Integration of production data into reservoir description here means using information from production data before reaching the reservoir simulation stage. By production data here we mean production history. There are two major sources of information that may be derived from production data: how fast can we produce (permeability and thickness) and how much can we produce (saturation and pore volume).
Production data is imported as Log Data. The file formats are the same as well log file formats. The only difference is that the depth column is replace by a time column. The format of the time column may be one of the following:
MM/DD/YY
MM-DD-YY
MMYY
MMDDYY
MMDDYYYY
MM/DD/YYYY
D-MMM-YY
D-MMM
MMM-YY
M/D/YY
YYYY-MM-DD
decimal
YYMM
Where decimal format is assumed to contain a decimal point, and YYMM format may contain up to 4 digits in the year portion (YYYY). All formats are automatically detected except YYMM, which numerically is identical to MMYY or MMDDYY. If the format is YYMM, it is necessary to select the YYMM format from the Options pull down menu Import Par panel, otherwise the time format selection is Autoread.
The internal time format after convert is year. Since KB is used as depth reference and time data is treated similar to depth, use 0 KB for time data. The XY coordinates should be the same as those for the well logs or transformed to be the same, otherwise it will be difficult to compare the maps.
Time data should be imported and mapped in a separate project to avoid mixing with well log data.
Horizon markers should not be used, otherwise they will be treated as time markers.
With Z Unit as time (Year), all graphs are the same as with well logs except for Data Trace and Traces graphs, where the time axis is horizontal.
Usually production data files contain a time column and several data columns. The data columns to be used include daily or monthly production rates of oil, gas, and water. A producing time column may also be required if the rates are daily rates, so that the correct cumulative production may be calculated excluding shut-in time.
After import, in the DataQC step, the data curves can be changed between rate and cumulative. Total production can be calculated by linearly combining the oil, gas and water rates, using Add then Combine with weighting factor 1 for oil and water and a conversion factor for gas. Water oil ratio and gas oil ratio can be calculated using Add then Normalize with Divide method. Zero and near zero values should be removed before division.
Intuitively we know that wells penetrating higher permeability area tend to have higher production. Of course, there are other factors influencing production, such as pressure drop and skin factor. The idea is to estimate the productivity index, which is proportional to permeability.
Reservoir pressure is different from location to location and changes with time. Sometimes it is possible to map the reservoir pressure and estimate the pressure drop. If the variation in pressure drop is small compared to the variation in production rate, we can use an average pressure drop to calculate productivity index from production history directly. If skin factors are known, such as from work-over or stimulation, the estimated productivity may be normalized by skin factor.
For simplicity, using the same average pressure drop for all wells, the total production rate is proportional to productivity index and proportional to permeability. Build a 3D model of the total production rate, where z is time instead of depth. In the GridOutput step, use GGraph then Top View to animate the change over time. Either select a particular time or map the average to compare with the k*h map from geological model. If they are not consistent, explanation will be required.
It is possible to constrain a 3D permeability model with production map. See Gridding Constrained by Mapped Average in General Topic under Selection of Models and Methods.
Water oil ratio and gas oil ratio can similarly be modeled and animated to gain understanding in the production process.
With an initial oil saturation model from geological modeling, the remaining oil after production can be derived from detailed reservoir simulation. A simplified faster estimation may be derived directly from production data, based on material balance.