Geology in Gulf of Thailand (GOT) is complicated because of fluvial deposition and multi-stacked reservoirs. The number of sands per well is ranging from three to more than twenty. Under logistic constraints and complex geology, commingled completion strategy is applied to produce these many small gas-condensate reservoirs. Most wells are on primary production without artificial lift. Therefore, water production can cause wells to load up and stop flowing. The gas well loading phenomenon is one of the serious problems that destroy value through production and reserves loss.

To reduce water production, we need to locate the water-producing sand(s) and perform water shut-off (WSO) job(s). Successful water source identification and WSO jobs require good production logging (PL) data. Well condition significantly affects the quality of PL data and WSO decision. On one hand, running PL too late, with high water-gas ratio (WGR), obviously yields insufficient time for WSO action before liquid loading. On the other hand, running PL too early, with low WGR, yields no WSO action. Because of low WGR, it is difficult to clearly identify which sand is the main water source. However, the optimum time for running PLs could not be determined by the available analytical tools.

This study uses a statistical approach, reviewing more than 350 PLs for both oil and gas wells in the GOT. Binary Logistic models are created for prediction of the probability of success (POS) to detect water entry. Based on the well conditions, the optimum time for running memory production logging tools (MPLT) can be estimated with the models. This study found that gas rate, oil rate, water rate, and number of perforated sands significantly affect the POS. Through wellwork prioritization, these models can optimize resource allocation by running MPLT at the right well conditions.

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