cover
Contact Name
Amalia Eka Rakhmania
Contact Email
amaliaeka.rakhmania@polinema.ac.id
Phone
+6282134134594
Journal Mail Official
eltek@polinema.ac.id
Editorial Address
Pusat Penelitian dan Pengabdian Masyarakat Politeknik Negeri Malang Graha Polinema Lantai 3 Jl. Sukarno Hatta No. 9, Malang, Indonesia, 65141
Location
Kota malang,
Jawa timur
INDONESIA
Jurnal ELTEK
ISSN : 16934024     EISSN : 23550740     DOI : doi.org/10.33795/eltek
Jurnal ELTEK merupakan media yang diterbitkan oleh Politeknik Negeri Malang sebagai sarana diseminasi dan publikasi artikel hasil penelitian dan artikel konseptual yang dilakukan oleh para peneliti, akademisi, praktisi dan industri di bidang Teknik Elektro. Artikel yang diajukan untuk diterbitkan pada jurnal ELTEK merupakan naskah asli dan belum pernah dipublikasikan secara tertulis pada majalah atau jurnal ilmiah dimanapun Jurnal ELTEK memuat artikel hasil penelitian dan konseptual dalam cakupan bidang ilmu Teknik Elektro yang meliputi: Ketenagalistrikan, Teknik Kendali, Kecerdasan Buatan, Teknik Komputer, Teknik Telekomunikasi, Teknik informatika, Teknik Biomedik, Elektronika Daya, Energi Terbarukan, Sistem Embedded, Jaringan Komputer, Sistem Operasi, Manajemen Data dan Sistem informasi, dan lain-lain. Jurnal ELTEK terbit dua kali dalam setahun masing-masing pada bulan April dan Oktober. Redaksi mengundang penulis dan peneliti untuk menyumbangkan artikel penelitian atau artikel konseptual kepada Editor Jurnal ELTEK. Editor menentukan apakah suatu artikel akan dimuat atau tidak dan memiliki hak untuk mengubah atau mengoreksi teks sejauh tidak mengganggu maksud atau kontennya.
Articles 86 Documents
Sistem prediksi produksi padi di Pulau Jawa menggunakan long short term memory berbasis website Shafa Zahira Nailatus Syahida; Hadiwiyatno; Hendro Darmono
JURNAL ELTEK Vol. 24 No. 2 (2026): October 2026
Publisher : Politeknik Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33795/eltek.v24i2.10868

Abstract

Padi merupakan komoditas pangan utama di Indonesia sehingga kestabilan produksinya penting bagi ketahanan pangan nasional. Selama periode 2021–2024, produksi padi di Pulau Jawa menurun sebesar 6,16%, sedangkan jumlah penduduk meningkat 2,71%. Penelitian ini bertujuan mengembangkan model Long Short Term Memory untuk memprediksi produksi padi empat bulan ke depan pada enam provinsi di Pulau Jawa dalam sistem berbasis web. Model menggunakan data produksi padi, luas panen, dan variabel iklim periode Januari 2018–Desember 2025. Penelitian meliputi preprocessing data, pengujian 216 kombinasi hyperparameter, training dengan optimizer Adam, serta evaluasi menggunakan galat absolut rata-rata, akar galat kuadrat rata-rata, dan persentase galat absolut rata-rata. Hasil penelitian menunjukkan bahwa setiap provinsi memiliki konfigurasi hyperparameter terbaik yang berbeda. Berdasarkan persentase galat absolut rata-rata, model menghasilkan akurasi sangat baik di Jawa Timur, baik di Jawa Tengah, Jawa Barat, dan Daerah Istimewa Yogyakarta, cukup di Banten, serta buruk di Daerah Khusus Ibukota Jakarta. ABSTRACT Rice is Indonesia's primary staple crop, making production stability essential for maintaining national food security. During the period of 2021–2024, rice production on Java Island decreased by 6.16%, while the population increased by 2.71%. This study aims to develop a Long Short-Term Memory model to forecast rice production for the next four months across six provinces on Java Island through a web-based system. The model utilizes historical data on rice production, harvested area, and climate variables from January 2018 to December 2025. The research includes data preprocessing, evaluation of 216 hyperparameter combinations, model training using the Adam optimizer, and performance evaluation using mean absolute error, root mean squared error, and mean absolute percentage error. The results indicate that each province requires a different optimal hyperparameter configuration. Based on the mean absolute percentage error, the model achieved very good accuracy in East Java, good accuracy in Central Java, West Java, and the Special Region of Yogyakarta, fair accuracy in Banten, and poor accuracy in Jakarta.
Semi-fragile watermarking berbasis DWT-Haar pada gateway IoT ESP32 untuk verifikasi integritas data detak jantung Syauqii Usaiyyah Kristanto; Rahmawati Hasanah; Teguh Ginanjar
JURNAL ELTEK Vol. 24 No. 2 (2026): October 2026
Publisher : Politeknik Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33795/eltek.v24i2.10869

