energies Article
Improvement of Transient State Response through Feedforward Compensation Method of AC/DC Power Conversion System (PCS) Based on Space Vector Pulse Width Modulation (SVPWM) Seok-Jin Hong 1 ID , Seung-Wook Hyun 1 , Kyung-Min Kang 1 , Jung-Hyo Lee 2 and Chung-Yuen Won 1, * 1 2
*
Information and Communication Engineering, Sungkyunkwan University, Suwon 16419, Korea;
[email protected] (S.-J.H.);
[email protected] (S.-W.H.);
[email protected] (K.-M.K.) Department of Electrical Engineering, Kunsan National University, Gunsan, Jeollabuk-do 54150, Korea;
[email protected] Correspondence:
[email protected]; Tel.: +82-31-290-7169
Received: 30 April 2018; Accepted: 3 June 2018; Published: 6 June 2018
Abstract: In a DC distribution system configured by AC/DC power conversion system (PCS), the voltage control performance of the AC/DC PCS determines the stability and reliability of the DC distribution grid. The DC voltage of grid is maintained by capacitor, thus transient voltage is an inevitable problem when a grid is connected with a high amount of load or renewable energy. Space vector pulse width modulation (SVPWM) is well known as a stable modulation method and is used in AC/DC PCS and many types of topologies, but a solution for the transient states issue of DC link has not clearly been studied. In this paper, a feedforward compensation method based on the mathematical model of SVPWM is proposed to solve the transient state problem in a DC distribution system. The proposed method is verified by simulation and experiment. AC/DC PCS with the proposed feedforward compensation method has more robust DC voltage control characteristics. Keywords: feedforward compensation; space vector pulse width modulation (SVPWM); robust DC grid; DC distribution; transient state response; improved response
1. Introduction As the fields of power semiconductor technology and power converter technology have advanced and interest in environmental pollution has increased, a distributed power source utilizing new and renewable energy has been attracting attention, and DC distribution has been actively studied for this purpose. DC distribution can be effectively connected to renewable energy sources and it can form an efficient power grid [1–3]. Therefore, it is being applied in many applications, such as smart grids, internet data center (IDC), commercial buildings, and etc. [4–6]. The DC distribution is configured as shown in Figure 1. Loads connected to the DC distribution are mostly nonlinear electronic loads such as lighting, computers, energy storage systems (ESS), and variable speed driving inverters for air conditioners, and these loads are sensitive to voltage fluctuations in the DC distribution [7–9]. Because AC/DC power conversion system (PCS) are essential in maintaining the existing AC system and constructing a separate DC distribution, research into how to apply various topologies in various fields has already been conducted [10,11]. The three-phase AC/DC PCS is particularly advantageous in that the power factor is controllable and that the power flow is bidirectional when compared to a diode rectifier. Therefore, AC/DC PCS is adopted in applications where less distortion is required in the AC current waveform in order to meet the stringent regulations for the electrical Energies 2018, 11, 1468; doi:10.3390/en11061468
www.mdpi.com/journal/energies
Energies 2018, 11, x Energies 2018, 11, 1468
2 of 16 2 of 16
is required in the AC current waveform in order to meet the stringent regulations for the electrical harmonic not only onlythe themagnitude magnitudeofofthe the harmonicquality qualityofofthe theAC ACsystem. system. Since Since the the AC/DC AC/DC PCS PCS can can control control not current, through the thephase phaseinformation informationofofthe the current,but butalso alsothe thephase, phase,the thepower power factor factor can can be controlled through ACsystem. system.Moreover, Moreover,the theDC DC link link voltage voltage can be controlled AC controlledby bycontrolling controllingthe theinput inputpower power[12,13]. [12,13].
AC Grid
+750[V] DC-Bus
DC AC
-750[V] DC-Bus
DC
DC DC
PC
DC DC
Office Machine Load
DC DC
EV
DC
DC AC
DC
PV-Module
Wind Power
DC DC
ESS
Renewable Energy
Figure 1. Configuration of DC distribution. Figure 1. Configuration of DC distribution.
Generally, space vector pulse width modulation (SVPWM) is applied to generate a switching vector pulse width modulation (SVPWM) is applied generatethe a switching signal Generally, in AC/DCspace PCS applying voltage oriented control (VOC) [14–16]. In to SVPWM, maximum signal in AC/DC PCS applying voltage oriented control (VOC) [14–16]. In SVPWM, the maximum magnitude of the voltage that each vector can output varies in proportion to the voltage of the DC magnitude of the voltage that each vector can output varies in proportion to the voltage of the DC link (DC distribution voltage in this case). Therefore, if the DC link voltage fluctuates, the magnitude link (DC distribution voltage in this case). Therefore, if the DC link voltage fluctuates, the magnitude of the output voltage fluctuates as well. This is particularly true in a DC distribution system that is of the output voltage fluctuates as well. This is particularly true in a DC distribution system that is constructed using AC/DC PCS where the load is centralized, and power is generated/stopped from constructed using AC/DC PCS where the load is centralized, and power is generated/stopped from various renewable energy sources or nonlinear load is frequently connected/disconnected. This also various renewable energy sources or nonlinear load is frequently connected/disconnected. This also causes a transient state of the DC distribution voltage, which refers to a fluctuation of the DC link causes a transient state of the DC distribution voltage, which refers to a fluctuation of the DC link voltage of the AC/DC converter. Also, DC-link voltage can be changed by AC voltage and current voltage of the AC/DC converter. Also, DC-link voltage can be changed by AC voltage and current dips [17–21]. A robust DC voltage control characteristic of AC/DC PCS is required to supply power dips [17–21]. A robust DC voltage control characteristic of AC/DC PCS is required to supply power to tothe thesensitive sensitive load (to voltage fluctuation) that is connected to the DC distribution in a stable load (to voltage fluctuation) that is connected to the DC distribution in a stable manner. manner. There have been several studies regarding the robust DC voltage control characteristics [22–24]. There have been several studies regarding the method robust DC voltage control In regards to robust DC distribution voltage, the of applying Droopcharacteristics Control to all [22–24]. power Inconverters regards toconnected robust DC distribution voltage, the method of applying Droop Control to all power to a DC distribution was studied in [25–28]. However, voltage ripple according converters connected to a DC distribution was studied in [25–28]. However, voltage ripple according to the droop curve always appears in voltage control of a DC distribution using droop control, and tostrict the droop curveisalways appears in voltage of a DC control, and regulation necessary for applying the control droop curve for distribution every deviceusing that isdroop connected to the strict regulation is necessary for applying the droop curve for every device that is connected to the DC distribution. Therefore, a study was conducted to improve the robustness of droop control using DC distribution. study was conducted tocontrol improve themust robustness of droop model predictiveTherefore, control in a[29]. However, a complex loop be constructed to control improveusing the model predictive control in [29]. However, a complex control loop must be constructed to improve control performance in order to implement the control method proposed in [29]. It is also impossible the control and performance order implement theDC control in power [29]. Itand is the also to predict model all ofinthe loadstoconnected to the grid inmethod terms ofproposed maintaining impossible to predict and model all of the loads connected to the DC grid in terms of maintaining supply of power. power and the supply of power.
