Pitting corrosion prediction from cathodic data: application of machine learning

Abstract

Purpose To constitute input data, the authors carried out electrochemical experiments. The authors performed voltammetric scans in a very cathodic potential region. The authors constituted an experimental table where for each experiment we note the current values recorded at a low polarization range and the pitting potential observed in the anodic region. This study aims to concern carbon steel used in a nuclear installation. The properties of the chemical solutions are close to that of the cooling fluid used in the circuit. Design/methodology/approach In a previous study, this paper demonstrated the effectiveness of machine learning in predicting the localized corrosion resistance of a material by considering as input data the physicochemical properties of its environment (Boucherit et al., 2019). With the present study, the authors improve the results by considering as input data, cathodic currents. The reason of such an approach is to have input data that integrate both the surface state of the material and the physicochemical properties of its environment. Findings The experimental table was submitted to two neural networks, namely, a recurrent network and a convolution network. The convolution network gives better pitting potential predictions. Results also prove that the prediction by observing cathodic currents is better than that obtained by considering the physicochemical properties of the solution. Originality/value The originality of the study lies in the use of cathodic currents as input data. These data contain implicit information on both the chemical environment of the material and its surface condition. This approach appears to be more efficient than considering the chemical composition of the solution as input data. The objective of this study remains, at the same time, to seek the optimal neuronal architectures and the best input data.

Application of machine learning in predicting pitting corrosion – electrochemical data around the open circuit potential

“…This work concludes a series of works on pitting corrosion prediction by neural networks through the consideration of different electrochemical parameters. At this stage, we can advance that the pitting potential prediction can be done by observing the composition of the solution (Boucherit et al , 2019); observing the current during cathodic polarizations (Boucherit and Arbaoui, 2021); or, as we saw it in this paper, by observing the evolution of the open-circuit potential. To be able to advance which of these observations is the most advantageous, it is necessary to resume the work carried out but under strictly identical conditions.…”

“…We explained in a previous publication the operating principle of one-dimensional convolution Conv1D (Boucherit and Arbaoui, 2021). The operating principle of GRU is illustrated in Figure 3.…”

“…This work showed that input data adjustment by taking into account chemical equilibria could influence the performance of the neural model. In the second work, we built neural networks that ignore the chemical composition of the solution and consider rather current answers to cathodic polarizations (Boucherit and Arbaoui, 2021). The work compares two neuronal models: a convolution approach and a recurrent approach.…”

Application of artificial intelligence (AI) in the area of corrosion protection

“…In terms of corrosion science research, the application of AI mainly focuses on corrosion protection materials and methods (Belayadi et al , 2019; Coşkun and Karahan, 2018; Feiler et al , 2020; Jafari et al , 2011; Jiménez-Come et al , 2012; Mousavifard et al , 2015; Wen et al , 2009; Zadeh Shirazi and Mohammadi, 2017), corrosion image recognition (Ahuja and Shukla, 2018; Hoang and Tran, 2019; Petricca et al , 2016; Tian et al , 2019; Yao et al , 2019) and corrosion life prediction (Al-Shehri, 2019; Alani and Faramarzi, 2014; Chae et al , 2020; Guzman Urbina and Aoyama, 2018; Zhi et al , 2020). Compared with traditional methods, AI shows unique advantages (Boucherit et al , 2022; Boucherit et al , 2019; Boucherit and Arbaoui, 2021). Its efficient and in-depth data processing method can rapidly advance the research process (Butler et al , 2018; Kumari et al , 2017; Wang et al , 2019; Zhao et al , 2021).…”