials of switches in the switch matrix are inconsistent, and at least n isolated driving power supplies are needed. If the auxiliary power supply scheme is adopted, the winding process of transformers is more complicated, and the reliability is not high. The battery self-driving technology is adopted, and the high-position battery is used to drive the low-position MOSFET switch so that the high-position switch matrix unit has no battery to drive. The highest switch needs to be driven by a low-level battery, so the P-channel basic unit is introduced into the switch matrix, as shown in Figure 5b. The five switches of the P-channel unit are all P-channel MOSFETs, which can be driven by a low-level battery voltage. In the figure, B_N and B_P refer to lithium battery monomers corresponding to N-channel cells and P-channel cells, respectively. Although the equivalent circuits and functions of the two basic units are the same, the physical structure of P-channel MOSFET leads to a higher on-resistance (R_DSon) than that of N-channel MOSFET of the same grade, and there are fewer models of P-channel MOSFETs, so the fewer P-channel basic units in the switch matrix, the better. Take a battery pack with 16 cells in series as an example; only the P-channel basic unit is used in the highest two units, that is, the P-channel MOSFET is used for the corresponding matrix unit of B₁₆ and B₁₅. The driving modes of the two units are described in detail below.

Figure 6 shows the driving of the bypass resistance switch. The lower N-channel MOSFET is driven by the battery voltage two bits higher than it, and the higher P-channel MOSFET is driven by the battery voltage two bits lower than it. The driving signals are all isolated by optocouplers. The driving current path is shown by the dotted line in this figure.

Figure 7 shows the driving circuit of bidirectional switches in the switch matrix. Two bidirectional switches on the same battery share a driving signal and are turned on and off at the same time. The driving mode is the same as the bypass resistance switch. The N-channel MOSFET S₅ is driven by the voltage of the third battery B₈ above it through the optical coupler O₅, and the P-channel MOSFET bidirectional switch S₁₄ is driven by the third battery B₁₁ below it through the optical coupler O₁₄. The driving current path is shown by the dotted line in this Figure.

4.3. Equalization Converter Design

The balanced converter adopts flyback topology. The input of the converter is the battery pack voltage. Taking a 16-string battery pack as an example, the input range is 54 V~65 V, and the output is clamped by the lowest battery pack voltage. Usually, the larger the charging current of the converter, the shorter the balancing time. However, considering the stress margin of components and the volume of the system, the design power of the converter needs to be balanced with the acceptable balancing time. Moreover, too much power will affect the heat dissipation design of the circuit, and too little power will not reflect the balance effect. If the discharge time of the battery pack is 120 min, it is appropriate to design an equilibrium converter to reduce the divergence of the battery to half within 120 min. The structure of the balanced flyback converter is shown in Figure 3 in the previous text.

The transformer of a flyback converter has three main functions: (1) electrical isolation; (2) realizing voltage matching by adjusting the transformation ratio; and (3) storing energy as an inductor. Because it involves the stability of the system, electromagnetic interference has a higher priority than system efficiency in design. Considering that the reverse recovery current of the output diode D₁ of the flyback converter in the discontinuous current mode (DCM) has become zero before it is turned on again, thus suppressing the electromagnetic interference (EMI) caused by the ringing phenomenon, the designed converter works in the discontinuous current mode in the system.

In the design, if the maximum output power of the designed converter is 10 W and the rated output current is 1.5 A, the components can be selected according to the above formula, and the primary MOSFET is 250 V/5 A with a voltage margin of 30%. The transformer adopts a ferrite core with an EI20 structure, which can meet the design requirements.

4.4. Equalization Time Analysis

The switch matrix equalizer works in the standby or discharge state of the battery pack, and supplements the minimum battery cell power of SOC by absorbing the battery pack power. Firstly, it is assumed that the operation time, ADC sampling delay, and voltage sampling error of the equalization controller are ignored. Obviously, the equalization time of a lithium battery is proportional to the equalization current, and the larger the equalization current, the shorter the equalization time. In addition, the equalization time is also related to the imbalance between batteries, and the larger the voltage difference (or SOC difference) between batteries, the longer the equalization time required. Next, the equalization time Δt_eq of the lithium battery is analyzed with the SOC maximum difference (ΔSOC) and equalization current I_eq as reference quantities.