Abstract

Perangkat wearable mengirimkan data detak jantung melalui kanal nirkabel yang tidak menyediakan jaminan bahwa nilai yang diterima sama dengan nilai yang dikirim. Perlindungan kriptografis konvensional sulit diterapkan pada gateway bersumber daya terbatas. Skema fragile watermarking juga tidak sesuai karena menolak setiap perubahan, termasuk kompresi yang sebenarnya tidak berbahaya. Penelitian ini mengimplementasikan mekanisme semi-fragile watermarking pada gateway tertanam untuk memverifikasi integritas data detak jantung. Setiap blok berisi 16 sampel data detak jantung yang didekomposisi menggunakan transformasi wavelet Haar tingkat satu. Watermark dibangkitkan dari koefisien aproksimasi melalui fungsi hash kriptografis, lalu disisipkan ke koefisien detail melalui modulasi indeks kuantisasi. Sistem divalidasi secara menyeluruh menggunakan smartwatch komersial berbasis sensor fotopletismografi (PPG) dengan antarmuka BLE yang diuji pada 8 subjek dalam 16 sesi. Sebanyak 247 blok yang terkumpul dievaluasi melalui 16 skenario ketahanan. Ekstraksi watermark berlangsung tanpa kesalahan pada kondisi transmisi normal. Kualitas sinyal rata-rata mencapai 41,59 desibel, akurasi deteksi manipulasi mencapai 96,1 persen, dan waktu komputasi rata-rata 87,16 milidetik per blok. Hasil tersebut menegaskan bahwa verifikasi integritas yang ringan layak dijalankan pada gateway tertanam berbiaya rendah tanpa perangkat keras kriptografis khusus.   ABSTRACT Wearable devices transmit heart rate data across wireless channels that offer no inherent guarantee that the values received match those that were sent. Conventional cryptographic protection is hard to deploy on resource-constrained gateway hardware. Fragile protection is unsuitable as well, because it rejects every modification, including harmless compression. This study implements a semi-fragile watermarking mechanism on an embedded gateway to verify heart rate data integrity. Each block of 16 samples of heart rate data is decomposed by a first-level Haar wavelet transform. A watermark is then derived from the approximation coefficients through a cryptographic hash function and embedded into the detail coefficients by quantization index modulation. The system was validated end to end on a commercial smartwatch with a photoplethysmography (PPG) sensor over BLE test by 8 subjects across 16 sessions. The resulting 247 blocks were evaluated under 16 robustness scenarios. Watermark extraction was error free under normal transmission. Average signal quality reached 41.59 decibels, tamper detection accuracy reached 96.1 percent, and computation averaged 87.16 milliseconds per block. These results confirm that lightweight integrity verification is feasible on low-cost embedded gateways without dedicated cryptographic hardware.
A comparative study of property-to-label and property-to-node refactoring in ontology-based graph data modeling Ade Hodijah; Sri Ratna Wulan; Jonner Hutahaean; Rahil Jumiyani
JURNAL ELTEK Vol. 24 No. 2 (2026): October 2026
Publisher : Politeknik Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33795/eltek.v24i2.10896

Abstract

Integrating Web Ontology Language with Neo4j through the neosemantics (n10s) framework enables semantic validation and automated reasoning that are not natively supported by Labeled Property Graphs. However, the choice of graph refactoring pattern influences query performance. This study compares two ontology-based graph refactoring approaches, namely Property-to-Label (P2L) and Property-to-Node (P2N), implemented in Protégé and integrated into Neo4j using n10s and Owlready2. Both models were evaluated for semantic validation and ontology reasoning using the PET ontology, while their query execution performance was benchmarked using 78 Cypher query variants derived from the IMDB dataset. The evaluation examined the relationship between execution time and reductions in WHERE clauses following schema refactoring. The results show that both approaches successfully enforce ontology constraints and infer missing classifications from incomplete data. Performance analysis indicates that both P2L and P2N improved query execution time, with the highest reduction percentage was 89% for queries with four total number of the WHERE. P2L exhibited a more consistent relationship between WHERE clause reduction and execution time improvement due to Neo4j's native label indexing.
Prototipe sistem pencatatan data suhu transformator gardu induk berbasis ESP32 dengan komunikasi nirkabel Rahman Azis Prasojo; Achsanul Khabib; Muhammad Rizqy Ramadhani; Veri Dian Firmansyah; Dio Ramadhan; I Gusti Ngurah Mahendrayana; Ary Tri Farismansyah
JURNAL ELTEK Vol. 24 No. 2 (2026): October 2026
Publisher : Politeknik Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33795/eltek.v24i2.11015

