Wicaksana, Rhema Adi Magiza (2022) Optimasi Sistem Manajemen Energi pada Mobil PHEV ITS Berkonfigurasi Seri dengan Backpropagation Neural Network Genetic Algorithm dan Backpropagation Neural Network Particle Swarm Optimization. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 02111740000115-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
02111740000115-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2024. Download (3MB) | Request a copy |
|
Text
02111740000115-Undergraduate_Thesis.pdf Restricted to Registered users only Download (3MB) | Request a copy |
|
Text
02111740000115-Undergraduate_Thesis.pdf - Updated Version Restricted to Registered users only Download (3MB) | Request a copy |
Abstract
Meningkatnya pemanasan global dan semakin sedikitnya kandungan minyak bumi menyebabkan perkembangan dan permintaan kendaraan hybrid electric vehicle (HEV) dan electric vehicle (EV) semakin besar. HEV dan EV lebih ramah lingkungan karena memiliki gas buang yang jauh lebih kecil dan bahkan tidak ada. HEV memiliki keunggulan pada jarak berkendaranya yang lebih jauh. Namun HEV juga memiliki kelemahan karena masih menghasilkan gas buang. Manajemen energi pada mobil hybrid sangatlah penting. Manajemen energi ini merupakan suatu algoritma yang bertujuan untuk mengurangi konsumsi bahan bakar serta memaksimalkan nilai SOC. Manajemen energi dengan strategi kontrol konvensional masih dirasa kurang maksimal dalam mengurangi konsumsi bahan bakar dan memaksimalkan SOC.
Penelitian ini berfokus pada optimasi strategi manajemen energi pada mobil PHEV ITS dengan konfigurasi seri dengan tujuan meminimalkan konsumsi bahan bakar dan memaksimalkan SOC dengan parameter input system status (engine on-off scenario), kondisi SOC aktivasi mesin pembakaran dalam, dan kecepatan kendaraan. Penelitian dilakukan dengan menggunakan backpropagation neural network untuk memprediksi respon dari konsumsi bahan bakar dan SOC. Prediksi tersebut digunakan untuk memperoleh parameter input dengan respon SOC maksimum dan konsumsi bahan bakar minimum dengan metode meteheuristik genetic algorithm (GA) dan particle swarm optimization (PSO).
Hasil optimum yang didapatkan dari percobaan menggunakan BPNN – GA dan BPNN – PSO didapatkan dengan parameter BPNN 2 hidden layer, 12 neuron, fungsi aktivasi tansig, dan MSE 0.0017, parameter input kecepatan adalah 36,33 km/jam, kondisi aktivasi SOC 48, 73%, dan RPM ICE 7499,9 RPM didapatkan output jarak tempuh 83,123 km dan konsumsi bahan bakar 6,0078L. Dengan metode PSO sedikit lebih baik dengan nilai optimum -0.185312759211082 dibanding nilai optimum GA -0.185312681975449. Optimasi pada setiap kecepatan konstan didapatkan hasil pada kecepatan 17 km/jam, maka kondisi aktivasi SOC pada 60% dan RPM 7500, tanpa ada kenaikan efisiensi kendaraan. Pada kecepatan 30 km/jam, kondisi aktivasi SOC pada 54,7% dengan RPM 7500, terdapat kenaikan efisiensi kendaraan sebesar 2,1%. Pada kecepatan 50 km/jam, kondisi aktivasi SOC pada 42% dengan RPM 7500, terdapat kenaikan efisiensi kendaraan sebesar 1,1%.
=====================================================================================================
The increasing global warming and the decreasing amount of petroleum have caused the development and demand for hybrid electric vehicles (HEV) and electric vehicles (EV) to increase. HEV and EV are more environmentally friendly because they have much smaller and even non-existent exhaust gases. HEV has the advantage of a longer driving distance. However, HEV also has a weakness because it still produces exhaust gas. Energy management in hybrid cars is very important. Energy management is an algorithm that aims to reduce fuel consumption and maximize the SOC value. Energy management with conventional control strategies is still considered less than optimal in reducing fuel consumption and maximizing SOC.
This study focuses on optimizing energy management strategies in PHEV ITS cars with a series configuration to minimize fuel consumption and maximize SOC with input system status parameters (engine on-off scenario), SOC conditions of internal combustion engine activation, and vehicle speed. The research was conducted using a backpropagation neural network to predict the response of fuel consumption and SOC. These predictions are used to obtain input parameters with maximum SOC response and minimum fuel consumption using the metaheuristics genetic algorithm (GA) and particle swarm optimization (PSO) methods.
The optimum results obtained from experiments using BPNN – GA, and BPNN – PSO were obtained with BPNN parameters 2 hidden layers, 12 neurons, tansig activation function, and MSE 0.0017, speed input parameters were 36.33 km/hour, SOC activation conditions 48, 73 %, and the ICE RPM is 7499.9 RPM, the output distance is 83.123 km and fuel consumption is 6.0078L. The PSO method is slightly better with the optimum value of -0.185312759211082 than the optimum value of GA -0.185312681975449. Optimization at each constant speed is obtained at a speed of 17 km/hour, the SOC activation conditions are at 60% and RPM 7500, without any increase in vehicle efficiency. At a speed of 30 km/hour, the SOC activation condition is at 54.7% with an RPM of 7500, there is an increase in vehicle efficiency of 2.1%. At a speed of 50 km/hour, the SOC activation condition is at 42% with an RPM of 7500, there is an increase in vehicle efficiency of 1.1%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | strategi manajemen energi, HEV, backpropagation neural network, genetic algorithm, particle swarm optimization, konsumsi bahan bakar, jarak tempuh |
| Subjects: | T Technology > TJ Mechanical engineering and machinery > TJ217 Adaptive control systems |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Mechanical Engineering > 21201-(S1) Undergraduate Thesis |
| Depositing User: | Rhema Adi Magiza Wicaksana |
| Date Deposited: | 16 Feb 2022 07:06 |
| Last Modified: | 16 Feb 2022 07:06 |
| URI: | http://repository.its.ac.id/id/eprint/94040 |
Actions (login required)
 |
View Item |