Research And Application Of Qualitative And Quantitative Evaluation Methods For Fractured-vuggy Reservoirs

Fractured-vuggy carbonate reservoirs have become one of the key exploration targets for oil and gas development.However,because this kind of reservoir has a strong heterogeneity,the development scales of fractures and vugs are different,the regularity of vertical and horizontal distribution is poor,and the combinations of pores,fractures and vugs are diverse,the qualitative and quantitative evaluation of fractured-vuggy carbonate reservoirs has always faced very complex technical problems,such as classification of fractured-vuggy reservoir development grades,identification of fractured-vuggy fillers and quantitative calculation of reservoir parameters.At present,the identification and classification of fractured-vuggy carbonate reservoirs are mainly through manual methods.This method is time-consuming and labor-intensive and the identification accuracy is low.Moreover,traditional porosity logging interpretation models and calculation methods often only interpret the total porosity of the reservoir,without calculating the actual effective porosity of the reservoir,which reduces the reliability of quantitative evaluation of the reservoir.In this paper,the Ordovician carbonate reservoir in an oil field in Xinjiang was taken as the research object,based on conventional logging data and geological data,the research on qualitative and quantitative evaluation methods for fracture-cavity reservoirs is carried out,and a set of qualitative and quantitative evaluation methods for fractured-vuggy reservoirs with reference and promotion value is established.In terms of qualitative evaluation of fractured-vuggy reservoirs,the predecessors have introduced some classification algorithms to automatically classify the development levels of fracture-cavity reservoirs.However,when dealing with such complex non-linear problems,it is sometimes difficult to achieve better application results due to the defects of the algorithm itself.Therefore,this paper introduces a novel convolutional neural network algorithm to learn the nonlinear relationship between the logging curve and the reservoir development level.Convolutional neural networks have powerful data feature recognition capabilities.When dealing with classification problems,the algorithm can effectively extract the characteristics of the logging curve by virtue of its sparse connection and weight sharing characteristics,so that the reservoir development level can be classified more accurately.In addition,this paper also applied this algorithm to identify the type of fractured-vuggy filling.Using conventional logging data and its corresponding filling type as the training data of the network model,multiple network models were designed for experiments and the optimal model and network parameters were selected.Finally,the algorithm was tested using the measured data in the study area.It shows that the algorithm can more accurately identify the type of filling,and effectively improve the accuracy of qualitative evaluation of fractured-vuggy reservoirs.In terms of quantitative interpretation of fractured-vuggy reservoirs,this paper focuses on the calculation method of reservoir porosity.The effective storage space of fracture-vuggy reservoirs is mainly secondary pores,that is,fractures and vugs.Therefore,in order to accurately calculate the effective porosity of the reservoir,research and discussion on the calculation methods of fracture porosity and vug porosity are carried out.In terms of quantitative calculation of porosity of large vugs,this paper attempts to use neutron,sonic logging curves and neutron-sonic intersection method to calculate its porosity,and finds that for large vugs filled with muddy and breccia,the porosity value calculated by the sonic logging curve is closer to the actual vug porosity.When the imaging data is lacking,the porosity calculated by the sonic logging curve can be approximated as the porosity value of the large vugs.