Usage of Grid-Connected Inverters (GCI) increased dramatically nowadays.
These systems are used in Active Power Filters (APF), static synchronous
var compensators (STATCOM), grid connected photovoltaic systems, grid
connection of wind turbines and in Fig. 1 general topology
of the grid connected inverter is shown. This simple topology is capable
of bidirectional real and reactive power flow. When viewed from the utility
side, the VSI can act as an ac source, a resistive load, an inductor or
capacitor all at the same KVA ratings.
Situation of the DC link depends on the application. For example, in
APF and STATCOM, that no real power flows between grid and inverter, it
is impossible to connect the separate DC power to the DC link. In Voltage
Source Inverters (VSI), there are two basic mechanisms by which the power
flow between GCI and grid can be controlled. The first method is through
the control of switching instance of inverter so as to produce a fundamental
50 Hz voltage in the output of inverter (Schauder, 1995; Mori, 1999).
In this method, the power flow is controlled by adjusting the amplitude
and phase of inverter output voltage relative to the line voltage. Since
the grid is invariably a rigid voltage source with very low line impedance,
power flow from the inverter to the grid, reduces to being simply current
flow control and voltage source inverters have been proposed for use as
current sources in number of applications (Moon, 1999; Borle and Nayar,
1995; Malesani and Kazmierkowski, 1993; Borle and Nayar, 1996; Jovcic
et al., 2006; Xiang-lian, 2006).
This study is aimed at both summarizing the main implementation refinements
which characterize the latest versions of the voltage source inverter
controllers and comparing the different performance of these two control
mechanisms, through simulation.
||General topology of the grid connected inverter
The organization of this study is as follow. This study initially focuses
on the identification of proper criteria for correctly comparing of the
voltage control and current control mechanisms. Then after a short description
of the principles of these methods and the presentation of the main refinements
the simulated system is briefly described. Finally the results of comparison
that is done according to the chosen criteria are discussed.
MATERIALS AND METHODS
Criteria for comparison of current and voltage controllers: This
study was done in lab of electronic, Khaje Nasir Toosi University for
4 month in 2007. The first criteria considered here, is the evaluation
of the inverter output current spectrum which identifies the distribution
of current harmonics in the different frequency ranges. This is very important
point for design of passive filters which smooths the modulation ripple
of current and for the evaluation of the capability of the different control
mechanisms to meet the International Electro Technical Commission (IEC)
standard, requirements. Current spectrum also determines the total harmonic
distortion of the output current.
GCI may be used for transient stability improvement, power oscillation
damping and increase of voltage stability limit and improvement of other
dynamic limitations of power transmission (Gyugi, 1994). In these and
many other applications (such as reactive power compensation of arc Furnaces)
transient response of the GCI should be as fast as possible and for this
reason transient response of inverter is chosen as a second criterion
Voltage control and current control of GCI: Power flow between
the grid and inverter can be controlled by adjusting the fundamental phase
and amplitude of Vpwm1 relative to Van. (Vpwm
is output voltage of inverter, Vpwm1 is output voltage
first harmonic of inverter and Van is grid line to neutral
voltage). For small angles of δ expressed in radians, the real and
reactive power flow can be approximated as:
In the above equations, V1 is fundamental voltage of grid; Vpwm
is fundamental output voltage of inverter, Xr is connection
impedance and δ is phase angle between V1 and Vpwm.
For control purpose it is apparent that the real power varies with δ
and the reactive power varies with (Vpwm1-V1). This realization
lets to the development of voltage phase and amplitude power flow control
(VPAC) which uses δ, Vpwm1 as the control variables.
Power flow control in ac-DC converters can be achieved through the direct
control of current in the inductor as shown in Fig. 5.
In this method in order to control active and reactive power flow, it
is enough to choose ac real current (IP) and ac reactive current
(Iq) as the controlled variables. In this shame, the real and
reactive power flow in the inverter can be expressed as:
SPWM voltage controller and hysteresis current controller: It
is well known that in order to obtain an approximately sinusoidal voltage
in the output of the inverter, usually SPWM voltage control is used. At
this method the firing pulses to the switches are constructed using a
reference sinusoid compared against a triangular wave. Figure
2 shows the basic control diagram for a single phase SPWM system.
