Klasifikasi Laju Produksi Pupuk Menggunakan Metode Convolutional Neural Networks (CNN) Untuk Meningkatkan Respons Operator

Prakoso, Reno Radix Prakoso (2025) Klasifikasi Laju Produksi Pupuk Menggunakan Metode Convolutional Neural Networks (CNN) Untuk Meningkatkan Respons Operator. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040211016-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (22MB) | Request a copy

Abstract

PT Petrokimia Gresik memiliki peran penting dalam memenuhi kebutuhan pupuk bersubsidi nasional. Permasalahan operasional yang diangkat dalam penelitian ini adalah keterbatasan operator pengantongan dalam memantau secara terus-menerus laju produksi pupuk yang bersifat fluktuatif. Keterbatasan pemantauan ini berisiko menyebabkan kekosongan silo, yang berdampak pada ketidakstabilan pembacaan load cell dan penurunan kualitas produk, serta berpotensi mengakibatkan tidak tercapainya target pengantongan. Untuk mengatasi masalah tersebut, penelitian ini merancang sebuah sistem bantuan operator berbasis Convolutional Neural Network (CNN) dengan arsitektur MobileNetV2 untuk mengklasifikasikan laju produksi secara real-time ke dalam empat kelas yaitu none, rendah, sedang, dan tinggi. Model yang dikembangkan berhasil mencapai akurasi keseluruhan 91,76% pada data uji. Hasil klasifikasi dari sistem ini disajikan kepada operator melalui modul notifikasi visual dan audio, yang terbukti sangat responsif dengan delay pengiriman data di bawah 1 detik. Dengan demikian, sistem ini berhasil berfungsi sebagai alat bantu yang menyediakan informasi laju produksi yang dapat ditindaklanjuti secara real-time untuk meningkatkan respons operator dan membantu operator dalam pengambilan keputusan pada kontrol translator.
======================================================================================================================================
PT Petrokimia Gresik plays a vital role in fulfilling the national subsidized fertilizer demand. The operational issue addressed in this research is the limited ability of bagging operators to continuously monitor the fluctuating fertilizer production rate. This monitoring limitation risks causing silo emptiness, which impacts the instability of load cell readings and product quality degradation, potentially leading to a failure in achieving bagging targets. To overcome these problems, this research designed an operator assistance system based on a Convolutional Neural Network (CNN) with MobileNetV2 architecture to classify the production rate in real-time into four classes: none, rendah, sedang, and Tinggi. The developed model successfully achieved an overall accuracy of 91.76% on the test data. The classification results from this system are presented to operators through visual and audio notification modules, which proved highly responsive with a data transmission delay of under 1 second. Thus, this system successfully functions as an aid that provides actionable, real-time production rate information to enhance operator responsiveness and assist operators in decision-making for translator control.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Network, Klasifikasi, Sistem Pengawasan, Convolutional Neural Network, Classification, Monitoring System.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.62 Decision support systems T Technology > T Technology (General) > T59.7 Human-machine systems. T Technology > TS Manufactures T Technology > TS Manufactures > TS155 Production control. Production planning. Production management
Divisions:	Faculty of Vocational
Depositing User:	Reno Radix Prakoso
Date Deposited:	04 Aug 2025 03:23
Last Modified:	04 Aug 2025 03:23
URI:	http://repository.its.ac.id/id/eprint/125838

Actions (login required)

View Item