Gaol, Erjuki M Lumban (2019) Perancangan Sistem Monitoring Kinerja Welder Berbasis Teknologi Wearable Device untuk Meningkatkan Produktivitas Pembangunan Kapal. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
04111240000083-Undergraduate_Theses.pdf Download (9MB) | Preview |
Abstract
Salah satu penyebab terjadi keterlambatan pembangunan kapal adalah kinerja dan produktivitas welder, yang dipengaruhi oleh sikap kerja, dan kedisiplin akan waktu, tugas dan tanggung jawabnya. Sistem monitoring yang diterapkan saat ini pun masih berupa sistem konvensional dengan memanfaatkan welding inspector untuk memeriksa hasil kerja welder. Sementara itu, perkembangan teknologi saat ini sudah digunakan untuk mengawasi pekerja pada konstruksi bangunan sipil dengan menggunakan wearable device dengan sensor akselerometer, giroskop, dan magnetometer yang terpasang di dalamnya. Oleh karena itu, dilakukan perancangan sistem yang dapat mengawasi kinerja seorang welder berbasis teknologi wearable device. Pertama, perancangan sistem dimulai dengan observasi lapangan untuk mengetahui aktivitas yang welder. Kedua, dilakukan persiapan alat berupa sensor MetaMotion dan software MetaBase untuk perekaman data. Ketiga, pengambilan data kegiatan pengelasan Welding Procedure Specification (WPS). Keempat, dilakukan analisis dan perancangan algoritma support vector machine (SVM) pengenalan aktivitas welder dengan menggunakan software matlab. Selain itu juga dilakukan analisis berdasarkan hasil pengamatan saat di galangan kapal. Kelima, dilakukan perencanaan konsep dan alur sistem monitoring dengan menggunakan wearable device. Hasil pengujian algoritma yang dibuat dapat memprediksi aktivitas secara benar dengan tingkat akurasi sebesar 96,00 % pada pengelasan WPS. Sedangkan pada pengujian dengan data welder di galangan sebesar 54,50 %. Hasil uji algoritma meningkat menjadi 78,10 % setelah dilakukan penggabungan data training pengelasan WPS dengan pengelasan di galangan kapal. Hasil penelitian menyimpulkan bahwa sistem monitoring berbasis wearable device dapat dilakukan dengan mengenali aktivitas yang dilakukan welder menggunakan algortima yang dirancang telah dirancang. Nilai akurasi sistem pengenalan dapat ditingkatkan dengan cara menambah variasi pada data latihnya.
==============================================================================================================================
One of factor in shipbuilding delays is the performance and productivity of the welder, which is influenced by work attitudes, discipline of time, duties and responsibilities. The monitoring system applied today is still in the form of a conventional system using a welding inspector to check the welder's work. Meanwhile, technological developments are currently being used for monitoring workers in the construction of civilian buildings using wearable devices with accelerometer, gyroscope and magnetometer sensors installed in them. Therefore, a system design that can monitor the performance of a welder based on wearable device technology is carried out. First, the system design starts with field observations to find out the welder's activities. Second, preparations were made in the form of a MetaMotion sensor and MetaBase software for data recording. Third, the data collection activities welding Welding Procedure Specification (WPS). Fourth, analysis and design of a support vector machine (SVM) algorithm for introducing welder activities using Matlab software. In addition, an analysis was also conducted based on observations at the shipyard. Fifth, conceptual planning and monitoring system flow is carried out using wearable devices. The results of testing the algorithm can predict the activity correctly with an accuracy of 96.00% in WPS welding. Whereas the test with data welder in the shipyard was 54.50%. The algorithm test results increased to 78.10% after combining WPS welding training data with welding at the shipyard. The results of the study concluded that the monitoring system based on wearable devices can be carried out by recognizing the activities carried out by welder using the algorithm designed. The accuracy value of the recognition system can be improved by adding variations to the training data.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | kinerja, sistem monitoring, support vector machine (SVM), wearable device, welder |
| Subjects: | V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering |
| Divisions: | Faculty of Marine Technology (MARTECH) > Naval Architecture and Shipbuilding Engineering > 36201-(S1) Undergraduate Thesis |
| Depositing User: | Erjuki Erjuki |
| Date Deposited: | 29 Jul 2024 03:42 |
| Last Modified: | 29 Jul 2024 03:42 |
| URI: | http://repository.its.ac.id/id/eprint/74477 |
Actions (login required)
 |
View Item |