Umam, Khatibul (2026) KLASIFIKASI KONDISI MINYAK ISOLASI TRANSFORMATOR TENAGA 60 MVA 150/20 kV PLN UPT CILEGON MENGGUNAKAN K-NEAREST NEIGHBORS. S1 thesis, Fakultas Teknik Universitas Sultan Ageng Tirtayasa.
![[img]](https://eprints.untirta.ac.id/style/images/fileicons/text.png) |
Text (Fulltext)
Khatibul Umam_3332200085_Fulltext.pdf Restricted to Registered users only Download (5MB) |
|
Text (Bab 1)
Khatibul Umam_3332200085_01.pdf Restricted to Registered users only Download (1MB) |
|
Text (Bab 2)
Khatibul Umam_3332200085_02.pdf Restricted to Registered users only Download (429kB) |
|
Text (Bab 3)
Khatibul Umam_3332200085_03.pdf Restricted to Registered users only Download (835kB) |
|
Text (Bab 4)
Khatibul Umam_3332200085_04.pdf Restricted to Registered users only Download (1MB) |
|
Text (Bab 5)
Khatibul Umam_3332200085_05.pdf Restricted to Registered users only Download (9kB) |
|
Text (Daftar Referensi)
Khatibul Umam_3332200085_Ref.pdf Restricted to Registered users only Download (83kB) |
|
Text (Lampiran)
Khatibul Umam_3332200085_Lamp.pdf Restricted to Registered users only Download (982kB) |
|
Text (Cek Plagiasi)
Khatibul Umam_3332200085_CP.pdf Restricted to Registered users only Download (5MB) |
Abstract
The K-Nearest Neighbors (KNN) algorithm is one of the artificial intelligence approaches that has proven to be effective in improving the accuracy of diagnosis of transformer conditions. KNN works on the principle of data proximity, where the classification is determined based on the majority of classes from a number of nearby neighbors. The advantage of this algorithm lies in its ability to recognize data distribution patterns without requiring certain assumptions, making it ideal for analyzing complex variations in the results of Dissolved Gas Analysis (DGA). This study aims to implement KNN in classifying the condition of transformer insulation oil based on DGA data and evaluating its performance compared to conventional interpretation methods. The analysis was carried out using the Matlab GUI interface according to IEEE C57.104 and IEC 60599 standards with the integration of interpretation methods such as TDCG, Doernenburg, Duval, and carbon ratios. The test results showed that the TDCG method indicated safe conditions, while the Duval triangle detected Partial Discharge (PD) and Thermal Fault 1 (T1). In the classification model, KNN provides high accuracy ≥ 84% to 95% at k=17 values. These findings confirm that the application of KNN, combined with appropriate interpretation methods, is able to improve diagnostic accuracy, support maintenance effectiveness, and reduce the risk of electrical system disruptions.
| Item Type: | Thesis (S1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
|||||||||
| Additional Information: | Algoritma K-Nearest Neighbors (KNN) merupakan salah satu pendekatan kecerdasan buatan yang terbukti efektif dalam meningkatkan akurasi diagnosis kondisi transformator. KNN bekerja dengan prinsip kedekatan data, di mana klasifikasi ditentukan berdasarkan mayoritas kelas dari sejumlah tetangga terdekat. Keunggulan algoritma ini terletak pada kemampuannya mengenali pola distribusi data tanpa memerlukan asumsi tertentu, sehingga sangat sesuai untuk menganalisis variasi kompleks pada hasil Dissolved Gas Analysis (DGA). Penelitian ini bertujuan mengimplementasikan KNN dalam mengklasifikasikan kondisi minyak isolasi transformator berdasarkan data DGA serta mengevaluasi kinerjanya dibandingkan metode interpretasi konvensional. Analisis dilakukan menggunakan antarmuka GUI Matlab sesuai standar IEEE C57.104 dan IEC 60599 dengan integrasi metode interpretasi seperti TDCG, Doernenburg, Duval, dan rasio karbon. Hasil pengujian menunjukkan metode TDCG mengindikasikan kondisi aman, sedangkan segitiga Duval mendeteksi Partial Discharge (PD) dan Thermal Fault 1 (T1). Dalam model klasifikasi, KNN memberikan akurasi tinggi ≥ 84% hingga 95% pada nilai k=17. Temuan ini menegaskan bahwa penerapan KNN, dikombinasikan dengan metode interpretasi yang tepat, mampu meningkatkan akurasi diagnosis, mendukung efektivitas pemeliharaan, serta mengurangi risiko gangguan sistem kelistrikan. | |||||||||
| Uncontrolled Keywords: | Dissolved Gas Analysis, GUI Matlab, KNN, Partial Discharges, Thermal Fault | |||||||||
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering | |||||||||
| Divisions: | 03-Fakultas Teknik > 20201-Jurusan Teknik Elektro | |||||||||
| Depositing User: | Mr Khatibul Umam | |||||||||
| Date Deposited: | 30 Jan 2026 09:10 | |||||||||
| Last Modified: | 30 Jan 2026 09:10 | |||||||||
| URI: | http://eprints.untirta.ac.id/id/eprint/57726 |
Actions (login required)
 |
View Item |