Pressure-Rate artificial neural network = PRANN

Specific type of the pressure-rate history analysis based on the Artificial Neural Network Regression.

This is the categorical concept and has many different implementations.

One normally split the solutions into the following sub-categories:

• Single-well
• Multi-well

and

• Pressure → Rate
• Rate → Pressure
• Rate → Rate
• Pressure → Pressure

The Rate property in the above table could be a:

• production or injection rate
• surface of subsurface rate
• total sandface rate  $//<![CDATA[ \begin{array}{l}q_t^{\uparrow}\end{array} //]]>$ or liuid rate $//<![CDATA[ \begin{array}{l}q_L^{\uparrow}\end{array} //]]>$ or component rate (oil, water or gas)

There is a large number of various permutations of input-output parameters which creates a big variety of model realizations and corresponding analysis.

One of the most useful Production Analysis can be conducted with the following PRANN design:

Producers BHP + Injector THP  →  Producers Oil/Water/Gas surface rate + Formation Pressure P_e

See Also

Petroleum Industry / Upstream /  Production / Subsurface Production / Field Study & Modelling / Production Analysis