Application of artificial intelligence in scale thickness prediction on offshore petroleum using a gamma-ray densitometer

Abstract This work presents a methodology to study the deposition of scale in pipelines of multiphase systems (oil/water/gas), commonly found in the petroleum industry. The scale prediction for the pipelines is done through an artificial neural network, trained by using simulated data obtained with MCNP6 code, and transmission measurements. The model considered only barium sulfate (BaSO4) as main scale's material. The transmission setup is composed of a 137Cs (662 keV) volumetric source and one NaI(Tl) detector placed around the pipe. The pulse height distributions recorded in the detectors are used as input data of the artificial neural network. The results showed that 94% of the scale thickness prediction error was within ±5%. The scale thickness in the oil industry's pipes can be calculated by the artificial neural network regardless of the presence of fluids with satisfactory results in water-gas-oil multiphase system with annular flow regime.