Integrated Capacitance Resistive Model for Reservoir Characterization in Primary and Secondary Recovery

Abstract

AbstractThis paper presents a new procedure to obtain unique estimates of interwell connectivities, pore volume, and productivity index in primary and secondary recovery. This method approaches the problem by analytically integrating the simplified continuity material balance equation. The proposed procedure uses linear multivariate regression on rate, initial reservoir pressure and bottom hole pressure (BHP) data. The method quantifies compressible pore volume – the capacitance, productivity index (PI), and degree of communication between well pairs – the resistance to flow. Hence, it is called the integrated capacitance-resistive (ICR) model.The ICR model fits cumulative production against rates, BHP, and cumulative water injected. For primary recovery, the model has been validated on synthetic field data generated by numerical simulator, covering the effects of initial oil saturation, heterogeneity, and number of wells. The model was also tested on a real field. For secondary recovery, the procedure was tested on synthetic fields with high permeability channels.In all primary recovery synthetic fields tested, good fit and less than 13% error of estimated pore volume and PI were observed. Real field application also shows good agreement with geological knowledge. For waterflooded synthetic fields, the estimated parameters are consistent with the geology and the results obtained from the capacitance resistive model (CRM) fit. The confidence intervals of the parameters are narrow enough to conclude regression coefficients are statistically significant. Hence, the ICR model can be used to quickly estimate reservoir properties and infer interwell communication in primary and secondary recovery from available data with high confidence.The ICR model is comparable to CRM in terms of number of inputs. Hence, both models can be used in large fields with hundreds of wells. However, the nature of the CRM's nonlinearity makes it difficult to establish confidence intervals of the model parameters and obtain a unique set of model parameters for a large number of wells. In the proposed method, a unique solution can always be obtained regardless of the number of parameters via linear regression. The simpler formulation of the ICR model would reduce the computation time and easily establish the confidence intervals of model parameters.