Identification of Corrosive Substances and Types of Corrosion Through Electrochemical Noise Using Signal Processing and Machine Learning

Abstract

Several systems in industries are subject to the effects of corrosion, such as machines, structures, and a lot of equipment. As consequence, the corrosion can damage structures and equipment, causing financial losses and accidents. Among the most common types is the localized corrosion, and it is present in most industrial processes and is the most difficult to detect. Such consequences can be reduced considerably with the use of methods of detection, analysis and monitoring of corrosion in hazardous areas, which can provide useful information to maintenance planning and accident prevention. In this work, we analyze some features extracted from electrochemical noise for the classification of different types of localized corrosion. Furthermore, we use some techniques to identify corrosive substances that may cause corrosion in materials. For both tasks, we apply signal processing and machine learning techniques. Experimental results show that the features obtained using wavelet transform and recurrence quantification analysis are effective to solve both tasks: the corrosion identification and the classification of substances. Almost all evaluated machine learning techniques achieved an average accuracy above 90%.