Article Details

A hybrid multi-criteria approach to GPR image mining applied to water supply system maintenance

Oleh   Silvia Carpitella [-]
Kontributor / Dosen Pembimbing : Silvia J. Ocaña-Levario ; Julio Benítez ; Antonella Certa a ; and Joaquín Izquierdo
Jenis Koleksi : Jurnal elektronik
Penerbit : Lain-lain
Fakultas :
Subjek : Mining & related operations
Kata Kunci : GPR, Radargrams, Water Supply Systems, FAHP, ELECTRE III
Sumber : Journal of Applied Geophysics 159 (2018) 754–764, https://doi.org/10.1016/j.jappgeo.2018.10.021
Staf Input/Edit : Resti Andriani   Ena Sukmana
File : 1 file
Tanggal Input : 2019-01-10 14:14:29

Generic placeholder image
2018_EJRNL_PP_SILVIA_CARPITELLA_1.pdf

Akses : Terbatas
» ITB


Data processing techniques for Ground Penetrating Radar (GPR) image mining provide essential information to optimize maintenance management of Water Supply Systems (WSSs). These techniques aim to elaborate on radargrams in order to produce meaningful graphical representations of critical buried components of WSSs. These representations are helpful non-destructive evaluation tools to prevent possible failures in WSSs by keeping them adequately monitored. This paper proposes an integrated multi-criteria decision making (MCDM) approach to prioritize various data processing techniques by means of ranking their outputs, namely their resulting GPR image representations. The Fuzzy Analytic Hierarchy Process (FAHP) is applied to weight various evaluation criteria, with the purpose of managing vagueness and uncertainty characterizing experts' judgments. Then, the ELimination Et Choix Traduisant la RÉalité III (ELECTRE III) method is used to obtain the final ranking of alternatives. A real case study, focusing on a set of four GPR images as outputs of different data processing techniques, is presented to prove the usefulness of the proposed hybrid approach. In particular, the GPR images are ranked according the evaluation of five criteria namely visualization, interpretation, identification of features, extraction of information and affordability. The findings offer a structured support in selecting the most suitable data processing technique(s) to explore WSS underground. In the presented case, two options, namely the variance filter and the subtraction methods, offer the best results.