Research Article
CNN-LSTM Model Optimized by Bayesian Optimization for Predicting Single-Well Production in Water Flooding Reservoir
Table 1
Initial data set. The italicized items are category features which need to be one-hot encoding.
| Well basic data | Reservoir data of well | Name | _coordinate | _coordinate | Depth | Unit | Layer | Start date | Flooding date | Position | Top depth (MD) | Bottom depth (MD) | Inject/production | Well type | Initial temperature | Initial pressure | Top depth (TVD) | Bottom depth (TVD) | Thickness of sand bodies | Initial pressure coefficient | Saturation pressure | Initial solution gas-oil ratio | Viscosity | Effective thickness | Porosity | Permeability |
| Monthly producing well production data | Time | Producing scheme | Production days | Nozzle | Pump diameter | Pump depth | Stoke | Jig frequency | Monthly liquid | Monthly oil | Monthly water | Monthly gas | Gas-oil ratio | Water cut | Cumulative oil | Cumulative water | Cumulative liquid | Recovery | Static pressure | FBHP (bottom hole flowing pressure) | Working fluid level |
| Monthly water-injecting well injection data | Injection methods | Injecting days | Pumping pressure | Casing pressure | Oil pressure | Injecting volume monthly | Cumulative injecting volume |
|
|