ADAPTIVE INTERPRETATION OF THE RESULTS OF HORIZONTAL WELL PRODUCTION TESTING USING FORECASTING MODELS

Abstract

The relevance of the research is caused by the need to interpret the results of testing horizontal well in the absence of the late-time radial flow on pressure build-up curves. The main aim of the research is to develop models and algorithms for adaptive identification and interpretation of testing horizontal wells with forecasting the late-time radial flow regime of pressure build-up curves. The methods. The authors have used the theoretical and practical developments in the field of adaptive interpretation of well testing, optimization methods, linear algebra, current research of system analysis, identification and adaptation of system. To solve the problems of identification and forecasting, the authors used evolutionary phenomenological models of pressure build-up curves with variable time-dependent parameters taking into account additional prior information on reservoir pressure. The pressure build-up curves were interpreted on the basis of the adaptive identification method. The problems were solved using the field data of testing horizontal wells of oil fields on the pressure build-up curves. The results. The authors investigated potentialities of evolutionary phenomenological models with variable parameters for identifying and forecasting the pressure build-up curve in the absence of the late-time radial flow regime. The results of interpretation of testing horizontal well show that the developed models and algorithms of adaptive identification and interpretation allow: reliable enough prediction of bottomhole pressure that has not recovered on the pressure build-up curve; determination of lateral permeability, skin factor, effective well length and the well shut-in time during the well testing, significant reduction of the well shut-in time.

Keywords

• interpretation
• identification
• adaptation
• well testing
• pressure build-up curve
• horizontal wells
• low-permeability oil reservoir
• phenomenological models
• a prior information