Because the SOC of a lithium-ion battery is linearly related to the open circuit voltage, the ΔSOC can be approximately replaced by the maximum voltage difference between batteries near ΔV, as shown in the following formula:

$$\Delta SOC={\displaystyle \frac{C\Delta V}{Q_{total}}},$$

(1)

where C is the remaining capacity of the battery cell and Q_total is the maximum charge of the lithium battery cell. According to the charging characteristics of the battery, the following formula can be deduced:

$$C={\displaystyle \frac{I_{eq}\mathsf{\Delta }{t}_{eq}}{\mathsf{\Delta }V}}.$$

(2)

Due to that characteristic of the switch matrix equalizer, the following equation can be obtained:

$$I_{eq}=I_{in}-I_{out}=I_{in}(1-\eta {\displaystyle \frac{V_{cell}}{V_{pack}}}),$$

(3)

where I_out is the balanced current flowing out of the lithium battery pack; I_in is the current flowing out of the lithium battery monomer; η is the efficiency of the balanced converter; and V_cell and V_pack are the voltages of the battery monomer and the battery pack, respectively.

Taking a battery pack composed of 16 battery cells in series as an example, (3) can be rewritten as follows:

$$I_{eq}=I_{in}\left(1-\eta {\displaystyle \frac{V_{cell}}{16V_{cell}}}\right)\approx I_{in}.$$

(4)

By simplifying (9)–(11), the following expression is derived:

$$\mathsf{\Delta }{t}_{eq}={\displaystyle \frac{C}{I_{eq}}}\mathsf{\Delta }V={\displaystyle \frac{Q_{total}}{I_{in}}}\mathsf{\Delta }SOC.$$

(5)

By analyzing the above results, the following conclusions can be drawn:

• The equalization control system can predict the equalization time through (5), adjust the optimized equalization algorithm in real time, and improve the equalization speed.
• By increasing the power of the equalization converter, the equalization time can be reduced.
• The influence of the efficiency of the switching power supply on the equalization time decreases with the increase in the number of battery cells.

5. Experimental Results and Discussion

5.1. Introduction of the Prototype

As shown in Figure 8, the experimental platform of the lithium-ion battery manager consists of four parts: charger, battery pack, BMS control panel, and monitoring platform. The charging and discharging devices mainly realize the charging and discharging function of the lithium-ion battery pack. The charger is simulated by a programmable DC power supply, and the discharger is simulated by a programmable electronic load. The test battery pack consists of 16 CALB LP2778102 aerospace-grade lithium-ion cells (China Aviation Lithium Battery Co., Ltd., Zhengzhou, China) in series configuration, with a rated capacity of 2.8 Ah. The battery capacity was appropriately reduced compared with the actual satellite platform to shorten the verification time. The charging and discharging parameters of the battery pack are as follows: charging cut-off voltage of 4.2 V and discharging cut-off voltage of 3.6 V. The detailed parameters of the battery are shown in

Table 1. The detailed parameters of the battery.

Parameter	CALB LP2778102
Chemistry	Lithium Nickel Manganese Cobalt Oxide (NMC)
Nominal Voltage	3.7 V
Capacity (Per Cell)	2.8 Ah
Cycle Life (80% DoD)	1200 cycles
Operating Temp.	−30 °C to +55 °C
Configuration in Study	16S (16 cells in series)
Total Pack Voltage	59.2 V (3.7 V × 16)
Dimensions (Per Cell)	27.8 × 78 × 102 mm
Weight (Per Cell)	420 g

.

The main electrical parameters of the system are as follows: (1) input voltage range of the equalization circuit is 55~64 V; (2) the balanced current is 1.5 A; (3) the converter switch is a 250 V/5 A MOSFET; (5) the transformer ratio is 60:10 and L_m is 600 μH; (5) switch matrix P-MOS is −150 V/−10 A; and (6) switch matrix N-MOS is 150 V/10 A. It should be noted that the actual switch matrix voltage stress does not need 150 V, and this model was chosen because it is easy to purchase on the market.

