Big Data Approach Application for Steel Pipelines in the Conditions of Corrosion Fatigue

Abstract

This paper presents results of the use of Big Data approach and neural network for the pipelines diagnosis problem. In this case the pipeline is in the conditions of crack growth of corrosion fatigue and exposed to hydrogen. It is proposed to use graphene protective coatings. The mathematical model for estimating the changes in the effective surface energy of WPL during plastic deformation, electrochemical overstrain, polarization potential and current density of the metal dissolution reaction at the top of the crack on the pipeline surface during its mechanical loading in an aqueous electrolyte solution is given. The dissolution of the metal is considered on the juvenile surface, taking into account the anode and cathode regions based on the approaches of surface physics and electrochemistry. An element of a mathematical model is a quality functional, taking into account information flows and a sensitivity coefficient. Functional quality is used to specify the feedback between the investment project methodology and risk estimates, as well as to optimize the information flows of enterprises and improve the system of protection of metallic underground pipelines that operate under conditions of corrosion fatigue. The purpose of this project is to improve the relevant regulatory and technical documents as well as software.

У роботі подано результати використання підходів Big Data та нейронних мереж для діагностування трубопроводів. Метою цього проекту є вдосконалення відповідних нормативів та технічних документів, а також програмного забезпечення.