Pengendalian Kecepatan Sistem Propulsi Hybrid Pada Model Kapal Trimaran Berbasis Neural Network

Munif, Muhammad Azmi Naufal (2021) Pengendalian Kecepatan Sistem Propulsi Hybrid Pada Model Kapal Trimaran Berbasis Neural Network. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

Text 04211640000051-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Kapal trimaran merupakan salah satu jenis kapal yang mampu berlayar dengan kecepatan service yang tinggi. Untuk mampu berlayar secara efisien, maka salah satu sistem propulsi yang dapat digunakan adalah sistem propulsi hybrid. Maka dari itu, kontrol sistem propulsi dibutuhkan untuk mengatur cara kerja dari sistem propulsi hybrid sehingga nantinya sistem propulsi dapat berjalan secara otomatis atau autospeed. Salah satu system kontrol yang memiliki keunggulan dalam mengolah data yang bersifat non-linear dan memiliki jumlah data yang banyak adalah Neural Network. Penelitian ini dilakukan untuk membuat kendali sistem propulsi hybrid berbasis neural network pada model kapal trimaran. Untuk membuat sistem kendali berbasis neural network, maka dibutuhkan data pendukung berupa output power serta kecepatan yang dihasilkan pada berbagai kondisi sistem propulsi hybrid selama beroperasi. Hasil yang didapatkan berupa sistem kendali neural network yang mampu memberikan nilai output kendali dengan nilai error yang kecil saat diberikan nilai input tertentu dengan nilai rata-rata eror hasil validasi pada throttle motor BLDC, throttle motor bakar, dan total output power masing-masing adalah 0.0064%, 0.0036%, dan 0.0068%. Di sisi lain, berdasarkan kurva perbandingan antara kecepatan dan power output sistem propulsi hybrid, terdapat temuan penelitian yang menunjukkan bahwa peningkatan nilai output power yang dihasilkan tidak selalu diikuti dengan peningkatan kecepatan kapal.
=============================================================================================
Trimaran ship is a type of ship that capable of sailing at high service speeds. To be able sail efficiently, one of the propulsion systems type that can be used is the hybrid propulsion system. Therefore, propulsion control system is needed to manage the workings of the hybrid propulsion system, so that the propulsion system can run automatically or autospeed. One of control system that has advantages in processing non-linear data and has a large amount of data is a Neural Network. This research is conducted to create a neural network for hybrid propulsion control system on trimaran ship models. To create a control system based on a neural network, supporting data is needed in the form of output power and the resulting speed under various conditions of the hybrid propulsion system during operation. The results is neural network control system that is able to provide control output values with a small error value when given a certain input value with the average error value of validation results on BLDC motor throttle, combustion motor throttle, and total output power respectively 0.0064 %, 0.0036%, and 0.0068%. On the other hand, based on the comparison curve between the speed and power output of a hybrid propulsion system, there is research finding that shows that an increase in the value of the output power produced not always followed by an increase in ship speed.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Kapal Trimaran, Sistem Propulsi Hybrid, Neural Network, Autospeed, Trimaran Ship, Hybrid Propulsion System, Neural Network, Autospeed
Subjects:	V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM365 Remote submersibles. Autonomous vehicles. V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM773 Ship propulsion, Electric
Divisions:	Faculty of Marine Technology (MARTECH) > Marine Engineering > 36202-(S1) Undergraduate Thesis
Depositing User:	Muhammad Azmi Naufal Munif
Date Deposited:	05 Mar 2021 00:45
Last Modified:	05 Mar 2021 00:45
URI:	http://repository.its.ac.id/id/eprint/83488

Actions (login required)

View Item