Print ISSN: 1681-6900

Online ISSN: 2412-0758

Keywords : Neuro


Prediction Fatigue Life of Aluminum Alloy 7075 T73 Using Neural Networks and Neuro-Fuzzy Models

Mustafa S. Abdullatef; Nazhat . AlRazzaq; Mustafa M. Hasan

Engineering and Technology Journal, 2016, Volume 34, Issue 2, Pages 272-283

In present paper the fatigue life of aluminum alloy 7075 T73 under constant amplitude loading is predicted using ANN and ANFIS models. Many neural networks models are used for this purpose and also different neuro-fuzzy models are built for predict fatigue life.Theclassical power law formula ismost common used to find fatigue behaviors of materials. In present study, two techniques are used to find coefficients of the formula linear and nonlinear regression. Forcomparison the fatigue life curves of soft computing methods are plotted together with two conventionalmethods. The neural network and neuro-fuzzy models give good results compared with two conventional methods. Also it is shown thatneural network model which is trained using Levenberg-Marquardt algorithm is best neural network modelscompared with other NNS models.Also, it is foundANFIS models with input trapezoidal membership function is best performance from other membership function types to predict fatigue life. It can be stated that neuro-fuzzy models are better models than neural network and conventional methods to predict fatigue life of the maintained alloy.

Study the Robustness of Automatic Voltage Regulator for Synchronous Generator Based on Neuro-Fuzzy Network

Abdulrahim Thiab Humod; Yasir Thaier Haider

Engineering and Technology Journal, 2015, Volume 33, Issue 3, Pages 612-627

Modern power systems are complex and non-li¬near and their operating conditions can vary over a wide range, and since neuro - fuzzy networkcan be used as intelligent controllers to control non-li¬near dynamic systems through learning, which can easily accommodate the non-linearity, time dependencies, model uncertainty and external disturbances.ANeuro-Fuzzy model system is proposed as an effective neural network controller model to achieve the desired robust Automatic Voltage Regulator (AVR) for Synchronous Generator (SG) to maintain constant terminal voltage. TheconcernedNeuro-fuzzy controller for AVRis examined on different models of SG andloads. The results show that the Neuro-Fuzzy -controllers have excellent responses for all SG models and loads in the view point of transientresponse and system stability compared with optimal PID controllers tuned by practical swarm optimization.They also show that the margins of robustness for Neuro-Fuzzy -controller aregreater thanPID controller.