In this paper, I present my research on the charge and protection circuit for a lithium iron phosphate traction battery pack. My work is motivated by the rapid development of electric power tools and the urgent need for cleaner energy storage systems. I focus on the system requirements for small and medium-sized power tools, where a reliable and intelligent battery management system is essential. Throughout this paper, I emphasize the importance of the traction battery pack and its management in practical applications. The system I designed operates at 48 V with a nominal capacity of 20 Ah, consisting of 15 series-connected cells. The maximum allowable charge and discharge currents are both 20 A, and the working temperature range is 0–50 °C.
My research begins with a brief introduction to lithium power batteries, then a description of the battery management system structure. The rest of the paper is devoted to the design of the charging section, based on a switch-mode power supply with active power factor correction, and the protection section, built around a microcontroller with advanced state-of-charge estimation and SMBus communication. I have included detailed calculations, component selections, and simulation results to verify the design.
1. Introduction and Research Motivation
With the increasing number of vehicles worldwide, the greenhouse effect and air pollution caused by exhaust emissions from fuel vehicles have become severe. It is reported that motor vehicles emit over 200 million tons of harmful gases into the atmosphere each year, which accounts for more than 60% of the total atmospheric pollution. The harmful substances in vehicle exhaust cause respiratory diseases and physiological disorders in humans. Moreover, some carcinogens can trigger cancer after prolonged exposure. Noise pollution is another major issue. In addition, petroleum is a non-renewable energy resource that will eventually be exhausted. China’s oil resources are quite limited, and it is predicted that by 2010 about 45% of the oil demand would have to be imported. The shortage of petroleum has become a worldwide problem.
In recent years, many countries have invested substantial resources in solving the problems of pollution and energy shortage caused by internal combustion engine vehicles. Developing battery-powered vehicles is one important solution. The energy stored in the battery can come from renewable sources such as wind, water, and solar energy. Electric motors are also more efficient than internal combustion engines, and they generate less noise. The Chinese government has listed the development of electric vehicles as a major scientific and technological project. Currently, electric bicycles and electric motorcycles are very popular in cities across China. However, the development of electric vehicles is still at the research stage. The complex structure and high performance requirements of automobiles are partially responsible, but the most serious bottleneck is the traction battery pack itself. Problems such as high cost, limited lifetime, unsatisfying driving range after one charge, and safety failures have hindered the commercialization of electric vehicles.
Valve-regulated lead-acid batteries are commonly used in the market because of their simple manufacturing process and low cost. However, they are bulky, heavy, have low specific energy, and a relatively low cycle life. Laboratory data show only about 300 cycles, while in practical use the number may be merely 80–120 cycles. Additionally, lead pollution is a serious concern. Due to the inherent limitations of lead-acid batteries, they cannot be applied to electric vehicles. In recent years, lithium-ion power batteries have made great progress. According to the different positive electrode compounds, lithium power batteries can be divided into several types. The question of which material is most suitable for large-scale application remains to be answered, but data from trial electric vehicles have shown that the battery is the most failure-prone component. Therefore, the charging and protection circuits should be tailored to the specific characteristics of the battery chemistry. Currently, lithium power batteries are not widely available in the market, partly because their own characteristics need improvement, but also because charging and protection circuits have not been redesigned for them. A well-designed battery management system can efficiently prevent battery failures, increase battery lifetime, ensure safety, and make the battery easier to use. My research on the charge and protection circuit of a lithium iron phosphate traction battery pack therefore has direct practical value for upgrading lead-acid batteries in electric bicycles and electric motorcycles, and it also contributes to the development of battery packs for larger vehicles.
