Queendarlyn Nwabueze, Bob L. Herd Department of Petroleum Engineering
The phenomenon of liquid loading is a dominant limitation in developed gas fields globally. Apparently, all gas wells will experience this depleting process in the subsequent phases of their production. The primary problem in dealing with liquid loading is the issue of forecasting its occurrence and accurately determining its onset. This paper is focused on developing an improved model for accurately predicting liquid loading in vertical gas wells as the available models often show variations.
In this paper, an improved model for predicting liquid loading was developed on the hypothesis that the liquid droplet is disk-shaped and retains its configuration throughout the wellbore. The developed model was established on the fundamental principles of Turner’s model but offers better prediction than the former. The model was validated with Turner’s well data using the commercial Microsoft statistical tool Excel®. The actual critical velocities and critical flowrates of 106 wells from Turner’s data set were compared with the evaluated critical velocities and flowrates from the new model and the existing Turner’s and Li’s models.
The error analysis carried out on the models showed that the models predicted the liquid loading status of the wells with average relative errors of 15.48%, 26.29% and 35.71%, with the improved model having the least error. The results obtained from this analysis indicate an improvement over the Turner’s and Li’s models. The improved model was applied to field data from Stubb Creek field in the Niger Delta to validate the efficiency of the model in detecting the liquid loading status of four (4) gas wells. The results obtained showed that the improved model detected the liquid loading status of the wells with the least percentage error of 10%. The analysis obtained using the data collected from Stubb Creek field revealed that the improved model gave a more accurate detection of liquid loading than the existing Turner’s and Li’s models. The improved model can be applied to gas wells with well head pressures lower than 500 psia and liquid/gas ratios within the ranges of (1-130 bbl/MMscf) to ensure the existence of a mist flow regime in the gas wells. The developed equations can also be applied in gas wells where annular flow regime and other flow geometries exist.
It has been theoretically established that liquid loading is an issue bound to occur in all natural gas wells during their productive life. Therefore, the results of this study will be beneficial to the industry as it would enable the early detection and mitigation of liquid loading. The resultant effect of the early detection of liquid loading is its possible avoidance and increase in gas recovery rate.