Energies 2018, 11, 1468
3 of 16
Energies 2018, 11, x
3 of 16
Therefore, the robustness robustnessand anddynamics dynamicscharacteristic characteristic Therefore,the themost mostreliable reliable way way to to increase increase the of of DCDC distribution is to improve the performance of AC/DC PCS and feedforward compensated to the output distribution is to improve the performance of AC/DC PCS and feedforward compensated to the through valuethe that wasthat obtained throughthrough modeling of the AC/DC PCS system [30]. [30]. output the through value was obtained modeling of the AC/DC PCS system Meanwhile, a control method for voltage stabilization of the DC link voltage is studied [19–21] Meanwhile, a control method for stabilization of the DC link voltage is studied in in [19–21] when voltage dips occur on the AC side voltage. The control method of improving control performance when voltage dips occur on the AC side voltage. The control method of improving control byperformance adding a feedforward to the d-q term axis current controller was applied in [20,21]. by adding compensation a feedforward term compensation to the d-q axis current controller was Inapplied this method, however, AC that however, was connected towas PCSconnected cannot control Power Factor the AC in [20,21]. In this method, AC that to PCSthe cannot control theofPower Factor of the AC side clearly. side clearly. thispaper, paper,aafeed feedforward forward compensation compensation method control InInthis methodisisproposed proposedfor forrobust robustDC DCvoltage voltage control characteristicofofAC/DC AC/DC PCS configuring configuring the applying thethe characteristic the DC DC distribution distributionnetwork. network.The Thesystem system applying proposed compensation method improved control performance not only in the transient state proposed compensation method improved control performance not only in the transient state occurring in the DC grid, in the AC voltage dip. The configuration the DC distribution is inoccurring the DC grid, but also in but the also AC voltage dip. The configuration of the DCofdistribution is reviewed reviewed and SVPWM to is analyzed to propose compensation method forthe improving the DC-link and SVPWM is analyzed propose compensation method for improving DC-link voltage control voltage control for applying the proposed method. The DC distribution system that is applied in the for applying the proposed method. The DC distribution system that is applied in the proposed proposed compensation method is modeled and compared with the conventional DC distribution compensation method is modeled and compared with the conventional DC distribution system model model in terms of stability. Moreover, simulation and experimentation carried to insystem terms of stability. Moreover, simulation and experimentation were carried outwere to verify theout validity verify the validity of the proposed method. of the proposed method. FeedforwardCompensation Compensation Method Method for 2.2.Feedforward for Transient Transient State StateOutput OutputVoltage Voltage The circuit diagram of the DC distribution system that is used in this paper The circuit diagram of the DC distribution system that is used in this paper is is shown shown in in Figure Figure 2. 2. The overall system consists of a bidirectional two-level AC/DC PCS module in series. Each module The overall system consists of a bidirectional two-level AC/DC PCS module in series. Each module is is configured with an initial charge circuit, LCL filter. The output voltage of each module is 750 Vdc configured with an initial charge circuit, LCL filter. The output voltage of each module is 750 Vdc and and the total voltage is 1500 Vdc. The initial charging circuit of each module was applied to limit the the total voltage is 1500 Vdc . The initial charging circuit of each module was applied to limit the inrush inrush current, which can cause damages of the IGBT and the output capacitor due to the low current, which can cause damages of the IGBT and the output capacitor due to the low impedance of impedance of the DC link capacitor at startup. The resistance (Rpre) that is applied to the initial the DC link capacitor at startup. The resistance (Rpre ) that is applied to the initial charging circuit was charging circuit was designed by the RC time constant with the DC link capacitor. The LCL filter has designed the RCand time constant value with the DC capacitor. has a smaller and a smallerby volume inductance than thelink L filter and theThe LC LCL filter,filter also having a large volume harmonic inductance value than the L filter and the LC filter, also having a large harmonic attenuation effect. attenuation effect. It is designed according to the considerations of system capacity, impedance, Itswitching is designed according to the considerations of system capacity, impedance, switching frequency, frequency, filter cut-off frequency, etc. As a tertiary passive filter, resonance may occur filter cut-off frequency, etc. As a tertiary passive filter, resonance may occur between the elements between the elements of the filter at specific frequencies on the LCL filter. In order to minimize this of the filter at specific frequencies on the LCL filter.but In the order to minimize this resonance, manyinstudies resonance, many studies have been conducted, passive damping method is applied this have been conducted, but the passive damping method is applied in this paper [31,32]. paper [31,32]. DC-Grid Ground +750[V] -750[V]
Charging MC1
3Phase AC/DC Converter
AC/DC Module AC Grid
Y-△ Transformer
Lg1
Lf1
S11
S12
MCCB1
S13 vdc1
Main MC1
Rd1
S14 ias1 ibs1 ics1
vab1 vbc1 vca1
S15
S16
Cf1
MCCB2
Charging MC2
3Phase AC/DC Converter
AC/DC Module Y-△ Transformer
Lg2
Lf2
S21
S22
Main MC2
Rd2
ias2 ibs2 ics2
S24
S25
DC Load -750[V]
S23 vdc2
vab2 vbc2 vca2
DC Load +750[V]
DC Load 1500[V]
S26
Cf2
Figure 2. Proposed configuration of a bipolar low voltage DC (LVDC) distribution system. Figure 2. Proposed configuration of a bipolar low voltage DC (LVDC) distribution system.
The top power conversion module and the bottom power conversion module control the output The top module and therespectively. bottom power module the output voltage +750power Vdc as conversion separate power converters, Theconversion voltage Vdc1 of the control upper capacitor voltage +750 V as separate power converters, respectively. The voltage Vdc1 of the upper dc Vdc2 of the lower capacitor do not affect each other because the closed circuitcapacitor and the voltage is not and the voltage Vdc2 of the lower capacitor do not affect each other because the closed circuit formed, even if the bipolar load power is unbalanced. Therefore, the simplicity of the control and is thenot
Energies 2018, 11, 1468
4 of 16
formed, even if the bipolar load power is unbalanced. Therefore, the simplicity of the control and the reliability of the system can be improved since the balancing control using a neutral point is no longer necessary. Due to various voltages, it can be used according to the purpose and the capacity of the load. AC distribution systems, allow up to a voltage variation of ±10% from the reference voltage as a steady voltage, thus AC loads have robust characteristics against voltage fluctuations. However, the loads that are in the next-generation DC distribution systems are mostly composed of electronic loads sensitive to voltage fluctuations. Therefore, a new compensation technique is proposed for transient conditions in DC distribution systems to satisfy high power quality and stable voltage regulation for sensitive loads when considering the distribution systems that are connected with irregular renewable energy generation. 2.1. Relation to Transient State and Voltage Vector In the AC/DC PCS, a PI voltage controller for DC link voltage control and a PI current controller for AC current control are used in series. The voltage of the DC link is indirectly controlled by controlling the 3-phase AC current and the power factor. The AC/DC PCS has an output dc link consisting of a capacitor, which causes an instantaneous undershoot due to dv/dt of the dc link capacitor when the load is connected or the power generation of the distributed power supply suddenly decreases. On the contrary, an overshoot occurs on the DC distribution voltage instantaneously when the connection of a load using a constant power is suddenly disconnected or when the distributed power has to regenerate the generated power to the system due to an instantaneous increase of power generation. As a method of minimizing the overshoot and undershoot, the transient response of the converter can be improved through the power feedforward control by receiving information of load and renewable energy generation power in real time. However, it is practically impossible to collect the information of the power input/output status of all the loads that are connected to the power distribution through such communication. Therefore, compensation for this transient condition must be made in an AC/DC PCS that forms a DC distribution. The voltage vector of AC/DC PCS using SVPWM is affected by fluctuations in the DC distribution voltage (i.e., DC link voltage of the AC/DC PCS) in this transient state when the converter controls the AC side pulse width modulation (PWM) output. Therefore, voltage vectors of SVPWM should be analyzed in both the steady state and transient state. 2.2. Space Vector Pulse Width Modulation (SVPWM) in Steady State The vector diagram of SVPWM in steady state is represented, as shown in Figure 3. The V1 to V6 are output voltage vectors from AC/DC PCS. These vectors are each phase-shifted by 60◦ and their magnitude is proportional to the DC link voltage of the converter to 2Vdc /3. The V0 and V7 vectors are referred to as zero voltage vectors because no actual voltage is output to the AC side load. The maximum output voltage of an inverter that can be linearly output in the SVPWM is equal to the radius of the circle that is inscribed in hexagon of Figure 3 [18–21].
