Influence of surface quality on residual stress of API 5L X80 steel submitted to static load and its prediction by artificial neural networks

• Published in: International journal of advanced manufacturing technology 2020-06, Vol.108 (11-12), p.3753-3764
• Main Authors: Silva, Danillo Pedro, Bastos, Ivan Napoleão, Fonseca, Maria Cindra
• Format: Article
• Language: English
• Subjects: Artificial neural networks; CAE) and Design; Compressive properties; Computer-Aided Engineering (CAD; Control equipment; Controlled rolling; Engineering; Exposure; High strength low alloy steels; Industrial and Production Engineering; Low carbon steels; Machining; Mechanical Engineering; Media Management; Microalloying; Neural networks; Original Article; Residual stress; Service loads; Shot peening; Static loads; Steel; Structural failure; Surface properties; Surface roughness; Surface treatment
• ISSN: 0268-3768; 1433-3015
• DOI: 10.1007/s00170-020-05621-2

Microalloyed low carbon steels with high mechanical strength and elevated toughness obtained by controlled rolling have been widely used in several equipment of oil and gas industry, ensuring safety and reliability. However, residual stresses are inherent to all manufacturing processes and their knowledge is of great importance, considering that in presence of corrosive environments, the joint effect of stress with service loads can lead to structural failure. In this work, the influence of surface quality obtained by machining, shot peening, and bristle blasting was studied on the residual stresses of API 5L X80 steel submitted to static loads, with and without the presence of a corrosive medium. Samples were submitted to static loading performed by proof rings and the residual stresses were analyzed by X-ray diffraction using sin 2 ψ method. The effect of input conditions surface treatment, exposure medium, and time on residual stress was analyzed via artificial neural networks. Results indicated that surface treatment and the exposure medium have greater influence on residual stress states than loading time, suggesting that the corrosion process along with coarse roughness affects significantly residual stresses of API 5L X80 subjected to static loads. Despite the presenting coarse surface roughness, shot peening was an effective treatment to generate and also maintain stable compressive residual stresses along loading time. Moreover, artificial neural networks with supervised training predicted in an effective way experimental residual stresses for the studied steel even under different conditions of surface treatments, exposure medium, and loading time.