Browsing by Subject "ANN"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    AI Based Modelling and Optimization of Turning Process
    (2012-08) Kulkarni, Ruturaj Jayant; El-Mounayri, Hazim; Anwar, Sohel; Wasfy, Tamer
    In this thesis, Artificial Neural Network (ANN) technique is used to model and simulate the Turning Process. Significant machining parameters (i.e. spindle speed, feed rate, and, depths of cut) and process parameters (surface roughness and cutting forces) are considered. It is shown that Multi-Layer Back Propagation Neural Network is capable to perform this particular task. Design of Experiments approach is used for efficient selection of values of parameters used during experiments to reduce cost and time for experiments. The Particle Swarm Optimization methodology is used for constrained optimization of machining parameters to minimize surface roughness as well as cutting forces. ANN and Particle Swarm Optimization, two computational intelligence techniques when combined together, provide efficient computational strategy for finding optimum solutions. The proposed method is capable of handling multiple parameter optimization problems for processes that have non-linear relationship between input and output parameters e.g. milling, drilling etc. In addition, this methodology provides reliable, fast and efficient tool that can provide suitable solution to many problems faced by manufacturing industry today.
  • Thumbnail Image
    Item
    AI on the Edge with CondenseNeXt: An Efficient Deep Neural Network for Devices with Constrained Computational Resources
    (2021-08) Kalgaonkar, Priyank B.; El-Sharkawy, Mohamed A.; King, Brian S.; Rizkalla, Maher E.
    Research work presented within this thesis propose a neoteric variant of deep convolutional neural network architecture, CondenseNeXt, designed specifically for ARM-based embedded computing platforms with constrained computational resources. CondenseNeXt is an improved version of CondenseNet, the baseline architecture whose roots can be traced back to ResNet. CondeseNeXt replaces group convolutions in CondenseNet with depthwise separable convolutions and introduces group-wise pruning, a model compression technique, to prune (remove) redundant and insignificant elements that either are irrelevant or do not affect performance of the network upon disposition. Cardinality, a new dimension to the existing spatial dimensions, and class-balanced focal loss function, a weighting factor inversely proportional to the number of samples, has been incorporated in order to relieve the harsh effects of pruning, into the design of CondenseNeXt’s algorithm. Furthermore, extensive analyses of this novel CNN architecture was performed on three benchmarking image datasets: CIFAR-10, CIFAR-100 and ImageNet by deploying the trained weight on to an ARM-based embedded computing platform: NXP BlueBox 2.0, for real-time image classification. The outputs are observed in real-time in RTMaps Remote Studio’s console to verify the correctness of classes being predicted. CondenseNeXt achieves state-of-the-art image classification performance on three benchmark datasets including CIFAR-10 (4.79% top-1 error), CIFAR-100 (21.98% top-1 error) and ImageNet (7.91% single model, single crop top-5 error), and up to 59.98% reduction in forward FLOPs compared to CondenseNet. CondenseNeXt can also achieve a final trained model size of 2.9 MB, however at the cost of 2.26% in accuracy loss. Thus, performing image classification on ARM-Based computing platforms without requiring a CUDA enabled GPU support, with outstanding efficiency.
  • Thumbnail Image
    Item
    Simulate Turning Process using ANN, Predict Optimum Control Factors to achieve Minimum Surface Roughness
    (Office of the Vice Chancellor for Research, 2012-04-13) Kulkarni, Ruturaj; El-Mounayri, Hazim
    Abstract Turning is a material removal process, a subtractive form of machining which is used to create parts of circular or rotational form of desired geometry/shape by removing unwanted material. Accuracy of any process depends on involvement of operational variables. The operating parameters that contribute to turning process are Cutting speed, Feed rate, Depth of cut. Vibrations, tool wear, tool life, surface finish and cutting forces etc are also in direct relation with values selected for process parameters. Hence to improve the efficiency of process and quality of the product it is necessary to control the process parameters. Surface roughness is the parameters with main focus, as it dictates the aesthetics and sometimes ergonomical characteristics of the product. The tests were carried out on AISI 4140 steel. 12 speed Jones and Lamson Lathe model was used for turning operation. The specimen with a diameter of 60mm, 500mm length and hardened 35 HRC is used. The tool used for this is one that is most commonly used for turning process DTGNR 163 C 0° Lead Angle 60° Triangle insert. It is product of Kennametal. Statistical Design of Experiments was used to reduce the total number of trials in order to save the time and resources without compromising the accuracy of prediction. These readings are used to train and validate the Neural Network. ANN is found to be very useful with simulations tasks which have complex and explicit relation between control factors and result of process. Neural Network was created using feed forward back propagation technique for simulation of the process using the Matlab Neural network toolbox. With assurance of accuracy of the predictive capabilities of the neural network, it was then used for optimization. Particle Swarm Optimization Algorithm, an evolutionary computation technique is used find out the optimum values of the input parameters to achieve the minimum surface roughness. The objective function used here is to minimize the surface roughness. Limits of the operational variables are used as constraints for developing the code for optimization algorithm. Keywords: Turning process, Surface roughness, Artificial Neural Network, Particle swarm optimization.