2. Lithium Power Battery Fundamentals
A power battery is an energy storage device that can provide significant power over a relatively long period. The most widely used power battery type used to be the valve-regulated lead-acid battery, which has the advantages of simple manufacturing and low cost, but suffers from large volume, heavy mass, low specific energy, and limited cycle life. Moreover, lead-acid batteries use lead as an electrode material, causing severe environmental pollution. Therefore, lithium power batteries, which are the successor to lithium-ion batteries introduced commercially in the 1990s, have become attractive. Conventional lithium-ion batteries have high specific capacity, small volume, and low weight, but they are not suitable for high-current discharge because of potential burn-out or explosion hazards. A traction battery pack, on the other hand, must provide much higher power. For example, an electric bicycle motor is usually below 200 W, an electric motorcycle motor is typically around 300–400 W, and electric vehicle motors may demand tens of kilowatts. A traction battery pack is not just a larger lithium-ion cell, because it must guarantee high capacity, high power, high reliability, and high safety. It differs from consumer lithium batteries in materials and manufacturing, and should not be treated simply as a scaled-up version.
The performance of lithium batteries depends on the internal materials, especially the cathode material, which is also a major cost factor. In commercial cells, the cathode material accounts for roughly 40% of the total cost. An ideal cathode material should have a high redox potential, high reversible lithium-ion capacity, stable structure during insertion and extraction, good electrical conductivity, good compatibility with electrolyte, and be inexpensive and environmentally friendly. Among the many cathode materials studied, LiCoO2 has high working voltage and stable discharge performance, but suffers from high cost and limited practical capacity. LiNiO2 has high capacity but poor thermal stability. LiMn2O4 is cheap and safe, yet suffers from capacity fading at high temperatures. LiFePO4 is a most promising material for traction battery pack applications because of its excellent safety, low cost, and acceptable power performance. Table 1 lists the typical properties of the most common cathode materials.
| Cathode material | Theoretical capacity (mAh/g) | Actual capacity (mAh/g) | Voltage range (V) | Safety | Cost |
|---|---|---|---|---|---|
| LiCoO2 | 274 | 140-150 | 3.7 | Fair | High |
| LiNiO2 | 274 | 190-210 | 2.5-4.2 | Poor | Medium |
| LiMn2O4 | 148 | 90-120 | 3.0-4.0 | Good | Low |
| LiMnO2 | 286 | 200 | 3-4.5 | Good | Low |
| LiFePO4 | 170 | 110-165 | 3.2 | Very good | Low |
LiFePO4 has an olivine crystal structure. Its nominal voltage is 3.2 V, the end-of-charge voltage is 3.6 V, and the discharge cut-off voltage is 2.1 V. The specific energy is about 60–75 Wh/kg. Experiments have demonstrated that this material can support high-rate discharge (5–10 C), good high-temperature performance, excellent structural stability, and long cycle life. After 500 cycles the discharge capacity remains greater than 95%. Compared with lead-acid batteries, LiFePO4 can be fast charged at 2 C, and its volume is about two-thirds of a lead-acid battery of the same capacity, while its weight is only one-third. It is environmentally friendly. The main drawback is the reduced capacity at low temperatures: at 0 °C the capacity drops to about 78% of the value at 23 °C, and at -20 °C only 65% remains. Despite this drawback, LiFePO4 is the most promising cathode for a traction battery pack in the near future. Applications range from light electric vehicles to large electric buses and energy storage systems.
3. Traction Battery Pack Management System
The battery management system (BMS) is responsible for dynamically monitoring the operating state of the traction battery pack and the individual cells. Its functions include precise measurement of the state of charge (SOC), implementation of overcharge and over-discharge protection, thermal management, balancing, and communication with host computers or chargers.
I divide the BMS into two main parts: the charging section and the protection (control) section. The charging section consists of a switch-mode power supply (SMPS) with a PWM controller, while the protection section is built around a microcontroller. The microcontroller acquires the battery voltage, current, and temperature, and also controls the reference voltage and current signals to the charger. The BMS handles the charge process, monitors safety conditions, and informs the user about the remaining capacity of the traction battery pack.