Energies 2018, 11, 1468
5 of 16
Energies 2018, 11, x
5 of 16
qs
V3 (0,1,0)
V2 (1,1,0)
steady state
Sector 2
Vqs*
Sector 1
T2
V*
Sector 3
V0 V7
V4 (0,1,1)
V1 (1,0,0)
T1
Vds 2Vdc/3
Sector 4
ds
*
Sector 6 Sector 5
V5 (0,0,1)
V6 (1,0,1)
Figure 3. Space voltage vector diagram at steady state.
Figure 3. Space voltage vector diagram at steady state.
The V0 and V7 vectors are referred to as zero voltage vectors because no actual voltage is output to theSVPWM AC side load. The maximum output of an thatvector can be in linearly in the the The selects two adjacent validvoltage vectors on inverter the space orderoutput to generate SVPWM is equal to the radius ofthe the time circle each that isvector inscribed in hexagonThe of Figure 3 [18–21]. command voltage and calculates is applied. applying time of two valid The SVPWM selects two in adjacent valid(1)–(3) vectorsusing on the vector period. in order to generate the vectors are calculated, as shown Equations thespace switching command voltage and calculates the time each vector is applied. The applying time of two valid ∗ | sin vectors are calculated, as shown in Equations using period. (π/3the − switching α) Ts |V(1)–(3)
T1 =
* 2VTdcV/3
·
sin(π/3)
sin( / 3 ) T1 2V / 3 (/ α3)) ∗ Tdc|V | sin( sin s
(1)
s
· * 2V sin(π/3) TsdcV/3 sin( ) T2 T0 =2V Tdcs − ) / 3( Tsin( 1 + T/23)
T2 =
(1)
(2) (2)
(3)
The time when two vectors adjacent voltage vector are applied(3)by the T0 to Tsthe (T1command T2 ) Equations (1)–(3) is proportional to the maximum voltage (=2Vdc /3) that each vector can output. The time when two vectors adjacent to the command voltage vector are applied by the Equations However, when T1 , T2 , and T0 are calculated to follow the reference voltage, according to (1)–(3) is proportional to the maximum voltage (=2Vdc/3) that each vector can output. However, when Equations if the voltagetoof the DC link changes, then T1 , T2 , toand T0 also(1)–(3), change. T1 , T2 , T1, T2, (1)–(3), and T0 are calculated follow the reference voltage, according Equations if the and Tvoltage transientthen voltage thechange. DC link PCS (i.e., the DC 0 are changed of the DCwhen link changes, T1, T2occurs , and T0in also T1,of T2,the andAC/DC T0 are changed when distribution voltage). If the voltage of the DC link increases, T , T , and T must be calculated transient voltage occurs in the DC link of the AC/DC PCS (i.e., the voltage). If the as 1 DC 2 distribution 0 voltage DC link increases, T1to , T2reach , and T 0 must be calculated smaller theresponse. steady state in the smaller than of thethe steady state in order the reference voltageasagain forthan a fast When order to reach the reference voltage for a be fastincreased response. when When the voltage to of the DC linkstate. voltage of the DC link decreases, T1 , T2again , T0 must compared steady decreases, 1, T2, T0 must be increased when compared to the steady state. Therefore, the time to apply Therefore, the Ttime to apply the adjacent vector in the currently used SVPWM should be changed the adjacent vector in change the currently used SVPWM should be changed according to the voltage change according to the voltage in the DC-link. in the DC-link.
2.3. SVPWM Analysis in Transient State
2.3. SVPWM Analysis in Transient State
In the steady state, the maximum value of V1 –V6 is fixed at 2Vdc /3. However, when an overshoot In the steady state, the maximum value of V1–V6 is fixed at 2Vdc/3. However, when an overshoot or undershoot occurs in the output voltage, as shown in Figure 4, the maximum voltage that can be or undershoot occurs in the output voltage, as shown in Figure 4, the maximum voltage that can be output by the six changes. When occurs,the themaximum maximum output voltage of each output by the vectors six vectors changes. Whenan anovershoot overshoot occurs, output voltage of each vectorvector increases by vby decreases whenanan undershoot occurs. relationship increases and decreasesby by vvundershoot when undershoot occurs. ThisThis relationship vovershootand overshoot undershoot is shown in the equations. is shown in following the following equations. v
v
V
(v
V )
overshoot vovershoot = vdcdc − dcVdc (vdcdc >dcVdc )
(4)
(4)
vundershoot = vdc − Vdc (vdc < Vdc )
(5)
vtransient = vovershoot + vundershoot
(6)
Energies 2018, 11, x
6 of 16
vundershoot v dc V dc ( v dc V dc )
Energies 2018, 11, 1468
(5) 6 of 16
vtransient v overshoot v undershoot
(6)
According maximum value valueof oftwo twoadjacent adjacentvectors vectorschanges changes when Accordingto toEquations Equations (4)–(6), (4)–(6), the maximum when anan overshoot the switching switchingtime timefor forgenerating generatingthe thereference reference voltage overshootororundershoot undershoot occurs. occurs. Therefore, Therefore, the voltage shouldbebechanged changedas asshown shownin in Equations Equations (7)–(9). should
V ∗* | sin(sin (/ π/3 3 − ) α) TTss|V T1 · T1 = 22 sin(π/3) (V(V +vvtransient ) ) sin( / 3) 3 3 dcdc transient
(7) (7)
TTss |V V *∗ | sin(sin )(α) T2 T= · sin(π/3) 2 2 (2Vdc +vvtransient )sin( / 3) dc transient ) 3 3 (V T0 = Ts − ( T1 + T2 ) T0 Ts (T1 T2 )
(8)
(9)
qs V3 (0,1,0)
Overshoot steady state
V0 V7
V4 (0,1,1)
V2 (1,1,0)
V3 (0,1,0)
Vqs*
T2
V1 (1, 0, 0)
ds
2V dc/3
V6 (1,0,1)
V0 V7
V4 (0,1,1)
V2 (1,1,0)
Vqs*
Vds*
T1
(a)
steady state Undershoot
V*
V5 (0,0,1)
(9)
qs
(8)
V*
T2
V1 (1,0,0)
vovershoot
ds
Vds*
T1
2Vdc/3
vundershoot
V6 (1,0,1)
V5 (0,0,1)
(b)
Figure 4. Changing of space voltage vector under transient state in (a) overshoot and (b) undershoot. Figure 4. Changing of space voltage vector under transient state in (a) overshoot and (b) undershoot.