Also for current control in the inverter, hysteresis method is usually
used since its implementation is easy. At this method the current have
two limits, top current and bottom current.
||The basic control diagram for a single phase SPWM system in voltage
||General format of control system with SPWM method
||Hysteresis current controller connected to network
The distance of top and bottom limits call Hysteresis (H) bound. Figure
shows the hysteresis current control (Brod
and Novotny, 1985). As mentioned before, transient response and current
spectrum of the inverter are two basic criteria for comparing current control
and voltage control; and it will be done through simulation. Moreover, the
following comparisons also could be made:
||In voltage control, null wire connection isn?t need; this method
prefers when null wire may not be accessible. Also in Hysteresis current
control, multiple three harmonics are injected to the grid.
||Since in current controlled inverter, output current
is directly controlled, there is inherent over current protection;
but in voltage controlled inverters external hardware is needed for
over current protection.
||According to Eq. 1, in voltage controlled inverters P is directly
related to δ. Therefore influence of phase measurement errors
will be high (Kazmierkowski and Dzieniakowski, 1994).
||Under voltage control, variation in the real or reactive power commands
will result in a cross variance in the other. It means that there
is a cross-coupling between real and reactive power control; but in
current control, active and reactive power could be controlled separately.
||For satisfactory operation of current control, DC link voltage should
be more than maximum peak of the grid voltage; but in voltage control
this condition is not necessary.
||In voltage control, DC link harmonics will appear in the output
current of inverter; but since in current control, DC link voltage
only determines slope of rise and fall of the output current, DC link
voltage variation or harmonics will be rejected.
||In current control it is possible to produce harmonic currents of
the nonlinear load via inverter in order to grid current be sinusoidal
(Active Power Filtering).
||Usually in current control, DC link consists of two separate capacitors;
and extra hardware is needed for voltage balancing of these separate
THE SIMULATED SYSTEM
||Voltage controlled inverter: In Fig. 2, grid connected three phase
voltage controlled inverter is shown. For transient response analysis,
it is better to suppose that the DC link is not connected to a DC
Source and is charged via absorbing an active power from the grid.
DC link voltage controller is shown in Fig. 6.
||Current controlled inverter:
Figure 5 shows connection of three phase current
controlled inverter to the grid. As mentioned before, in this topology
null wire of the grid should be connected to the middle of DC Link.
Total DC link voltage controller and DC link voltage balancer are
shown in Fig. 7 and 8, respectively.
In both control methods, it is supposed that the grid characteristics,
coupling inductors and DC link capacitors are the same as shown below.
|Grid phase to neutral voltage:220 Vrms
|Coupling reactance: 500 μH
|DC link capacitors (C1 = C2):1000 μF
In voltage control, switching frequency of the inverter is 2 KHz. In
hysteresis current control, hysteresis band determines the switching frequency
of inverter. If fixed hysteresis band current control be used, switching
frequency of the inverter will not be constant during a cycle (Bose, 1990;
Kazemi and Jalilian, 2006). Some methods are proposed for controlling
of the hysteresis band, to obtain a fixed switching frequency. Implementation
of hysteresis band controller needs powerful processors and usually tends
to be unstable (Malesani and Tenti, 1990; Kale and Ozdemir, 2003). For
this reason in this study current control inverter is implemented with
fixed hysteresis band to obtain an average 2 KHz switching frequency.
||Current controlled GCI
||Capacitor voltage controller in voltage controlled GCI
||Capacitor total voltage controller in current controlled GCI
||Capacitor voltage balancer in current controlled GCI
||Capacitor voltage in Voltage Control GCI
||Capacitor voltage in Current Control GCI
Simulation results: Voltage controlled and current controlled GCI
have been simulated with matlab/simulink in static synchronous compensators
(STATCOM). First we consider the transient response. As mentioned before,
these voltage and current controllers are examined in reactive power compensator.
In these systems usually DC link do not connect to a separate DC power and
voltage of the DC link is controlled via absorption of a bit active power.
