Setiariawan, Telly Fahrul (2017) Simulasi Kontrol Tegangan pada AVR Generator Berbasis Jaringan Saraf Tiruan. Undergraduate thesis, Institut Teknnologi Sepuluh Nopember.
![[thumbnail of TELLY FAHRUL S_4215105002.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
TELLY FAHRUL S_4215105002.pdf Restricted to Repository staff only Download (1MB) | Request a copy |
Abstract
Generator merupakan sumber tenaga listrik untuk mensuplai kebutuhan listrik di kapal seperti penghasil panas untuk tangka, penghasil gerak untuk pompa – pompa atau kompresor, kemudian untuk penerangan dikapal dan lain lain. Begitu besarnya kebutuhan listrik pada kapal oleh sebab itu maka tegangan pada generator harus di atur sedemikian rupa agar tetap stabil. Kestabilan tegangan generator di atur oleh AVR (Automatic Voltage Regulator). Akurasi AVR konvensional yang ada di kapal mempunyai nilai eror ±2.5% dengan waktu 1.5 detik untuk menstabilkan tegangan. Pada tugas akhir ini komponen thyristor pada AVR akan di gantikan dengan control Jaringan Saraf Tiruan yang dianalisa oleh software Mathlab Simulink dengan harapan hasil yang lebih baik. Kontrol Jaringan Saraf Tiruan membutuhkan data pembelajaran dimana data tersebut digunakan untuk menyusun JST. Semakin banyak data yang didapat makan makin baik konstrol JST. Susunan JST dapat dilakukan dengan cara menyusun node yaitu dalam arti menentukan nilai PI dari fariasi tiap beban yang tebaik sehingga menghasilkan respon tegangan paling baik dan eror tegangan paling sedikit. Kemudian menyusun jaringan dalam arti menyusun berapa jumlah layer yang digunakan sebelum masuk ke langkah training. Dalam hal ini penulis menggunakan nilai default dari program Mathlab Simulink untuk menentukan desain jaringan. Setelah mendesain jaringan langkah selanjutnya yaitu melakukan desain pelatihan dimana nilai pembelajaran atau nilai ambang dalam arti nilai training, validation dan testing dimasukan secara berurutan yaitu 70%, 15% dan 15% dari 900009 sampel data. Desain pelatihan tersebut dirtaining sehingga menghasilan data atau grafik nilai eror terbaik. Dalam hal ini pengulangan training mencapai 127 pengulangan dengan nilai eror terbaik yaitu 0.011703. Dari hasil analisa simulasi pengaturan tegangan menggunakan AVR (Automatic Voltage Regulator) berbasis Jaringan Saraf Tiruan didapat nilai eror dari fariasi beban 720 kW, 992 kW, 608 kW, 240 kW dan 880 kW secara berurutan yaitu 0.15%, 0.525%, 1.0165%, 0.6509% dan 0.319%. Dimana nilai tersebut jauh dari nilai ambang yaitu 2.5%. Nilai respon waktu dari hasil analisa simulasi pengaturan tegangan menggunakan AVR (Automatic Voltage Regulator) berbasis Jaringan Saraf Tiruan didapat nilai respon tegangan dari fariasi beban 720 kW, 992 kW, 608 kW, 240 kW dan 880 kW secara berurutan yaitu 1.031 s, 1.041s, 0.9098s, 0.7346s dan 1.043s. nilai tersebut jauh dari nilai ambang yang diinginkan yaitu kurang dari 1.5s
============================================================================================
Generator is a source of electric power to supply the needs of electricity on the ship such as heat producers for tangka, generator motion for pumps or compressors, then for lighting on the vessel and others. So great the need for electricity on the ship therefore the voltage on the generator must be set in such a way as to remain stable. The stability of the generator voltage is set by the AVR (Automatic Voltage Regulator). Accuracy of conventional AVR on board has eror ± 2.5% with 1.5 seconds to stabilize voltage. In this final project thyristor component on AVR will be replaced with Neural Network control which analyzed by Mathlab Simulink software in hope of better result. Artificial Neural Network Control requires learning data where the data is used to compile an NN. The more data you get the better the JST constrol. The arrangement of NN can be done by arranging the node that is in the sense of determining the value of PI from the fariation of each of the best load so as to produce the best voltage response and the least voltage error. Then compile the network in the sense of compiling how many layers are used before going into training step. In this case the author uses the default value of the Mathlab Simulink program to determine the network design. After designing the network the next step is to do the training design where the learning value or the threshold value in terms of training value, validation and testing are included in sequence that is 70%, 15% and 15% of 900009 data samples. The training design is dirtaining so that it produces the best value data or graph of errors. In this case the repetition of training reached 127 repetitions with the best error value of 0.011703. From the simulation result of voltage regulation using AVR (Automatic Voltage Regulator) based on Artificial Neural Network obtained error value from fariation load 720 kW, 992 kW, 608 kW, 240 kW and 880 kW respectively ie 0.15%, 0.525%, 1.0165%, 0.6509 % And 0.319%. Where the value is far from the threshold value of 2.5%. The response time value of the simulation result of voltage regulation using AVR (Automatic Voltage Regulator) based on Artificial Neural Network obtained value of voltage response from load fariation 720 kW, 992 kW, 608 kW, 240 kW and 880 kW respectively ie 1,031 s, 1,041s, 0.9098s, 0.7346s and 1.043s. The value is far from the desired threshold value of less than 1.5s.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSSP 623.850 3 Set s |
| Uncontrolled Keywords: | Voltage, AVR, Generator, Neural network, Tegangan, Jaringan saraf tiruan |
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK201 Electric Power Transmission T Technology > TK Electrical engineering. Electronics Nuclear engineering T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.5 Modulation (Electronics), Demodulation (Electronics) V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering |
| Divisions: | Faculty of Marine Technology (MARTECH) > Marine Engineering > 36202-(S1) Undergraduate Thesis |
| Depositing User: | Telly Fahrul Setiariawan |
| Date Deposited: | 16 Jan 2018 05:08 |
| Last Modified: | 16 Jan 2018 05:08 |
| URI: | http://repository.its.ac.id/id/eprint/45238 |
Actions (login required)
 |
View Item |