Energy Storage System with Voltage Equalization Strategy for Wind Energy Conversion

In this paper, an energy storage system with voltage equalization strategy for wind energy conversion is presented. The proposed energy storage system provides a voltage equalization strategy for series-connected lead-acid batteries to increase their total storage capacity and lifecycle. In order to draw the maximum power from the wind energy, a perturbation-and-observation method and digital signal processor (DSP) are incorporated to implement maximum power point tracking (MPPT) algorithm and power regulating scheme. In the proposed energy storage system, all power switches have zero-voltage-switching (ZVS) feature at turn-on transition. Therefore, the conversion efficiency can be increased. Finally, a prototype energy storage system for wind energy conversion is built and implemented. Experimental results have verified the performance and feasibility of the proposed energy storage system for wind energy conversion.


Introduction
A large amount of carbon dioxide produced by combustion of fossil fuels are believed to be responsible for trapping heat in the Earth's atmosphere, resulting in serious greenhouse effects and environmental pollution.This effect is being linked to changes in the Earth's climate.To reduce emissions of carbon dioxide, the demand for renewable energy sources has increased significantly.The typical renewable energy sources include solar, wind and geothermal energies, which have the features of cleanliness and freedom [1], therefore developing renewable and clean energy sources to substitute for fossil fuels has been an important topic.Currently, wind is one of most widely utilized renewable OPEN ACCESS energies.Wind turbine technology has been undergoing a dramatic development and now is the world's fastest growing energy source [2][3][4][5][6].However, due to the instability and intermittent characteristics of wind energy, it cannot provide a constant or stable power output.Thus, maximum power point tracking (MPPT) and energy storage are very important technologies.Many MPPT methods have been developed, mostly dominated by a simple perturbation and observation method [7][8][9][10].The method requires the measurement of only a few parameters thus facilitating an MPPT control.As a result, it is often applied to the wind turbine energy for enhancing power capacity.
Batteries are essential elements for wind energy storage applications.Lead-acid batteries connected in series and parallel to supply higher voltage and capacity are utilized in many practical applications, because large amounts of energy can be stored at low cost.However, the differences between battery charge and discharge rates, residual capacities and internal resistance voltage usually cause imbalances in series-connected lead-acid batteries, resulting in a decrease of the total storage capacity and battery lifecycle.Thus, voltage equalization charging techniques for series-connected batteries are becoming important.In general, there are several charging techniques to overcome the voltage imbalances in series-connected batteries.The evolution of charging strategies is as follows: (1) Dissipative shunting resistor, (2) Multi-winding transformer, and (3) Multiple transformers.
The dissipative shunting resistor is an effective voltage equalization method because of its reliability and simplicity [11].Figure 1 shows a voltage equalization circuit with dissipative resistor.However, the circuit has disadvantages that the dissipative resistor method has poor efficiency and needs many MOSFETs, resulting in complex control.In the multi-winding transformer equalization circuit, a shared transformer has a single magnetic core with secondary taps for each battery, as shown in Figure 2. The multi-winding transformer must be customized according to the number of batteries, which results in this circuit complex and high cost [12].Figure 3 shows the multiple transformers equalization circuit.In this circuit, several transformers can be used with the same result by coupling the primary windings instead of coupling via a single magnetic core.Compared to the multi-winding transformer scheme, this method is better for modular design, but it is still expensive.In this paper, an energy storage system with voltage equalization strategy for wind energy conversion is presented, as shown in Figure 4.It is composed of a phase-shift full-bridge converter, a current-doubler rectifier and a non-dissipative voltage equalization circuit.The phase-shift full-bridge converter is incorporated with perturbation and observation method to draw the maximum power from the wind turbine energy.In addition, by phase-shift PWM approach, all power switches of full-bridge converter can be operated with a ZVS feature at turn-on transition [13].In order to reduce charging current ripples of series-connected batteries, the current-doubler rectifier with an interleaved method is implemented.Finally, the non-dissipative voltage equalization circuit is in cascade connection to achieve voltage equalization of series-connected batteries [14,15].
The operational principles of the proposed energy storage system are described in Section 2. The MPPT wind energy and control scheme algorithm is decribed in Section 3. Experimental results obtained from a 600 W prototype with the proposed energy storage system for wind energy conversion are presented in Section 4. Finally, a conclusion is given in Section 5.

