Design and Implementation of a High Power Dual-module Parallel Charging System

Abstract This paper presents a high-power dual-module parallel charging system which consists of a power converter and digital control software. The power converter consists of an active power factor correction (APFC) circuit and a resonant voltage fed full-bridge converter (RVFFBC). The APFC circuit is used for AC/ DC conversion and power factor correction for improved power quality. The RVFFBC is used to provide stable power for charging the battery by converting the high voltage of the power stage to a lower value. The system software was designed by combining the power control loop into a voltage and current loop to achieve current sharing of the dual-module charging system and to overcome the effect of differences in hardware components. Finally, this paper demonstrates the implementation of a high-power dual-module charging system with a widely varying AC input voltage ranging from 85 ~ 265 Vac. Experimental results verified that the maximum power factor of the system and the Total Harmonic Distortion (THD) were 0.99 and 2.8% respectively. The proposed charging system possesses very high expansibility. The parallel dual-module output DC current of the system is 70 A and the total output power is 4 kW.


Introduction
With the increase in the use of devices such as laptops, mobile phones, tablet PCs, etc., lithium batteries have gradually replaced lead-acid batteries as the power source of choice. Similarly, with the increase in the demand of battery capacity for use in applications such as electric vehicles (EV) and uninterruptible power systems (UPS), the characteristics of the charging systems of these batteries have assumed great importance [1]- [4], with a great deal of work being carried out in this area. The desired features of a good charging system are an improved power factor and low harmonic pollution.
Traditional AC-DC converters consist of a bridge rectifier and a filter capacitor. The advantages of such converters are simplicity of the circuit structure and ease of control. However, they exhibit serious nonlinear distortion, which not only causes the power factor to be reduced, but also affects the device and power network equipment. It also affects the quality of the overall electrical supply system. This leads to poor power grid stability and power loss. Therefore, a power factor correction (PFC) circuit is generally used to improve these problems. The PFC circuit can be divided into two: active power factor correction (APFC) circuits and passive power factor correction (PPFC) circuits [5]- [7]. The correction capability of PPFC circuits is low and they are unable to meet standard requirements. APFC circuits use the Pulse Width Modulation (PWM) scheme, which leads to the input current and voltage waveforms having approximately the same phase. APFC circuits therefore meet the standard requirements and specifications. Current APFC circuits can achieve power factors in the range of 0.98~1, which enhances the power conversion efficiency and power quality of high-power high-current charging systems.
Typical charging systems which provide stable power source is designed by DC / DC converter circuit, it convert a high DC voltage to a low DC voltage in order to match the voltage of the battery pack. Commonly-used step-down DC/DC converter circuits such as buck, buck-boost, etc., suffer from the inherent disadvantages of having large current ripples and low output powers. To overcome these deficiencies, a full bridge converter circuit with a center tap rectifier filter circuit is proposed in this paper. Such a circuit has a highly stable output voltage and current.
In general, to improve the system's expansibility and usability, the charging system consists of several modules to increase the output power [8]- [13], [19]- [22]. The key point of multi-module parallel operation is current-sharing. If the output current is not shared equally among modules, it will result in one module getting easily damaged, which can lead to system shutdown. Poor current sharing and complicated control are the main drawbacks of traditional parallel charging systems. Therefore a dual-module high power charging system in parallel operation that exhibits better current-sharing and control properties is presented in this paper. Fig. 1 shows the proposed single module charging system architecture. It can be divided into three parts. The first is the APFC circuit [14], [15] which is used to improve the power quality. The second is the RVFFBC which is used to provide stable power for charging the power battery. The third is a rectifier filter circuit with a center tap [16]- [18]. The PWM signal is generated by the microcontroller and the D/A converter circuit is used to convert the MCU signals to control the output current of the UC3846 IC. Each part will be illustrated in the following sub-sections.

Active Power Factor Correction (APFC)
A commonly-used PFC circuit architecture includes a buck, buck-boost, boost, SEPIC converters, etc. Table 1 gives a comparison of the features of various converters. To meet the specifications of the proposed single module charging system, the Boost converter was used, as shown in Fig. 2. The IC L4981 was used to control the boost converter to achieve a power factor of about 0.99 with a widely varying AC input voltage (85Vac-265 Vac). The PFC circuit can operate in the continuous conduction mode (CCM) and can be used in charging systems of up to 500 W.

