Prediction of Corrosion Rates in Structural Steel Using Artificial Neural Networks
Journal Title: International Journal for Research in Applied Science and Engineering Technology (IJRASET) - Year 2015, Vol 3, Issue 2
Abstract
A phenomenal outcome for the prediction of corrosion in steel was proposed with the learning ability of artificial neural network using MATLAB software. The prediction of corrosion rate has become an important challenge for the Indian steel industry as well as for the engineering community. This paper presents the studies carried out towards the prediction of corrosion rates by using artificial neural networks (ANN), in which training of 406 sets of data using Levenberg-Marquardt algorithm obtained from experimental data. The training sets have been developed for three levels of corrosion such as mild, moderate and severe through ANN and resulted in a trend of an incremental parabolic curve. The input parameters considered were equivalent to simulate corrosion of structural steel exposed to atmospheric, marine or chemical environment. The correlation statistics (R) in ANN has proved to be 90%. The test results have been validated to confirm the efficacy of developed ANN model for prediction of corrosion rate.
Authors and Affiliations
Sharon John, Umesha P. K, Cinitha A, M. G. Rajendran
Keywords
Related Articles
- EP ID EP19585
- DOI -
- Views 289
- Downloads 5
Molecular Characterization of Medicinal Plant – Cassia Fistula L by Matk Gene
The studies on taxonomy, ethnobotany and antimicrobial potential are imperative for plants with traditional medicinal value. Two important reasons for plant barcoding is to identity unknown specimens to known species an...
Next Generation Mobile Network in India and Impact on the Indian Entrepreneurship EcoSystem
The year 2016 was perhaps a watershed in the history of the telecommunications industry in India with the launch of the Reliance Jio. Reported to be the largest ever investment for a single brand in the history of the s...
An Efficient Algorithm for Approximate String Matching
The approximate string matching is the technique of finding strings that match a pattern approximately (rather than exactly). Most often when we need to match a pattern exact matching is not possible, due to insufficien...
Signature Recognition Using Matlab
Signature has been a distinguishing feature for person identification. Even today, an increasing number of transactions, especially related to financial and business are being authorized via signatures. Hence, the need...
Crack detection and crack propagation through vibration monitoring in sugar mill roller shaft
The effects of vibration in mill roller shaft during sugar cane crushing which is one of the crack possibility phenomena were analysed. Inducer tests under various unbalanced conditions performed to clarify the effect o...