Operational Principles
For convenience of illustration and analysis, the proposed energy storage system shown in Figure 4 is divided into two operational modes.One is charging mode, the other is voltage equalization mode, as shown in Figure 5. Based on the above assumptions, the operational principle of the proposed energy storage system for charging mode and voltage equalization mode can be explained stage by stage as follows:

Operational Principles of Charging Mode
In the proposed energy storage system, the operation principles of the charging mode over one switching period can be divided into eight stages.Figure 6 shows conceptual voltage and current waveforms relative to key components of the proposed energy storage system at charging mode.Figure 7 shows its equivalent circuits.

Stage 1
[Figure 7(a), t 0 ≤ t < t 1 ]: At time t 0 , active switches M 1 and M 4 are conducted, the primary current i pri flows through the path M 1 -L r -T r -M 4 .At secondary side of transformer, a positive voltage V sec crosses the secondary winding of transformer T r .The rectifier diode D r1 is reversely biased and D r2 is conducting.During this interval, inductor current i L1 flowing through the path B 1 -B 2 -D r2 -Ns-L 1 is linearly increased, and inductor current i L2 flowing through the path At time t 2 , the current i pri forces the body diode D 3 of M 3 conducting and creating a ZVS feature for M 3 .At secondary side of transformer, the rectifier diodes D r1 and D r2 are forward biased.Inductor currents i L1 and i L2 flow through the rectifier diodes D r1 and D r2 , respectively.

Stage 4
[Figure 7(d), t 3 ≤ t < t 4 ]: At time t 3 , the active switch M 3 is turned on under ZVS transition.While the active switch M 4 , rectifier diodes D r1 and D r2 are maintained conducting.During this interval, the inductor current i L1 and i L2 continue decreasing linearly.

Stage 5
[Figure 7(e), t 4 ≤ t < t 5 ]: At time t 4 , the active switch M 4 is turned off and M 3 is maintained conducting.The parasitic capacitor C 4 of the M 4 is charged to V in , while parasitic capacitor C 2 of the M 2 is discharged down to zero.

Stage 6
[Figure 7(f), t 5 ≤ t < t 6 ]: At time t 5 , the current i pri forces the body diode D 2 of M 2 conducting and creating a ZVS feature for M 2 .During this interval, the inductor currents i L1 and i L2 continue decreasing linearly.

Stage 7
[Figure 7(g), t 6 ≤ t < t 7 ]: At time t 6 , the active switch M 2 is turned on under ZVS transition and the active switch M 3 is maintained conducting.When the energy of resonant inductor L r is released to zero, the primary current i pri begins reversely flowing through the path M 2 -T r -L r -M 3 .At secondary side of transformer, a negative voltage V sec crosses the secondary winding of transformer T r .The rectifier diode D r2 is reversely biased and D r1 is conducting.During this interval, inductor current i L2 flowing through the path B 1 -B 2 -D r1 -N s -L 2 is linearly increased, and inductor current i L1 flowing through the path B 1 -B 2 -D r1 -L 1 is linearly decreased.

Stage 8
[Figure 7(h), t 7 ≤ t < t 8 ]: At time t 7 , the active switch M 3 is turned off and M 2 is maintained conducting.The parasitic capacitor C 3 of the M 3 is charged to V in , while parasitic capacitor C 1 of the M 1 is discharged down to zero.

