BP神经网络对3C钢腐蚀性能的预测分析

BP Neural Networks Used in Prediction and Analyses of 3C Steel Corrosion Function

利用三层BP神经网络,根据已有的3C钢在不同海水环境参数下的腐蚀速度数据,建立了3C钢在海洋环境中腐蚀速度的人工神经网络模型;并分析预测了海水环境参数与腐蚀速度之间的关系。预测结果表明:在近海条件下,温度和含氧量越低,氧化还原电位越高,腐蚀速度越慢;在弱碱性条件及盐度大于25ppt时,腐蚀速度比较慢。此环境参数下,3C钢材料的腐蚀较小。本文的预测结果能很好地反映出海水环境参数对腐蚀速度的影响。

The neural network model for 3C steel corrosion speed in the seawater environment was set up by three layers of BP neural networks, according to the 3C steel corrosion speed data under different seawater environment parameters. The relation between the seawater environment parameters and corrosion speed was analyzed and forecast. The forecast results show that, under the coastal waters condition, the lower the temperature and oxygen content, and the higher the.redox potential, the slower the corrosion rate is; In the condition of slightly alkaline and salinity greater than 25 ppt, the corrosion rate is slow. And the corrosion to 3C steel materials is comparatively minor under those environment parameters. The forecast results of this paper reflect the influence of seawater environmental factors on the corrosion rate well.