pISSN : 1598-706X / eISSN : 2288-8381
한국주조공학회지 (34권3호 107-111)
Estimation of Permeability of Green Sand Mould by Performing Sensitivity Analysis on Neural Networks Model
Estimation of Permeability of Green Sand Mould by Performing Sensitivity Analysis on Neural Networks Model
Subba Reddy. N†, Yong-Hyun Baek, Seong-Gyeong Kim and Hur Bo Young
Gyeongsang National University
Permeability is the ability of a material to transmit fluid/gases. It is an important material property and it depends on mould parameters such as grain fineness number, clay, moisture, mulling time, and hardness. Modeling the relationships among these variable and interactions by mathematical models is complex. Hence a biologically inspired artificial neural-network technique with a back-propagation-learning algorithm was developed to estimate the permeability of green sand. The developed model was used to perform a sensitivity analysis to estimate permeability. The individual as well as the combined influence of mould parameters on permeability were simulated. The model was able to describe the complex relationships in the system. The optimum process window for maximum permeability was obtained as 8.75-10.5% clay and 3.9-9.5% moisture. The developed model is very useful in understanding various interactions between inputs and their effects on permeability.
Green sand mould, Permeability, Neural networks, Sensitivity analysis