篇名 | Computational Intelligence for Corrosion Rate Prediction of Refinery Cooling Water Plant |
---|---|
卷期 | 29:3 |
作者 | Chen Xuan 、 Feng Dan 、 Lei Jing |
頁次 | 001-011 |
關鍵字 | corrosion rate prediction 、 deep learning 、 refinery cooling water plant 、 restricted boltzmann machines 、 support vector machine 、 support vector regression 、 EI 、 MEDLINE 、 Scopus |
出刊日期 | 201806 |
DOI | 10.3966/199115992018062903001 |
In order to predict the corrosion rate of refinery cooling water plant effectively, this paper proposes a novel model, which combines the Restricted Boltzmann Machines (RBM) and Support Vector Regression (SVR). The SVR has been confirmed the effectiveness for corrosion rate prediction, but it has not been used to predict the corrosion rate of the refinery cooling water plant. Moreover, by the aid of Deep Learning (DL) approaches that can model high-level abstractions in the data, the hybrid model is implemented by integrating the RBM with SVR to predict the corrosion rate of the refinery cooling water plant. The proposed model can reduce the dimensionality of data space and preserve the effective features of refinery cooling water plant. Compared with only using SVR model, the combination of the RBM and SVR model can achieve higher prediction accuracy and more efficient and effective. Prediction accuracy is evaluated using the actual Industrial data. According to the comparison results, the hybrid model proposed was better than the previous models in predicting the corrosion rate of the refinery cooling water plant. The hybrid model proposed is a promising and practical methodology for realtime tracking of corrosion in refinery cooling water plant system. The factory can use the hybrid model to schedule maintenance process that leads to risk reduction of structure failure and maintenance cost.