Page 132 - Renewable Energy Devices and System with Simulations in MATLAB and ANSYS

Overview of PV Maximum Power Point Tracking Techniques 119

in this section, require periodic reinitialization of their execution. This is indispensable in order
to be able to detect a possible displacement of the global MPP position, due to a change of either
the meteorological conditions or the shading pattern on the surface of the PV array (e.g., change
of the shadow shape due to sun movement). This results in power loss until the MPPT algorithm
converges to the new global MPP and also imposes the need of a high sampling rate in order to be
able to quickly detect the global MPP changes. The application of the PV array reconfiguration
method requires knowledge of the configuration of the PV source but provides a high efficiency,
since the PV system is able to produce more power than at the global MPP without reconfiguration.
Due to their inherent randomness, the evolutionary, stochastic, and chaos-based MPPT algorithms
do not guarantee convergence to the global MPP under any partial shading conditions. Thus, they
exhibit a lower energy yield. However, they have the advantage of not requiring detailed knowledge
of the PV system operational characteristics. The global MPPT methods that are based on numeri-
cal optimization algorithms either do not guarantee convergence to the global MPP or convergence
can be accomplished after a large number of search steps, both resulting in a reduction of the
PV-generated energy. The robustness of the PV array reconfiguration, evolutionary, stochastic, and
chaos- and numerical optimization–based MPPT algorithms may easily be degraded by external
disturbances affecting the correctness of the decisions taken during the global MPP tracking process
with respect to the direction toward which the global MPP resides. In such a case, possible wrong
estimations may only be recovered at the next reexecution of the corresponding MPPT algorithm.
The DMPPT techniques require knowledge of the PV source configuration but they are able to
extract the maximum possible energy from the PV source at the cost of a significantly higher hard-
ware complexity. Most of the other global MPPT algorithms, which have been presented in Section
5.4.6, require knowledge of the operational characteristics of the PV source. Also, since a search
process is applied periodically during their execution in order to detect possible changes of the
global MPP position (e.g., due to a change of meteorological conditions), the power produced by
the PV source is reduced and also a high sampling rate is required. Their performance in terms of the
remaining metrics considered in Table 5.3 depends on the specific technique applied in each case.
The evolutionary, stochastic, and chaos- and numerical optimization–based MPPT methods, as
well as some of the global MPPT approaches presented in Section 5.4.6, do not operate by using
information about the electrical characteristics of the PV source, so their accuracy is not affected by
the PV module aging. The robustness of DMPPT architectures to the PV module aging depends on
the type of method that has been employed for performing the MPPT process.
Among the global MPPT methods presented earlier, the PV array reconfiguration approach is the
least suitable for operation in combination with a constant-power-mode controller [5], due to the low
power adjustment resolution achieved by controlling the configuration of the PV modules within a
PV array, instead of directly controlling a power converter.

5.5 SIMULATION EXAMPLES
®
®
The MATLAB /Simulink model shown in Figure 5.21 (filename: “MPPT.mdl”) simulates the
operation of the P&O MPPT process applied to a buck-type DC/DC converter with a constant out-
put voltage. The buck converter is assumed to operate in the continuous conduction mode and its
model has been implemented in an embedded MATLAB function. Also, an ideal PV array has
®
been modeled as a controlled current source, with an output current calculated through an embed-
ded MATLAB function using the single-diode model, according to [15]. A series resistance is
®
connected to the output of the controlled current source, in order to form a nonideal PV array. The
parallel resistance of the PV array is assumed negligible. The P&O MPPT process has been imple-
®
mented in the corresponding embedded MATLAB function, based on [2]. Initially, the PV array
open-circuit voltage, short-circuit current, number of solar cells in series, solar cell temperature, and
ideality factor, as well as the P&O MPPT process perturbation step and the DC/DC converter output
®
®
voltage, must be set to the desired values. The “scope” of the MATLAB /Simulink model provides