KLASIFIKASI KONDISI SHADING DAN NON-SHADING PADA PANEL SURYA 240W MENGGUNAKAN METODE RANDOM FOREST

Ananto, Fany Azzahra (2026) KLASIFIKASI KONDISI SHADING DAN NON-SHADING PADA PANEL SURYA 240W MENGGUNAKAN METODE RANDOM FOREST. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221107-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Kondisi shading dapat menyebabkan perubahan karakteristik keluaran dan penurunan performa sistem photovoltaic (PV). Proyek akhir ini bertujuan menerapkan sistem klasifikasi kondisi operasional pada dua panel surya monocrystalline dengan total daya 240 Wp yang dipasang secara seri. Klasifikasi dilakukan menggunakan metode Random Forest berdasarkan parameter tegangan masukan (V_in), arus masukan (I_in), dan daya masukan (P_in) sebagai fitur klasifikasi. Data diperoleh melalui pengambilan kurva I–V dan P–V menggunakan metode PWM sweeping dengan variasi duty cycle 0% hingga 100% pada kondisi Non-shading, Light shading, dan Heavy shading. Data hasil pengukuran terlebih dahulu melalui tahap preprocessing dan augmentasi menggunakan gaussian noise untuk meningkatkan jumlah dan variasi sampel pada setiap kelas. Dataset akhir terdiri atas 100 sampel yang dibagi menjadi 80% data pelatihan dan 20% data pengujian untuk pembentukan dan evaluasi model Random Forest. Hasil pengujian menunjukkan bahwa model memperoleh accuracy sebesar 85%, weighted precision sebesar 90%, weighted recall sebesar 85%, dan weighted F1-score sebesar 84,45%. Berdasarkan confusion matrix, sebanyak 17 dari 20 data pengujian berhasil diklasifikasikan dengan benar, dengan seluruh kesalahan klasifikasi terjadi pada kelas Light shading yang diprediksi sebagai Heavy shading.
====================================================================================================================================
Shading conditions can alter the output characteristics and reduce the performance of photovoltaic (PV) systems. This study aims to implement an operating-condition classification system for two monocrystalline solar panels connected in series with a total capacity of 240 Wp. Classification is performed using the Random Forest method based on input voltage (V_in), input current (I_in), and input power (P_in) as classification features. The data are obtained from I–V and P–V curves acquired using the PWM sweeping method by varying the duty cycle from 0% to 100% under Non-shading, Light shading, and Heavy shading conditions. The measured data are then subjected to preprocessing and augmentation using Gaussian noise to increase the number and variability of samples in each class. The final dataset consists of 100 samples, which are divided into 80% training data and 20% testing data for the development and evaluation of the Random Forest model. The test results show that the model achieves an accuracy of 85%, a weighted precision of 90%, a weighted recall of 85%, and a weighted F1-score of 84.45%. Based on the confusion matrix, 17 out of 20 testing samples are correctly classified, with all misclassifications occurring in the Light shading class, which is predicted as Heavy shading.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Sistem Photovoltaic, Shading, Klasifikasi Shading, PWM Sweeping, Random Forest, Photovoltaic System, Shading, Shading Classification, PWM Sweeping, Random Forest
Subjects:	L Education > L Education (General) L Education > LB Theory and practice of education T Technology > T Technology (General) T Technology > T Technology (General) > T56.8 Project Management T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.74 Linear programming T Technology > TK Electrical engineering. Electronics Nuclear engineering T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK531 Current and voltage waveforms
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Fany Azzahra Ananto
Date Deposited:	10 Aug 2026 01:42
Last Modified:	10 Aug 2026 01:42
URI:	http://repository.its.ac.id/id/eprint/144220

Actions (login required)

View Item