Stage 9
[Figure 7(i), t 8 ≤ t < t 9 ]: At time t 8 , the current i pri forces the body diode D 1 of M 1 conducting and creating a ZVS feature for M 1 .At secondary side of transformer, the rectifier diodes D r1 and D r2 are forward biased.Inductor currents i L1 and i L2 flow through the rectifier diodes D r1 and D r2 , respectively.
Stage 10 [Figure 7(j), t 9 ≤ t < t 10 ]: At time t 9 , the active switch M 1 is turned on under ZVS transition.While the active switch M 2 , rectifier diodes D r1 and D r2 are maintained conducting.During this interval, the inductor currents i L1 and i L2 continue decreasing linearly.
Stage 11 [Figure 7(k), t 10 ≤ t < t 11 ]: At time t 10 , the active switch M 2 is turned off and M 1 is maintained conducting.The parasitic capacitor C 2 of the M 2 is charged to V in , while parasitic capacitor C 4 of the M 4 is discharged down to zero.

Operational Principles of Voltage Equalization Mode
In the proposed energy storage system, operation principles of the voltage equalization mode over one switching period can be divided into eight stages.Figure 8 shows conceptual voltage and current waveforms relative to key components of the proposed energy storage system at volatge equalization mode.Figure 9 shows its equivalent circuits.

Stage1
[Figure 9(a), t 0 ≤ t < t 1 ]: At time t 0 , active switches M 5 is conducted and M 6 is turned off.While the battery B 1 begins discharging, the inductor current i LB flowing through the path At time t 1 , the active switch M 5 is turned off and M 6 is maintained off.The parasitic capacitor C 5 of the M 5 is charged to V B1 + V B2 , while the parasitic capacitor C 6 of M 6 is discharged down to zero.

Stage 3
[Figure 9(c), t 2 ≤ t < t 3 ]: At time t 2 , the The inductor current i LB forces the body diode D 6 of M 6 conducting and creating a ZVS feature for M 6 .During this interval, the inductor current flowing through the path B 2 -D 6 -L B is linearly decreased.In the stage, the battery B 2 is charged in a short time to create a voltage equalization feature.

Stage 4
[Figure 9(d), t 3 ≤ t < t 4 ]: At time t 3 , the active switch M 6 is turned on under ZVS transition.During this interval, the inductor current i LB continues decreasing linearly.

Stage 5
[Figure 9(e), t 4 ≤ t < t 5 ]: When the inductor current i LB is decreased to zero at time t 4 .The battery B 2 begins discharging and inductor current i LB flowing through the path B 2 -L r -M 6 is linearly increased.

Stage 6
[Figure 9(f), t 5 ≤ t < t 6 ]: At time t 5 , the active switch M 6 is turned off and M 5 is maintained off.The parasitic capacitor C 6 of the M 6 is charged to V B1 + V B2 , while parasitic capacitor C 5 of the M 5 is discharged down to zero.

Stage 7
[Figure 9(g), t 6 ≤ t < t 7 ]: The inductor current i LB forces the body diode D 5 of M 5 conducting and creating a ZVS feature for M 5 .During this interval, the inductor current i LB flowing through the path B 2 -D 6 -L B is linearly decreased.
Stage 8 [Figure 9(h), t 7 ≤ t < t 8 ]: At time t 7 , the active switch M 5 is turned on under ZVS transition.During this interval, the inductor current i LB continues decreasing linearly.When the inductor current i LB is decreased to zero at time t 8 , the voltage equalization operation of battery charger over one switching cycle is completed.

Control Scheme
In order to achieve the best energy transfer and control scheme, an MPPT control algorithm and a voltage equalization strategy must be integrated into the proposed energy storage system.In this section, an MPPT algorithm that uses the perturbation and observation method to achieve the best energy utilization of the wind turbine energy and a voltage equalization method will be explained in detail.Furthermore, to achieve an optimal stability and safety for the proposed energy storage system, the functions of under-voltage, over-voltage, over-current, and over-temperature protection circuits are also illustrated.

