Farrukh, Muhammad Aqil Farrukh (2026) Evaluasi Fitur Handcrafted Dan Model Speech Pralatih Untuk Estimasi Kepribadian Big Five Berbasis Audio Pada Protokol Strict Split Serta Analisis Fine-Tuning Berbasis LoRA. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5025221158-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5025221158-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (13MB) | Request a copy |
Abstract
Perkembangan sistem energi terdistribusi memerlukan konverter yang mampu mengatur aliran daya dua arah sekaligus menyediakan pemantauan kondisi operasi secara real-time. Penelitian terdahulu telah membahas konverter Dual Active Bridge (DAB), pengendalian phase-shift, dan SCADA pada sistem DC, tetapi penerapan konverter DAB tegangan rendah dengan kontrol Fuzzy-PI yang terintegrasi langsung dengan SCADA masih perlu dikaji melalui implementasi dan pengujian yang menyeluruh. Penelitian ini bertujuan merancang, mengimplementasikan, dan mengevaluasi Bidirectional DC-DC Converter DAB pada mode 12 V ke 24 V dan 24 V ke 12 V. Pelaksanaan penelitian meliputi studi pustaka, perancangan perangkat keras dan perangkat lunak, simulasi DAB, implementasi prototipe, karakterisasi sensor, validasi PWM dan phase-shift, pengujian kontrol dan efisiensi, serta pengujian komunikasi. Pembahasan dibagi menjadi dua fokus. (1) Pada sistem daya DAB, STM32G4 dan HRTIM digunakan untuk menghasilkan delapan PWM komplementer pada 20,16 kHz dengan dead-time 504 ns, sedangkan Fuzzy-PI mengatur pergeseran fasa berdasarkan umpan balik arus. Model simulasi menggunakan parameter yang sama dengan firmware dan dibandingkan dengan gelombang arus aktual. Sensor INA219 menghasilkan rata-rata error tegangan 0,48% pada sisi 12 V dan 0,28% pada sisi 24 V, serta rata-rata error arus 0,32% dan 0,29% setelah titik arus sangat rendah dikecualikan. Pada setpoint 650 mA dan beban 34 $\Omega$, kontrol mencapai tegangan tunak 21,8 V terhadap target 22,1 V, error tunak 1,36%, overshoot 0%, rise time 12,5 ms, dan settling time 25,0 ms. Efisiensi maksimum mencapai 85,05% pada arah 12 V ke 24 V dan 85,26% pada arah 24 V ke 12 V. (2) Pada sistem SCADA, modul WIZnet, Modbus TCP, UART, dan antarmuka web digunakan untuk monitoring telemetri, pencatatan data, pemilihan arah daya, dan pengiriman setpoint. Pengujian monitoring menghasilkan 599 sampel selama 119,77 s dengan success rate API dan connected rate 100%, tanpa putus koneksi, serta update rate 4,994 Hz. Namun, pembaruan perintah yang terverifikasi hingga RAM STM32 baru mencapai 66,67%. Dengan demikian, sistem daya DAB telah mampu melakukan konversi dua arah dengan respons cepat dan efisiensi yang relatif seimbang, sedangkan SCADA telah memenuhi fungsi monitoring tetapi jalur perintah Modbus TCP–WIZnet–UART masih perlu disempurnakan agar deterministik.
================================================================================================================================
The development of distributed energy systems requires converters capable of regulating bidirectional power flow while providing real-time monitoring of operating conditions. Previous studies have addressed Dual Active Bridge (DAB) converters, phase-shift control, and SCADA for DC systems; however, a low-voltage DAB converter with Fuzzy-PI control directly integrated with SCADA still requires comprehensive implementation and experimental evaluation. This study aims to design, implement, and evaluate a DAB Bidirectional DC-DC Converter operating from 12 V to 24 V and from 24 V to 12 V. The research stages include a literature review, hardware and software design, DAB simulation, prototype implementation, sensor characterization, PWM and phase-shift validation, control and efficiency tests, and communication tests. The discussion is divided into two focuses. (1) For the DAB power system, an STM32G4 and HRTIM generate eight complementary PWM signals at 20.16 kHz with a 504 ns dead time, while the Fuzzy-PI controller regulates the phase displacement using current feedback. The simulation uses the same parameters as the firmware and is compared with the measured inductor-current waveform. INA219 characterization yields average voltage errors of 0.48% on the 12 V side and 0.28% on the 24 V side, and average current errors of 0.32% and 0.29%, respectively, after excluding the very-low-current point. At a 650 mA setpoint with a 34 Ω load, the controller reaches 21.8 V against a 22.1 V target, with a 1.36% steady-state error, 0% overshoot, a 12.5 ms rise time, and a 25.0 ms settling time. Maximum efficiency reaches 85.05% in the 12 V-to-24 V direction and 85.26% in the 24 V-to-12 V direction. (2) For the SCADA system, a WIZnet module, Modbus TCP, UART, and a web interface provide telemetry monitoring, data logging, power-flow direction selection, and setpoint transmission. The monitoring test records 599 samples over 119.77 s with 100% API success and connection rates, no disconnection events, and a 4.994 Hz update rate. However, only 66.67% of command updates are verified at the STM32 RAM. Therefore, the DAB power system achieves bidirectional conversion with fast response and relatively balanced efficiency, while SCADA fulfills the monitoring function but the Modbus TCP–WIZnet–UART command path requires further improvement to become deterministic.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | audio-only, Big Five, ChaLearn First Impressions V2, eGeMAPS, HuBERT,LoRA, self-supervised learning, strict split, WavLM. |
| Subjects: | Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA9.58 Algorithms |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | Muhammad Aqil Farrukh |
| Date Deposited: | 28 Jul 2026 01:42 |
| Last Modified: | 28 Jul 2026 01:42 |
| URI: | http://repository.its.ac.id/id/eprint/138000 |
Actions (login required)
 |
View Item |