Trend and Pattern of Studies on Corrosion and Corrosion Science: Fundamentals of Corrosion Science and Application of Artificial Intelligence and Machine Learning

Abstract

Corrosion is the process by which metallic material either degrades or loses its initial quality due to several factors, such as chemical reactions in the environment and high concentrations of salts, leading to the failure of the component. Corrosion of metals not only deteriorates components, but in addition it poses risks to life and the safety of both materials and humans. This corrosion process affects infrastructure and several sectors such as production, manufacturing, and transportation. The information and data in this review were assembled, edited, and refined to improve comprehension, readability, and application of corrosion science utilizing a variety of articles such as journals, conference papers, theses, books, and other credible literature published between 1800 and May 2026 (226 years) from standard databases such as Web of Science (WoS), Google Scholar, and Multidisciplinary Digital Publishing Institute (MDPI). The study revealed that a total of 1,091,722 (4831 articles per year) and 73,202 (324 articles per year) articles were published between 1800 and May 2026 on corrosion and corrosion science, application of Artificial Intelligence (AI), Artificial Neural Network (ANN) and Machine Learning (ML) to corrosion and corrosion science, respectively. The trend and pattern of these publications were exponential for both focused keywords. It was revealed that the forms of corrosion are erosion, cavitation, fretting, microbiological, atmospheric, fatigue, waterline, selective leaching (selective), dry (chemical), wet (electrochemical), galvanic, pitting, stress (cracking), and intergranular. The techniques for monitoring or detecting corrosion can be grouped into Non-Destructive Testing (NDT), Visual, Destructive Testing (DT), and Analysis Techniques. The use of AI, ANN, and ML has made predicting corrosion possible and faster, without stress. It was concluded that premature prediction and suitable corrosion control and monitoring can decrease economic losses, and that traditional approaches utilised in corrosion prediction and detection systems are time-consuming and challenging to execute in inaccessible areas