Artificial Neural Network Simulation Model for Predicting Oil and Gas Pipeline Corrosion Rate in Nigerian Niger Delta

Abstract

This study appraised the significance of all the factors affecting corrosion of oil and gas pipeline in Niger
Delta region of Nigerian using artificial neutral network (ANN) technique in order to ascertain the root
cause of incessant unexpected pipe failure in this region. Operational and process data of sixty oil and gas
transmission pipelines used by six oil and gas companies from 2000 to 2010 for onshore and offshore
applications at different oil and gas fields located in this region were sampled and used for both ANN
model development and parametric analysis. Results revealed that the predictions of ANN simulation
developed in this study which approximates the actual corrosion rates by over 99% are greater than those
of conventional De-Waard based simulations presently used by most Nigerian oil and gas companies.
Thus, the inadequate prediction of oil and gas pipeline corrosion rate resulting to incessant unexpected
pipeline failure in this region. Parametric analysis of the model showed that corrosion rate of carbon steel
pipelines used for oil and gas transportation in this region varies linearly with temperature, flow pressure,
CO2 partial pressure, pipe length and pH value while the effects of the pipe age, flow velocity, density,
viscosity, chloride sand flow and pipe diameter are non-linear. The used of models/simulations whose
predictions are relatively less than the actual corrosion rate of oil and gas pipeline by Nigerian oil and gas
companies should be discouraged in order to eliminate catastrophic pipeline failure and its resulting oil
spillage/environmental degradation in this region.