Article Per Year (5 Year)
p-Index From 2019 - 2024
0.23
P-Index
This Author published in this journals
All Journal Majalah Ilmiah Teknologi Elektro
Muhammad Rizalul Wahid
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Electrical & Electronics Engineering Energy Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Analisis Komputasi Paralel pada Image Encoding Framework untuk Konversi Citra Data Deret Waktu Sistem Kontrol Industri Helmy Rahadian; Muhammad Rizalul Wahid; Zaenal Arifin
Jurnal Teknologi Elektro Vol 22 No 2 (2023): (Juli - Desember) Majalah Ilmiah Teknologi Elektro
Publisher : Program Studi Magister Teknik Elektro Universitas Udayana

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/MITE.2023.v22i02.P06

Abstract

Sensors in industrial control systems send a series of data each time, known as time series data, to the controller. The data contains important information for the controller to determine the control signal for the actuator. The appearance of anomalies in time series data can be detected using the Convolutional Neural Network (CNN) method utilizing image encoding techniques such as Gramian Angular Field (GAF) and Markov Transition Field (MTF). This technique converts time series data into images through data preparation, encoding, and conversion. Dividing extensive data into many smaller segments requires repeated encoding and conversion processes. Repeated processes that are done serially take a long time, which slows down the detection of anomalies and the responses that must be taken. This research applies parallel computation with Joblib and Mpire libraries on the GAF and MTF image encoding provided by the Python-based pyts library. The n_jobs configuration determines the number of CPU logical cores used to execute the program. According to the number of CPU logic cores of the computer, applying the value of n_jobs = 8 can save an average processing time of 63% (Joblib) and 49% (Mpire), which theoretically will be able to detect anomalies that occur at least every 62.73 ms (Joblib) and 86.20 ms (Mpire) compared to 167.51 ms in serial computing.
Co-Authors Helmy Rahadian Zaenal Arifin