A fully functional BMS should perform the following tasks: monitoring and statistical analysis of battery parameters; state-of-charge estimation; consistency compensation among series cells; temperature control; battery safety protection; and communication. Protection generally includes over-voltage and under-voltage protection, charge and discharge over-current protection, short-circuit protection, voltage equalization, and temperature protection. All these functions are required for the safe operation of a lithium traction battery pack.
4. Design of the Charging Section
The charging circuit must convert the AC mains input to a suitable DC voltage and current for charging the traction battery pack. A switch-mode power supply is preferred over a linear supply because of its higher efficiency, smaller size, and lighter weight. The general structure is shown below. The output voltage is sampled and compared with a reference to generate an error signal. The error signal controls the pulse width modulator, which adjusts the duty ratio to regulate the output voltage.
4.1 Switching Power Supply Topology Selection
Several types of DC-DC converters exist: buck, boost, buck-boost, single-ended forward, single-ended flyback, push-pull, half-bridge, and full-bridge. The choice depends on output power and other requirements. The comparison is summarized in Table 2.
| Topology | Isolation | Merits | Drawbacks | Output power range (W) |
|---|---|---|---|---|
| Buck | No | Simple | Not for high current | −50 |
| Boost | No | Simple | Not for high current | −30 |
| Buck-boost | No | Simple | Inverted output | −30 |
| Single-ended forward | Yes | Reliable, simple drive | Unidirectional core excitation | 50-200 |
| Single-ended flyback | Yes | Simple, low cost | High switch voltage stress | 20-100 |
| Push-pull | Yes | Bi-directional core | Flux imbalance | 100-500 |
| Half-bridge | Yes | Low switch voltage stress | Floating driver | 100-700 |
| Full-bridge | Yes | Low switch voltage stress, high power | Four switches, complex drive | 500-2000 |
The required charging parameters for my traction battery pack are: maximum output voltage about 55 V, maximum charge current up to 10 A (I limited the current to 0.5 C to prolong cell lifetime). The maximum output power is therefore 55 V × 10 A = 550 W. Since the power exceeds 75 W, a power factor correction (PFC) stage is mandatory to limit harmonic pollution on the utility grid.
I adopted a two-stage architecture: a boost-type active PFC preregulator followed by a DC-DC converter. The PFC stage shapes the input current to follow the input voltage and boosts the rectified voltage to 380 V. For the DC-DC stage, I chose a parallel single-ended forward converter also known as a dual interleaved forward converter. Two identical forward converters are interconnected in parallel to double the output power. With alternate switching of the two main transistors, the effective ripple frequency is doubled, which allows a smaller output filter. I used two high-frequency transformers, each delivering about 275 W.
4.2 Control Strategy
The PWM control method can be voltage-mode or current-mode. Voltage-mode control is simple but has slow dynamic response to input voltage changes. Current-mode control offers faster response and easier loop compensation. Among current-mode controls, average current-mode control is particularly suitable for PFC because it provides low distortion and good noise immunity. For the PFC section, I employed the average current-mode controller operating in continuous conduction mode (CCM). The inductor current is sensed and averaged by a current-error amplifier, which makes the input current closely track the sinusoidal reference, thus realizing a near-unity power factor. For the DC-DC section, I used a current-mode controller with both voltage and current feedback to achieve constant-current and constant-voltage charging profiles. The topology of a boost PFC with average current-mode control is well known.
4.3 PFC Circuit Design
I selected the integrated controller UCC28517, which provides a rising-edge PFC PWM signal and a falling-edge PWM signal for the DC-DC stage. Using a single chip reduces the component count and simplifies the timing relation between the two power stages. The operating frequency of the PFC stage was set to 100 kHz.
Rectifier bridge selection: The maximum DC voltage after rectification is
$$V_{DC,\max}=\sqrt{2}V_{AC,\max}\approx 375\ \text{V}$$
Assuming an efficiency of 0.85 and maximum output power of 550 W, the maximum input current is
$$I_{DC,\max}=\frac{\sqrt{2}P_{OUT,\max}}{\eta V_{AC,\min}}\approx 7.04\ \text{A}$$
I selected a KBPC1010 bridge rectifier, rated at 1000 V and 10 A, leaving a sufficient safety margin.
