Jurnal Ilmiah Teknik Elektro
Jetri, Volume 19, Nomor 2, Februari 2022

Sistem Inspection Double Sided Tape Menggunakan Jaringan Syaraf Tiruan Untuk Pengecekan Otomatis

Ketty Siti Salamah (Universitas Mercu Buana)
Muhammad Afif Assidiq (Universitas Mercu Buana)
Imelda Uli Vistalina Simanjuntak (Universitas Mercu Buana)



Article Info

Publish Date
21 Jan 2022

Abstract

One of the quality demands in the automotive industry is found in the meter cluster product of a car, namely Double Sided Tape. Installation of Double Sided Tape so as not to cause light leakage on the display meter cluster. An inspection system is needed that can replace the role of humans in checking automatically so that it can eliminate human-error problems. To ensure that the level of inspection system is the same as that of humans, they can use Artificial Neural Networks for the decision-making process. In supporting the checking system, a matlab and a webcam camera are needed which functions to analyze the digital image captured from the camera. The results of testing the accuracy of single data samples OK and NG is an accuracy of 95%. The length of learning time obtained to be able to distinguish which Card Board has been installed with Double Sided Tape or not is 291.2541 seconds. From the experiments that have been carried out, it can be seen that the factors that affect accuracy and RMSE are the number of epochs and the learning rate, the most optimal learning rate obtained is 0.3, and the epoch value is 5000.Tuntutan kualitas pada industri otomotif salah satunya yang terdapat pada produk meter cluster sebuah mobil yaitu Double Sided Tape. Pemasangan Double Sided Tape agar tidak menimbulkan kebocoran cahaya pada display meter cluster. Diperlukan suatu sistem inspection yang bisa menggantikan peran manusia dalam pengecekannya secara otomatis sehingga dapat menghilangkan masalah human-error. Untuk memastikan level inspection system tersebut sama dengan yang dilakukan manusia bisa menggunakan Jaringan syaraf Tiruan untuk proses pengambilan keputusan. Dalam menunjang sistem pengecekan diperlukan matlab dan kamera webcam yang berfungsi menganalisa hasil tangkap gambar citra digital dari kamera. Hasil dari pengujian akurasi data tunggal sample OK dan NG adalah akurasi 95%. Lama waktu learning yang didapat untuk dapat membedakan mana Card Board yang sudah terpasang Double Sided Tape atau tidak yaitu 291.2541 detik. Dari percobaan yang telah dilakukan dapat diketahui faktor yang mempengaruhi akurasi dan RMSE adalah jumlah epoch dan learning rate, nilai paling optimal learning rate yang didapat adalah 0.3, dan nilai epoch 5000.

Copyrights © 2022






Journal Info

Abbrev

jetri

Publisher

Subject

Computer Science & IT Control & Systems Engineering Electrical & Electronics Engineering Energy

Description

Jetri is a scientific journal aims to publish high quality and up to date articles in electrical engineering field. Its scope includes (but not limited to): - Power Systems: nonrenewable and renewable energy power generation, power transmission and distribution, power conversion, protection system, ...