•  
  •  
 

Article Type

Research Paper

Abstract

Wax precipitation from crude oil remains a major flow assurance issue in the oil industry. Accurate determination of the Wax Appearance Temperature would prevent sudden clogging of production facilities. Existing WAT predicting models shows inadequacy in accurately predicting the WAT and thus pose impending challenges as some either underestimates or overestimates the WAT. Using Multiple Regression Analysis, five new semi-empirical models are developed to give a more reliable estimation of WAT. These comprise of two compositional models, one GOR based model and two hybrid models. Using published data, these models were compared with existing models to determine their accuracy and reliability. The coefficient of determination (R2) for the compositional and GOR based models are 0.9931 and 0.9944, which indicates a minimum deviation between the experimental and WAT. In comparison with existing models using five and three sample datasets, it was observed that the four parameter composition model outperformed all other models, followed by the new GOR based model and the Hybrid-1 model developed from this study with AAD% of 0.2114, 0.3244 and 0.3567 respectively. The existing models ranked 4th, 7th and 8th in performance with AAD % of 0.4242, 1.1481 and 5.9764 compared to the models developed in this work. Moreover, it was also observed that the new GOR model which accommodates a wider range of GOR (1 to 1228 scf/stb) outperformed the existing model, and was in good agreement with measured data obtained from Niger Delta. The composite (hybrid) models also did well in accurately predicting the WAT, however, more needs to be done to improve their predictive capability compared to the four parameter composition model. These new models show good prospects in predicting the WAT and as such should be deployed for accurately predicting WAT and managing wax deposition problems in oilfields.

Keywords

Waxy crude; Wax Appearance Temperature (WAT); Composition; Gas-Oil ratio; Flow assurance, Pour-point, multiple regression

Download

Share

COinS