Pengembangan Adaptive Single-Diode Model Berbasis Estimasi Dinamis Parameter Internal Menggunakan Data Operasional REIDI ITS

Simatupang, Benny (2026) Pengembangan Adaptive Single-Diode Model Berbasis Estimasi Dinamis Parameter Internal Menggunakan Data Operasional REIDI ITS. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009211051-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Model Single Diode Model (SDM) konvensional yang umum digunakan dalam pemodelan photovoltaic (PV), seperti PVsyst dan CEC model, menetapkan resistansi seri (Rs) dan resistansi paralel (Rp) sebagai konstanta hasil kalibrasi pada Standard Test Condition (STC), meskipun studi eksperimental menunjukkan bahwa kedua parameter tersebut sesungguhnya bergantung pada kondisi irradiance (G) dan temperatur (T) operasi. Penelitian ini mengembangkan Adaptive Single Diode Model (Adaptive SDM) yang memodelkan Rs dan Rp sebagai fungsi dinamis Rs(G,T) dan Rp(G,T), diestimasi dari data operasional PLTS REIDI ITS Cluster 3 periode Juli–September 2025 menggunakan pendekatan estimasi parameter berbasis window waktu (60 menit) dengan regularisasi Tikhonov, dilanjutkan dengan regresi polinomial kuadratik terhadap 582 titik data hasil estimasi, dan divalidasi secara komparatif terhadap SDM konvensional pada empat komponen kelistrikan (tegangan string, daya inverter, daya cluster, dan cross-check SCADA). SDM konvensional tervalidasi baik pada kondisi STC (kesalahan maksimum 0,046%) dan pada level daya cluster (R²=0,9585; nMAE=4,59%), namun menunjukkan keterbatasan pada level tegangan string (R² negatif) akibat asumsi titik operasi maximum power point (MPP) ideal. Hasil validasi Adaptive SDM menunjukkan pola campuran—perbaikan signifikan pada komponen tegangan string namun sedikit penurunan pada komponen daya cluster dan analisis sensitivitas yang mengisolasi kontribusi masing-masing parameter mengungkapkan bahwa perbaikan akurasi terutama didorong oleh dinamisasi Rp, sementara dinamisasi Rs justru menurunkan akurasi pada kondisi irradiance tinggi akibat lonjakan nilai yang belum sepenuhnya dapat dijelaskan secara fisis. Skenario Rp dinamis saja terbukti menjadi satu-satunya konfigurasi yang secara konsisten mengungguli SDM konvensional pada seluruh komponen yang diuji, sehingga penelitian ini merekomendasikan pendekatan tersebut untuk aplikasi praktis yang memprioritaskan akurasi estimasi daya.
===================================================================================================================================
Conventional Single Diode Model (SDM) approaches commonly used in photovoltaic (PV) modeling, such as PVsyst and the CEC model, treat series resistance (Rs) and parallel resistance (Rp) as constants calibrated under Standard Test Conditions (STC), even though experimental studies show that both parameters actually vary with operating irradiance (G) and temperature (T). This study develops an Adaptive Single Diode Model (Adaptive SDM) that represents Rs and Rp as dynamic functions, Rs(G,T) and Rp(G,T), estimated from operational data of the REIDI ITS Cluster 3 PV plant collected from July to September 2025 using a sliding time-window parameter estimation approach (60-minute windows) with Tikhonov regularization, followed by quadratic polynomial regression over 582 estimated data points, and validated comparatively against the conventional SDM across four electrical components (string voltage, inverter power, cluster power, and SCADA cross-check). The conventional SDM was successfully validated under STC conditions (maximum error of 0.046%) and at the cluster power level (R²=0.9585; nMAE=4.59%), but showed limitations at the string voltage level (negative R²) due to its assumption of an ideal maximum power point (MPP) operating condition. Validation of the Adaptive SDM revealed a mixed pattern—significant improvement in the string voltage component alongside a slight decrease in the cluster power component—and a sensitivity analysis isolating each parameter's contribution further showed that the accuracy gains were primarily driven by making Rp dynamic, while making Rs dynamic instead degraded accuracy under high-irradiance conditions due to a sharp increase not yet fully explainable on physical grounds. The Rp-dynamic-only configuration proved to be the only scenario that consistently outperformed the conventional SDM across all evaluated components, leading this study to recommend that configuration for practical applications prioritizing power estimation accuracy.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Single-Diode Model, Adaptive SDM, Resistansi Seri, Resistansi Pararel, Estimasi Parameter Dinamis, PLTS REIDI ITS,Series Resistance, Parallel Resistance, Dynamic Parameter Estimation, REIDI ITS PV Plant
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1001 Production of electric energy or power T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1056 Solar power plants. Ocean thermal power plants T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1087 Photovoltaic power generation
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Benny Simatupang
Date Deposited:	05 Aug 2026 08:55
Last Modified:	05 Aug 2026 08:55
URI:	http://repository.its.ac.id/id/eprint/143824

Actions (login required)

View Item