Prediction of optimal mild steel weld parameters using the Adaptive Neuro-Fuzzy Inference System (ANFIS) technique

Welding is one of the major operations in many industries as it provides a durable means of joining metals and ensuring that diverse equipments are created to meet the growing needs of the manufacturing industries. To enhance the production of these diverse equipments, studies are continually been performed to identify improved means of obtaining reliable joints. This study applies the Adaptive Neuro-Fuzzy Inference System (ANFIS) technique, in improving the predictability of the optimal weld characteristics for a mild steel welded joints, with focus on tensile strength and hardness as responses. From the study, the variation in tensile strength and hardness as a result of the process parameter effects is illustrated, and it reveals the optimal tensile strength, and hardness is obtained at the combined input parameters: 170 Amp, 20 V, 24 l/min, and 2.2 mm for the tensile strength and 220 Amp, 20 V, 20 l/min, and 2.4 mm for the hardness.