Prediction system for seawater corrosion of steel based on accumulating of corrosion data and analyzing of artificial neural network

DU Cuiwei，LI Xiaogang，Gao jin＊

School of Material Science and Engineering, University of Science and Technology Beijing, Beijing 100083, China

Abstract

According to the corrosion data obtained from China aqueous corrosion sites, corrosion rules of carbon steel and low alloy steel were analyzed and studied under the seawater environments in Qingdao, Xiamen, Zhoushan, Yulin by artificial neural network, regression analysis, grey system theory and computer technology. These four sites represented the typical seawater environment of China. The corresponding models were set up based on the data and further the seawater corrosively was determined.

The corrosion rules of carbon steel and low alloy steel in seawater were studied with neural network technology and mathematic method. Firstly, the main influencing factors on metal corrosion in seawater environments were analyzed with regression analysis and network technology. Seawater temperature, marine growth adhesion and pH were concluded as the main factors. Based on that, the corrosion rates of seawater of carbon steel and low alloy steel were estimated. Then, mathematic prediction models and BP neural network models were set up separately. The BP artificial neural network models without environmental factors, showed the relationship between metal corrosion rates and alloy elements, and the artificial neural network models considering environmental factors, showed the relationship between metal corrosion rates and water corrosion factors. Mathematic models such as regression models and grey models that show the relationship between the corrosion rate and environmental factors were also built up. Based on the forecast result of 16^th year corrosion rate of low carbon steel and low alloy steel in different seawater corrosion sites, the comparison of the different models was made. Furthermore, the optimum prediction method for carbon steel and low alloy carbon steel was obtained, which provided a good reference to corrosion decision-making and the relevant information for corrosion evaluation.

Key Words： Carbon steel and low alloy steel, seawater corrosion, artificial neural network

* Contract Email: g.jin@163.com

** Acknowledgements

The authors are grateful for the financial support from the National R&D Infrastructure and Facility Development Program of China (registration number: 2005DKA10400), from the National natural science fund project, No: 50499336-10