Optimalisasi Kecepatan Putaran Spinner Minyak Berbasis Adaptive Neuro-Fuzzy Inference System (ANFIS) untuk Penirisan Minyak Keripik Pisang

Prasistya, Pungky Alvina (2026) Optimalisasi Kecepatan Putaran Spinner Minyak Berbasis Adaptive Neuro-Fuzzy Inference System (ANFIS) untuk Penirisan Minyak Keripik Pisang. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2042221027-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Kandungan minyak berlebih pada keripik pisang dapat menurunkan kualitas produk dan memperpendek daya simpannya. Proses penirisan menggunakan spinner konvensional umumnya bekerja pada kecepatan putaran konstan tanpa memperhitungkan perbedaan massa bahan, sehingga penggunaan energi menjadi kurang efisien. Penelitian ini bertujuan merancang dan mengimplementasikan sistem kontrol kecepatan spinner peniris minyak berbasis Adaptive Neuro-Fuzzy Inference System (ANFIS) untuk mengatur kecepatan putaran motor berdasarkan massa awal dan target massa akhir dengan waktu penirisan konstan selama 60 detik. Model ANFIS dikembangkan menggunakan MATLAB dan diimplementasikan pada Raspberry Pi yang terintegrasi dengan Human-Machine Interface (HMI). Pengujian dilakukan menggunakan pisang kepok dengan variasi massa awal 100 gram, 200 gram, 300 gram, 400 gram, dan 500 gram. Sistem berbasis ANFIS diuji sebanyak tiga kali pada setiap variasi massa, kemudian dibandingkan dengan spinner konvensional dan spinner berbasis Fuzzy Logic yang menggunakan spesifikasi mekanik, elektrik, dan komponen yang sama. Hasil pelatihan model ANFIS menunjukkan nilai Root Mean Square Error (RMSE) sebesar 15,474 rpm, sedangkan hasil implementasi menghasilkan akurasi rata-rata sebesar 90,11% dengan error rata-rata sebesar 9,89%. Kecepatan putaran spinner berbasis ANFIS berada pada rentang 1175–1303 rpm, sedangkan spinner konvensional menghasilkan kecepatan putaran yang relatif konstan sekitar 1500 rpm. Selain itu, sistem berbasis ANFIS menghasilkan efisiensi penirisan sebesar 9,7%, 5,2%, 6,8%, 4,7%, dan 5,9% pada variasi massa 100 gram, 200 gram, 300 gram, 400 gram, dan 500 gram. Hasil pengujian menunjukkan bahwa sistem berbasis ANFIS memiliki efisiensi penirisan yang lebih tinggi dengan konsumsi daya listrik yang lebih rendah dibandingkan spinner konvensional maupun spinner berbasis Fuzzy Logic.
==================================================================================================================================
Excessive oil content in banana chips can reduce product quality and shorten shelf life. Conventional oil-draining spinners generally operate at a constant rotational speed without considering variations in material mass, resulting in inefficient energy utilization. This study aims to design and implement an Adaptive Neuro-Fuzzy Inference System (ANFIS)-based speed control system for an oil-draining spinner to automatically regulate the motor rotational speed based on the initial mass and target final mass with a constant draining time of 60 seconds. The ANFIS model was developed using MATLAB and implemented on a Raspberry Pi integrated with a Human-Machine Interface (HMI). Experiments were conducted using Kepok bananas with initial masses of 100 g, 200 g, 300 g, 400 g, and 500 g. The ANFIS-based system was tested three times for each mass variation and compared with a conventional spinner and a Fuzzy Logic-based spinner employing identical mechanical, electrical, and hardware specifications. The ANFIS training process achieved a Root Mean Square Error (RMSE) of 15.474 rpm, while the implementation results obtained an average accuracy of 90.11% with an average error of 9.89%. The rotational speed of the ANFIS-based spinner ranged from 1175 to 1303 rpm, whereas the conventional spinner operated at a nearly constant speed of approximately 1500 rpm. Furthermore, the ANFIS-based system achieved draining efficiencies of 9.7%, 5.2%, 6.8%, 4.7%, and 5.9% for initial masses of 100 g, 200 g, 300 g, 400 g, and 500 g, respectively. The experimental results also demonstrate that the ANFIS-based system provides higher draining efficiency with lower power consumption than both the conventional spinner and the Fuzzy Logic-based spinner.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Spinner, Penirisan Minyak, Kontrol Kecepatan, ANFIS, Spinner, Oil Draining, Speed Control, ANFIS
Subjects:	Q Science Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > TJ Mechanical engineering and machinery
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	Pungky Alvina Prasistya
Date Deposited:	31 Jul 2026 06:07
Last Modified:	31 Jul 2026 06:07
URI:	http://repository.its.ac.id/id/eprint/140578

Actions (login required)

View Item