Resonant Voltage Fed Full Bridge Converter (RVFFBC)
The architecture of the RVFFBC is shown in Fig. 3. The resonance is generated by the resonant transformer, the inductor Lr, and the parasitic capacitor of the power switches. Thus, zero voltage switching (ZVS) of the power switches is achieved as shown in Fig. 4. It also greatly improves the overall efficiency and stability of the system and reduces Electro-Magnetic Interference (EMI).  The operation of the RVFFBC is divided into positive and negative half-cycles. As the operation is similar in each half-cycle, only the positive half-cycle is described in this paper. The positive half-cycle operation mode is divided into five intervals t0 -t5 as shown in Fig. 5. The operation first enters the energy clamping stage, and then enters the Mode 1 ~Mode 5 interval operation, it is illustrated as follows.

Energy clamping interval
This interval is shown in Fig. 5(a), which is used to improve the energy imbalance between the resonant inductor and the transformer. Clamping diodes D1 and D2 improve the oscillation generated by the reverse recovery time of the rectifier diode and stabilizes the unstable energy state between the transformer and the resonant inductor Lr.

Mode 1: Energy transfer interval [t0-t1]
The switches Q 1 and Q 3 are turned on as shown in Fig. 5(b). Vin is equal to the sum of the inductor voltage V Lr and the transformer's primary voltage V AB. The input current flows from switch Q 1 to Q 3 and energy is transferred to the transformer and the resonant inductor Lr.

Mode 2: Resonant state interval [t1-t2]
The switch Q 3 is kept in the on-state and Q 1 is turned off, as shown in Fig. 5(c). The current direction is the same as in Mode 1. The parasitic capacitor C Q1 of Q 1 is charged and the parasitic capacitor C Q4 of Q 4 is discharged. Therefore, the voltage across Q 4 is reduced.

Mode 3: Energy hold interval [t2-t3]
The inductor Lr releases energy and the switch Q 3 is kept in the on state, as shown in Fig. (d). After the discharge of C Q4 to turn on D Q4 , the voltage across Q 4 is reduced. Thus, Q 4 is turned on under ZVS.

Mode 4: Resonant state interval [t3-t4]
The switch Q 4 is kept in the on-state, as shown in Fig. 5(e). The resonant effect is generated by the resonant inductor and the parasitic capacitor of C Q3 and C Q2 . The switch Q 2 is turned on under ZVS at this moment.

Mode 5: Energy transfer interval [t4-t5]
The switches Q 2 and Q 4 achieve ZVS under the resonant effect, and the current flow direction is from Q 2 to Q 4 . The positive half cycle operation is over and the device enters the negative half-cycle at this time, as shown in Fig. 5

Rectifier Filter Circuit
Fig . 6 shows the rectifier and filter circuit with the center tap transformer double winding, which provides the positive and negative half-cycles. Therefore, the output current ripple has a frequency which is twice that of the switching frequency of the power switches. This results in a stable DC output voltage and a small current ripple.

Parallel Operation of Dual-module Charging System
The proposed architecture of the dual-module charging system is shown in Fig. 7. To achieve current sharing, the master microcontroller outputs a reference current to adjust the output of the other module. Thus, the parallel operation dual module charging system is implemented by the current sharing scheme, and its control block is shown in Fig. 8. The outer loop is the feedback voltage V bat that regulates the output voltage and the inner loop is the feedback current I out that improves the dynamic response. In addition, the output power loop, which is generated by the product of the feedback voltage and current, is used to establish the common reference power P ref for achieving the same output power of each module. Fig. 8 shows the hardware of the dual-module charging system.  Fig. 9 shows a system program flowchart which contains the following three subroutines: the charge mode program, the protection program, and the constant charge power program. These are explained as follows. Table 2 shows the charging/discharging specification of the battery cell. It can be seen that the charging method of the battery cell is constant voltage-constant current (CV-CC), and the battery cell's maximum charging current and voltage were 3 A and 3.65 V, respectively. The charging voltage of the charging system must be determined according to the battery cell's charging curve, as shown in Fig. 10. Therefore, 16 battery cells were connected in series to constitute a battery pack whose maximum charging voltage was 58.4 V.  The charging procedure consists of three steps which are explained below. Fig. 11 shows a flowchart of the charging mode program. First, it is determined whether the battery voltage is > 29V. If it is, then the battery may be damaged or the wiring may be abnormal. The output power command (Pcmd) must be set to low. If the battery voltage is < 29V, it represents that the battery can be charged properly.

Constant Current: CC
The initial value of the output current is 0, and considering the battery capacity, the output power command (Pcmd) which according to the 1/10 setting value of maximum current (Imax) gradually increase, it can ensure the output current of different battery capacity sets the value of the maximum current to 1/10 of its value to ensure that the output current matches the capacity of a different battery. When the output power command (Pcmd) increases to the maximum current (Imax), the charging system output current (Iref) corresponds to the maximum output current command (Imax). This value is maintained until the battery voltage reaches 58.4 V. Then, the charging mode changes.

