Abstract

Water production from gas wells is a major issue in the Gulf of Thailand. Production Logging (PL) is the main way to try to discover the water source and hence shut it off. Sometimes there are clear indications of water entry such as density increases and holdup sensors showing more water entering the flow. For many wells, though, there is no clear water entry and the water rate is a result of the calculations that are very dependent on slippage velocities derived from various flow models. Emeraude from "Kappa Engineering" (Kappa from now on) allows the use of any one of ten different models to derive the slippage from. The problem is that there is no clear way to decide which model is best and the effect on the results of using different models is significant, often moving the water split between zones.

This study uses the surface rates to compare with the rates calculated from the PL data using each of the ten different models. Some 300 PL cases are used to allow a statistical analysis to determine the most precise, accurate with minimum error, models. The best three from this statistical approach go forward to a Discriminant Analysis where the critical factors which governed when a model works best are ascertained. Then the Fischer Discriminant Functions could be calculated for these three models so that this could be used to select which model to use for future Production Log Analysis.

The study comes up with a systematic way to select which model to use. It found the models that were not typically being used in the past should be used so this was quite a big change. The logic was applied to some past PL examples and the recommended models are found to give results more consistent with known shut-off's that the original analysis had done. The method seems to have value and should improve the effectiveness of future water shut-off actions though the number of wells the study has been applied to is small and time will allow further verification of the technique.

You can access this article if you purchase or spend a download.