Print ISSN: 1681-6900

Online ISSN: 2412-0758

Keywords : Artificial Neural Network (ANN)

Intelligent controller Design based on wind-solar system

Areeg F. Hussein; Hanan A. R. Akkar

Engineering and Technology Journal, 2021, Volume 39, Issue 2A, Pages 326-337
DOI: 10.30684/etj.v39i2A.1761

This paper presents an Intelligent controller designed to mastery the output power flow from the Solar System, the Wind system, the sum of the two systems or from the battery system, according to the Maximum power point tracking algorithm, to ensure the continuity of the output power at fast time response. The proposed controller has been designed using MATLAB m-file and trained with the different number of hidden neurons using two different algorithms to get as fast a response time with minimum Mean Square Error (MSE) as possible which resulted in six hidden neurons using Levenberg-Marquardt training algorithms.

Prediction of Surface Roughness of Mild Steel Alloy in CNC Milling Process Using ANN and GA Technique

Hind H. Abdulridha

Engineering and Technology Journal, 2020, Volume 38, Issue 12, Pages 1842-1851
DOI: 10.30684/etj.v38i12A.1579

In this paper, Analysis Of Variance (ANOVA), Artificial Neural Network (ANN), and Genetic Algorithm (GA) have been studied to predict the effect of milling parameters on the Surface Roughness (Ra) during machining of mild steel alloy. The milling experiments carried out based on the Taguchi design of experiments method using (L16) orthogonal array with 3 factors and 4 levels. The influence of three independent variables such as spindle speed (910, 930, 960, and 1000 rpm), feed rate (93, 95, 98, and 102 mm/min), and Tool Diameter (8, 10, 12, and 14 mm) on the Surface Roughness (Ra) were tested and analyzed with (ANOVA) to predict the response which indicates that spindle speed was the most significant factor effecting on Surface Roughness (Ra). Artificial Neural Network (ANN) and numerical methods are used widely for modeling and predict the performance of manufacturing technologies. Neural Network technique with 2 hidden layers, 10 neurons size, 1000 epochs, and Trainlm transfer function is used to predict the result. The Genetic Algorithm (GA) has been utilized to find optimal cutting conditions during a milling process.
From the results, the optimal value of spindle speed is (930 rpm), feed-rate is (95 mm/min) and tool diameter is (8 mm). This network structure is capable of predicting the Surface Roughness (Ra) well to optimize the milling parameters. Artificial Neural Network (ANN) predicted results indicate good agreement between the experimental and the predicted values

Prediction the Influence of Machining Parameters for CNC Turning of Aluminum Alloy Using RSM and ANN

Hind H. Abdulridha; Aseel J. Helael; Ahmed A. Al-duroobi

Engineering and Technology Journal, 2020, Volume 38, Issue 6, Pages 887-895
DOI: 10.30684/etj.v38i6A.705

The main objective of this paper is to develop a prediction model using Response Surface Methodology (RSM) and Artificial Neural Network (ANN) for the turning process of Aluminum alloy 6061 round rod. The turning experiments carried out based on the Central Composite Design (CCD) of Response Surface Methodology. The influence of three independent variables such as Cutting speed (150, 175 and 200 mm/ min), depth of cut (0.5, 1 and 1.5 mm) and feed rate (0.1, 0.2 and 0.3 mm/rev) on the Surface Roughness (Ra) were analyzed through analysis of variance (ANOVA). The response graphs from the Analysis of Variance (ANOVA) present that feed-rate has the strongest influence on Ra dependent on cutting speed and depth of cut. Surface response methodology developed between the machining parameters and response and confirmation experiments reveals that the good agreement with the regression models. The coefficient of determination value for RSM model is found to be high (R2 = 0.961). It indicates the goodness of fit for the model and high significance of the model. From the result, the maximum error between the experimental value and ANN model is less than the RSM model significantly. However, if the test patterns number will be increased then this error can be further minimized. The proposed RSM and ANN prediction model sufficiently predict Ra accurately. However, ANN prediction model is found to be better compared to RSM model. The artificial neutral network is applied to experimental results to find prediction results for two response parameters. The predicted results taken from ANN show a good agreement between experimental and predicted values with the mean squared error of training indices equal to (0.000) which produces flexibility to the manufacturing industries to select the best setting based on applications.

Fault Diagnosis in Wind Power System Based on Intelligent Techniques

Kanaan A. Jalal; Lubna A. Abd alameer

Engineering and Technology Journal, 2018, Volume 36, Issue 11A, Pages 1201-1207
DOI: 10.30684/etj.36.11A.11

Wind energy is one of the most important sources as well as being environmentally friendly and sustainable. In this paper, different types of faults of Doubly-Fed Induction Generator (DFIG) have been studied based on Artificial Neural Network (ANN), Particle Swarm Optimization (PSO) and Field Programmable Gate Array. To simulate the wind generators model MATLAB/Simulink program has been used. Artificial Neural Network (ANN) is trained for detection the faults and (PSO) technique is used to get the best weights. After the training process, the network was transformed into a Simulink program and then converted into the Very High Speed Description Language (VHDL) for downloading on the (FPGA) card, which in turn is used to detect and diagnosis the presence of faults where it can be re-programmed with high response and accuracy.