GPR-based adaptive neuro fuzzy inference system [ANFIS] model for diesel pipeline corrossion / Amir Hamzah Hijazi Abidin Jali Hijazi

This research stimulated the investigation of using the Adaptive Neuro Fuzzy Inference System (ANFIS) model in detecting corrosion of diesel pipeline. This problem highlighted to retrieve the permittivity on contaminated soil because there were several common factors that affected dielectric permittivity such as temperature and moisture content which were causing difficulty in GPR data interpretation. This research aims to investigate the capability of ANFIS in locating the underground diesel pipeline corrosion based on GPR data. The aim of this research can be achieved by identifying three objectives for these studies. The objective is to identify the parameters involved in designing ANFIS, to identify the capabilities of ANFIS in for modeling the GPR data, and to produce an analysis of the technique for locating the corrosion in the underground pipeline. The study area was at UiTM Perlis. The experimental site designed and filled with dry sand and Jitra soil with corroded pipeline installed and diesel is inserted through the pipeline. GPR measurement carried out by using an 800MHz antenna while the temperature and soil moisture reading recorded to analyze the relationship between temperature and soil moisture towards permittivity. Dielectric permittivity of underground contaminant retrieved based on dielectric contrast in radargram, GPR signal amplitude, time travel, and electromagnetic wave velocity. The dielectric permittivity reading then analyzed using the ANFIS method. MATLAB is used to produce a model for a corroded pipeline based on the ANFIS method. This study aimed to understand and utilize the use of ANFIS in producing a model for detecting corrosion on the underground pipeline.