Nurullah, Fajar Perdana (2019) Multi response optimization of end milling process of combo glass fiber reinforced polymer (GFRP) composites by using combined methods of back propagation neural network and genetic algorithm (BPNN-GA). Masters thesis, Sepuluh Nopember Institute of Technology.
![[thumbnail of Master thesis]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text (Master thesis)
02111750010003-Master_Thesis.pdf - Published Version Restricted to Repository staff only Download (8MB) |
Abstract
Peningkatan penggunaan fiber reinforced polymer (FRP) pada industri pertahanan, kedirgantaraan, dan antariksa menjadikan proses pemesinan material ini semakin penting. Glass fiber reinforced polymer (GFRP) merupakan salah satu jenis FRP yang banyak digunakan, terutama pada aplikasi kedirgantaraan. Dalam proses manufaktur komponen berbahan GFRP diperlukan proses end milling untuk menghilangkan material berlebih sehingga diperoleh komponen yang memenuhi spesifikasi produk. Salah satu cara untuk mengevaluasi kinerja proses end milling adalah dengan menganalisis sifat mampu mesin (machinability) dari GFRP serat combo. Beberapa indikator machinability pada proses end milling meliputi gaya potong, kekasaran permukaan, dan delaminasi. Ketiga respons tersebut diharapkan memiliki nilai sekecil mungkin (smaller is better). Untuk mencapai kondisi tersebut, diperlukan penentuan level parameter proses end milling yang optimal melalui proses optimasi terhadap ketiga respons tersebut. Optimasi multirespons dalam penelitian ini dilakukan melalui tahap pemodelan dan optimasi. Pemodelan menggunakan metode Back Propagation Neural Network (BPNN), sedangkan optimasi dilakukan menggunakan Genetic Algorithm (GA). Variabel respons yang dianalisis meliputi gaya potong, kekasaran permukaan, dan delaminasi. Parameter proses yang divariasikan adalah kecepatan spindel, kecepatan makan, dan kedalaman potong. Pahat yang digunakan adalah pahat karbida berdiameter 6 mm. Kecepatan spindel divariasikan pada 3000 rpm, 4000 rpm, dan 5000 rpm, sedangkan kecepatan makan divariasikan pada 500 mm/menit, 750 mm/menit, dan 1000 mm/menit. Kedalaman potong divariasikan sebesar 1 mm, 1,5 mm, dan 2 mm. Rancangan percobaan menggunakan desain faktorial penuh 3 × 3 × 3 dengan tiga kali replikasi. Material yang digunakan adalah komposit glass fiber reinforced polymer (GFRP) serat combo. Hasil penelitian menunjukkan bahwa arsitektur jaringan BPNN terbaik yang diperoleh melalui optimasi menggunakan metode GA terdiri atas 3 neuron pada input layer, 10 neuron pada hidden layer pertama, 8 neuron pada hidden layer kedua, 7 neuron pada hidden layer ketiga, dan 3 neuron pada output layer (3–10–8–7–3). Model BPNN tersebut mampu memprediksi ketiga respons dengan kesalahan relatif sebesar 0,0627 dibandingkan hasil eksperimen. Kombinasi parameter proses end milling yang mampu meminimalkan gaya potong, kekasaran permukaan, dan delaminasi secara simultan adalah kedalaman potong 1 mm, kecepatan spindel 4889 rpm, dan kecepatan makan 791 mm/menit.
==================================================================================================================================
The increasing use of reinforced fiber composites (FRP) in the defense, aerospace and aerospace industries causes the machining process of this material becomes very important. Combo glass fiber reinforced polymer (GFRP) composites is one of the FRP materials that is often found, especially in aerospace applications. In machining of combo GFRP composites, an end milling process is needed to remove excess material, in order the components meet the product specifications. The performance of end milling process can be determined by evaluating the machinability of combo GFRP. Several of the machinability criteria of combo GFRP in the end milling process are cutting force, surface roughness and delamination. The target of these three machinability criteria is the smaller the better. In order to achieve this target, the levels of the end milling process parameters must be determined correctly. The levels of these parameters can be determined by optimizing the three machinability criteria of combo GFRP in the end milling process.
Multi-response optimization in this study was conducted by modeling and optimization. Modeling was performed by using back propagation neural network (BPNN) method, while the optimization used genetic algorithm (GA) method. The response parameters or output of the end milling processes studied were cutting force, surface roughness, and delamination. The parameters of the end milling process varied are spindle speed, feeding speed, and depth of cut. The end mill cutter material was made of solid carbide with a diameter of 6 mm. The spindle speed variations were 3000 rpm, 4000 rpm, and 5000 rpm. The feeding speed levels varied were 500 mm/min, 750 mm/min, and 1000 mm/min. The levels of depth of the cut were 1 mm, 1.5 mm, and 2 mm. The experimental design was a full factorial 3 x 3 x 3 with 3 replications. The composite material used was combo glass fiber reinforced polymer (GFRP). The best BPNN network architecture obtained by conducting optimization using GA method has 3 neurons in the input layer, 10 neurons in the first hidden layer, 8 neurons in the second hidden layer, 7 neurons in the third hidden layer, and 3 neurons in the output layer (3-10- 8-7-3). This BPNN model is able to predict three responses accurately with the relative errors between the results of the predictions and the experimental results is 0.0627. The levels of end milling process parameters that can minimize cutting force, surface roughness, and delamination simultaneously are depth of cut, spindle speed, and feeding speed of 1 mm, 4889 rpm, and 791 mm/minute, respectively.
| Item Type: | Thesis (Masters) |
|---|---|
| Subjects: | T Technology > TJ Mechanical engineering and machinery T Technology > TS Manufactures > TS176 Manufacturing engineering. Process engineering (Including manufacturing planning, production planning) |
| Divisions: | Faculty of Industrial Technology > Chemical Engineering > 24101-(S2) Master Thesis |
| Depositing User: | Nurullah Fajar Perdana |
| Date Deposited: | 21 Jul 2026 08:08 |
| Last Modified: | 21 Jul 2026 08:08 |
| URI: | http://repository.its.ac.id/id/eprint/69790 |
Actions (login required)
 |
View Item |