Modeling and mining dual-rate sampled data in corrosion potential online detection of low alloy steels in marine environment

With the rapid development of the Internet of Things technology, the use of front-end sensors makes the corrosion potential online detection of low alloy steels in marine environment a reality, thereby obtaining oceans of corrosion data. However, the monitoring frequency of seawater environmental data is usually much lower than corrosion potential data, which produces the dual-rate sampled corrosion datasets. In this paper, firstly, a method based on comprehensive index value (CIV) is proposed to process the dual-rate sampled corrosion data, which retains more information of original data compared with the mean value method. Secondly, the relevance vector regression model named CIV-RVR is established to predict the corrosion potential, which outperforms other modeling methods like artificial neural networks and support vector regression. Moreover, key corrosion resistance elements including Cr, Ni, Mo, P, Cu, Si and V of experimental steels are determined by Spearman correlation analysis, which have been pointed out as positive elements by previous studies. Finally, the effects of key corrosion resistance elements on seawater corrosion potential are quantitatively mined and visualized by applying the CIV-RVR model. The results mined by the proposed model show that elements Cu and P have a positive synergistic effect and can help to promote the corrosion potential of the micro-alloy steel, which is consistent with accepted conclusions. It can be concluded that the CIV-RVR model proposed in this paper can be well applied to model and mine dual-rate sampled data in corrosion potential online detection of low alloy steels under the marine environment.