SPE Western Regional Meeting · 2010 · 12 citations · 23 references
Natural ResourcesPetroleum EngineeringWater Resources ManagementForecastingHydrologic ModelingTime Series AnalysisProducer TheoryReservoir ForecastingAutoregressive ModelsAutoregressive ProcessesInterwell CommunicationLinear Difference EquationsPerformance PredictionPerformance Prediction ModelsProducer RelationshipsProduction Relationships
Abstract In this paper we introduce injector-producer models to characterize communications among wells in waterfloods, using only injection and production data. In other proposed modeling approaches, the producers are often assumed to be independent of each other and to be only influenced by nearby injectors. We present an improved technique based on a multivariate autoregressive model with exogenous inputs (M-ARX), which models the interaction among all injectors and producers (including producer to producer interactions) in a region, using a set of linear differential/difference equations. By coupling all injection/production rates in a region under analysis, the M-ARX model outperforms other models in terms of prediction performance, while requiring a relatively small number of parameters to be estimated. Also, when the rates of some producers are artificially constrained, e.g., due to shutting-in of a producer, the retraining period of our M-ARX model is significantly shorter, as compared to other models. Moreover, model parameters can be interpreted as the producer-to-producer interactions. Validation under various scenarios using numerical simulated data shows that the M-ARX model outperforms the capacitance model (CM) in terms of forecasting ability and is much more robust to noise than the finite-impulse-response (FIR) model. The M-ARX model provides estimates of the interwell connectivities that are highly consistent with the results obtained from both CM and FIR models. In addition, the M-ARX model can be extended to higher orders easily, making it possible to handle more complex and heterogeneous scenarios.
23
System identification—Theory for the user
Naresh K. Sinha · Automatica · 1989 · 9.2K citations
The Estimation of the Order of an ARMA Process
E. J. Hannan · The Annals of Statistics · 1980 · 502 citations · Full text