Boost inductor calculation: The minimum AC input voltage (85 V AC) yields the largest duty cycle. The duty cycle is
$$D_{\max}=1-\frac{\sqrt{2}V_{AC,\min}}{V_C}$$
where VC is 380 V. This gives Dmax ≈ 0.52. I chose a current ripple factor K = 0.2. The boost inductance is:
$$L = \frac{\sqrt{2}V_{AC,\min} D_{\max}}{\eta K P_{OUT,\max} f_{PFC}} \approx 2.44\ \text{mH}$$
Using a core with an AL-value of 200 nH/N2 at an air gap of 0.8 mm, the number of turns is:
$$N = \sqrt{\frac{L}{AL}} \approx 110\ \text{turns}$$
Power switch and diode: The PFC MOSFET must sustain over 380 V and 7.04 A. I selected 2SK1249, rated at 500 V and 15 A. The boost diode must have fast reverse-recovery because of CCM operation. I selected MUR860, rated at 600 V and 8 A.
Output capacitor: For a hold-up time tHU of 60 ms, the capacitance is:
$$C = \frac{2 P_{OUT,\max} t_{HU}}{V_C^2 – V_{\min}^2} \approx 580\ \mu\text{F}$$
I used a 600 μF/400 V electrolytic capacitor, with a 0.22 μF film capacitor in parallel for high-frequency filtering.
4.4 DC-DC Converter Design
The DC-DC stage operates at 200 kHz. The period is
$$T = \frac{1}{f_s} = 5\ \mu\text{s}$$
For the forward converter, the maximum duty cycle Dmax was set to 44% to avoid the risk of transformer saturation. Thus:
$$t_{on,\max} = T D_{\max} = 2.2\ \mu\text{s}$$
With a maximum output voltage of 55 V, forward voltage drop of the output rectifier UF = 0.5 V, and inductor-winding drop UL = 0.2 V, the minimum secondary voltage is:
$$U_{s,\min} = \frac{(U_{o,\max}+U_L+U_F)T}{t_{on,\max}} = 126.6\ \text{V}$$
The turns ratio is Np/Ns = 0.633 at a minimum DC input of 200 V. For the transformer core, I used two EI-E40 cores each of effective area S = 148 mm2, and maximum flux density Bm = 0.2 T. The secondary turns are:
$$N_s = \frac{U_{s,\min} t_{on,\max}}{B_m S}\times 10^2 = 9.4\ \text{turns}$$
I took Ns = 10, and the primary turns Np = Ns/0.633 ≈ 15.8, hence 16 turns. The secondary current is 6.64 A and the primary current is 4.15 A. Because of the skin effect at high frequencies, I used multiple thin wires in parallel for each winding: the primary was wound with four AWG22 wires in parallel, and the secondary with ten AWG24 wires in parallel.
Output filter: I selected the output inductor ripple current ΔIL = 1 A. The inductance is:
$$L_o = \frac{(U_{s,\min}-U_{o,\max}-U_F) t_{on,\max}}{\Delta I_L} \approx 71\ \mu\text{H}$$
Using an AL-value of 200 nH/N2, the number of turns is N ≈ 19. For the output capacitor, with a ripple voltage of 10 mV, I obtained:
$$C_o = \frac{\Delta I_L T}{8 \Delta U_o} \approx 111\ \mu\text{F}$$
I used a 120 μF capacitor rated at 110 V, together with a ceramic capacitor.
Power switches: The DC-DC transistor must block 380 V and carry more than the primary peak current. I selected IRF460 MOSFETs (500 V, 20 A) for two switches.