Constant Voltage: CV
When the battery reaches the maximum voltage (Vmax), it represents the battery as being under pseudo saturation. In order to charge the battery to saturation, 1/10 of Imax is taken as the charging system shutdown command. It must be ensured that the battery voltage does not exceed Vmax. While the battery current gradually decreases and the battery voltage gradually stabilizes, the charging system will immediately shut down when the charge current is less than 1/10 of Imax.  Fig. 12 shows a flowchart of the protection program. It can be seen that the protection items of the charging system include input voltage, battery voltage, charger temperature, and charge current.

Constant Charge Power Program
Constant power control possesses the function of limiting the maximum value of the system power; its output control is to change the rating of the current regulator. Such a control scheme can achieve a fast response of the DC current regulation loop. In addition, it can also substantially prevent transients and fluctuations generated by the severe variation of the feedback circuit. Fig. 13 shows a flowchart of the constant charge power program which consists of three protection functions. First, consider the security protection. When the instantaneous output exception occurs, that is, when the product of battery voltage Vbat and output current Iout is greater than the constant power, it must be immediately determined whether the battery voltage is > 58.4V or < 29V. If it does, the output power command Pcmd is immediately set to 0. This avoids charging the system by overcurrent caused by a short circuit.
The maximum temperature protection is called the drop power function. We must detect whether the current temperature Tnow reaches the over-temperature; the temperature must reach 70°C before entering the computing equation, and the maximum temperature Tmax is designed to be 80°C; if it does, it means that the temperature of the charging system is higher than 70°C. The charging system is turned off when the temperature reaches 80°C.
Finally, the last function of the constant charge power program is overload protection. To calculate the current battery voltage Vbat and the maximum current Imax, if the output power of the charging system is greater than the maximum power, then maximum power output is maintained. This approach can limit the maximum output power of the charging system and avoid overload and damage.  In this paper, a dual-module charging system using a parallel architecture with a digital microcontroller P89LPC938 is implemented. Its maximum output current is 70 A and the output power is 4 kW. Table 3 shows the system's electrical specifications and component specifications.

Single Module Verification
This section starts by testing the effectiveness of a single-module. Fig. 14(a)-(d) shows the input and output measured waveforms operating under a light load and a full load, respectively. The light load input power was 336.7 W and the output power was 290.1 W. The full load input power was 2096.2 W and the output power was 1904 W. Fig. 15 shows the efficiency curve of the single module under different operating powers. It can be seen that the highest efficiency was 89.7% and the full load efficiency was 89%.  Fig. 16 shows that the total output current under light load was 10 A and that the current error was 0.2 A. Fig. 17 shows that the total output current under full load was 70 A, and that the current error was 0.4 A. Fig. 18 shows the current error comparison of the dual-module under different operating powers. It can be seen that the proposed dual module charging system exhibits good current sharing behavior.

Power Factor and THD Measurement
Fig . 19 shows the input/output voltage and current of a single module. It can be seen that the output voltage and current were 53.7 V and 34.3 A, respectively. Fig. 20 shows the input/output voltage and current of a dual module where it can be seen that the output voltage and current were 54.4 V and 70.3 A, respectively. Table 4 shows that the measured PF and THD values of the single-module were 0.98 and 3.3%, respectively, and that those of the dual module were 0.99 and 2.8%, respectively. In addition, Fig. 21 shows odd harmonics. It can be seen that the proposed charging system meets the requirements of the IEC61000-3-2 standard.

Charging Curve Verification
The charging curve of a battery cell is shown in Fig. 10. The first stage is CC charging mode in which the charging system charges the battery pack with a maximum charge current of Idc ＝ 35 A. When the battery voltage was gradually increased to 58.4 V, the charging system switched to the second stage which is the CV charging mode. When the charging current gradually decreased, the battery voltage maintained a maximum voltage of 58.4 V. When the charge current dropped to zero, the charging system was shut down. Fig. 22 shows the measured charging curve of a single module. It can be seen that when the battery voltage reached the maximum voltage of 58.4 V, the system switched to CV charging mode and the output voltage was maintained at 58.4 V until the charging current was less than or equal to the set stopping point and the charging system stopped the output. The charging curve of the dual module can be similarly obtained. It is shown in Fig. 23.

Conclusions
In this paper, a dual module parallel charging system is presented, which can be used for AC voltages varying from 85 V to 265 V. The maximum PF was 0.99, the THD was 2.8%, and the specifications are consistent with the requirements of the IEC61000-3-2 standard. The RVFFBC significantly improves the overall efficiency, system stability, and EMI. The dual module charging system can achieve a maximum output current of 70 A, and its current error is less than 2%. The proposed charging system has a high expansibility and utility and is suited for the requirements of high power charging.