Well Logging Geological Modeling And Data Analysis For Carbonate Reservoir: A Case Study Of The Mishrif Formation,R Oil Field In The Middle East

The carbonate reservoir is rich in oil and natural gas resources.However,carbonate reservoirs have the characteristics of complex genesis,diversified diagenetic processes,various pore types,and multiple solutions of petrophysical properties.How to evaluate carbonate reservoirs more accurately and effectively is the hot spot among academic field,and the objective of industry research as well.One of the most important method for reservoir evaluation is to conduct comprehensive research with the knowledge of both reservoir geology and geophysics.In general,well logging geological model is an essential method connecting geological model with geophysical or seismic model and data sets,which is the effective procedure to describe reservoir characteristics and layers'distribution patterns.Take the Mishrif Formation in R oilfield of the Middle East as an example,based on thin sections,petrophysical data derived from core analysis,conventional well logs as well as micro-resistivity image logs and nuclear magnetic resonanc(NMR)logging data,this paper discuss the method of well logging geological modeling and data interpretation technique,which considered the relationship between rock composition,cements,matrix,pore types and their petrophysical responses.The results show that when the content of bioclasts decreases along with micrite increasing can lead to the increase of natural gamma value.Dissoluted intergranular pores,moldic pores and intraparticle pores result in the increase of acoustic transit time(AC),increasing neutron value(CNL),and declining of bulk density(DEN),and the separation between shallow and deep resistivity curves.As calcite cements distributed in intergranular or intraparticle pores,can cause lower CNL and AC,as well as the smaller difference between shallow and deep resistivity value.The oil-bearing or cemented layers show as bright background in static electrical images,while block or patch light image symbolize bioclast and cements in dynamic images.The vugs and dissolved pores show as dark spots in their images,and the dark areas indicate the micrite or matrix components.The porous layers characterize higher magnitude NMR T₂ spectrum with the dominant peak located in long relaxation time,while cementation enhance the morphology of lower T₂ spectrum with short or medium relaxation time peaks.The classification of reservoir types in the Mishrif Formation is based on depositional settings,intensity of both cementation and dissolution,and comprehensive logging responses.Eight reservoir types have been identified,including dissolved reef,dissolved shoal,weak cemented lagoon,etc.Six parameters characterizing rock fabrics and pore types have been selected to establish reservoir prediction models based on the double hidden layer BP neural network algorithm.The prediction results were tested by both core features and micro-electrical images and NMR data.The findings show that the lower and middle parts of progradation sequence in the Mishrif Formation develop thick dissolved shoals presenting continuously lateral distribution pattern,while dissolved reefs distribute separately in the upper part of the sequence with lagoon dissolution reservoirs located above.The upper of the Mishrif Formation develop dissolved shoal reservoirs with cemented or dense reef reservoirs located above.The reservoir geological evaluation model and data interpretation method mentioned in this paper have made the foundation for reservoir geological modeling and fine evaluation in futher work.