Wind Turbine MPPT Algorithm
For experimental convenience, instead of natural wind, a controllable dc motor is used to drive the wind turbine to simulate the actual operating situation under a natural air-stream.Figure 10 shows the emulator simulation system for the wind energy generator.In Figure 10, the dc motor is controlled by different values of dc voltage V dc,motor to provide different limited maximum driving power for driving the wind turbine.When the output power of the wind turbine is small, the dc motor will request small power from the dc voltage V dc,motor to drive the wind turbine.When the speed of the wind turbine increases with generated power, its output voltage along with the rotational speed will be decreased so that the dc motor will generate a larger torque to drive it.Since the limited maximum driving power of the dc motor is predetermined by the dc motor controller, the dc motor can only provide a limited maximum torque corresponding to the predetermined limited maximum driving power to drive the wind turbine.As a result, the wind turbine can only generate a limited maximum power for the load.Figure 11 shows the typical output power characteristic curves of the wind turbine under different output voltages.The output power characteristic curves imply that the wind turbine will generate different maximum output powers for different wind speeds.In order to draw the maximum power from the wind energy, the MPPT algorithm adopts the perturbation and observation method.Figure 12 illustrates the flow chart of the MPPT algorithm for the wind energy generator.First of all, the terminal voltage V wind and current I wind of wind turbine are measured.The output power of wind turbine P wind can be obtained from the product of V wind and I wind .The power operating point location can be determined by a perturbation in the wind turbine output power, as shown in Figure 11.The parameters n − 1 and n of Figure 12 indicate the measured quantities before and after the perturbation, respectively.

Control Strategy of Voltage Equalization
Figure 13 shows the proposed non-dissipative charging and voltage equalization strategies.It consists of three operational modes: (1) the charging mode; (2) the inactive rest mode; and (3) the voltage equalization mode.During the active-charge mode, a phase-shift full-bridge converter with continued charging constant currents is applied.Thus, a rest period is used to provide a relaxed time of the series-connected batteries.Finally, in the voltage equalization mode, a series of sharp triangular currents is applied to achieve the voltage equalization and reduce internal temperature of the series-connected batteries.

Protection Circuits
To achieve an optimal stability and safety for the proposed energy storage system, the functions of under-voltage, over-voltage, over-current, and over-temperature protection circuits are required.All of the protection signals are also realized on the TMS320F240 microcontroller.The conceptual control block diagram of the proposed energy storage system for wind energy conversion is shown in Figure 14.In Figure 14, the control circuits of the proposed energy storage system are implemented on microcontroller along with auxiliary analog circuits.The front-end is used phase-shift full-bridge converter by microcontroller to draw maximum power from the wind turbine, while the rear-end is used non-dissipative voltage equalization circuit to control the voltage equalization of the batteries.The output voltages and currents of the wind turbine are sensed, and sent to the microcontroller where the MPPT algorithm will determine the reference currents Iwind,ref and V wind,ref corresponding to the maximum power of wind turbine at that moment for wind-turbine output current Iin and V in to keep up with.By comparing these signals, the PWM comparator will generate desired gate driving signals of active switches (M 1 , M 2 , M 3 , and M 4 ) realize the MPPT functions.