Therefore, the switching time calculated from Equations (1)–(3) and (7)–(9) has an error. The Therefore, time calculated from Equations (1)–(3) and (7)–(9) has an error. The error error in Sector the 1 isswitching shown in Equations (10) and (11). in Sector 1 is shown in Equations (10) and (11). T T1_ err T1 T1_ transient Tas Tbs s (10) vtransient Ts T1_err = T1 − T1_transient = −( Tas − Tbs ) (10) vtransient T T2 _ err T2 T2 _ transient Tbs Tcs s (11) vtransient Ts T2_err = T2 − T2_transient = −( Tbs − Tcs ) (11) vtransient When the switching time is calculated as in the conventional method in Equations (1)–(3), the When the switching time is calculated as in the conventional method in Equations (1)–(3), response characteristic due to the error is worsened when the transient state occurs. Therefore, it is the response characteristic due to error is worsened when the transient state occurs. Therefore, it is desired to minimize this error bythe improving the transient response characteristics. desired to minimize this error by improving the transient response characteristics. 2.4. Feedforward Compensation Method of Output Voltage 2.4. Feedforward Compensation Method of Output Voltage The output of SVPWM is compared with the carrier wave in order to determine the switching The output of SVPWM is compared with ordercapacitor to determine switching pattern. The y-axis of the carrier wave refers to the the carrier voltage wave of the in output (= Vdc ).the Therefore, pattern. The y-axis of the carrier wave refers to the voltage of the output capacitor (=Vdc ). Therefore, when an overshoot or undershoot occurs, the maximum output value of each vector changes in the when an overshoot or undershoot occurs, the maximum output value of each vector changes in the space voltage vector diagram, so that an error occurs in the output voltage of the three-phase inverter, space voltage vector diagram, so that an error occurs in the output voltage of the three-phase inverter, according to Equations (1)–(11). according to Equations (1)–(11). The relation between switching frequency and carrier waveform in a micro controller unit (MCU) is shown in Figure 5. Generally, a method of changing the value (y-axis) of the carrier waveform is used to compensate for the error. However, when generating the carrier waveform in the MCU,
Energies 2018, 11, x Energies 2018, 11, x
7 of 16 7 of 16
The relation between switching frequency and carrier waveform in a micro controller unit The is relation switching frequency and carrier waveform in a micro controller unit (MCU) shownbetween in Figure 5. Generally, a method of changing the value (y-axis) of the carrier Energies 2018, 11, 1468 in Figure 5. Generally, a method of changing the value (y-axis) of the carrier 7 of 16 (MCU) is shown waveform is used to compensate for the error. However, when generating the carrier waveform in waveform used to compensate forwaveform the error. However, when generating the carrier in the MCU,isthe slope of the carrier is fixed according to the dividing ratiowaveform of the MCU. the MCU, the slope of the carrier waveform is fixed according to the dividing ratio of the MCU. Therefore, maximum size (y-axis) the carrier waveform changes, period the slope of if thethe carrier waveform is fixedof according to the dividing ratio ofthe theswitching MCU. Therefore, Therefore, if well. the maximum size (y-axis) of(1)–(3) the carrier waveform changes, time the switching period changes as According to Equations and (7)–(9), the switching is affected by if the maximum size (y-axis) of the carrier waveform changes, the switching period changes as one well. changes as well. According to frequency, Equations (1)–(3) and (7)–(9), using the switching time is affectedcannot by onebe period (= ) of the carrier so compensation the carrier waveform T s According to Equations (1)–(3) and (7)–(9), the switching time is affected by one period (=Ts ) of the period (= T s ) of the carrier frequency, so compensation using the carrier waveform cannot be accomplished atsothe same switching The switching frequency is changed if same the carrier frequency, compensation usingfrequency. the carrier waveform cannot be accomplished at the accomplished at the same switching Theanswitching frequency is changed ifwhich the conventional compensation method isfrequency. usedis when or undershoot occurs, switching frequency. The switching frequency changed ifovershoot the conventional compensation method is conventional compensation method is used when an overshoot or side undershoot occurs, which aggravates the total harmonic distortion (THD) characteristic of the ac current. used when an overshoot or undershoot occurs, which aggravates the total harmonic distortion (THD) aggravates the total harmonic distortion (THD)method characteristic of the ac current. Therefore, a feedforward compensation for changing theside reference voltage is proposed, characteristic of the ac side current. Therefore, a feedforward compensation method for changing the reference voltage is proposed, as shown in Figure 6, without compensation changing the carrier wave a fixedthe switching frequency). space Therefore, a feedforward method for(i.e., changing reference voltage isThe proposed, asvoltage shown in Figure 6, without changing the carrier wave (i.e., a fixed switching frequency). The space vector diagram of the proposed output voltage compensation method in Figure is as shown in Figure 6, without changing the carrier wave (i.e., a fixed switching frequency). The6b space voltage vector diagramin of the proposed voltage compensation method Figure 6b is illustrated, as shown Figure 7. It showsoutput the case in Sector 1 (In that time, = ). The v v c*o m p ein transient n voltage vector diagraminofFigure the proposed output method Figure 6b is illustrated, illustrated, as shown 7. It shows the voltage case in compensation Sector 1 (In that time, in v c*o m p e n = vtransient ). The ∗ value of T 1 , T 2 , T 0 is changed because the value of the reference voltage is changed by the feed as shown in Figure 7. It shows the case in Sector 1 (In that time, vcompen = vtransient ). The valueforward of T1 , T2 , value of T1, T2, T0 is changed because the value of the reference voltage is changed by the feed forward T0compensation. is changed because the value of the reference voltage is changed by the feed forward compensation. compensation.
Figure5.5.The Therelation relationbetween betweenswitching switching frequency frequency and and carrier Figure carrier wave wave in inmicro microcontroller controllerunit unit(MCU). (MCU). Figure 5. The relation between switching frequency and carrier wave in micro controller unit (MCU). vovershoot vovershoot
vovershoot vovershoot
0
0
0
vovershoot /2 vovershoot /2
0 [V] [V]
[V] [V]
0
Carrier
0
ȣ
ȣ
Carrier
0
Vdc
Vdc
ȣ
ȣ
0
vovershoot /2 vovershoot /2
Carrier
vovershoot /2 v overshoot /2 (a)
Vdc
Vdc
vovershoot /2 vovershoot /2
Carrier
(b) (a) (b) Figure 6. Two cases for the realization of the compensation value: (a) altered carrier wave: Figure 6. Two cases isfor the realization of the compensation value: wave: (a) altered carrier value wave:is compensation carrier wave; altered reference Figure 6. Twovalue cases applied for the in realization of and the (b) compensation value: (a)compensation altered carrier wave: compensation value is applied in carrier wave; and (b) altered reference wave: compensation value is applied in reference compensation value iswave. applied in carrier wave; and (b) altered reference wave: compensation value is applied in reference wave. applied in reference wave.