4.5 Feedback Loops
For the PFC voltage loop, the crossover frequency is placed below twice the line frequency to attenuate the second-harmonic ripple. The voltage error amplifier used a type-2 compensation network with component values: Rf = 16.7 kΩ, Cf = 110 nF, CZ = 1.1 μF. The current loop of the PFC must have a high bandwidth to track the reference current. I set the current-sense resistor to 2 kΩ and the feedback network comprised Rf = 27 kΩ, Cp = 100 pF, CZ = 470 nF.
The DC-DC control loop contains two operational amplifiers: one for voltage regulation and one for current regulation. Their outputs are combined with an OR-diodes configuration. When the battery voltage is low, the current loop is active and the circuit delivers a constant current. When the battery reaches the preset voltage, the voltage loop takes over and the circuit provides constant-voltage charging. This dual-loop scheme is very practical for charging a traction battery pack.
4.6 EMC Considerations
Switch-mode converters generate electromagnetic interference (EMI) that must be suppressed. I added an input EMI filter using common-mode inductors and X/Y capacitors. The high-frequency transformer was designed to minimize leakage inductance and was properly shielded. The power MOSFETs have RC snubber circuits to absorb voltage spikes caused by leakage inductance. Printed circuit board layout followed standard EMC practices: short traces, proper grounding, separation of power and control circuits, and minimized loop areas.
4.7 Simulation Results
I simulated the designed converter in PSPICE. The PFC stage was simulated at the rated conditions. The gate drive waveform of the PFC switch was stable at 100 kHz; the inductor current showed a low ripple level, and the output voltage settled at 380 V after the start-up transient. The DC-DC stage simulation showed correct switch drive signals at 200 kHz with a duty cycle near 44%. The DC output voltage was stable at 55 V when the charger was operated in constant-voltage mode. These results verify that the proposed charging circuit can provide the required voltage and current to charge the traction battery pack safely.
5. Protection (Control) Section Design
The protection section must ensure that the traction battery pack remains within its safe operating area. It performs voltage monitoring, current monitoring, residual-capacity estimation, and communication with external devices. For the control core, I selected the C8051F020 microcontroller, a fully integrated mixed-signal system-on-chip. This MCU contains a 12-bit A/D converter with analog multiplexer, on-chip temperature sensor, two 12-bit D/A converters, digital crossbar, multiple serial buses including SMBus/I2C, and 64 KB of flash memory. Its capability and ease of development make it an excellent choice for a BMS.
5.1 Hardware Design
Voltage detection: The total battery bus voltage is scaled down using a resistive divider and filtered before being applied to the ADC. The protection limits were set as follows. The battery pack with 15 cells has a full-charge voltage of 3.6 V per cell, so the overcharge detection threshold is set at 54.5 V ± 0.1 V. Overcharge release occurs at 53.5 V ± 0.1 V. The overcharge detection delay is 2 s to prevent false tripping. The discharge cutoff voltage of the cells is 2.1 V each, so the pack undervoltage threshold is 31.5 V ± 0.3 V. After charging, the release voltage is set to 37.5 V. The undervoltage delay is 500 ms. Short-circuit protection is triggered when the voltage between the battery terminals falls below 5 V, with a very short delay of 10 μs. The nominal charge-discharge conversion voltage is set to 52.5 V for switching from constant-current to constant-voltage charging.
Current detection: Accurate current sensing is essential for both protection and state-of-charge estimation. I considered three methods: resistive shunt, current transformer, and Hall sensor. In high-current circuits, the shunt dissipates considerable power, while current transformers suffer from core saturation in single-ended converters. I therefore selected a magnetic balance hall current sensor. This sensor works on the principle shown: the primary current IP creates a magnetic field that is balanced by a compensating current IS in the secondary coil. The equilibrium condition is:
$$N_P I_P = N_S I_S$$
By measuring the voltage across the sense resistor in the secondary path, the primary current can be deduced. This method offers excellent galvanic isolation, high accuracy, and wide bandwidth. The measured current is compared in the software with preset thresholds. The charge over-current limit is 20 A, and the discharge over-current limit is also 20 A, matching the capability of the traction battery pack. The delay for over-current protection is 50 ms.
