IPARS Introduction
The Implicit Parallel Accurate Reservoir
Simulator (IPARS) provides a framework and a growing number
of physical models suitable for research and practical calculations.
Both oil reservoirs and aquifers can be simulated with the
program. IPARS runs on parallel and single processor computers
and can solve problems involving a million or more grid elements.
It can handle multiple fault blocks with unaligned grids and
problems that involve different physical models in various
regions of the reservoir. This document is a second draft
of a user’s manual for IPARS; it covers the capabilities,
configuration, and application of the simulator.
The elements of IPARS can be combined in
many different ways (more than 50) so the first task a user
faces is to decide which elements he needs and which he does
not need. He may want to track a contaminant in a faulted
aquifer using a pc running Microsoft Windows and a multigrid
linear solver, or he may want to conduct a history matching
study on a 30 well unfaulted oil reservoir using a network
of workstations running Linux with line successive over relaxation
as the linear solver. The IPARS configurations for these two
problems are quite different. Fortunately, IPARS is portable
with respect to machines and operating systems, and includes
a compiler preprocessor that automates configuration.
Preparation of input data is the next user’s
task. The simulator employs a free-form keyword input with
default values supplied for most of the input variables. There
are two categories of input, initial data and transient data.
In general, the input variables are different for the two
categories. However, some variables can be input in both categories;
for example, a well may be initially specified to be a production
well but half way through the simulation may be converted
to a water injection well. Various physical models and elements
of the IPARS framework each require different sets of input
variables; these sets are documented separately herein.
Execution of the simulation comes next. Inevitability,
the first attempt at execution will produce error messages
due perhaps to misspelled keywords or omission of required
data. If at all possible, IPARS will process all of the initial
data before terminating due to an error condition. Execution
may be broken up into time segments so that intermediate results
can be examined and transient data can be modified if necessary.
IPARS can be restarted with the original input data; a specially
prepared restart file is not required.
Both during and after execution, the user
will want to visualize results. IPARS currently supports a
commercial plotting package, Tecplot, by writing plot files
in the appropriate format. Eventually, other plotting packages
may be supported. |