Developing Rock Physics Relationships To Distinguish Fractures From Matrix Pores To Characterize A Coalbed Methane Reservoir

Porosity is one of the literally simple but practically complicated features of the porous coal reservoirs.In geological formations,there are generally two types of porosity known as matrix porosity?herein referred to as pore-only porosity?and fracture porosity?herein referred to as fracture-only porosity?.Characterizing matrix/fracture porosity plays a leading role in better understanding of such reservoir properties as porosity,permeability,and saturations,and also provides improved perception of fluid flow conditions inside the hydrocarbon-bearing formations.Moreover,it is definitely necessary for economical production of petroleum from many low-permeability reservoirs,and could offer availability of an optimized well completion program.Therefore,at the first step of this research,a comprehensive review on the fracture characterization studies conducted in recent years was carried out.Our agenda was helping researchers in extending the existing reservoir characterization methods for appropriate usage in unconventional and particularly coal resources.This is because some techniques for quantifying and qualifying fractures have been investigated in conventional sandstone and carbonate reservoirs,but the reality for unconventional resources,including CBM reservoirs,is that such techniques are still poorly developed.?Chapter 2?Our main scheme in this study was distinguishing between matrix porosity and fracture porosity in coalbed methane reservoirs based on a novel approach which is the joint use of NMR transverse relaxation?T₂?measurements and Levenberg-Marquardt?LM?algorithm.However,even though we had access to NMR measurements in our research study,the absence of NMR measurements in different hydrocarbon bearing formations is a common problem which is caused by the high cost of NMR measurements and its particular measurement challenges such as it is not possible to run this log in producing wells.This problem is despite the fact that NMR is one of the most reliable techniques in obtaining permeability and porosity information of hydrocarbon reservoirs in today's oil and gas industry.Such a problem would appear as a significant problem on the way of our introduced method.Therefore,we established a novel approach for simulating fluids'relaxation phenomenon mathematically and synthesizing artificial NMR T₂ relaxation curves in order to guarantee that the absence of NMR data in a desired field would not question the applicability of our matrix/fracture porosity characterization scheme.This method was based on implementation of artificial intelligence and included application of fuzzy inference systems?FIS?,adaptive neuro-fuzzy inference systems?ANFIS?,and artificial neural networks?ANN?to estimate NMR T₂ relaxation curves from widely accessible seismic attributes.These attributes were from the categories of amplitude,frequency,and phase of the seismic traces and our study made use of them in a depth range of about 100 meters using more than 600 depth points.Our novel methods method provides the feasibility of simulating the relaxation phenomenon within the pores and fractures and proved to be a reliable approach in the case of our study.?Chapter 3?On the other hand,we also investigated capability of LM algorithm in determining porosity in a selected well of a coal reservoir using the data obtained from well logging sets.These data included elastic properties?compressional and shear wave velocities?and saturation information acquired from the measured logs which had been corrected for environmental effects as well as borehole enlargement effect.This way,we made sure the selected inversion algorithm would fulfill our requirements in its further application in our porosity characterization task inside coal samples.In other word,we tried to evaluate every single algorithm as well as obtained results to assure their accuracy in all step of the current research.?Chapter 4?After all,we got to the next phase of this study which was distinguishing between two types of porosity based on the novel approach of employing NMR T₂ measurements and rock physics relationships.In this approach,NMR T₂ curves of 34 water-saturated coal samples,prepared and processed in laboratory,were measured and the pore size distribution inside them were investigated.Subsequently,matrix and fracture porosity were separated based on a threshold T₂ relaxation time which was achieved through our particularly designed experiments?this in different from the common T₂ cutoff?.Once finished with this step,the 3D structure of the entire reservoir was extracted and using the recorded data inside drilled wells,our established models were upscaled and the independent contour maps of fracture porosity and matric porosity were drawn over the entire reservoir area.Throughout this research,the T₂measurements were performed in laboratory using our NMR machine.After that,a rock physics scheme based on Levenberg-Marquardt algorithm was applied to independently estimate NMR matrix porosity and NMR fracture porosity from the samples mechanical properties including compressional velocity?Vp?,shear velocity?Vs?,bulk modulus?K?,shear Modulus?G?,Young's Modulus?E?,and Poisson's ratio???which were all carefully measured in laboratory.Afterwards,characterization of either of abovementioned two types of porosity were comprehensively explored and the obtained achievements were listed.This novel approach provides that information about porosity condition of the media which is not obtainable with either of NMR or rock physics methods when they are used individually.In oil and gas industry,rock physics is the tool to map the elastic-domain information of the reservoir media to the rock-properties domain which would include permeability,porosity,as well as fluid distribution and fluid characterization.However,application of this valuable approach in coalbed methane reservoirs is challenged by the complicated nature of methane and water's existence within the coal seams.The gas inside the coal exists as adsorbed methane in its internal surfaces,free methane in fractures,or as a solid solution;likewise,the water exists as free water in the cleats or moisture forming an integral part of the coal structure.Therefore,this study established a novel discussion investigating application of rock physics relationships in order to determine the porosity condition of coal reservoirs.The approach was used to successfully investigate pore-only porosity and fracture-only porosity of aforementioned coal samples for which the NMR porosity and the mechanical parameters were measured and calculated in the laboratory.According to our obtained results,rock physics modeling would be considered as a reliable technique in quick or deeper exploration of unconventional coal reservoirs.?Chapter 5?After mathematical modeling proved effective in different stage of the present research including simulating nuclear magnetic relaxation phenomenon,porosity inversion using conventional well logging data,and also distinguishing between pore/fracture porosity using core samples,we took one further step and established a novel alternative method for the complicated seismic porosity inversion which is based on using seismic surveying data.In this novel method,in order to avoid the inevitable problems of seismic inversion process,such as tuning effect,misinterpretation and mismatch of the data,we used two statistical algorithms,i.e.Bag and LSBoost ensembles,to estimate porosity from seismic attributes in both two dimensional and three dimensional scales.For these ensembles,the most significant parameter was number of weak learners,or trees,that was determined by a trial and error method examining the performance of 200 various ensembles.The seismic attributes were also form the categories of frequency,amplitude and phase as these three characteristics of a seismic trace would represent all the geology-related information recorded inside that trace.This stage also led to satisfyingly reliable results and proved capability of the introduced method.?Chapter 6?...