System and Method for Predicting Fluid Behavior in an Unconventional Shale Play

A system for generating a PVT model capable of predicting well behavior across a play comprising

• a memory comprising
• an application;
• and
• a data store;
• and
• a processor that according to said application in said memory
• obtains, relating to each well of a subset of wells,
• a measured API gravity;
• and
• a measured gas-to-oil ratio (GOR);
• and
• one or more lab experiments, said lab experiments measuring one or more PVT characteristics;
• trains a PVT model to match said measured API gravities and said measured GOR with said PVT characteristics;
• inputs a PVT input into said PVT model, said PVT input related to an additional hydrocarbon sample, said PVT input comprising an API gravity and a GOR;
• and
• receives a PVT output from said PVT model, said PVT output based on said API gravity and said GOR.

A method for generating a PVT model capable of predicting well behavior across a play comprising

• obtaining sets of lab measurements from hydrocarbon samples of a subset of a plurality of wells, each set of said sets associated with a subset well of said subset of said plurality of wells, said lab measurements comprising
• a measured API gravity;
• a measured gas-to-oil (GOR) ratio;
• a measured composition;
• and
• one or more lab experiments;
• training a model using said lab measurements by
• tuning an equation of state by
• dividing each of said measured compositions into component groupings, one or more groups of said component groupings comprising variable attributes;
• and
• adjusting said variable attributes to match said one or more lab experiments;
• adjusting Peneloux correction factors such that said measured API gravities and said measured GORs of said lab measurements match calculated API gravities and calculated GORs;
• and
• creating a composition model, said composition model a function of a variable API gravity and a variable GOR, further said composition model a composition model constituent of a PVT model such that when said PVT model receive a PVT input comprising an API gravity and a GOR, said PVT model generates PVT output, said PVT output comprising a composition table generated using said composition model.

A system for generating a PVT model capable of predicting well behavior across a play comprising

• a memory comprising
• an application;
• and
• a data store;
• and
• a processor that according to said application in said memory
• obtains sets of lab measurements from hydrocarbon samples of a subset of a plurality of wells, each set of said sets associated with a subset well of said subset of said plurality of well, said lab measurements comprising
• a measured API gravity;
• a measured gas-to-oil (GOR) ratio;
• a measured composition;
• and
• one or more lab experiments;
• and
• trains a model using said lab measurements by
• tuning an equation of state by
• dividing each of said measured compositions into component groupings, one or more groups of said component groupings comprising variable attributes;
• and
• adjusting said variable attributes to match said one or more lab experiments;
• adjusting Peneloux correction factors such that said measured API gravities and said measured GORs of said lab measurements match calculated API gravities and calculated GORs;
• and
• creating a composition model, said composition model a function of a variable API gravity and a variable GOR, further said composition model a composition model constituent of a PVT model such that when said PVT model receives a PVT input comprising an API gravity and a GOR, said PVT model generates a PVT output, said PVT output comprising a composition table generated using said composition model.