When system starts, DC link voltage charges from zero up to the reference
value. Figure 9 and 10 show this transient
response for voltage and current controllers (voltage controller has SPWM
controller and current controller has hysteresis controller). As it is obvious,
settling time of this response for voltage controller is about ten times
more than current controller.
||Transient response in voltage control GCI
||Transient response in current control GCI
In other word, transient response of current controller is faster than voltage
controller. Next method for comparing transient response of GCI is considering
situation of response when the GCI goes from capacitive mode to inductive
mode or vice versa. This is shown in Fig. 11
It confirms that transient response of current controller is faster than
voltage controller. High speed of response in the Hysteresis control at
this research is explained in below. Hysteresis band current control is
used very often because of its simplicity of implementation. Also, besides
fast response current loop, the method does not need any knowledge of load
parameters, current regulator techniques based on the Hysteresis control
together with switch logic are presented. However, the current control with
a Hysteresis band has the disadvantage that the PWM frequency varies within
a band because peak-to-peak current ripple is required to be controlled
at all points of the fundamental frequency wave.
The output current spectrum is another more important criterion for comparing.
This spectrum for voltage and current controller is shown in Fig.
13 and 14. However, in voltage controllers, current
harmonics appear around switching frequency and its integer multiples,
But in hysteresis current control, harmonics are spread in wide range
and also these harmonics appear in low frequencies too. This problem may
increase size and cost of output ac filter. Although some methods have
been proposed for solving of this problems in hysteresis current control,
but these methods usually suffer from instability problem and do not be
||Harmonic spectrum in SPWM control method
||Harmonic spectrum in Hysteresis control method
If switching frequency of the inverter increase, output current harmonics
may satisfy IEC harmonic standards and filtering may not be needed. At the
SPWM control if the frequency be high, the switches switching losses are
high and around switching frequency, the harmonic be high. Therefore at
the SPWM voltage control cannot increase frequency than switching frequency,
although the using PI controller can recovery this problem. From other differences
at Hysteresis current control can adjust Hysteresis bound and other parameters
cannot adjust, but SPWM voltage control can adjust amplitude and frequency
the triangle signal.
Among the various PWM techniques, the Hysteresis band current control
is used very often because of its simplicity of implementation. Also,
besides fast response current loop, the method does not need any knowledge
of load parameters. However, the current control with a fixed Hysteresis
band has the disadvantage that the PWM frequency varies within a band
because peak-to-peak current ripple is required to be controlled at all
points of the fundamental frequency wave. The method of adaptive Hysteresis-band
current control PWM technique where the band can be programmed as a function
of load to optimize the PWM performance is described in (Malesani and
Tenti, 1990). In this study has been result that the Hysteresis current
control is very simplicity of implementation.
In voltage controllers, current harmonics appear around switching frequency
and its integer multiples. But in hysteresis current control, harmonics
are spread in wide range and also these harmonics appear in low frequencies
too. This problem may increase size and cost of output ac filter (Kale
and Ozdemir, 2003). For transient voltage stability a large number of
disturbances, some with pre fault outages were considered (Kale and Ozdemir,
2003). Results of this study show harmonics appear in the spectrum of
frequency in the Fig. 13, 14.
The transient voltage recovery times were tabulated for the lowest bus voltage
to recover to 0.95 pu of its pre fault value. Transient security simulations
indicated that the duration of transient voltage recovery increased significantly
for some disturbances when Holly Power Plant was removed. The worst situation
occurred at Holly 138 kV bus, indicating it as the best location for installation
of dynamic shunt compensation (John et al., 2005). Figure
9, 10 shown at this research the STATCOM inject or absorb
the reactive power to network that the capacitor voltage is higher or lower
than network voltage and when the capacitor reach to steady state. Figure
11, 12 show change of state STATCOM that hysteresis current
control is faster than voltage controller.
This study has discussed the differences in two most popular control
methods of GCIs. The comparison is performed by simulating a typical reactive
power compensator. When switching frequency of the inverter is not high
enough (in large scale GCI), filtering of the output current may be difficult
and expensive. For this reason, in these applications, Voltage Control
Regardless of undesirable harmonic spectrum, if null wire connection
be accessible, Hysteresis current controller, which could be implemented
easily, will be a good choice because of its transient response.
The authors express their gratitude to the research council of Khaje
Nasirodin Toosi University for their financial supports.