A Pragmatic Approach to Rock Typing in Carbonate Formations with Limited MICP Data and Several Vintages of Wireline Logs

Abstract

Abstract Objectives The development of a Reservoir Rock Typing (RT) scheme for a heterogeneous and challenging reservoir is described. RT is an essential method to link capillary pressure data, conventional core analysis and logs. A valid RT scheme permits proper allocation of permeability and water saturations to reservoir grid cells resulting in accurate performance predictions allowing optimization of development options. Methods The approach taken was to carefully select key wells with maximum data. These were usually the most recently drilled.Analysis proceeded in two directions from the logs via electrofacies analysis using Self Organizing Maps and from the capillary pressure curves.The limited number of the latter meant the results are based on the logs plus conventional core analysis (CCA) porosity and permeability data. Saturation height curves per rock type are provided by association to MICP clusters. Challenges arise when older wells have very different (typically poorer) datasets than newer wells.Logging tools have evolved over time especially those used in rock typing such as bulk density, neutron porosity and gamma ray. In this study logs span 6 decades starting from the 1960's. Techniques of Pc measurement and the quality of data have also varied over time.MICP is a technique used to derive Pc curves. A drawback is that mercury does not wet the rock and may not accurately mimic reservoir fluids at high entry pressures. Some older and less well measured curves dropped to zero at high entry pressure. Integrating air/water porous plate Pc measurements was also a challenge.The Pc curve data analysis used all measurements that passed a set of quality control checks. Results Well defined clusters were generated by means of Self Organizing Maps (SOM) and were used to define ten electrofacies. Two of the electrofacies were very similar and were merged during the assignment of electrofacies to the MICP clusters to provide nine rock types. The rock types were compared to lithofacies determined in an associated Sedimentological study. Good correlations were found – particularly on an individual layer basis -- giving good confidence on the log-core derived answers. Permeability and saturation from height were generated along the logged wells and a good match to core permeability and log saturation was obtained over most intervals. Novel Information The most common approach to developing an RT scheme is to use a supervised training method to associate electrofacies to MICP clusters. This was attempted as a first approach but failed as cluster definition was poor due to the complexity of the rocks and limited availability of MICP data. This paper demonstrates an alternative approach that can succeed even in very heterogeneous rocks. In this paper the following terms are used to describe groupings of similar rocks, each made using different methods and assumptions: Facies Association (FA) – based on geological examinations and methodsElectrofacies (EF) – based on electric log signatures but not yet linked to the core properties of porosity and permeabilityReservoir Rock Type (RT) – a grouping of rocks with consistent log properties but differing from EF in that they are linked to permeability, saturation height etc.