Prediction of corrosion resistance of some dental metallic materials applying artificial neural networks

Abstract

Numerous studies have been done on the corrosion process of different dental metallic materials, especially using a comparative method and analyzing the electrochemical phenomena. Simultaneously, the effects of the corrosion process have been quantify by different physical quantities, such as corrosion rate, corrosion resistance, polarization resistance, corrosion current density etc. These experimental data can be used to model the corrosion process and, subsequently, to perform predictions with the aim to analyze or to control the process. In this work, a series of experimental data about corrosion of some titanium-based dental materials in artificial saliva were obtained through electrochemical impedance spectroscopy (EIS) tests and used as a database to develop a model of corrosion process by artificial neural networks. The process parameters taken into account were chemical compositions of the materials (cp–Ti, NiTi, NiTiNb), immersion time, pH, NaF content, and albumin content. The corrosion resistance of the metallic materials was evaluated by polarization resistance determined by EIS tests. Neural networks were developed and applied for evaluating the corrosion resistance of the alloys, depending on the process parameters. The predictions provided by the model are useful to understand the contribution of each parameter in the process and possible ways to control it.