Abstract:
Flux Cored Arc Welding (FCAW) is an advanced arc welding process that uses a
continuously fed tubular electrode wire with internal flux to generate heat and protect
the weld fromcontamination. In the oil and gas industry, ensuring high-quality pipeline
welds is critical for structural integrity and public safety. Depth of Penetration (DoP)
strongly influences weld joint strength, while excessive Heat Affected Zone (HAZ)
can compromise material properties. This study applied Artificial Neural Networks
(ANNs) to predict and optimize FCAW process parameters to maximize DoP and
minimize HAZ in pipeline welding. The research investigated welding speed, torch
angle, contact-tip-to-work distance, welding current, arc voltage, and heat input as key
variables. Pipe samples of ASTM A335 Grade B– API 5L Schedule 40 were prepared.
The experiments were designed using the Taguchi method. Welding was performed
using an MMA/MIG/TIG 200 Prescott FCAW machine. An ANN model was designed
and trained in MATLAB, with optimization achieved through Stochastic Gradient
Descent (SGD) with momentum. Confirmatory welding trials were conducted at the
optimizedparameters. WeldqualitywasassessedusingVickersMicrohardness, Charpy
Impact, tensile testing, and microstructural examination. The ANN model achieved
an optimum DoP of 7.66 mm and HAZ of 2.90 mm, with prediction accuracies of
99.9948% and 99.9966% respectively, compared to validation results of 7.70 mm
DoP and 2.91 mm HAZ at a heat input of 1.13 kJ/mm. Mechanical testing showed
a Yield Strength of 276.23 MPa, Ultimate Tensile Strength (UTS) of 483.37 MPa,
Engineering Strain of 0.2026 mm/mm, Fusion Zone (FZ) average hardness of 230.7
HV, and Impact toughness of 1.3598 J/mm2. Optimum FCAW process parameters
of heat input (1.13 kJ/mm), welding current (126 A), arc voltage (21.5 V), welding
speed (115 mm/min), torch angle (450), and contact-tip-to-work distance (5 mm),
produced the most desired microstructure, acicular ferrite. The microstructure had
refined grains, optimal ferrite-pearlite balance, effective tempering, and minimized
brittle phase formation across the Coarse Grained Heat Affected Zone (CGHAZ), Fine
Grained Heat Affected Zone (FGHAZ), Inter-Critical Heat Affected Zone (ICHAZ),
and Sub-Critical Heat Affected Zone (SCHAZ). The ANNmodeleffectively optimized
FCAWparameters, producing welds with deep penetration, narrow HAZ, and superior
mechanical properties. This approach offers a reliable predictive tool for improving
pipeline weld quality, supporting safer and more durable infrastructure in the oil and
gas industry
