Deteksi Kebocoran Pipa Air Menggunakan Data Arus Air Berbasis Time-Series

Hikmatullah, Kenji (2023) Deteksi Kebocoran Pipa Air Menggunakan Data Arus Air Berbasis Time-Series. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05111840000074-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (991kB) | Request a copy

Abstract

Di era revolusi industri saat ini, air merupakan suatu kebutuhan wajib bagi siapapun. Agar air dapat selalu digunakan, perlu adanya suatu mekanisme distribusi air yang andal. Salah satu masalah yang kerap terjadi pada distribusi air adalah adanya kebocoran pada pipa air. Kebocoran ini dapat menimbulkan kerugian biaya hingga membuat air bersih menjadi lebih langka. Berdasarkan hasil riset pada tahun 2018 di Australia, diperkirakan 12% air hilang karena kebocoran. Salah satu cara untuk mencegah masalah yang dapat timbul tersebut adalah dilakukan deteksi kebocoran pada pipa air mengggunakan data arus air. Deteksi kebocoran menggunakan data arus air yang berbasis time-series dapat menghasilkan performa yang lebih baik dibanding menggunakan data pada satu titik waktu. Dengan menggunakan data berbasis time-series, besar arus air normal dapat dianggap berbeda pada waktu yang berbeda. Pada penelitian ini, dilakukan percobaan simulasi dan deteksi lokasi kebocoran. Pada mulanya, peta jaringan pipa air di ITS digambar di aplikasi EPANET. Kemudian, timing dan pattern-nya diatur sedemikian rupa sehingga terdapat pola demand yang berulang setiap hari dalam sepekan. Pola ini digunakan untuk menghasilkan data arus air berbasis time-series. Setelah peta dibuat, dengan bantuan script berbasis Python, dilakukan simulasi kebocoran sebesar 0.5 LPS pada setiap junction, dimana kebocoran pada suatu junction merepresentasikan satu kasus kebocoran. Kemudian, menggunakan data arus air berbasis time-series dari hasil simulasi dan script berbasis Python, dilakukan uji coba deteksi lokasi kebocoran. Uji coba dilakukan menggunakan dua metode, yaitu Euclidean Distance dan Correlation Coefficient. Dari hasil uji coba yang telah dilakukan menggunakan metode Euclidean Distance, didapatkan hasil rata-rata confidence level sebesar 100%. Sedangkan metode Correlation Coefficient, didapatkan rata-rata confidence level sebesar 99.85%.
==============================================================================================================================
In the current revolution industry era, water is a must for everyone. So that water can always be consumed, there must be a reliable water distribution mechanism. One of the problems that often occurs in water distribution is leaks. These leaks can cost money and make clean water even more scarce. Based on research on 2018 in Australia, it is estimated that 12% water is lost due to leakage. One way to prevent problems that can arise is to do leak detection in water pipes using water flow data. Detection based on time series monitoring data is more convincing than one-time point data. By using time-series-based data, normal water flow patterns can be assumed to be different at different times. In this research, leak is simulated and the location is then predicted. At first, the water pipe network map at ITS is drawn on the EPANET application. Then, the timing and pattern are set in such a way that there is demand pattern that repeats every day of the week. This pattern is used to produce time-series based water flow data. After the map is created, with the help of Python-based script, leak simulation of 0.5 LPS is carried out at each junction, where a leak at a junction represents one leak case. Then, using time-series based flow data from the simulation results and Python-based script, the leak location is predicted. The prediction were carried out using two methods, namely Euclidean Distance and Correlation Coefficient. Based on the prediction result using Euclidean Distance method, the average confidence level is 100%. While the Correlation Coefficient method obtained an average confidence level of 99.85%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Deteksi Kebocoran Pipa, Arus Air, Data Berbasis Time-Series, Pipe Leak Detection, Water Flow, Time-Series Based Data.
Subjects:	T Technology > T Technology (General) > T57.62 Simulation
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Kenji Hikmatullah
Date Deposited:	09 Feb 2023 16:40
Last Modified:	09 Feb 2023 16:40
URI:	http://repository.its.ac.id/id/eprint/96622

Actions (login required)

View Item