petroleum_engineering

Main Folder: Decline Curve Analysis

sub Folder: fbProphet_Decline

fbprophet_DCA2.0.py

Summary:
Does decline curve analysis for any number of wells downloaded as production time series csv from drillinginfo (download the production time series csv from drillinginfo for any number of wells and leave it exactly as is). Only thing to change is path to file. The most important final output is stored in the 'prediction' variable, which will contain a dataframe for each well in the csv. One forecast figure and one DCA figure will be saved for each well.

• Libraries needed:
• pandas
• matplotlib
• numpy
• fbprophet
• Data Format:
• Production Time Series csv from DrillingInfo AS-IS (any number of wells)
• Example uses '2wellsGrady Production Time Series.csv'
• Uncomment:
• m.plot(fcst) to plot each forecast (not a good idea for lots of wells)
• Figures:
• DCA [API#].png are the final DCA (Actual+Forecast) figures created and saved by the program (example used two wells)
• fcst [API#].png are the forecast plots generated with fbprophet (example used two wells)

sub Folder: fbProphet_Decline

ARPS_decline.py

Summary:
Does exponential, hyperbolic, and harmonic DCA

Folder: Reservoir Fluid Characterization

Res_Fluid_Clustering.py

Summary:
Uses KMeans clustering to characterize reservoir fluid types. This particular example creates three clusters and then further divides the cluster with the lowest average initial GOR because it is dealing specifically with oil wells in Grady County producing from the Woodford. Folium is used to map each well on an interactive map in the browser.

• Libraries needed:
• pandas
• matplotlib
• numpy
• sklearn
• folium
• Data Format:
• Wells Table csv from DrillingInfo AS-IS (any number of wells)
• Example uses 'WellsTableGrady.csv'
• Figures:
• res_fluid_clustermap.html is the output of the example ran using wells in Grady County clustered into four fluid types