Amanah, Ludviatul (2026) Pengembangan Desain Kontrol Adaptif Arus Histerisis pada Bidirectional Voltage Source Inverter (VSI) Berbasis Artificial Neural Network (ANN). Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6022232006-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6022232006-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (14MB) | Request a copy |
Abstract
Peningkatan kebutuhan energi listrik dari pengguna kendaraan listrik serta kebutuhan akan energi terdistribusi mendorong pengembangan sistem konversi daya dua arah yang efisien, salah satunya adalah bidirectional VSI. Penerapan kontrol Fixed-HCC kurang efektif karena memiliki respon yang kurang cepat dan menghasilkan frekuensi switching yang tidak stabil dan menyebabkan menurunkan kualitas daya. Solusinya, penelitian ini mengusulkan kontrol Adaptive Hysteresis Current Control (AHCC) berbasis Artificial Neural Network (ANN) pada single phase bidirectional VSI dengan topologi full-bridge. Metode ini menggunakan ANN untuk menentukan lebar pita histerisis secara adaptif berdasarkan perubahan kondisi operasi sistem, yaitu tegangan DC-link (Vdc-link), tegangan grid sesaat (Vs), dan laju perubahan arus referensi inverter (m). Proses pelatihan ANN dilakukan secara offline menggunakan algoritma Levenberg-Marquardt dengan 4.092 sampel data yang diperoleh dari simulasi sistem AHCC konvensional. Hasil simulasi menunjukkan pada mode charging, rentang instantaneous frequency switching dengan kontrol AHCC-ANN adalah 5.263 Hz– 25 kHz lebih sempit jika dibandingkan dengan kontrol Fixed HCC sebesar 9,225 kHz –50 kHz. Sedangkan mode discharging, rentang kontrol AHCC-ANN juga lebih kecil yaitu sebesar 5.263 Hz–25 kHz dibandingkan kontrol Fixed-HCC sebesar 11.1 kHz–50 kHz. Selanjutnya frekuensi switching rata-rata dari kontrol AHCC-ANN pada mode charging dan discharging lebih mendekati frekuensi switching referensi 15 kHz yaitu sebesar 13,02 kHz dan 13,37 kHZ, dibandingkan dengan kontrol Fixed-HCC sebesar 19,96 kHz dan 20,25 kHz. Selain itu metode yang diusulkan mampu menekan harmonisa dan menghasilkan nilai THD arus grid dan arus inverter yang lebih rendah pada mode charging dan discharging, masing-masing 2,72% dan 2,09% dibandingkan dengan kontrol Fixed-HCC, masing-masing 3,20% dan 2,31%. Selain itu sistem mampu merespon perubahan mode chargiung ke discharging dengan cepat dan mampu beroprasi sesuai dengan kondisi SOC baterai serta dapat mengalirkan daya baterai menuju beban yang bervariasi dengan tetap menjaga kestabilan tegangan DC-link agar menghasilakan adaptive hysteresis band dan memungkinkan arus inverter mengikuti arus referensi dengan tepat.
===================================================================================================================================
The increasing need for electrical energy from electric vehicle users and the need for distributed energy encourage the development of efficient two-way power conversion systems, one of which is a bidirectional VSI. The application of Fixed-HCC control is less effective because it has a slower response and results in unstable switching frequencies, leading to decreased power quality. The solution, this study proposes Adaptive Hysteresis Current Control (AHCC) based on Artificial Neural Network (ANN) on a single-phase bidirectional VSI with a full-bridge topology. This method uses an ANN to determine the hysteresis bandwidth adaptively based on changes in system operating conditions, namely DC-link voltage (Vdc-link), instantaneous grid voltage (Vs), and inverter reference current change rate (m). The ANN training process was carried out offline using the Levenberg-Marquardt algorithm with 4,092 data samples obtained from conventional AHCC system simulations. The simulation results showed that in the charging mode, the instantaneous frequency switching range with AHCC-ANN control was narrower, 5.263 Hz – 25 kHz, when compared to the Fixed HCC control of 9,225 kHz – 50 kHz. As for the discharging mode, the AHCC-ANN control range is also smaller, which is 5,263 Hz–25 kHz compared to the Fixed-HCC control of 11.1 kHz–50 kHz. Furthermore, the average switching frequency of the AHCC-ANN control in charging and discharging modes is closer to the 15 kHz reference switching frequency of 13,02 kHz and 13,37 kHz, compared to the Fixed-HCC control of 19,96 kHz and 20,25 kHz. In addition, the proposed method can suppress harmonization and produce a lower THD value of grid current and inverter current in charging and discharging modes, 2.72% and 2.09%, respectively, compared to Fixed-HCC controls, 3.20% and 2.31%, respectively. In addition, the system can respond quickly to the change in charge mode to discharging. It can operate according to the SOC conditions of the battery, and can deliver battery power to varying loads while maintaining the stability of the DC-link voltage to produce an adaptive hysteresis band and allow the inverter current to follow the reference current appropriately.
Actions (login required)
 |
View Item |