Generally, control of a three-phase inverter (or converter) generates a switching pattern using the output of the current controller. However, the proposed control method measures variations in the output voltage that it is used to compensate for the reference voltage. The compensated value is then used in the voltage controller. However, the control period of the current controller is 10 times faster than the voltage controller because of the control frequency bandwidth, and the actual switching pattern is generated using the output of the current controller. Therefore, the current controller also
Energies 2018, 11, 1468
8 of 16
must perform the same compensation in order to quickly respond in the frequent transient state. The current compensation value of the current controller is obtained by the voltage variation and the capacitance of the DC-link. Figure 8 shows the overall control block diagram of the proposed system. The voltage and the current compensation values for transient state were applied to each controller. In Figure 8, ωe is voltage angular speed of AC side for VOC and phase locked loop (PLL). ωe377 is fixed value of AC grid (60 Hz). iqe means active current and ide means reactive current. ved , veq are the values synchronous d-q transformed of input voltage. vdc is DC-link voltage, idc is DC-link current. Ri and Li are resistor and inductance Energies 2018, 11, x of converter side, respectively. 8 of 16
qs V2 (1,1,0)
Sector 1 vovershoot _ qs
vovershoot T2
* qs
V
V*
T1
* ds
ds
vovershoot _ ds
V
V1 (1,0,0)
Figure 7. The altered reference voltage in Sector 1 of space vector pulse width modulation (SVPWM).
Figure 7. 11, The Energies 2018, x altered reference voltage in Sector 1 of space vector pulse width modulation (SVPWM). 9 of 16
Generally, control of a Pre-Charging three-phase inverter (or converter) generates a switching pattern using MC 3 Phase the output of the current controller. However, the proposed control method measures variations in AC/DC Converter the output voltage that it is used to compensate for the reference voltage. The compensated LOAD value is LCL Filter L L f g va then used in the voltage controller. However, the control period of the current controller is 10 times vb faster than the voltage controller because of the control frequency bandwidth, and the actual vc switching pattern is generated the output of theiascurrent controller. Therefore, the current ibs ics Main using MC Rd controller valso must perform the same compensation in order to quickly respond in the frequent C f ab vbc vca vdc * vas transient state. The current compensation value of theisdcurrent controller is* Tobtained by the voltage Tas* Tbs cs s Park Clark v i bs Phasecapacitance Positive q variation and the of the DC-link. TransforSVPWM Transforvdc Estimation Sequence vcsoverall mation Figure 8 shows the control block mation diagramθe of the proposed system. The voltage and the e v compen 8, is VIn dc vapforvbptransient vcp current compensation values applied to each Figure istate iewere V sds* controller. V sqs* d q e s *
vd ie PI voltage angular speed of AC side Park Clarke for VOC and phase locked dloop (PLL). e 377 vsq Current TransforTransfore e Controller v e grid (60 Hz).mation current and id means reactiveied*current. θe active mation iq means d
*
PI is fixed value Vof AC dc *
Voltage
, Controller vqe are the values
θe i d c is DC-link current. R i synchronous v dc is DC-link voltage, ved d-q veq transformed of input voltage. veq Power ieq ieq* and Li are resistor and inductance of converter side, respectively. ved Calculation P Active QReactive 0
ωe
PLL PI Controller
ʃ
θe
ωe377
PF Calculation
PF
ieq*_compen PF PI Controller
C
d dt
Proposed Feedforward Compensation
PF*
Figure Overallcontrol controlblock blockdiagram diagram that that applied of of a module. Figure 8. 8. Overall applied the theproposed proposedcompensation compensationmethod method a module.
3. Simulation In order to verify the feasibility of the proposed feed forward compensation method, the simulation was performed using PSIM (PowerSIM, Rockville, MD, USA). The parameters of the system applied to the simulation are shown in Table 1. Table 1. Parameter of proposed AC/DC pulse width modulation (PWM) converter. Parameters
Value
Unit
Energies 2018, 11, 1468
9 of 16
3. Simulation In order to verify the feasibility of the proposed feed forward compensation method, the simulation was performed using PSIM (PowerSIM, Rockville, MD, USA). The parameters of the system applied to the simulation are shown in Table 1. Table 1. Parameter of proposed AC/DC pulse width modulation (PWM) converter.
LCL Filter
Parameters
Value
Rated Power
50
Unit kW
Input voltage (VLL.rms )
380
Vrms
Source frequency
60
Hz
Switching frequency
5
kHz
Source Side Inductance
120
µH
Filter Capacitance
50
µF
Converter Side Inductance
500
µH
DC-link Capacitance
10,200
µF
DC-link Voltage
750
Vdc
Relation to Transient State and Voltage Vector Entire operation sequence is as shown in Figure 9. PCS starts operation at 0.2 s and the loads are connected to 0.4 s and 0.6 s in 25 kW increments. An AC voltage dip occurs at 0.8 s for the verified dynamic characteristic of the DC grid. The same value gains of PI voltage and current controller are applied to compare the proposed method to the conventional method. An AC voltage 70% dip (0.7 p.u) occurs, as shown in Figure 9. Energies 2018, 11, x
10 of 16
Vdc_reference Vdc_conventional Vdc_proposed 790 780
50% Load
100% Load
50% Dip
770 760 750 740 730
Vas Vbs Vcs 600 400 200 0 -200 -400 -600
ILOAD 80 60 40 20
25kW Load
AC voltage 50% dip
50kW Load
0 -20 0.2
0.4
0.6
0.8
1
Time (s)
Figure 9. Comparison of the conventional method with the proposed control.