Charge termination is determined during constant-voltage charging by monitoring the current. When the current drops below 200 mA, the battery is effectively full. The termination delay is 3 s. A separate fuse is placed in series with the battery for a final safety cutoff if the electronic protection fails.
Internal resistance detection: The internal resistance of a battery can indicate its state of health. I described both DC and AC methods. The DC method applies a large current for a few seconds but can polarize the cell. The AC method uses a small 1 kHz excitation current and derives the resistance from the resulting AC voltage. While more accurate, the DC method can damage the battery; the AC method is safer but may be affected by ripple. For practical BMS implementation, I used a simplified AC approach to infer approximate internal resistance, but I did not rely on it for SOC estimation because of inherent inaccuracies.
Temperature detection: Temperature sensing is required to prevent over-temperature conditions and to provide temperature compensation for SOC estimation. I used an AD590 integrated temperature sensor with a 4–30 V operating range, capable of measuring temperatures from −55 °C to 150 °C with an error of ±0.5 °C. The sensor output is a current proportional to absolute temperature, which is converted to a voltage and read by an ADC channel.
Cell voltage monitoring: To monitor individual cell voltages in a long series string, I used a voltage-controlled current-source circuit that converts the differential cell voltage into a current signal. With proper resistor matching and a suitable load resistor, the output current is proportional to the voltage of the selected cell. The loaded voltage can then be directly applied to the microcontroller’s ADC. The mathematical relation for the circuit is approximately:
$$I_L \approx \frac{U_1-U_2}{R_1} \propto U_{cell}$$
This design is immune to common-mode voltages and allows accurate measurement of each individual cell.
5.2 Software Design
State of Charge Estimation for the Traction Battery Pack
State of charge estimation is the most challenging part of the software. I reviewed several methods: open-circuit voltage (OCV), internal resistance, neural networks, and coulomb counting (ampere-hour integration). OCV-based estimation requires the battery to rest, which is not practical during dynamic operation. Internal resistance methods lack monotonicity over the SOC range and vary with aging. Neural networks have high computational cost and need extensive training data. Coulomb counting is straightforward and suitable for microcontroller implementation, but it accumulates errors and does not account for battery aging, temperature, and discharge rate unless compensation is added.
Thus, I chose an improved ampere-hour integration method with several compensation factors. The fundamental equation is:
$$SOC = SOC_{initial} – \frac{Q_{use}}{Q_E} = \frac{Q_E – \sum_{j} I_j \Delta t}{Q_E}$$
where Quse is the number of ampere-hours consumed, Ij is the discharge current in the j-th time interval, Δt is the sampling interval, and QE is the full charge capacity under standard conditions. However, variations in temperature, discharge current, self-discharge, and battery aging alter the effective capacity. To model these effects, I introduced compensation factors.
Self-discharge compensation: A leakage current term Ig represents energy loss due to self-discharge and internal reactions. The compensated current If is Idischarge + Ig during discharge and Icharge − Ig during charge. Ig is modeled as a linear function:
$$I_g = \eta + \mu (T – T_0)$$
with temperature-dependent coefficients obtained experimentally.
Temperature compensation: for a lithium iron phosphate traction battery pack, the available capacity at low temperatures is significantly reduced. The corrected SOC at temperature T is:
$$SOC_T = SOC \left[ 1 – \nu (T – T_0) \right]$$
where ν is a temperature coefficient that may be piecewise constant over different ranges.
Aging compensation: The available capacity changes as the battery cycles. I defined the aging factor:
$$A_F = \frac{Ah_{ref} – Ah_{pre}}{Ah_{ref}}$$
where Ahref is the maximum capacity over life and Ahpre is the capacity at the current stage of life.
