Comparison and verification of production prediction models

Empirical and analytical decline curve analysis models are used frequently to estimate reserves and to predict production in petroleum reservoirs. There have been many such models. In this study, three representative models (exponential, harmonic, and the mechanistic Li–Horne models) were chosen to predict and match oil and gas production from both core samples and reservoirs with different permeability. According to the experience of reservoir engineers, the exponential model tends to underestimate reserves and production rates while the harmonic model has a tendency to overpredict the reservoir performance. This has been verified using the experimental data in the core samples in which the true values of reserves were known. The values of recovery were estimated using the three models and the results were compared. The comparison demonstrated that the recovery at an assumed economic limit predicted using the Li–Horne model was greater than the exponential model but was less than the harmonic model. It was also found that the reserves predicted using the harmonic model were greater than one pore volume in some cases, which is physically impossible. The model predictions using the experimental data of recovery in the core samples were also compared to the true values. The results demonstrated that the Li–Horne model had the best estimation of the recoverable recovery at an assumed economic limit, compared to the exponential and the harmonic models in the cases studied.