Research And Application Of Algorithms For Automatic Layering Of Logging Curves

In recent years,with the improvement of oil recovery technology,the demand for analysis and interpretation of logging information has also increased.At present,oil recovery are mainly divided by based on experience,but artificial stratification is time-consuming and laborious,and each person has its own unique classification criteria.The error rate is high and there is no theoretical basis for this stratification method.Therefore,many people are now trying to improve the processing and analysis capabilities of logging curves through the study of some theoretical algorithms to restore the formation information as comprehensively as possible.The division of the stratum depends on the various index values of the logging curve.Therefore,the division of the stratigraphic layer is the division of the logging curve.However,due to the large number of index values of the logging curve,the regularity of the logging curve band is difficult to explore.So seeking the stratification of the logging curve band has become an important research in geology and mathematics.Logging curve stratification is an important process in logging interpretation.Based on the sufficient research on the existing hierarchical algorithm,this paper studies the algorithm of three logging curve stratification and applies it to the specific well to verify the feasibility of the algorithm.(1)The logging curve is a pattern of band morphology,so the second idea in this paper is from the logging curve.The shape characteristics(band unit length,amplitude,and off-axis)are set to stratify the logging curve band.First,find the baseline of the band(that is,the line composed of the mean).This baseline divides the curve into several bands,according to the band.The length of the unit is roughly divided into large,medium and small bands.Because the difference between the large and medium bands is still large,we conduct a secondary search for the large and medium bands,and then the second stage,and finally use the three in the geology.The principle of high and low,find out the hydrocarbon-bearing layer of the separated layer system.(2)Try to use the logging curve band as the signal band to deal with the research,that is,introduce the EMD(empirical mode decomposition)algorithm and the eigenmode function,and decompose the logging curve band into multiple eigenmode functions and one The sum of the residual quantities,and the extreme points of the high-order eigenmode function are selected as the marking layer of the geological strata;(3)Constructing a comprehensive index function(combining multiple log curves into a curve in some way),even if the adjacent two layers of geological strata are between their inter-layer variances as large as possible and intra-layer variance As small as possible to highlight the stratified position of the geological formation.Then,using the support of algebraic theory,plus the theoretical derivation of the positive definite quadratic quotient,the analytical parameters of the comprehensive index function constructed by us are solved analytically.The algorithm is applied to the related logging curves and good results are obtained.The application results show that the method proposed in this paper has application value for stratigraphic division.