5.2. Performance Verification

(1)

Converter performance

Figure 9 shows the waveform when the balanced converter works, and the converter works in intermittent current mode, where (a) is the driving waveform of the switch, and the designed switching duty ratio is 30%; (b) is the voltage between the drain and source of the switch, with v_ds of 180 V at the maximum, with a design margin of 75%; (c) is the waveform of the output rectifier diode; (d) is the current waveform of the primary switch, and the current peak I_p is 1.2 A; and (e) and (f) are output voltage waveforms. The output voltage value is clamped by the battery, and the voltage ripple is 10 mV, which will not affect the battery life. Experiments show that the selected devices meet the stress requirements of the converter, and the performance of the converter meets the design requirements.

Because the design of the solar power generation power of the satellite power supply system is much larger than the load demand power, the excess energy will be converted into heat energy and discharged into space, and the equalizer efficiency is not the focus of attention during charging, so only the efficiency curve of the discharge equalization converter is tested as shown in Figure 10. When the difference between the battery cell voltage with the worst performance and the average voltage is relatively large, the converter adopts 1.5 A constant current control. At this time, this load is set at 60% of the load of the converter, ignoring the energy loss in the battery, and the balancing efficiency is 89%. When the difference between the cell voltage and the average voltage of the battery with the worst performance is not large, the equalizing converter adopts constant voltage control, and the efficiency is slightly lower, ranging from 73% to 89%, but the equalizing power is very small, and the loss is within the controllable range.

(2)

Unbalanced system test

The charge/discharge characteristics of the battery pack were evaluated with both charge and discharge equalization disabled, as shown in Figure 11. To ensure graphical clarity, the number of displayed battery cells was reduced in these illustrations. The lithium battery was charged at 2 A and discharged at 1.5 A, with voltage sampling conducted at 2-min intervals. The 2-min sampling interval was strategically selected to address the following three design requirements: (1) maintaining state of charge (SOC) variation per sampling cycle within the ±0.15% measurement resolution; (2) accommodating the computational constraints of radiation-hardened processors to ensure system stability in radiation-intensive orbital environments; and (3) minimizing energy consumption for satellite power budget compliance through optimized voltage sampling frequency.

Before the test, the initial voltage of each battery cell is maintained at a relatively close voltage value, the initial charging voltage is about 3.65 V, and the initial discharging voltage is about 4.15 V. Both charging and discharging actions begin after the circuit runs stably for 25 min. As can be seen from Figure 11, the charging and discharging of the tested battery pack are inconsistent. The initial charging imbalance is 4.2%, and it diverges to 6.8% at the end of charging. The voltage imbalance during discharge is more serious, from the initial 4.0% to the final 13.6%. Unbalanced battery voltage will lead to incomplete charging and discharging of the battery pack, resulting in a waste of capacity.

(3)

Charge equalization test

The charge equalization performance was evaluated under constant-current charging conditions (2 A) with the balancing circuit enabled. As shown in Figure 12a, the measured voltage profiles demonstrate improved voltage convergence characteristics during the constant-voltage charging phase when compared with the unbalanced case in Figure 11a. The SOC difference among cells decreased from an initial 4.0% to 1.9% through passive balancing operation, with the balancing system effectively mitigating cell-to-cell variations during the charging process.

(4)

Discharge equalization test

The battery pack is in a standby state. The battery pack voltage was 53.23 V, and the average voltage of lithium-ion battery cells was 3.802 V. Cell₇ was artificially discharged to 3.696 V, and the voltages of the other batteries were close. The voltage of each cell is sampled regularly, and the measurement frequency is measured once every two minutes, too. At the 40th minute, the discharge equalization circuit starts to equalize Cell₇ until the program judges that Cell₇ meets the equalization conditions. Finally, the relationship between equilibrium voltage and time is drawn as shown in Figure 12b.

As can be seen in Figure 12b, the initial time point Cell₇ differs from the average voltage by 110 mV, and the estimated SOC difference is 17%. The voltage difference exceeds the programmed threshold, and the equalization circuit is started. After about 60 min, the maximum pressure difference is less than the set value, and the equalization circuit is closed. The final pressure difference is about 25 mV, and the estimated SOC difference is 6%. After repeated tests, the adjustment error and equalization time are within the acceptable range.

