CD Skripsi

Deteksi Kebocoran Dan Sumbatan Pipa Menggunakan Sinyal Getaran Pada Pipa Bertekanan Dengan Sensor Adxl 335

---

ABSTRACT
Pipes are widely used for the transportation of liquids from one place to another, commonly found in domestic and industrial areas. Piping systems are becoming increasingly important for energy supply, economic activities, industry, the social sector, and other aspects of life in most countries because they are widely used for the distribution of water, oil, gas, and others. Pipeline damage can be caused by natural disasters, corrosion, third-party damage, mechanical failure, and others, thus requiring special monitoring and management actions for cost-saving and environmental reasons. One of the most reliable ways to detect early symptoms of blockage and leakage in pipeline systems is vibration analysis, which is currently the most commonly used method. In this study, damage detection was carried out on a pressurized pipe attached to the wall in the form of blockages and leaks using vibration signals with the ADXL335 accelerometer sensor based on Arduino. The accelerometer captures vibration signals generated from fluid flow driven by an electric pump. The data obtained by the accelerometer is then recorded and processed into frequency and amplitude values, with frequency values ranging between 51 – 52.5 Hz. This study uses three machine learning models for classifying the types of damage and detecting the position of the damage in the pipe, namely Decision Tree, K-Nearest Neighbors, and Support Vector Machine. The K-Nearest Neighbors-based model has the highest accuracy rate with a peak value of 95%, followed by the Decision Tree model with a highest accuracy of 90%, while the Support Vector Machine model cannot be used as a classification model due to its very low accuracy of around 55%. Damage position can also be detected using the regression method in machine learning with the Gaussian Process Regression model, which has the highest error percentage of 4.71%.
Keywords : Pipe Damage, Vibration Signal, Frequency, Amplitude, Machine Learning

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2007113874

Penerbit

Pekanbaru : Universitas Riau – F.TEKNIK – mesin., 2025

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2007113874

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

TEKNIK MESIN

Info Detail Spesifik

-

Pernyataan Tanggungjawab

daus

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar