TransGen is a system for estimating and outputting either single or two-phase faulted transmissibility data suitable for inclusion in flow simulation models.
One of the design goals was to minimise the amount of work required when preparing input and using output from TransGen. To achieve this goal, the program mimics the behaviour of the
Eclipse keyword system and adopts other
Eclipse styles and conventions. TransGen (via the
WizGen module) constructs and uses its own run instruction file (<Project>.TGDATA file) which conforms to and uses
Eclipse keywords. For details of the TransGen instruction file click
here.
TransGen will only take an
Eclipse format cellular model as input, it does not take horizon and fault files from a seismic interpretation package. Wells are positioned by the indices of the cells they connect with, not by well trajectories in UTM or geographical coordinates. For TransGen:-
Only an eight corner-point geometry is supported
Only a Cartesian Coordinate system can be used. TransGen does not accept radial coordinates.
Only one Eclipse 'reservoir' is allowed per model (i.e. Eclipse keyword NUMRES = 1 which means only one coordinate system per model)
Only a subset of (the most common) Eclipse keywords is used (TransGen generates an error message for unrecognised keywords)
Inactive cells are used in fault seal potential/permeability calculations and therefore property data must be entered for all cells
TransGen uses the offset of cell corners to estimate displacement. It can find the faults from the positions of offset cells, and a
FAULTS file is not required, but cell-centred geometries cannot be used. TransGen assumes the slip vector is oriented parallel to the
COORD lines. There are various methods to inactivate regions of the reservoir model in
Eclipse so that grid-blocks take no part in the
Eclipse simulation. TransGen similarly calculates and displays the properties of connections in the active regions of the model. However, as material may be incorporated into faults in the 'active' regions from the inactive regions outside, TransGen still relies on the cell geometry and shale content (or Net-to-Gross) of cells which may have been inactivated. Although they take no part in the simulation, they cannot be ignored. In order to calculate fault seal potential (i.e. SGR and/or CSP in "Basic project" mode), property data have to be supplied for inactive cells, and mis-shapen cells outside the active region could result in an incorrect computation. In this way, TransGen is more demanding of the geometrical integrity and property data of the model than
Eclipse.
HINT:- See the Technical Description for a detailed description of what TransGen does.
Eclipse files can be written by cellurisation packages in an order which departs from the
Eclipse standard.
Eclipse essentially reduces a reservoir model to a matrix of transmissibilities, so mirroring of the data does not result in incorrect results. It
does matter for visualisation purposes however, when it is difficult to check the geology of a mirrored model. TransGen will accept input data in any row-ordered format. The data can be read into and visualised in the
ViewGen module and if the model is incorrectly oriented, selecting an alternative input layout in the
WizGen module will allow the model to be re-visualised correctly. Buckled, concave or 'inside-out' cells have an incorrect ordering. The
ViewGen module of TransGen identifies such geometrical defects and treats all incorrectly ordered cells as defective and inactivates them.
FileGen
The
FileGen module within TransGen is an
Eclipse file pre-processor. It has been written to simplify
Eclipse files and sort out the input data needed by TransGen by simply removing keywords TransGen does not use (or recognise). In the best cases, when the structure of the
Eclipse input files is straightforward, this could be done manually, but can still be done more reliably using the FileGen module.
However,
Eclipse is very flexible in the way input files can be constructed. Input data can be read in and then overwritten or modified later, by the same or different keywords, resulting in what is essentially a 3D worksheet. This may be done in an attempt to obtain a better simulation result, but the result of many changes over time can be a very complicated and obscure
Eclipse run file. In the worst case, sections of data may be made totally redundant, different parameters may be made interdependent through COPY keywords and modified by a chain of other keywords whose order is critical. These changes can be made over several files, which themselves may be too large to edit, making manual editing very difficult. In these cases, the
FileGen module is essential to enable straightforward preparation of the TransGen input files.
However there is no substitute for understanding how the the
Eclipse run file has been constructed and why modifications were made, but FileGen will simplify the task. What FileGen does is:-
read one (or several) Eclipse input files, following all the links by INCLUDE keywords to other secondary files and compile a single temporary file containing all the data.
It then searches through this temporary file and finds all the locations of the Eclipse keywords that TransGen recognises. Multiple occurrences of keywords and modifications to keywords will be identified in their order of occurrence. Keywords which modify or copy data arrays are also identified. The list and number of keywords provides an outline of the structure of the input data.
Each of the types of data identified by keywords can be written to a separate files (or combined if required). The data includes relevant BOX statements, which define the spatial range of the keyword.
FileGen extracts all the data that TransGen requires from the
Eclipse input data. All that remains to complete the TransGen instruction file is to add or modify TransGen specific keywords to control the TransGen run. This is done interactively, step-by-step, in the
WizGen wizard.
Minimum data requirements
ESSENTIAL ECLIPSE DATA
In order to load and view a model, the data associated with the following keywords are required:-
DIMENS (or SPECGRID) - to define the number of columns, rows, and layers
COORD - to define the map position of the cell corners
ZCORN - to define the depth of the cell corners
UNITS (LAB or FIELD or METRIC) - to define the units
END - to end the file
Additionally, in order to view cell and fault properties and generate transmissibility multipliers, the data associated with the following Eclipse keywords are required:-
PERMX and PERMY - to define cell permeabilities (The PERMX data can be replicated in the PERMY array using the COPY keyword)
NTG or TGVS or both - to define the effective Vshale content of the cells
MULTX and MULTY - X and Y direction cell transmissibility multipliers optionally used in transmissibility expressions
ECLIPSE KEYWORDS WHICH MODIFY DATA ARRAYS
It these are present in the original run file, they should also be included in the TransGen run file to ensure that TransGen uses the same data as that used by
Eclipse.
ADD
COPY
EQUALS
MULTIPLY
Data arrays may be reset (wholly or in part) by reading the same keyword a second time. e.g. PERMX may reset the array initially set by a previous PERMX keyword.
ECLIPSE KEYWORDS WHICH CONTROL WHICH CELLS ARE ACTIVE
It these are present in the original run file, they should also be included in the TransGen run file to ensure that TransGen identifies the same cell/cell connections as
Eclipse.
ACTNUM
MINPV
MINPVV
PORO
OPTIONAL KEYWORDS for visualisation purposes
PERMZ
FAULTS
TGWELL (New TransGen keyword)
TGSTRLNE (New TransGen keyword)
ECLIPSE DEFAULTS
Eclipse assumes:
A cell with a zero porosity is inactive
A cell with a net pore volume of less than 1.0E-06 is inactive (unless redefined by the MINPV keyword)
A cell with zero permeabilities (in X,Y and Z) is inactive.
A cell with zero NTG is inactive
A faulted connection has no fault zone property associated and is fully open,
i.e. its transmissibility is the unfaulted transmissibility.