Implementasi Text Mining Pada Pengelolaan Database Maintenance Sistem Generator Di Pembangkit Listrik

Wicaksono, Bagas Wahyu (2024) Implementasi Text Mining Pada Pengelolaan Database Maintenance Sistem Generator Di Pembangkit Listrik. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5010201013_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (6MB) | Request a copy

Abstract

Generator systems atau sistem pembangkit, merupakan salah satu bagian dari peralatan listrik yang dimiliki oleh PT XYZ. Evaluasi risiko telah dilakukan menggunakan metode Failure Mode, Effect, and Criticality Analysis (FMECA), serta menggunakan perangkat lunak Computerized Maintenance Management Systems (CMMS) untuk mencatat seluruh kegiatan corrective maintenance(CM), merencanakan jadwal pemeliharaan dan melacak riwayat pemeliharaan. Namun, data dari laporan FMECA dan work order CMMS sering kali mengalami inkonsistensi penulisan oleh operator yang berdampak pada database yang sangat bervariasi. Data yang tidak terstruktur serta variasi pengisian yang tinggi menyebabkan kesulitan dalam pengambilan keputusan oleh tim maintenance dilapangan untuk dapat menentukan equipment yang tergolong kritis dan perlu ditangani terlebih dahulu. Selain itu, banyaknya work order yang tidak tertangani mengakibatkan terjadinya penundaan pekerjaan dan menumpuknya work order maintenance dari periode satu ke periode berikutnya karena terbatasnya jumlah operator untuk menangani kegagalan dan perlunya proses konfirmasi pekerjaan pada top management. Penelitian ini bertujuan untuk menerapkan model text mining pada jenis data yang bersifat semi-struktur dan tidak terstruktur untuk mendapatkan hasil berupa identifikasi high maintenance equipment, standar data mode kegagalan, penyebab kegagalan, efek kegagalan dan penanganan kegagalan pada Generation System. Model text mining yang digunakan untuk menangani permasalahan ini yaitu Latent Dirichlet Allocation (LDA), salah satu model yang sering digunakan dan sangat populer dalam ranah topik modeling dengan kemampuannya menangani database tidak terstruktur, bahasa yang bervariasi, menangkap kandungan makna, dan mengelompokkan teks ke dalam topik yang lebih koheren. Output model dievaluasi menggunakan coherence score untuk mengetahui parameter optimal yang dapat digunakan pada setiap database maintenance. Hasil menunjukkan model LDA mampu menangani database maintenance yang mengalami inkonsistensi penulisan dengan mengimplementasikan eksperimen pada parameter alpha, beta, dan number of topics. Model mampu membentuk 69 standar mode kegagalan, 44 standar efek kegagalan, 37 standar penyebab kegagalan, dan 66 standar penanganan kegagalan serta mengetahui 41 equipment yang termasuk dalam kategori high maintenance. Worksheet FMECA terstandarisasi dibangun menggunakan software Ms.Excel memanfaatkan fitur data validation untuk membuat pengisian dengan mekanisme dropdown.
=================================================================================================================================
Generator systems or generation systems are one part of the electrical equipment owned by PT XYZ. Risk evaluation has been carried out using the Failure Mode, Effect, and Criticality Analysis (FMECA) method, as well as using Computerized Maintenance Management Systems (CMMS) software to record all corrective maintenance (CM) activities, plan maintenance schedules and track maintenance history. However, data from FMECA reports and CMMS work orders often experience writing inconsistencies by operators that impact highly variable databases. Unstructured data and high filling variations cause difficulties in decision-making by maintenance teams in the field to be able to determine equipment that is classified as critical and needs to be handled first. In addition, the large number of work orders that are not handled results in work delays and the accumulation of work orders for maintenance from one period to the next due to the limited number of operators to handle failures and the need for a work confirmation process at top management. This study aims to apply a text mining model to semi-structured and unstructured data types to obtain results in the form of identification of high maintenance equipment, standard data modes of failure, causes of failure, failure effects and failure handling in Generation Systems. The text mining model used to deal with this problem is Latent Dirichlet Allocation (LDA), one of the most frequently used and very popular models in the field of modeling topics with its ability to handle unstructured databases, varied languages, capture meaning content, and group text into more coherent topics. The model output is evaluated using a coherence score to determine the optimal parameters that can be used in each maintenance database. The results show that the LDA model is able to handle database maintenance that experiences writing inconsistencies by implementing experiments on alpha, beta, and number of topics parameters. The model is able to form 69 standard failure modes, 44 standard failure effects, 37 standards for failure causes, and 66 standards for handling failures and knows 41 equipment that is included in the high maintenance category. Standardized FMECA worksheets are built using Ms.Excel software utilizing the data validation feature to make filling with a dropdown mechanism.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Penambangan Teks, Sistem Pembangkit, Basis Data Pemeliharaan, Text Mining, Latent Dirichlet Allocation(LDA), Generation Systems, Database Maintenance
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > TJ Mechanical engineering and machinery > TJ174 Maintenance and repair of machinery T Technology > TJ Mechanical engineering and machinery > TJ778 Gas turbines T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK2797 Motor-generator sets. Cascade
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User:	BAGAS WAHYU WICAKSONO
Date Deposited:	17 Jul 2024 08:14
Last Modified:	17 Jul 2024 08:14
URI:	http://repository.its.ac.id/id/eprint/108407

Actions (login required)

View Item