The battery pack removes the equalization circuit and repeatedly charges and discharges for three rounds so that the imbalance between battery cells is randomly distributed. Then, connect the lithium-ion Battery Management System and start to adjust the voltage between the battery cells at 40 min. The voltage data of five typical battery cells are captured every two minutes, in which Cell₅ and Cell₃ are the two battery cells with the highest voltage, Cell₇ is the battery cell with the middle voltage, and Cell₄ and Cell₉ are the battery cells with the lowest voltage. Record 200 min of data and draw a V-t curve as shown in Figure 13a.

It can be seen from Figure 13a that the maximum voltage difference is 130 mV when the five sampled cells are in the initial state, and the difference converted into SOC is 22.2%. When the lithium-ion battery manager starts to work, it first transfers energy to Cell₉, with the lowest voltage. Due to the onboard microcontroller’s computing limitation, the equalizer cannot operate when multitasking is needed. Therefore, each working stage of the equalizer is set to 30 s in the program, and then it is scanned again in 2 min to determine whether equalization is needed. At the 90th minute, the voltages of Cell₄ and Cell₉ are equal, and the balanced current alternately flows into them. It was not until the 110th minute that Cell₄ became the minimum voltage battery cell, and the equalizer began to act only on Cell₄. Further observation shows that the closer the voltage value is to the monomer with the average voltage, the smaller the change rate is. However, Cell₅, with the highest initial voltage, has the largest decline slope, and the energy is continuously transferred outward during the whole process. As time goes on, the voltage difference in the battery pack tends to converge. By the end of the measurement, the maximum voltage difference is 70 mV, and the SOC difference is reduced to 6.2%. It should be pointed out that if the measurement is not stopped, the battery voltage difference will continue to decrease. In addition, there are several measurement bumps in the curves of Cell₃ and Cell₇ because the communication of the lower computer was interrupted during the measurement process, which delayed the sampling time point, which was caused by the relaxation voltage of the lithium battery, and the bump voltage was less than 5 mV, which was within the allowable error voltage range.

The test shows that the addition of a discharge equalization circuit can improve the consistency of the battery pack. Redistribution of energy in battery cells will not increase system energy consumption. By compensating for the low-voltage monomer, the working time of the whole battery pack can be increased, and the charging times can be reduced, thus prolonging the life of the battery pack and further improving the reliability of the whole system.

The battery is artificially unbalanced, and the voltage of Cell₉ is the highest in the battery pack, with a voltage of 4.15 V, which is close to the fully charged state. The battery Cell₁₂ has the lowest voltage of 3.66 V, which is close to the lowest allowable voltage. The maximum voltage difference in the battery is 0.49 V. The initial state of the circuit is standby equalization. At the 25th minute, the load is connected to discharge the battery, and the discharge current is 1 A.

Select the voltage sampling data of six typical battery cells in the battery pack and draw the V-t curve as shown in Figure 13b. In the whole test process, because the voltage of Cell₁₂ is the lowest, the active equalization circuit always supplements it, which is equivalent to bypassing Cell₁₂ in the discharge circuit. It can be seen from the experimental results that the voltage of Cell₁₂ is basically stable when other batteries are discharged, and there will be no overdischarge. If there is no active equalization circuit, the voltage of Cell₁₂ will decrease rapidly, and when it is discharged to the maximum allowable cell voltage of 3.6 V, the whole battery pack will be shut down, which will lead to the waste of stored energy in other batteries. Experiments verify the effectiveness and necessity of discharge equalization.

6. Conclusions

Most of the traditional energy storage systems of LEO satellites only have a passive equalization function, and all of them have poor performance, which affects the reliability of power supply systems. The power supply of the LEO satellite has the characteristics of a short orbit period and many cycles, and the power of the solar panel is much greater than the power required by the load in design. Combining the advantages of high reliability and good active balancing performance of lithium batteries, this paper proposes an active and passive balancing topology based on a switch matrix. This topology can quickly store energy by using simple and reliable passive equalization when the energy storage system needs to be charged quickly during the illumination period. In the shadow period, when the lithium battery is in a standby or discharge state, active equalization is adopted to prolong the battery life. The topology is simple, easy to control, and highly reliable. In this paper, the working principle and design method of topology are described in detail, and an experimental platform is built to test its performance. The experimental results show its effectiveness and practicality.