English abstract
Renewable energy plays an increasingly important role in achieving
sustainable development, ensuring energy security, and controlling greenhouse gas
emissions. Solar energy is widely used source for electricity production due to
continuous advancements in solar power technology and decreasing costs.
Additionally, advancements in PV panel and inverter designs have made it possible to
invest in low-insolation areas.
Partial shading conditions (PSCs) can significantly reduce the energy output of
photovoltaic (PV) systems. Conventional and advanced MPPT systems often fail to
operate PV systems at their peak performance during PSCs due to the conduction mode
of bypassi diodes, which can trap the PV system at a low power point. However, global
peak searching tools can enable the operation of PV systems at the GMPP. It is
important to note that frequent use of these tools may decrease the output of PV
systems, as they force the system to operate outside its power region during the
scanning of the I-V curve to determine the GMPP. Therefore, global peak searching
tools should only be deployed when PSCs occur.
In this thesis, a simple and accurate method for detecting PSCs is proposed by
monitoring the sign of voltage changes. The method predicts a PSC if the sign of
successive voltage changes remains the same for a certain number of successive
changes. The proposed method was tested on two types of PV array configurations
(series and series-parallel) with various emulated shading patterns. The method
successfully and timely identified all emulated shading patterns and outperformed a
detection method based on monitoring the normalized change in power. It can be used
to trigger GMPP searching techniques for improving the output of PV systems under
PSCs.
Machine learning tools, such as fuzzy logic control and neural networks, have
been used in numerous studies to enhance conventional MPPT methods and optimize
PV systems under PSCs. However, these methods are complex to implement as they
require intensive calculations and exhibit limitations in ensuring GMPP operation
under all PSC conditions. To address this, a light and fast GMPP searching method
based on Bald Eagle Searching (BES) is proposed in this thesis. The BES method
locates the maximum value in three stages: selecting space, searching in space, and
swooping. The first stage of the BES method is utilized to design the proposed GMPP
method. MATLAB results demonstrate that the BES technique outperforms Cuckoo
Search (CS) and Particle Swarm Optimization (PSO) methods, reducing search time
significantly and successfully finding the GMPP in all simulated PSC cases. The
proposed method is simple, easy to implement, and only requires a single tuning
parameter, distinguishing it from PSO and CS methods. A validation method using a
real-time digital simulator (RTDS) confirms the capability of the proposed method to
find the GMPP within a reasonable time.
MPPT methods aim to maximize the output power of PV systems under
changing meteorological conditions. The performance of these methods depends on
the algorithm's complexity and the number of variable inputs used to obtain the MPP
value. However, they tend to oscillate around the MPP during steady-state operations,
resulting in energy waste. Moreover, traditional MPPT methods do not perform
optimally under PSCs. To address these issues, modifications are proposed to the
global maximum power point bald eagle search-based (GMPP BES) method to make
it function as an MPPT method as well. The modifications enable the GMPP BES
method to detect minor changes in insolation and temperature by monitoring the PV
array output voltage and trigger the search for the suitable MPP voltage accordingly.
The RTDS simulation results demonstrate the modified GMPP BES method's ability
to accurately and timely locate MPP values under the changes in insolation and
ambient temperature. The results show that the proposed method outperforms the
perturb and observe (P&O) method in responding to changes in insolation and ambient
temperature, and it effectively reaches correct MPP values with minimal oscillation
around the MPP. Therefore, the proposed method is considered a practical solution for
solar farms aiming to harvest large amounts of energy.
The suggested GMPP tracking method is implemented at the Sultan Qaboos
University Hybrid Station Lab using a real experimental test platform. Various
experimental conditions, including switching between partial shading and no shading
situations, as well as changes in temperature and irradiance, are conducted to evaluate
the performance of the proposed GMPP tracking method. Based on the experimental
results, the GMPP tracking method demonstrates excellent tracking performance
under diverse operating conditions.
