Use of Artificial Neural Networks for Predicting Crude Oil Effect on Carbon Dioxide Corrosion of Carbon Steels

Abstract

The role of crude oil on carbon dioxide (CO 2 ) corrosion has gained special attention in the last few years because of its signifi cance when predicting corrosion rates. However, the complexity and variability of crude oils makes it hard to model its effects, which can infl uence not only wettability properties but also the corrosiveness of the associated brine. This study evaluates the usefulness of artifi cial neural networks (ANN) to predict the corrosion inhibition offered by crude oils as a function of several of their properties that have been related in previous studies to the protectiveness of crude oils, i.e., nitrogen and sulfur contents, resins and asphaltenes, total acid number, nickel and vanadium content, etc. Results showed that neural networks are a powerful tool and that the validity of the results is closely linked to the amount of data available and the experience and knowledge that accompany the analysis.

Models for predicting corrosion

“…An overall approach of this type of model is shown in Fig. 21.7 (Hernandez et al, 2006;Kumar and Buchheit, 2004). The model consists of several input values connected to hidden nodes through weighing factors.…”

Using a heuristic approach to derive a grey-box model through an artificial neural network knowledge extraction technique

“…The iterative scheme of the grey-box model is shown in (2). For example, to begin the procedure for the greybox model, an ANN will be trained using all independent variables (x 1 , x 2 ,…, x n and y 1 , y 2 ,…, y n ) to model the dependent variable (z).…”

“…Therefore, these models are more accepted by decision makers because they offer a transparency that ANNs do not provide. However, ANNs have historically outperformed their regressive counterparts in many forms of science and engineering including: physical [1], chemical [2], communication [3], financial [4], and ecological [5] systems.…”

Artificial Neural Networks for Corrosion Rate Prediction in Gas Pipelines

“…Using a number of data input and output characteristics of the fluid, the result of corrosion rate model will be able to predict the corrosion rate in the range of the entered input data. With models that are adaptive and can adjust to changes in the dynamic data, one will be able to obtain predictions with a great degree of accuracy.Research methodology on the subject using crude oil was developed by Hernandez and Nesic (2005). Some crude oil characteristic associated with the corrosion rate are, among others, sulfur, vanadium, nickel contents, total acid number and API.…”