Figure 9. Comparison of the conventional method with the proposed control. The expanded waveform at the moment of applied load is as shown in Figure 10. When the 25 kW load is applied at 0.4 s, an undershoot occurs in the DC grid voltage. This transient state (undershoot) cannot be avoided because the DC grid voltage is maintained by the capacitor. This undershoot becomes worse as a larger load is applied at a time. When the conventional method is applied, about 6.4 V undershoot occurs. When the same load is applied with the proposed method, the magnitude of undershoot is reduced to about 3.9 V, as shown in Figure 10a. The current control characteristic of the d-q axis when the undershoot occurs is
Energies 2018, 11, 1468
10 of 16
The expanded waveform at the moment of applied load is as shown in Figure 10. When the 25 kW load is applied at 0.4 s, an undershoot occurs in the DC grid voltage. This transient state (undershoot) cannot be avoided because the DC grid voltage is maintained by the capacitor. This undershoot becomes worse as a larger load is applied at a time. When the conventional method is applied, about 6.4 V undershoot occurs. When the same load is applied with the proposed method, the magnitude of undershoot is reduced to about 3.9 V, as shown in Figure 10a. The current control characteristic of the d-q axis when the undershoot occurs is as shown in Figure 10b. In this paper, the proposed compensation method is not applied to the d-axis current controller. Thus, d-axis current has the same control characteristic. The current control characteristic of q-axis is improved by proposed compensation method. Therefore, the dynamic characteristic of DC voltage control is improved. Figure 11 shows the dynamic characteristic of DC grid voltage on AC voltage 70% dip. In applying constant load, if an AC voltage dip occurs, DC grid voltage droops, as shown in Figure 11a. This situation is similar to when the load is applied on a DC grid. Therefore, this situation can be improved Energies 2018, 11, xby the same compensation method. 11 of 16 Energies 2018, 11, x
11 of 16
Vdc_reference Vdc_conventional Vdc_proposed
752
50% Load Vdc_proposed Vdc_reference Vdc_conventional
752 750
50% Load
750 748
3.9 V droop 3.9 V droop
748 746
6.4 V droop 6.4 V droop
746 744 744
Iq_Ref_conventional Iq_Real_conventional Iq_Ref_proposed Iq_Real_proposed
(a) (a)
80 Iq_Ref_conventional Iq_Real_conventional Iq_Ref_proposed Iq_Real_proposed
80 60 60 40
Proposed Method (Reference, Real) Proposed Method (Reference, Conventional Method Real) (Reference, Real) Conventional Method (Reference, Real)
40 20 20 0 0
0.38
0.4
0.38
0.4
0.42
0.44
0.42
Time (s)
0.44
0.46
0.48
0.5
0.52
0.46
0.48
0.5
0.52
Time (s) (b) (b) Figure Figure10. 10.VVdcdcand andq-axis q-axisresponse responsecharacteristic characteristic(expanded). (expanded).(a) (a)Comparison ComparisonofofDC DCvoltage voltagedynamic dynamic Figure 10.atVload dc and q-axis response characteristic (expanded). (a) control Comparison of DC voltage dynamic responses connection; and (b) comparison q-axis current characteristic. responses at load connection; and (b) comparison q-axis current control characteristic. responses at load connection; and (b) comparison q-axis current control characteristic.
755 755 750 750 745 745
Vdc_reference Vdc_conventional Vdc_proposed Vdc_reference Vdc_conventional Vdc_proposed
7.4 V droop 7.4 V droop
10.2 V droop 10.2 V droop
740 740 735 735
(a) (a) Vas Vbs Vcs Vas Vbs Vcs
600 600 400 400 200 200 0 0 -200 -200 -400 -400 -600 -600
AC voltage AC 50%voltage dip 50% dip
Figure 11. Cont.
(b) (b)
750 7.4 V droop
745
10.2 V droop
740
Energies 2018, 11, 1468
11 of 16
735
(a) Vas Vbs Vcs AC voltage 50% dip
600 400 200 0 -200 -400 -600
(b) Iq_Ref_conventional Iq_Real_conventional Iq_Ref_proposed Iq_Real_proposed
300 250
100% Load (25kW)
200 Proposed Method (Reference, Real)
150 100
Conventional Method (Reference, Real)
50 0 0.78
0.8
0.82
0.84
0.86
0.88
0.9
Time (s)
(c) Figure 11. Dynamic response of DC grid voltage to AC voltage dip. (a) Comparison of DC voltage Figure 11. Dynamic response of DC grid voltage to AC voltage dip. (a) Comparison of DC dynamic responses at voltage dip occurrence; (b) occurrence of AC voltage dip by 50%; and (c) voltage dynamic responses at voltage dip occurrence; (b) occurrence of AC voltage dip by 50%; comparison q-axis current control characteristic. and (c) comparison q-axis current control characteristic.
4. Experimental Result Energies 2018, 11, x 12 of 16 The hardware configuration for applying the proposed compensation method is as shown in 4. Experimental Result Figure 12. In this paper, two AC/DC modules are applied. The capacity of each module is 50 kW hardware configuration applying the proposed compensation method is as shownvoltage in and the switching The frequency is set to 5forkHz. Each module controls the output to 750 Vdc . Figure 12. In this paper, two AC/DC modules are applied. The capacity of each module is 50 kW and The output of each module is configured in series. Therefore, the final distribution voltage is ±750 Vdc . the switching frequency is set to 5 kHz. Each module controls the output voltage to 750 Vdc. The output of each module is configured in series. Therefore, finalvarious distribution voltage is ±750 Vdc. The loads that are connected to the proposed system cantheuse voltages.
The loads that are connected to the proposed system can use various voltages.
Figure 12. Hardware configuration.
Figure 12. Hardware configuration.
Figure 13 shows the experimental waveforms that are not applied to the proposed compensation method of 3-phase AC/DC PCS. A soft start is applied for minimizing exceed current inflow to capacitor during initial operation. 7 kW load is connected after output voltage is stable as 750 Vdc (in ±1% during 2 s). Figure 13b shows the expanded waveform of the reference and feedback voltage of the voltage controller when the load is applied, as in Figure 13a. The reference voltage of the voltage controller is 750 Vdc and it can be confirmed that the output voltage instantaneously decreases when the load is applied. Figure 13c shows the waveform of the q-axis reference and feedback of the current controller when the load is applied. At the moment of undershoot, the reference value of the current controller
Energies 2018, 11, 1468
12 of 16
Figure 13 shows the experimental waveforms that are not applied to the proposed compensation method of 3-phase AC/DC PCS. A soft start is applied for minimizing exceed current inflow to capacitor during initial operation. 7 kW load is connected after output voltage is stable as 750 Vdc (in ±1% during 2 s). Figure 13b shows the expanded waveform of the reference and feedback voltage of the voltage controller when the load is applied, as in Figure 13a. The reference voltage of the voltage controller is 750 Vdc and it can be confirmed that the output voltage instantaneously decreases when the load is applied. Figure 13c shows the waveform of the q-axis reference and feedback of the current controller when the load is applied. At the moment of undershoot, the reference value of the current controller changes, but the output (=feedback) of the controller shows a slower response than the reference. Since the current controller and the voltage controller are operated separately and the actual switching waveform is generated using the output of the current controller, the two controllers must each be compensated for, respectively. Figure 13d shows the expanded waveform of the output voltage of the 3-phase AC/DC PCS in Figure 13a. When the load is applied, it can be confirmed that an undershoot of Energies about 8.9 occurs for 136 ms. 2018,V11, x 13 of 16 Analog Signal
Digital Signal
1400 V
770 V
1200 V
765 V
1000 V 800 V
Reference Voltage [Digital Signal]
760 V 755 V
600 V
750 V
400 V
745 V
200 V 0V
AC/DC Module Output Voltage [Analog Signal]
740 V
Feedback Voltage [Digital Signal]
735 V
[1s/div] -200 V
730 V
(a) 770 V 765 V
Reference Voltage [Digital Signal]
760 V 755 V 750 V 745 V 740 V
Feedback Voltage [Digital Signal]
735 V
[50ms/div]
730 V
(b) 40 A
Reference Q-Axis [Digital Signal]
30 A 20 A 10 A 0A -10 A
Feedback Q-Axis [Digital Signal]
-20 A -30 A
[50ms/div] -40 A
(c) 770 V
Figure 13. Cont.
