Design and Experimental Results of Battery Charging System for Microgrid System

Nowadays, many countries have paid attention to renewable energy due to fossil fuel crisis and its related environmental pollution. In particular, following the government supply business for renewable energy industry, the private sectors drive the stable power supply by using renewable sources for both microgrid system and standalone application. Battery charging and discharging control system of microgrid system are critical to extend lifetime of standalone photovoltaic system. Corresponding to this demand, this paper presents the development of battery charging and discharging system based on battery modeling, SOC (state of charge) estimation, and its implementation for 5 kW. As a result, the conversion efficiency shows 96.35% with over 95% charging performance.


Introduction
A microgrid system is defined as an autonomous energy grid system to operate in parallel with or independently of the traditional grid [1,2].In accordance with the growing industry trend on renewable energy, microgrid systems integrated with renewable energy resources are becoming widespread in the world [3][4][5].This type of microgrid system consisted of several renewable energy resources like photovoltaic generation and a battery backup system like lead-acid battery system [6][7][8].In particular, battery system is the key component to stabilize the microgrid system as a concrete grid voltage function.Typically, lead-acid battery is mostly used for storage system, but it has relative short lifetime for a stable operation.Thus, the battery charging and discharging system are regarded as one of the most essential systems for the long lifetime operation [9][10][11][12].The battery systems for both microgrid system and standalone PV generation are different from the conventional system for UPS (uninterruptable power supply), because the charging and discharging profiles are depending on the environmental condition like irradiance.Corresponding to this demand, this paper presents the development of battery charging and discharging system based on battery modeling, SOC (state of charge) estimation, and its implementation.
This paper consisted of three sections.Firstly, the modelling of battery and battery charger/discharger is presented.Secondly, based on the modeling, the simulation of the target system is conducted with topology design for 5 kW.Lastly, with the proposed control algorithm, the final system output is implemented to verify the proposed algorithm and hardware.

Battery Modeling
It is required to have appropriate battery modeling to implement a battery charging and discharging system.Typically, this modeling can be classified into three categories.One is an electrochemical modeling, and it is too complicated to apply even though it shows the highest accuracy [13].The second one is a mathematical modeling and it is acceptable to apply under some limited condition [14].The third one is based on the electrical modeling by using voltage source, resistor, and capacity and it is practical for battery charging system implementation.In particular, this paper uses the electrical modeling for battery as in Figure 1.In order to model battery electrical characteristic, it is assumed that SOC function is used to improve the modeling accuracy as the following equation: where Ah rate is the battery rating and Ah init is the initial battery charging state.

System Configuration
Battery system is installed to regulate the DC bus among PV generation, wind turbine, and local loads in Figure 2. Specifically, the battery charger is implemented by using 5 kW bidirectional buck-boost converter.As shown in Figure 3, the operating mode between charging mode and discharging mode is determined by monitoring DC bus voltage and current.In other words, when it is the charging mode, the charger is operated as buck converter.While it is the other mode, the charger is operated as the boost converter.

Buck Converter for Charging Mode.
Buck converter is called a step-down converter, at which the input voltage magnitude V 1 is decreased from the output voltage V 2 by the on/off control of switch  1 (Figure 4(a)).When the switch  1 is on during , where the switching duty ratio is  and the switching period is designated as , the inductor voltage V  is determined by the following equation.During the short switching period , both V 1 and V 2 are assumed as constant values.
When the switch  1 is off during (1 − ), the inductor voltage V  is determined by the following equation.
Under the steady state condition for the continuous current control, the net inductor current change during one switching cycle is equal to zero as following equation.
Thus, the output voltage V 2 is equal to V 1 .
Since the average capacitor ( 2 ) current is equal to zero at the steady state, the average inductor current should be equal to the load current. is representing the load resistance for the 5 kW battery charger converter.Since the nominal battery bank voltage is 240 V, the nominal load resistance can be calculated by the power definition.This is because the load current through  is limited by the controller.Thus,  is used for the converter design.Thus, The maximum inductor current and the minimum inductor current can be calculated from the inductor average current and the net inductor current change.
For the continuous current mode of inductor current, the minimum inductor current  ,min should be more than zero.Thus, the minimum inductance can be calculated from the fixed switching frequency .
From the designed switching 12 kHz and 5 kW converter power capacity, the minimum inductance  min can be calculated by 877 uH.
For the capacitance design, the capacitance  2 can be calculated by the designed voltage ripple ratio.
When the current variation is 40% and the ripple voltage ΔV 2 is designed as 1 V, the designed capacitance  2 can be designed as 87.5 uF.

Boost Converter for Discharging Mode.
Boost converter is called a step-up converter, at which the input voltage magnitude V 2 is decreased from the output voltage V 1 by the on/off control of switch  2 (Figure 4(b)).When the switch  2 is on during , where the switching duty ratio is  and the switching period is designated as , the inductor voltage V  is determined by the following equation.
When the switch  1 is off during (1 − ), the inductor voltage V  is determined by the following equation.
Under the steady state condition for the continuous current control, the net inductor current change during one switching cycle is equal to zero as following equation.
Since the average capacitor ( 1 ) current is equal to zero at the steady state, the average inductor current should be equal to the load current.Thus, The maximum inductor current and the minimum inductor current can be calculated from the inductor average current and the net inductor current change.
For the continuous current mode of inductor current, the minimum inductor current  ,min should be more than zero.Thus, the minimum inductance  can be calculated from the fixed switching frequency .
From the designed switching 12 kHz and 5 kW converter power capacity, the minimum inductance  min can be calculated by 884 uH.
For the capacitance design, the capacitance  2 can be calculated by the designed voltage ripple ratio.
When the ripple voltage ΔV 1 is designed as 1 V, the designed capacitance  1 from (8) can be designed as 400 uF.

Experimental Results
The proposed battery charging system was implemented by using buck converter as designed in the previous section.The electrical specification is listed in Table 1. Figure 5 shows the implemented system front-view.
Figure 6 shows the experimental waveforms for both CC and CV modes.DC link voltage is implemented by DC power supply and batteries are stacked to generate 240 V. Buck-boost converter is implemented for battery charging operation.The charging current is designed as 3 A, while the maximum charging current is limited as 10 A, and the temperature is controlled to be less than 50 ∘ C. When the battery voltage reaches the nominal battery voltage, the charging mode   moves from constant current (CC) mode to constant voltage (CV) mode as shown in Figure 6.As a result, SOC as a charging performance index, is estimated over 95% with 96.35 conversion efficiency.

Conclusion
This paper presents the development of battery charging and discharging system based on battery modeling, SOC estimation, and its implementation.As a result, the conversion efficiency shows 96.35% with over 95% charging performance.Based on the presented design and experimental results for battery charging system, the battery backup system can be expected to be of high efficiency with long lifetime.

Figure 4 :
Figure 4: Battery charger operational circuit diagram.(a) Buck converter for charging mode; (b) boost converter for discharging mode.