Abstract

Transformator daya memerlukan pemantauan suhu secara kontinu untuk menjaga keandalan operasi.  Penelitian ini mengembangkan sistem otomatis pencatat data suhu transformator berbasis mikrokontroler ESP32 dan komunikasi nirkabel. Sistem membaca sinyal analog 4–20 mA dari parameter Oil Temperature Indicator (OTI), Winding High Voltage (WHV), dan Winding Low Voltage (WLV), kemudian menampilkan data pada LCD, mengirimkannya ke ThingSpeak, serta menyimpannya secara otomatis ke Google Sheets. Pengujian akurasi menggunakan generator sinyal 4–20 mA menghasilkan nilai Mean Absolute Percentage Error (MAPE) sebesar 0,6% sebelum uji keandalan dan maksimum 0,8% setelah uji keandalan. Pengujian komunikasi menunjukkan bahwa penggunaan antena eksternal mampu menurunkan data loss dari 41 menjadi 5 data serta menurunkan persentase data terlambat dari 10,70% menjadi 6,35%. Pengujian selama tujuh hari menghasilkan persentase kelengkapan data hingga 99,88% dengan rata-rata latency sekitar 1 detik. Hasil penelitian menunjukkan bahwa sistem mampu melakukan monitoring suhu transformator secara otomatis, real-time, dan andal. Abstract Power transformers require continuous temperature monitoring to maintain operational reliability. This study develops an ESP32-based wireless temperature data logging system for transformer monitoring. The system acquires 4–20 mA analog signals from the Oil Temperature Indicator (OTI), Winding High Voltage (WHV), and Winding Low Voltage (WLV), displays the measurements on an LCD, transmits data to ThingSpeak, and automatically stores the records in Google Sheets. Accuracy testing using a 4–20 mA signal generator resulted in a Mean Absolute Percentage Error (MAPE) of 0.6% before the reliability test and a maximum of 0.8% after the test. Communication testing showed that an external antenna reduced data loss from 41 to 5 records and decreased delayed data from 10.70% to 6.35%. A seven-day reliability test achieved data completeness of up to 99.88% with an average latency of approximately 1 second. The results demonstrate that the proposed system provides automatic, real-time, and reliable transformer temperature monitoring.
Reference-based degradation screening of 150-kV insulators using GWO-optimized one-class SVM under limited puncture-test data Giovanni Dimas Prenata; Ahmad Ridhoi
JURNAL ELTEK Vol. 24 No. 2 (2026): October 2026
Publisher : Politeknik Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33795/eltek.v24i2.11136

Abstract

This study proposes a reference-based degradation screening framework for 150-kV ceramic, glass, and polymer insulators using a Grey Wolf Optimizer–One-Class Support Vector Machine (GWO-OCSVM) under limited puncture-test data. The dataset comprised 72 observations, including 36 replacement samples used exclusively as reference data and 36 existing samples reserved for screening and evaluation. Input features were constructed from normalized string position and voltage ratio relative to the material-specific reference mean. Model development employed nested leave-one-position-out validation, while the OCSVM parameters ν and γ were optimized by GWO using reference data only. REF-GWO-OCSVM achieved an AUC of 0.8299, improving over REF-OCSVM at 0.7535, while increasing the mean anomaly-score gap from 0.0560 to 0.1447. The final consensus parameters were ν=0.1338 and γ=1.8257. A conservative P95=0.1970 threshold reduced reference rejection to 2/36 while flagging 9/36 existing samples. The results demonstrate that the proposed method enables reference-based anomaly screening without using existing-condition data during training or hyperparameter optimization.
Using edge computing and Large Language Models (LLM) for constructing a local agent Setya Widyawan Prakosa; Naufal Ainur Rizal; Tias Kurniati
JURNAL ELTEK Vol. 24 No. 2 (2026): October 2026
Publisher : Politeknik Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33795/eltek.v24i2.11288

Abstract

The growing demand for artificial intelligence systems that are responsive, privacy-preserving, and capable of operating autonomously is driving the adoption of Large Language Models (LLMs) on edge devices. However, deploying LLMs as the core of agents on devices with limited resources still faces computational and memory constraints. This research develops an LLM-based local agent architecture that runs locally on a Raspberry Pi without relying on cloud services. The system integrates local LLM inference with agentic functions, including intent processing, reasoning, tool invocation, and interaction with hardware via GPIO. Several quantized Small Language Models (SLMs) were evaluated on a Raspberry Pi 5 with 16 GB of RAM using CPU-based inference. The evaluation included total latency, Time to First Token (TTFT), token generation rate, peak RAM usage, peak CPU utilization, and peak temperature. The results showed significant performance differences among the models. Qwen 2.5 0.5B Q3 produced the lowest total latency of 23.9 seconds, the lowest TTFT of 269.0 ms, and the highest generation rate of 12.1 tokens/s. However, the results indicate that inference speed alone is not sufficient to determine the most suitable model as the core of a local agent. Most models reached peak CPU utilization above 95%, while RAM usage and thermal characteristics varied across configurations. Based on a combination of inference performance and resource usage, models in the 1B–1.7B range specifically LLaMA 3.2 1B Q4, Qwen 2.5 1.5B Q4, and SmolLM 1.7B Q4 show potential as a balanced set of candidates between responsiveness and model capacity. The research results demonstrate the feasibility of deploying LLMs as local agents on low-cost edge devices and provide a basis for selecting models based on a balance between inference efficiency, resource constraints, and agentic capabilities.