Rancang Bangun Double Active Continuous Single Axis Solar Tracking Berbasis Kontrol ANFIS Type-2 pada Photovoltaic

Wulandari, Vinni Nuriski (2026) Rancang Bangun Double Active Continuous Single Axis Solar Tracking Berbasis Kontrol ANFIS Type-2 pada Photovoltaic. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009221031_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Perkembangan konsep smart farming mendorong kebutuhan akan sumber energi mandiri yang andal dan efisien untuk mengoperasikan berbagai perangkat pertanian modern. Photovoltaic merupakan salah satu solusi yang menjanjikan, namun efisiensi konversi energinya sangat dipengaruhi oleh orientasi photovoltaic terhadap arah datangnya cahaya matahari. Penelitian ini mengembangkan sistem double active continuous single-axis solar tracker berbasis kontrol Adaptive Neuro-Fuzzy Inference System (ANFIS) Type-2 pada dua photovoltaic monokristalin 315 WP sebagai sumber energi mandiri untuk smart farming system. Sistem dirancang dengan aktuator ganda berupa motor DC power window 12 V yang dikendalikan driver BTS7960, mikrokontroler ESP32, sensor LDR sebagai masukan arah cahaya pada mode aktif, sensor MPU6050 sebagai umpan balik posisi photovoltaic sekaligus dasar mode pasif berdasarkan Sun Position Algorithm (SPA), serta sensor PZEM-017 untuk pemantauan daya DC. Kontrol ANFIS Type-2 menggunakan dua masukan berupa error dan delta error tegangan LDR, sembilan fungsi keanggotaan generalized bell, serta keluaran berupa sinyal PWM untuk menggerakkan motor secara kontinu mengikuti pergerakan matahari. Hasil evaluasi model menunjukkan performa yang sangat baik dengan R² sebesar 0,9990, MAE 2,2683, RMSE 4,2465, MSE 18,0324, MAPE 6,17%, dan mean error 0,0000, yang menunjukkan kemampuan prediksi model sangat akurat tanpa bias sistematis. Pengujian kendali juga menunjukkan ANFIS Type-2 memberikan akurasi tracking yang lebih baik dibandingkan ANFIS Type-1, dengan nilai MAPE menurun dari 4,90% menjadi 3,80% pada PV1 dan dari 5,89% menjadi 4,69% pada PV2, disertai penurunan MAE, MSE, dan RMSE pada kedua photovoltaic. Pada uji set point, ANFIS Type-2 mencapai settling time 4 detik atau 75% lebih cepat dibandingkan ANFIS Type-1 yang memerlukan 16 detik, dengan steady-state error sebesar 0,05% pada PV1 dan 0,11% pada PV2 serta maximum overshoot di bawah 2%. Pengujian performansi menunjukkan sistem solar tracker menghasilkan energi bersih sebesar 1.505,51 Wh atau meningkat 32,57% dibandingkan PV fixed, sedangkan sistem switching menghasilkan energi bersih sebesar 1.629,74 Wh atau meningkat 43,51%, sehingga membuktikan bahwa strategi switching aktif–pasif mampu memberikan performa energi terbaik sebagai penyedia energi mandiri untuk aplikasi smart farming.
===================================================================================================================================
The development of smart farming has increased the demand for reliable and efficient autonomous energy sources to operate modern agricultural systems. Photovoltaic (PV) photovoltaics are a promising solution; however, their energy conversion efficiency strongly depends on their orientation relative to the sun. This study develops a double active continuous single-axis solar tracker based on an Interval Type-2 Adaptive Neuro-Fuzzy Inference System (ANFIS) controller using two 315 Wp monocrystalline PV modules as an autonomous power source for smart farming applications. The proposed system employs dual 12 V DC power window motors driven by BTS7960 motor drivers, an ESP32 microcontroller, Light Dependent Resistor (LDR) sensors for active tracking, MPU6050 sensors for photovoltaic position feedback and passive tracking based on the Sun Position Algorithm (SPA), and a PZEM-017 sensor for DC power monitoring. The ANFIS Type-2 controller utilizes two inputs, namely LDR voltage error and delta error, nine generalized bell membership functions, and a PWM output to continuously control photovoltaic movement. Model evaluation demonstrates excellent prediction performance with R² = 0.9990, MAE = 2.2683, MSE = 18.0324, RMSE = 4.2465, MAPE = 6.17%, and a mean error of 0.0000, indicating highly accurate prediction capability without systematic bias. Compared with ANFIS Type-1, the proposed controller consistently achieved lower MAE, MSE, RMSE, and MAPE values, with MAPE decreasing from 4.90% to 3.80% on PV1 and from 5.89% to 4.69% on PV2, demonstrating improved tracking accuracy. In field set-point tests, the proposed controller reached steady state within 4 s, representing a 75% reduction in settling time compared with ANFIS Type-1 (16 s), while maintaining steady-state errors of 0.05% and 0.11% for PV1 and PV2, respectively, with maximum overshoot below 2%. Performance evaluation showed that the conventional active solar tracker produced a net energy of 1,505.51 Wh, representing a 32.57% increase over the fixed PV system, whereas the proposed active-passive switching strategy generated 1,629.74 Wh, corresponding to a 43.51% improvement. These results demonstrate that the proposed switching-based ANFIS Type-2 solar tracker provides superior tracking accuracy and energy harvesting performance, making it a reliable autonomous energy solution for smart farming applications.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	ANFIS Type-2, Double Active, Photovoltaic, Single-Axis Solar Tracker, Smart Farming
Subjects:	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:	Vinni Nuriski Wulandari
Date Deposited:	03 Aug 2026 07:07
Last Modified:	03 Aug 2026 07:07
URI:	http://repository.its.ac.id/id/eprint/142243

Actions (login required)

View Item