Compositional Simulation of Well Performance for Fractured and Multiple Fractured Horizontal Wells in Stratified Gas Condensate Reservoirs.
The work for this paper has been carried out to improve development concepts in gas condensate reservoirs in the North Sea. While early North Sea oil fields were characterized by very large dimensions, thick homogeneous layers, and high permeabilities, newer discoveries have not been as blessed. Fields in the Haltenbanken area, such as Asgard have large extensions and reserves, but are found in deeper water and deep formations, characterized by hot reservoir temperatures of more than 300 degrees Fahrenheit, pressures from 7,500 to 13,000 psi, and most importantly, in heterogeneous, low permeable rocks. The need to make these high cost discoveries economic and accurate prediction of reservoir performance is more important than ever.
In this context, the application of advanced well technologies such as multi-laterals and hydraulic fracturing takes on more significance. The primary advantages of fracturing are increasing productivity from low permeable rocks and improving the connectivity of the well in stratified formations (Abou-Sayed et. al. (1)). In gas condensate reservoirs, however, fracturing has additional advantages: stimulation reduces pressure drawdown and thus leads to less liquid dropout. Non-Darcy effects are minimized and the well will suffer less productivity reduction once liquid blocking occurs (Setari et. al. (2)).
Simulation models developed for these applications differ from conventional approaches. The correct representation of geological features is normally the main gridding criterion. In fractured well simulation, however, the size and aspect ratio of the grid cells have a strong influence in predicting the flow behavior for different well types, fluid compositions, and recovery processes correctly. The goal of this paper is to determine practical simulation guidelines in modeling fractured well productivity in gas condensate fields.
Several analytical methods are available to predict the productivity and performance for vertical fractured wells. New models for horizontal wells have been developed by Babu/Odeh and Joshi, and for multiple fracture wells by Babu, Horne and Temeng, and others. Some of them have now the capability to model stratified reservoirs. The complex equations in these methods can be handled easily with the new faster PCs. Programs such as MEPFRAC/HYFRAC, a Mobil internal analytical software, can predict the productivity in time of multiple fractured wells with different fracture sizes in a stratified reservoir in only a few minutes.
Unfortunately, different analytical models were found to have individual restrictions. One limitation of all analytical models is in the definition of fluids because they permit only one set of PVT data and no compositional modeling at all. Therefore, most of the methods are only highly accurate for dry gas or almost incompressible fluids but cannot model gas condensate characteristics. In most cases, simulation and analytical model results agree only during certain flow periods depending on the assumptions used in the analytical model. Another limitation is that either uniform flux or infinite conductivity descriptions are applied in the individual methods.
After comparing the benefits of a number of analytical models with simulation results, our conclusion is that no analytical method is available to predict fractured well productivity reductions due to liquid effects in gas-condensate systems with sufficient detail. The only complete tool to predict this is fully-compositional simulation. Even a black oil representation is inadequate when finely-gridded models are used, since incorrect predictions result if improper fluid models are selected. See Coats and Hwang.
Fracture representation in coarse simulation models depends on the selection of well productivity corrections derived from analytical techniques or locally fine gridded simulations. The necessary correction depends upon the timeframe, or flow regime, which needs to be treated most accurately. Figure 1 illustrates the timeframes and flow regimes for one example and the drainage patterns along fractures for infinite flow and uniform-pressure fracture systems. The flow periods begin with linear flow in the fracture and progress to pseudo-radial flow as the drainage area increases. P. 175