Laporkan Masalah

Optimasi Kinerja PLTS Nusa Penida Melalui Analisa Kehilangan Energi dan Albedo Enhancement

REZA ADISETIA SAPUTRA, Dr. Eng. Ir. Dwi Joko Suroso, S.T., M.Eng., IPP.

2026 | Tesis | MAGISTER TEKNIK FISIKA

Pembangkit Listrik Tenaga Surya (PLTS) merupakan salah satu teknologi utama dalam mendukung transisi energi bersih di Indonesia. Namun, kinerja PLTS dapat menurun akibat pengaruh kondisi lingkungan, topografi, serta ketidakseimbangan performa antar inverter. Permasalahan tersebut ditemukan pada PLTS Hybrid Nusa Penida yang memiliki kondisi lahan berbukit dan tidak merata. Penelitian ini bertujuan untuk menganalisis kehilangan energi dan mengoptimalkan kinerja PLTS Nusa Penida melalui pendekatan machine learning, analisis SHAP, dan metode Albedo enhancement. Data yang digunakan merupakan data operasional aktual tahun 2024 dengan resolusi 30 menit, meliputi produksi inverter, iradiasi, temperatur, dan kecepatan angin. Hasil penelitian menunjukkan bahwa model XGBoost memberikan performa terbaik dalam mendeteksi kondisi underperformance inverter. Analisis SHAP mengidentifikasi iradiasi, temperatur, waktu operasi, dan karakteristik inverter sebagai faktor utama yang memengaruhi performa sistem. Inverter 16 terdeteksi sebagai inverter dengan tingkat underperformance tertinggi yang diduga dipengaruhi oleh kondisi topografi dan self-shading. Implementasi Albedo enhancement menggunakan material reflektif mampu meningkatkan iradiasi pantul dan berpotensi meningkatkan produksi energi pada area pengujian. Dengan demikian, kombinasi machine learning, SHAP, dan Albedo enhancement dapat digunakan sebagai pendekatan untuk mendukung optimasi kinerja dan operasional PLTS.

Solar Photovoltaic Power Plant (PV Power Plant) is one of the key technologies supporting the clean energy transition in Indonesia. However, PV system performance may decrease due to environmental conditions, topography, and performance imbalance among inverters. This issue was identified at the Nusa Penida Hybrid PV Power Plant, which is located on hilly and uneven terrain. This study aims to analyze energy loss and optimize the performance of the Nusa Penida PV Power Plant through a machine learning approach, SHAP analysis, and the Albedo enhancement method. The data used in this study consist of actual operational data from 2024 with a 30-minute resolution, including inverter production, irradiance, temperature, and wind speed. The results show that the XGBoost model provides the best performance in detecting inverter underperformance conditions. SHAP analysis identifies irradiance, temperature, operating time, and inverter characteristics as the main factors influencing system performance. Inverter 16 was detected as the inverter with the highest level of underperformance, which is suspected to be influenced by topographical conditions and self-shading. The implementation of Albedo enhancement using reflective material was able to increase reflected irradiance and has the potential to improve energy production in the test area. Therefore, the combination of machine learning, SHAP, and Albedo enhancement can be used as an approach to support PV performance and operational optimization.

Kata Kunci : PLTS, Machine learning, XGBoost, SHAP, Underperformance Inverter, Energy Loss, Albedo enhancement, Nusa Penida.

  1. S2-2026-548350-abstract.pdf  
  2. S2-2026-548350-bibliography.pdf  
  3. S2-2026-548350-tableofcontent.pdf  
  4. S2-2026-548350-title.pdf