Based on Vibration and Improved GRNN Identify Eggshell Crack

Jian Yang; He Pan; Zhan Wu Peng; Xiu Ying Li

doi:10.4028/www.scientific.net/AMM.472.404

Paper Titles

Advanced Performance Estimation Model for Lithium-Ion Power Battery Used in Pure Electric Vehicles
p.379

Verification of Control System by Physical Simulation
p.389

Research on the Detail Design of Aircraft Wings
p.394

Exploration of Forecasting the State of Metal Oxide Volt-Sensitive Protectors Based on Improve Recursive Neural Networks
p.398

Based on Vibration and Improved GRNN Identify Eggshell Crack
p.404

Study of Establishment and Operation of Salt-Gradient Solar Pond
p.409

Irradiance and Temperature Dependence Characterization of Vacuum Glass BIPV Modules
p.413

AR Technology and its Application in Biomedical Engineering
p.421

Multi-Radius Density Clustering Algorithm Based on Outlier Factor
p.427

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vol. 472Based on Vibration and Improved GRNN Identify...

Based on Vibration and Improved GRNN Identify Eggshell Crack

Abstract:

When knocking at crack eggs and nondestructive egg, a flexible piezoelectric film sensor shall be adopted to obtain frequency domain characteristic signals. Consequently, the frequency domain characteristic curves of nondestructive eggs show obvious main frequency domain values, while those of crack eggs show multiple and irregular peak values. In accordance with equal interval frequencies, the first 16 maximum and minimum amplitude values are selected orderly as input vector training Generalized regression neural network .Its Predicted output vector is the egg cracks.. In order to improve prediction accuracy, smooth factor of GRNN neural network is calculated by Particle Swarm Optimization. The experimental result showed that accurate prediction rate of GRNN is up to 94%.