Estimation of the Water Gas Ratio from Wellhead Temperature
Australian School of Petroleum,
The University of Adelaide
Water-gas ratio (WGR) is a key parameter in the optimisation of production from gas wells. The WGR is a measure of the volume of water produced per unit of gas production and is strongly related to gas production rate, gas well operability, and the ultimate recovery from a well and/or field. Currently the only method employed by Santos Ltd. in the Cooper-Eromanga basins to accurately measure the WGR of producing gas wells, is to use mobile well-head separator tests, known as Liquid Evaluations Tests (LET's).
This research project investigated the feasibility of developing a method to estimate WGR from flowing wellhead temperature (WHT). It was initially envisaged that predictive charts to allow estimation of WGR from WHT would be produced. The project considered the premise that if models can accurately simulate field observations, then they can be used in a predictive role. After the previous work into WGR estimation had been reviewed, nodal analysis models were constructed for wells in Field A. Using these models, simulations were run to compare the modelled WHT to that observed in the field, for a known WGR.
Matches between simulated and actual WHT's were generally poor, and consequent sensitivity studies were conducted into model input variables in order to understand their relationship to WHT. The key variables to WHT were lithology, annular liquid level, time since production began, reservoir temperature, depth and water gas ratio. Minor variables were the downhole equipment, the cement depth, reservoir permeability, drainage area, reservoir pressure and condensate gas ratio. The geothermal gradient was also identified as a variable of importance, and a study into its nature was conducted.
Sources of error and uncertainty were identified, with recommendations made to decrease their impact. It is possible that upon addressing these errors, the quality of matches between field observations and simulations will significantly improve. Whilst the theoretical basis of using heat transfer and nodal analysis concepts to estimate WGR from WHT is sound, predictive charts are not considered viable given the poor matches seen in this study and the simplifying assumptions that would be required in the construction of charts. As such it is proposed that estimations of WGR be initially made from nodal analyses constructed specifically for the well of interest, with charts to be considered in the future after the technique has been confirmed by comparing the results from predictive computer simulations to the actual results from upcoming LET's.