Condition of ZVS
In order to reduce the switching losses of the active switches (M 1 , M 2 , M 3 and M 4 ), the phase-shift full-bridge converter is necessary to store enough energy in the resonant inductor L r to achieve ZVS at switches turn-on transition.The ZVS condition for four active switches can be obtained from the following inequality: When the value of the parasitic capacitors is equal ( , the inequality can be expressed as follows: If the value of active switches (M 1 ~M4 ) is selected so that a proper resonant inductor L r can be obtained.In inequality (2), for determining resonant inductor L r , we can plot the curves showing the relationship between resonant inductor L r and input currents i pri under different input voltages, as illustrated in Figure 15.The resonant inductor should be selected from the gray area for achieving ZVS.Similarly, the non-dissipative voltage equalization circuit is also necessary to store enough energy in the inductor L B to achieve ZVS at switches turn-on transition.Usually, the function of inductor L B is designed to achieve the voltage equalization of the series batteries.The relationship of the inductor L B and battery voltage can be obtained as following equation: where D and f s are the duty ratio and switching frequency of the active switches (M 5 and M 6 ).From (3), the proper value of the inductor L B can be determined to achieve the voltage equalization of the series Figure 17 shows measured output currents, voltages and their corresponding powers from start-up to the steady state for wind turbine energy with MPPT algorithm.From Figure 17, it can be observed that the perturbation and observation method for the wind turbine energy can be always obtained from maximum power point.Figure 18 shows measured primary current and voltage waveforms of transformer.Figure 19 shows measured current waveforms of inductors L 1 and L 2 , from which it can be seen that the inductor currents i L1 and i L2 have an interleaved feature.Thus, the ripples of output current can be reduced.Figure 20 shows power switches M 1 and M 4 to illustrate a ZVS feature, respectively.Figure 21 shows measured voltage and current waveforms of inductors L B for voltage equalization of series-connected batteries.Figure 22 shows measured voltage and current waveforms of power switch (M 5 or M 6 ), from which it can be seen that the switch (M 5 or M 6 ) has a ZVS feature at turn-on transition.Figure 23 shows the comparison between voltage equalization measurements of series-connected batteries B 1 and B 2 under the condition of voltage imbalance (V B1 > V B2 and V B1 < V B2 ), from which it can be seen that the series-connected batteries can be improved in a short time by the proposed voltage equalization strategy.

Conclusions
In this paper, an energy storage system with voltage equalization strategy for wind energy conversion is proposed.All power switches of the proposed energy storage system have a ZVS feature at turn-on transition.Hence, the switching losses of the power switches can be reduced.In order to draw maximum power from the wind turbine energy, a simple perturbation and observation method is incorporated to realize maximum power conversion.To adopt a cost effective of the proposed energy storage system, the MPPT algorithms and protected circuits consist of a digital signal processor (DSP) and analog circuits to implement maximum power conversion and protect the system.Thus, the control circuit of the proposed energy storage system is compact and programmable.Experimental results have verified that the proposed energy storage system is suitable for wind energy conversion.

Figure 4 .
Figure 4. Circuit structure of the proposed energy storage system for wind energy conversion.

Figure 5 .( 1 )
Figure 5. Two equivalent circuits of the proposed energy storage system: (a) charging mode; (b) voltage equalization mode.

Figure 6 .
Figure 6.Driving signals of switches (M 1 ~M4 ) and conceptual current and voltage waveforms of key components for charging mode.

Stage 12 [
Figure 7(l), t 11 ≤ t < t 12 ]: At time t 11 , the current i pri forces the body diode D 4 of M 4 conducting and creating a ZVS feature for M 4 .During this interval, the inductor currents i L1 and i L2 continue decreasing linearly.When active switch M 4 starts conducting again at the end of Stage 12, the charging operation of energy storage system over one switching cycle is completed.

Figure 7 .
Figure 7. Equivalent circuits of operating stages for charging mode.

Figure 8 .
Figure 8. Driving signals of switches (M 5 and M 6 ) and conceptual current and voltage waveforms of key components for voltage equalization mode.

Figure 9 .
Figure 9. Equivalent circuits of operating stages for voltage equalization mode.

Figure 10 .
Figure 10.The emulator simulation system for the wind energy generator.

Figure 11 .
Figure 11.Typical output power characteristic curves of the wind turbine under different driving power from the dc motor.

Figure 12 .
Figure 12.Flow chart of perturbation and observation method for MPPT algorithm.

Figure 13 .
Figure 13.The proposed non-dissipative charging and voltage equalization strategies.

Figure 14 .
Figure 14.Conceptual control block diagram of the proposed energy storage system for wind energy conversion.

Figure 15 .
Figure 15.Plots of the ZVS region relating to resonant inductor L r and input current ipriunder different input voltage.