765 V 760 V
AC/DC Module Output Voltage 755 V 750 V 745 V 740 V
8.9 V 136 ms
10 A 0A -10 A
Feedback Q-Axis [Digital Signal]
-20 A -30 A
Energies 2018, 11, 1468
[50ms/div]
13 of 16
-40 A
(c) 770 V 765 V 760 V
AC/DC Module Output Voltage 755 V 750 V
8.9 V
745 V 740 V
136 ms
735 V
[50ms/div] 730 V
(d) Figure 13. Experimental waveforms without the proposed method. (a) Reference and real voltage; (b) Figure 13. Experimental waveforms without the proposed method. (a) Reference and real voltage; reference and real voltage of DC Bus (expanded); (c) reference and real q-axis current in synchronous (b)frame; reference real voltage of DC Bus (expanded); (c) reference and real q-axis current in synchronous andand (d) output voltage of module 1 (expanded). frame; and (d) output voltage of module 1 (expanded).
Figure 14 shows the experimental waveform of a 3-phase AC/DC PCS using the proposed output Figurefeed 14 shows the experimental a 3-phase AC/DC PCS using proposed output voltage forward compensationwaveform method. of Experiments were carried outthe under the same conditions, as in Figure 13. voltage feed forward compensation method. Experiments were carried out under the same conditions, as Energies in Figure 13. 2018, 11, x 14 of 16 Analog Signal
Digital Signal
1400 V
770 V
Feedforward Compensation Voltage (+750 Offset) [Digital Signal]
1200 V 1000 V 800 V
Reference Voltage [Digital Signal]
765 V 760 V 755 V
600 V
750 V
400 V 200 V 0V
745 V
AC/DC Module Output Voltage [Analog Signal]
740 V
Feedback Voltage [Digital Signal]
735 V
[1s/div] -200 V
730 V
(a) 770 V
Feedforward Compensation Voltage (+750 Offset) [Digital Signal]
765 V 760 V 755 V
Reference Voltage [Digital Signal]
750 V 745 V
Feedback Voltage [Digital Signal]
740 V 735 V
[50ms/div] 730 V
(b) 40 A
Reference Q-Axis [Digital Signal]
30 A 20 A 10 A 0A -10 A -20 A
Feedback Q-Axis [Digital Signal]
Feedback Compensation Current [Digital Signal]
-30 A
[50ms/div] -40 A
(c) 770 V
Figure 14. Cont.
765 V 760 V
AC/DC Module Output Voltage 755 V 750 V
4.5 V 745 V 740 V
75 ms
10 A 0A -10 A -20 A
Feedback Q-Axis [Digital Signal]
Feedback Compensation Current [Digital Signal]
-30 A
Energies 2018, 11, 1468
[50ms/div]
14 of 16
-40 A
(c) 770 V 765 V 760 V
AC/DC Module Output Voltage 755 V 750 V
4.5 V 745 V 740 V
75 ms
735 V
[50ms/div] 730 V
(d) Figure 14. Experimental waveforms with the proposed method. (a) Reference and real voltage; (b) Figure 14. Experimental waveforms with the proposed method. (a) Reference and real voltage; reference and real voltage of DC Bus (expanded); (c) reference and real q-axis current in synchronous (b) reference and real voltage of DC Bus (expanded); (c) reference and real q-axis current in synchronous frame; and (d) output voltage of module 1 (expanded). frame; and (d) output voltage of module 1 (expanded).
Figure 14b shows the output of the voltage controller using the proposed compensation method Figure the output of the voltage controller using the proposed method when the 14b loadshows is applied. The undershoot is reduced when compared to that compensation shown in Figure 13, when the load is applied. Theofundershoot reducediswhen to thatcompensation shown in Figure because the reference voltage the voltage is controller 750 Vdccompared and the forward value13, is applied the load is applied. shows theisoutput of the controller using the because the when reference voltage of the Figure voltage14ccontroller 750 Vdc andcurrent the forward compensation compensation The response characteristic is improved the section where the controller voltage value is applied method. when the load is applied. Figure 14c shows thein output of the current variation occurs by the compensation to current controller based on Figure 8 when overshoot (orthe using the compensation method. The response characteristic is improved in the section where undershoot) occurs as compared to the resultstoincurrent Figure controller 13c. Figurebased 14d shows the output voltage of voltage variation occurs by the compensation on Figure 8 when overshoot
(or undershoot) occurs as compared to the results in Figure 13c. Figure 14d shows the output voltage of a three-phase AC/DC PWM converter. When the load was applied, about a 4.5 V undershoot occurred for 75 ms. When the proposed feed forward compensation method is applied, the magnitude of undershoot is reduced by 50% and the response is improved by 45%. 5. Conclusions In this paper, a new feed forward compensation method is proposed to improve the transient state response for the PCS configuring a DC distribution network in the DC distribution system. The degradation cause of the transient response is analyzed when SVPWM is used, and the feed forward compensation method is proposed to improve the response characteristic. The proposed method is investigated in order to solve the problems to be applied to the MCU, and a solution for this problem is suggested. The proposed method is applied to the bipolar low voltage DC (LVDC) system. It is also confirmed that the stability of the system is improved when the proposed method is applied through a modeling of the system. Simulation and experimentation were carried out by applying the proposed compensation method and the feasibility of the proposed method was verified. When the proposed method is applied to the LVDC system connected with various loads and renewable energy, it is possible to construct a robust distribution with an improved transient response characteristic. Author Contributions: S.-J.H. designed the experiment; S.-J.H., S.-W.H. and K.-M.K. performed the experiment; S.-J.H., K.-M.K. and J.-H.L. analyzed the theory. S.-J.H. wrote the manuscript. J.-H.L. and C.-Y.W. participated in research plan development and revised the manuscript. All authors have contributed to the manuscript. Acknowledgments: This work was supported by “Human Resources Program in Energy Technology” of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), granted financial resource from the Ministry of Trade, Industry & Energy, Korea (No. 20164030200980). Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2.
Wu, T.F.; Chang, C.H.; Lin, L.C.; Yu, G.R.; Chang, Y.R. DC-bus voltage control with a three-phase bidirectional inverter for DC distribution systems. IEEE Trans. Power Electron. 2013, 28, 1890–1899. [CrossRef] Yu, X.; She, X.; Zhou, X.; Huang, A.Q. Power Management for DC Microgrid Enabled by Solid-State Transformer. IEEE Trans. Smart Grid 2014, 5, 954–965. [CrossRef]
Energies 2018, 11, 1468
3. 4. 5. 6. 7. 8. 9.
10.
11.
12.
13.
14.
15.
16.
17.
18. 19. 20.
21. 22.
