Optimasi Parameter Proses Friction Stir Welding Terhadap Kedalaman Shoulder Spesimen Las AA6061-T651 Menggunakan Metode Backpropagation Neural Network-Genetic Algorithm

Ghosa, Gede Gosali Sunu (2022) Optimasi Parameter Proses Friction Stir Welding Terhadap Kedalaman Shoulder Spesimen Las AA6061-T651 Menggunakan Metode Backpropagation Neural Network-Genetic Algorithm. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
02111840000024-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 April 2024.

Download (4MB) | Request a copy

Abstract

Saat ini, penggunaan las jenis friction stir welding (FSW) pada industri manufaktur di Indonesia masih sangat minim dikarenakan kurangnya pengembangan dan masih adanya kelemahan pada penggunaan las jenis FSW seperti membutuhkan sistem clamping khusus, hanya cocok untuk jenis sambungan yang sederhana (butt weld), dan besarnya tekanan pada spindle yang dapat merusak mesin las. Maka dari itu, penelitian mengenai optimasi parameter proses FSW terhadap kedalaman shoulder spesimen las AA6061-T651 ini dilakukan sebagai bentuk pengembangan terhadap penelitian FSW. Penelitian diawali dengan studi literatur untuk menemukan parameter proses dan output yang dapat memengaruhi hasil FSW. Kedalaman shoulder pada spesimen las diukur menggunakan dial indicator dimana pengukuran dilakukan pada 5 titik di spesimen las. Selanjutnya dilakukan pengujian analysis of variance (ANOVA) menggunakan Minitab 19 untuk menentukan signifikansi pengaruh kecepatan putar tool, kecepatan pengelasan, tilt angle, dan sudut bahu cekung terhadap kedalaman shoulder spesimen las. Setelah parameter proses yang berpengaruh diperoleh, maka dilakukan pengidentifikasian nilai optimal parameter proses FSW yang dapat menghasilkan kedalaman shoulder minimum menggunakan metode optimasi backpropagation neural network-genetic algorithm (BPNN-GA) pada MATLAB. Berdasarkan hasil pengujian ANOVA, kecepatan putar tool, kecepatan pengelasan, dan tilt angle merupakan parameter proses yang tidak berpengaruh signifikan terhadap hasil kedalaman shoulder spesimen las, sedangkan sudut bahu cekung merupakan parameter proses yang berpengaruh signifikan terhadap hasil kedalaman shoulder spesimen las. Berdasarkan dari hasil optimasi menggunakan metode BPNN-GA, nilai optimum sudut bahu cekung yang dapat menghasilkan kedalaman shoulder paling minimum pada spesimen las adalah 9°. Namun hasil optimasi ini tidak valid dikarenakan BPNN tidak cukup baik memprediksi hubungan antara sudut bahu cekung dengan kedalaman shoulder hasil FSW. ================================================================================================ Currently, the use of friction stir welding (FSW) in the manufacturing industry in Indonesia is still minimal due to the lack of development and there are still weaknesses in the use of FSW, such as the need for a special clamping system, only suitable for simple connection types (butt welds), and the magnitude of the pressure on the spindle that can damage the welding machine. Therefore, this research regarding the optimization of the FSW process parameters towards the shoulder depth of the AA6061-T651 welding specimen was carried out as a form of development of the FSW research. The research begins with a literature study to find the process parameters and outputs that can affect FSW results. The shoulder depth on the weld specimen is measured using a dial indicator where measurements are made at 5 points on the weld specimen. Next, analysis of variance (ANOVA) was tested using Minitab 19 to determine the significance effect of tool rotational speed, welding speed, tilt angle, and concave shoulder angle towards the shoulder depth of the weld specimen. After the influential process parameters are obtained, identification of the optimal value of the FSW process parameters that can produce the minimum shoulder depth was conducted using backpropagation neural network-genetic algorithm (BPNN-GA) on MATLAB. Based on the ANOVA test results, tool rotation speed, welding speed, and tilt angle are process parameters that have no significant effect on the results of the weld specimen shoulder depth, while the concave shoulder angle is a process parameter that has a significant effect on the weld specimen shoulder depth results. Based on the optimization results using BPNN-GA method, the optimum value of the concave shoulder angle that can produce the minimum shoulder depth on the weld specimen is 9°. However, the results of this optimization are not valid because BPNN is not good enough to predict the relationship between the angle of the concave shoulder and the depth of the shoulder from the FSW results.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: FSW, ANN, optimization, shoulder depth, BPNN-GA, JST, optimasi, kedalaman shoulder, BPNN-GA
Subjects: T Technology > TS Manufactures
T Technology > TS Manufactures > TS227 Welding.
Divisions: Faculty of Industrial Technology and Systems Engineering (INDSYS) > Mechanical Engineering > 21201-(S1) Undergraduate Thesis
Depositing User: Gede Gosali Sunu Ghosa
Date Deposited: 11 Feb 2022 00:49
Last Modified: 11 Feb 2022 00:49
URI: https://repository.its.ac.id/id/eprint/93652

Actions (login required)

View Item View Item