Abstract
The Engineering industry is constantly exploring an effective and fast-solving method for complicated engineering problems. The adaptation of artificial intelligent technology can diminish the time-consuming of conventional analysis methods, especially in offshore engineering. For that reason, this study is pursued to build a prediction model to predict the residual strength of API 5L X42 subsea pipelines. An artificial neural network is used as an analytical medium in developing the prediction model. Three (3) physical shapes of corrosion data with diverse corrosion level are designed as input data based on the corroded subsea pipelines of true 2009 historical inspection data of South China Sea. The output data are obtained from the finite element analysis to produce the burst pressure data. The performance model is evaluated using mean squared error (MSE) and mean absolute error (MAE) which results in 9.13 x 10-5 and 0.005499 respectively for the optimum model. The predicted output shows significant similarity in line with the finite element output for validation purposes. This model is expected to provide quick prediction reliability of subsea pipelines to the engineers and reduce or eliminate massive analysis work.
Highlights
The structural reliability of pipelines system is one of the major contributors to the efficiency of productivity in the oil and gas industries
The performance of the artificial neural network (ANN) model is monitored by the value of mean square error (MSE) and mean absolute error (MAE)
A total of 50 models are trained continuously resulting in different MSE and MAE values. '0' MSE value denotes a perfect model without error
Summary
The structural reliability of pipelines system is one of the major contributors to the efficiency of productivity in the oil and gas industries. Corrosion for underwater pipelines especially subsea pipelines is categorized into seven types of defects, that is axial grooving, axial slotting, circumferential grooving, circumferential slotting, general, pinhole and pitting corrosion All these types of corrosion reduce the strength of the pipelines which may put the service to a complete stop if no repair measures are taken. A failed pipeline is identified by determining the remaining burst pressure of the corroded pipelines and comparing with the Estimate Repair Factor, ERF. Several established codes such as ASME B31G, Modified ASME B31G, DNV RP F101, Civil Engineering and Architecture 10(1): 334-344, 2022
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
