Comparative analysis of different PV technologies under the tropical environments.

Area of Science:

• Renewable Energy
• Photovoltaics
• Materials Science

Background:

• Tropical regions present unique challenges for solar photovoltaic (PV) performance due to environmental factors.
• Evaluating diverse PV technologies is crucial for optimizing solar energy deployment in these climates.
• Understanding power production fluctuations is key for grid integration and stability.

Purpose of the Study:

• To compare the field performance of six different solar PV technologies under tropical conditions.
• To analyze power production variability and ramping behavior of these PV systems.
• To develop predictive models for PV system ramping using machine learning.

Main Methods:

• Utilized three years of performance data from a 1.2 MW experimental solar farm.
• Assessed performance using standard indices: Array Yield, Reference Yield, Capture Loss, Performance Ratio, and Efficiency Ratio.
• Employed probabilistic modeling (Generalized Logistic Distribution) and machine learning (ANN, SVM, kNN) to analyze power ramps.

Main Results:

• Amorphous silicon and HIT PV systems exhibited higher Performance and Efficiency Ratios with lower capture losses.
• Copper Indium Selenium (CIS) PV systems demonstrated the least power ramps, while HIT systems showed the most fluctuations.
• Machine learning models achieved over 96% Normalized Root Mean Square Error (NRMSE) in predicting PV system ramping.

Conclusions:

• Amorphous silicon and HIT technologies are promising for tropical solar PV applications.
• PV system ramping characteristics vary significantly, impacting grid management strategies.
• Accurate prediction of power fluctuations using machine learning is feasible and valuable for solar farm design and operation.