Discharge rate compensation using Peukert’s equation: Peukert’s law relates discharge time and current:
$$I^n t = K$$
Given two discharge tests with different constant currents I1 and I2, the empirical constant n is found from
$$n = \frac{\lg t_1 – \lg t_2}{\lg I_2 – \lg I_1}$$
and K = I1nt1. For a current Ij, the effective capacity QE,j at that current is:
$$Q_{E,j} = K I_j^{1-n}$$
Combining all these compensations, the instantaneous SOC under a variable discharge profile is:
$$SOC = \frac{Q_E – \sum_j \frac{I_j + I_{g,j}}{Q_{E,j}} \Delta t}{Q_E} \left[ 1 – A_F \right] \left[ 1 – \nu (T – T_0) \right]$$
This expression incorporates the action of self-discharge, temperature, aging, and rate-dependent capacity.
Correction using open-circuit voltage: Because the ampere-hour integration method accumulates errors, I introduced a reset procedure. When the traction battery pack is at rest for a specified period, the open-circuit voltage reflects the equilibrium state. The SOC is then recomputed from the OCV-SOC curve and compared with the value from integration. If the error exceeds a preset threshold, the SOC estimate is corrected to match the OCV value. This method provides good long-term accuracy without adding significant computational burden.
SMBus Communication
The designed protection section communicates through the System Management Bus (SMBus). SMBus is a two-wire serial bus based on I2C principles. It provides low-cost, reliable communication between the BMS and external devices such as chargers, displays, or PC adapters. The data rate is 100 kbps, which is sufficient for real-time status monitoring. Only two lines are needed: SDA (serial data) and SCL (serial clock). Both lines are open-drain and require pull-up resistors. The bus operates with start and stop conditions, a 7-bit slave address, direction bit, acknowledgement (ACK), and one or more data bytes. Figure below illustrates the typical SMBus data transfer sequence.
A typical SMBus transmission consists of a START condition, a slave address followed by a read/write bit, an ACK from the recipient, then data bytes with ACK or NACK, and finally a STOP condition. In cases where multiple masters attempt to write simultaneously, arbitration occurs automatically on the SDA line. Because all devices use an open-drain output, they form a wired-AND configuration. The device sending low while another sends high loses arbitration and resets itself. This non-destructive arbitration ensures that no data is lost.
Timing monitoring is also defined: every SMBus device should detect a low SCL period greater than 25 ms as a timeout, and perform communication reset. This robust protocol provides suitable communicational support for the traction battery pack management system. I chose C8051F020 because it integrates an SMBus controller module directly. I configured the SMBus in the master mode to report data to a host every second, but it can also be reprogrammed as a slave for receiving charger commands.
6. Conclusions and Future Work
In this paper, I have investigated the design of a charging and protection circuit for a 48 V, 20 Ah lithium iron phosphate traction battery pack. The charging circuit uses an active PFC stage followed by a parallel single-ended forward DC-DC converter. The PFC controller and the PWM controller are integrated in a single UCC28517 chip. I selected the circuit topology, calculated the important magnetic components, and verified the design by simulation. The simulation results confirm that the charger meets the traction battery pack requirements. The protection circuit is built around the C8051F020 microcontroller. It monitors individual cell voltages, pack voltage, current, and temperature. The protection functions include overcharge, over-discharge, overcurrent, short circuit, and temperature limits. For SOC estimation, I proposed an improved ampere-hour integration algorithm that includes self-discharge, temperature, aging, and discharge-rate compensation factors. The SOC read at rest is corrected using the open-circuit voltage method. An SMBus communication interface allows the BMS to exchange information with external devices. The entire design provides a practical, reliable, and low-cost solution for improving the lifetime and safety of a lithium iron phosphate traction battery pack.
In future work, I plan to optimize the compensation coefficients with more experimental data; improve the thermal design of the power stage; explore soft-switching techniques for higher efficiency; and design a fully automatic equalization circuit for the cells. With gradual improvements in battery materials and electronics, the next generation of traction battery packs will enable electric vehicles to be more competitive, cleaner, and safer.