INTRODUCTION
The effective stress in a hydrocarbon reservoir increases as a result of the decline in pore pressure associated with the withdrawal of fluids. Under such conditions, certain porous rocks experience a sudden increase in compressibility. The phenomenon that leads to such an irreversible deformation, is often termed as "pore collapse" and is believed to be a leading cause for abrupt reduction in production. It is usually accompanied by subsidence of the ground surface and may result in sand production, wellbore instability, etc. [ 1].
For a safe and efficient management of the production facilities in the fields involving pore collapse-prone reservoir rocks, an effective simulation is warranted. In this paper, different aspects of numerical simulation are outlined. As a field example, a nonlinear finite element (NLFE) idealization of the Ekofisk field, where the problems of reservoir compaction and surface subsidence have been reported, is carded out, followed by a brief discussion of the numerical results. It has to be noted here that the input data used here for the simulation does not represent the actual practice after the year 1986.
DIFFERENT ASPECTS OF A SIMULATION TASK
Some important aspects which have to be given attention while carrying out the simulation of extremely complex phenomena of reservoir compaction and surface subsidence in an optimized yet efficient way, are outlined below.
A subsiding field encompasses a number of zones, involving the reservoir and its surroundings, which contribute to its subsidence. The main cause of the subsidence is the reservoir compaction resulting from the increase in the effective stress due to pore pressure depletion. The most desirable simulation would obviously require a three-dimensional idealization of the field. A simpler and computationally more efficient alternative is to adopt a plain strain or axisymmetric idealization of the domain. The factors that govern the selection of a suitable spatial idealization are: geometry of the field/reservoir, pore pressure depletion zones and depletion scenarios, amount of field data available, the level of sophistication desired, depending upon the simulation goals, and computer time and available storage facilities.
Rocks in general exhibit material nonlinearity particularly when the porosities are high. The rate of loading may also influence the material behavior. For the areas surrounding the reservoir, it is common to assume a linear material behavior for computational efficiency. When dealing with a nonlinear material along with finite elements, care should be taken to account for the different stages of material behavior which may create potential numerical problems.
A depleting reservoir is subjected to both the self-weight and an increase in effective stress due to withdrawal of hydrocarbons from the reservoir matfix. Since the nature of these two loadings is different, separate approaches are generally adopted in subjecting the reservoir matrix to these loadings. The first step in simulating the field conditions is to subject the whole domain to selfweight. Since the hydrocarbon in the producing regions of the reservoir is under high pressure before production, the state of stress caused by this internal pressure should also be accounted for at this stage. This step ensures that the state of stress in the field corresponds to the initial, preproduction stage.
