Artificial neural network based production forecasting for a hydrocarbon reservoir under water injection
NEGASH Berihun Mamo,YAW Atta Dennis
Table 2 Features extracted based on available input and output terms.
Feature Form Physical meaning
1 WIMP(k-1)-
WIP(k-1)
One time step delayed pressure difference along the wellbore
2 THT-GLR Random combination
3 WIMP-WIP The pressure difference along the injection wellbore
4 WIP/WIMP The pressure ratio of bottomhole and surface
5 THP/GLR Related to gas lifting required to bring fluid to surface
6 CP-PMP Pressure loss due to friction and hydrostatic column
7 (CP-PMP)THT Random combination
8 OPR(k-1) One time step delayed oil flow rate
9 OPR(k-2) Two time step delayed oil flow rate
10 GPR(k-1) One time step delayed gas flow rate
11 GPR(k-2) Two time step delayed gas flow rate
12 WPR(k-1) One time step delayed water flow rate
13 WPR(k-2) Two time step delayed water flow rate