Multi Information Fusion Tracking Algorithm With Application To Sand Body Connectivity

In the process of reservoir development effect,recovery and the remainder of the direct factors are many,in which the interwell sand body distribution,lateral continuity and connectivity is one of the important factors.Due to the unclear understanding of reservoir deposition,especially the complex fluvial reservoir sedimentary character:istics and sand body connectivity understanding is not clear,or because the injection-production relationship imperfect region still have a lot of surplus oil cannot be produced,therefore in most oilfields have entered to dig give priority to the development phase of high water cut,the production difficulty and development costs continue to rise,the comprehensive utilization of seismic,logging,development and production of fine understanding of dynamic data,reservoir,geological feature recognition,refined recognize sand body connectivity,guide the oilfield scientific and targeted to develop development adjustment scheme,to maintain the steady production of oil field has important significance.The traditional mathematical statistics in the classical theory,can only reflect the large scale changes,cannot express the actual geological phenomenon the special key details of requirements.These parameters are variable and random,is difficult to determine,continuous known theory to express.Information fusion technology and comprehensive utilization of oilfield multidisciplinary information,decreased due to the lack of effective information in multiple solutions,enhance system capacity and adaptive fault-tolerant ability,improve the results of scientific validity and reliability.In this paper,firstly,establish the layer structure,utilization of reservoir lithology indicator curve of natural potential and natural gamma logging curve of the composite of multi information fusion using a variety of genetic algorithm for stratigraphic correlation.Secondly,establish the breakdown structure,the petrographic period compared to subdivision of sedimentary unit or single sand layer,at the completion of each joint well section formation contrast,will be extended to the entire field block comparison results of oil-bearing strata space,identified each layer of single sand body in the reservoir space distribution.Finally,using the information fusion technology,combined with the well log and seismic data of sand body trace,and building a three-dimensional model.Using probabilistic neural network and information fusion of geological statistics of reservoir lateral prediction and reservoir fine description.