15 of 16
Mackay, L.; Blij, N.H.V.D.; Ramirez-Elizondo, L.; Bauer, P. Towards the Universal DC Distribution System. Electr. Power Compon. Syst. 2017, 45, 1032–1042. [CrossRef] Kim, H.S.; Ryu, M.H.; Baek, J.W.; Jung, J.H. High-Efficiency Isolated Bidirectional AC-DC Converter for a DC Distribution System. IEEE Trans. Power Electron. 2013, 28, 1642–1654. [CrossRef] Sechilariu, M.; Wang, B.; Locment, F. Building integrated photovoltaic system with energy storage and smart grid communication. IEEE Trans. Ind. Electron. 2013, 60, 1607–1618. [CrossRef] Dragicevic, T.; Lu, X.; Vasquez, J.C.; Guerrero, J.M. DC Microgrids—Part II: A Review of Power Architectures, Applications, and Standardization Issues. IEEE Trans. Power Electron. 2016, 31, 3528–3549. [CrossRef] Whaite, S.; Grainger, B.; Kwasinski, A. Power Quality in DC Power Distribution Systems and Microgrids. Energies 2015, 8, 4378–4399. [CrossRef] Ryu, M.H.; Kim, H.S.; Baek, J.W.; Kim, H.G.; Jung, J.H. Effective test bed of 380-V DC distribution system using isolated power converters. IEEE Trans. Ind. Electron. 2015, 62, 4525–4536. [CrossRef] Shin, S.C.; Lee, H.J.; Kim, Y.H.; Lee, J.H.; Won, C.Y. Transient response improvement at startup of a threephase AC/DC converter for a DC distribution system in commercial facilities. IEEE Trans. Power Electron. 2014, 29, 6742–6753. [CrossRef] Chen, W.; Dai, W.; Wang, Z.; Zhang, G.; Yan, Y.; Xia, C. Optimal space vector pulse width modulation strategy of neutral point clamped three-level inverter for output current ripple reduction. IET Power Electron. 2017, 10, 1638–1646. [CrossRef] Hintz, A.; Prasanna, U.R.; Rajashekara, K. Comparative Study of the Three-Phase Grid-Connected Inverter Sharing Unbalanced Three-Phase and/or Single-Phase systems. IEEE Trans. Ind. Appl. 2016, 52, 5156–5164. [CrossRef] Huang, Y.; Wang, D.; Shang, L.; Zhu, G.; Tang, H.; Li, Y. Modeling and Stability Analysis of DC-Link Voltage Control in Multi-VSCs with Integrated to Weak Grid. IEEE Trans. Energy Convers. 2017, 32, 1127–1138. [CrossRef] Chen, W.W.; Zane, R.; Corradini, L. Isolated Bidirectional Grid-Tied Three-Phase AC-DC Power Conversion Using Series-Resonant Converter Modules and a Three-Phase Unfolder. IEEE Trans. Power Electron. 2017, 32, 9001–9012. [CrossRef] Jamma, M.; Barara, M.; Akherraz, M.; Enache, B.A. Voltage oriented control of three-phase PWM rectifier using space vector modulation and input output feedback linearization theory. In Proceedings of the 8th International Conference on Electronics, Computers and Artificial Intelligence (ECAI), Ploiesti, Romania, 30 June–2 July 2016; pp. 1–8. Kadri, R.; Gaubert, J.P.; Champenois, G. An improved maximum power point tracking for photovoltaic grid-connected inverter based on voltage-oriented control. IEEE Trans. Ind. Electron. 2011, 58, 66–75. [CrossRef] Trinh, Q.N.; Wang, P.; Choo, F.H. An Improved Control Strategy of Three-Phase PWM Rectifiers under Input Voltage Distortions and DC-Offset Measurement Errors. IEEE J. Emerg. Sel. Top. Power Electron. 2017, 5, 1164–1176. [CrossRef] Jung, T.; Gwon, G.; Kim, C.H.; Han, J.; Oh, Y.; Noh, C.H. Voltage Regulation Method for Voltage Drop Compensation and Unbalance Reduction in Bipolar Low-Voltage DC Distribution System. IEEE Trans. Power Deliv. 2018, 33, 141–149. [CrossRef] Choe, S.; Son, Y.K.; Sul, S.K. Control and Analysis of Engine Governor for Improved Stability of DC Microgrid against Load Disturbance. IEEE J. Emerg. Sel. Top. Power Electron. 2016, 4, 1247–1258. [CrossRef] Kakigano, H.; Miura, Y.; Ise, T. Low-Voltage Bipolar-Type DC Microgrid for Super High Quality Distribution. IEEE Trans. Power Electron. 2010, 25, 3066–3075. [CrossRef] Huang, M.; Peng, Y.; Chi, K.T.; Liu, Y.; Sun, J.; Zha, X. Bifurcation and Large-Signal Stability Analysis of Three-Phase Voltage Source Converter Under Grid Voltage Dips. IEEE Trans. Power Electron. 2017, 32, 8868–8879. [CrossRef] Bongiorno, M.; Svensson, J. Voltage dip mitigation using shunt connected voltage source converter. IEEE Trans. Power Electron. 2007, 22, 1867–1874. [CrossRef] Davari, M.; Mohamed, Y.A.R.I. Dynamics and robust control of a grid-connected VSC in multiterminal DC grids considering the instantaneous power of DC-and AC-side filters and DC grid uncertainty. IEEE Trans. Power Electron. 2016, 31, 1942–1958. [CrossRef]
Energies 2018, 11, 1468
23.
24. 25. 26. 27.
28. 29. 30. 31. 32.
16 of 16
Davari, M.; Mohamed, Y.A.R.I. Robust Vector Control of a Very Weak-Grid-Connected Voltage-Source Converter Considering the Phase-Locked Loop Dynamics. IEEE Trans. Power Electron. 2017, 32, 977–994. [CrossRef] Davari, M.; Mohamed, Y.A.R.I. Robust multi-objective control of VSC-based DC-voltage power port in hybrid AC/DC multi-terminal microgrids. IEEE Trans. Smart Grids 2013, 4, 1597–1612. [CrossRef] Nasirian, V.; Davoudi, A.; Lewis, F.L.; Guerrero, J.M. Distributed Adaptive Droop Control for DC Distribution Systems. IEEE Trans. Energy Convers. 2014, 29, 944–956. [CrossRef] Tahim, A.P.N.; Pagano, D.J.; Lenz, E.; Stramosk, V. Modeling and stability analysis of islanded DC microgrids under droop control. IEEE Trans. Power Electron. 2015, 30, 4597–4607. [CrossRef] Lu, X.; Guerrero, J.M.; Sun, K.; Vasquez, J.C. An Improved Droop Control Method for DC Microgrids Based on Low Bandwidth Communication with DC Bus Voltage Restoration and Enhanced Current Sharing Accuracy. IEEE Trans. Power Electron. 2014, 29, 1800–1812. [CrossRef] Federico, I.; Jose, E.; Luis, F. Master-slave DC droop control for paralleling auxiliary DC/DC converters in electric bus applications. IET Power Electron. 2017, 10, 1156–1164. [CrossRef] Li, G.; Du, Z.; Shen, C.; Yuan, Z.; Wu, G. Coordinated Design of Droop Control in MTDC Grid Based on Model Predictive Control. IEEE Trans. Power Syst. 2018, 33, 2816–2828. [CrossRef] Simes, M.G.; Farret, F.A. Designing Power Electronic Control Systems, 1st ed.; Wiley-IEEE Press: Hoboken, NJ, USA, 2017; ISBN 9781119058458. Jayalath, S.; Hanif, M. Generalized LCL-Filter Design Algorithm for Grid-Connected Voltage-Source Inverter. IEEE Trans. Ind. Electron. 2017, 64, 1905–1915. [CrossRef] Rockhill, A.A.; Liserre, M.; Teodorescu, R.; Rodriguez, P. Grid-Filter Design for a Multimegawatt Medium-Voltage Voltage-Source Inverter. IEEE Trans. Ind. Electron. 2011, 58, 1205–1217. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
