• Korean Management Science Review
• /
• v.11 no.2
• /
• pp.133-146
• /
• 1994

This study compares the performance of neural networks and ordinary least squares regression with quality-control processes. We examine the applicability of neural networks because they do not require any assumptions regarding either the functional from of the underlying process or the distribution of errors. The coefficient of determination($R^2$), mean absolute deviation(MAD), and the mean squared error(MSE) metrics indicate that neural networks are a viable and can be a superior technique. We also demonstrate that an assessment of the magnitude of the neural notwork input layer cumulative weights can be used to determine the relative importance of predictor variables.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.23 no.1
• /
• pp.45-52
• /
• 2003

A number of stress corrosion cracks in turbine rotor disk keyway in power plants have been found and the necessity has been raised to detect and evaluate the cracks prior to the catastrophic failure of turbine disk. By ultrasonic RF signal analysis and using a neural network based on bark-propagation algorithm, we tried to evaluate the location, size and orientation of cracks around keyway. Because RF signals received from each reflector have a number of peaks, they were processed to have a single peak for each reflector. Using the processed RF signals, scan data that contain the information on the position of transducer and the arrival time of reflected waves from each reflector were obtained. The time difference between each reflector and the position of transducer extracted from the scan data were then applied to the back-propagation neural network. As a result, the neural network was found useful to evaluate the location, size and orientation of cracks initiated from keyway.