×
The submission system is temporarily under maintenance. Please send your manuscripts to
Go to Editorial ManagerThe present work aims to build mathematical models based on experimental data to estimate the mechanical properties of submerged arc weldment. AISI 1020 low carbon steel plates 16mm thickness were welded according to orthogonal array in order to establish the relationship between input parameters (welding current, Arc voltage and welding speed) and output parameters (ultimate tensile stress, yield stress, impact energy and hardness) by submerged arc welding (SAW) process. The relationship between input and output parameters for the welding process are conducted using two suitable mathematical models the first one based on regression analysis, while the second one based on multi input single output ANFIS model for estimation of some mechanical properties of the welded plates. It was found that ANFIS results are closer to the experimental results than regression results. The optimal parameters (which give a maximum value of ultimate tensile strength (UTS), yield stress and impact energy; 446 MPa, 318 MPa and 213 J) are welding current is (380 Amp), Arc voltage is (25 V) and welding speed is (40 cm/min), while the maximum value of hardness number is (228 HV), when current welding is (380 Amp), Arc voltage is (25 V) and welding speed is (25 cm/min).
An intelligent and anticipatory speed controller for internal combustion engines was designed theoretically and examined experimentally. This design was based on the addition of a torque loop to the main speed loop. The model can sense the external load with the help of a load cell and send this signal to a soft computing unit for analysis and processing. This scheme will improve the ability of anticipation of controller since it treats the factors that affect the speed, not the speed itself. The experimental design was implemented using two types of actuating techniques; an intelligent throttling actuator and an intelligent injection actuator. The signal was analyzed by using intelligent techniques such as fuzzy logic, neural network and genetic algorithm. The experimental data were used to train the neural and the Adaptive Neuro–Fuzzy Inference System. The comparison of the results obtained in this work with other available models proved the efficiency and the robustness of the present model.