Research On Automatic Processing Technology Of Isotope Injection Profile Logging Data

Isotope water injection profile logging is one of the most important monitoring methodsof water injection profile logging in water drive oilfield, which plays a dominant role in thedynamic monitoring of injecting water reservoir. The core job of the logging data is tocalculate the water injection ate and master the water absorption condition of the layers, sothat we can gain the fundamental premise for the analysis about single well,injection-production well cluster and the block analysis.This paper focuses on how to using isotope water injection profile logging data toexplain the water absorption condition of water injection layers efficiently and accurately.There are two core steps for data interpretation,‘Tool Mark’ and ‘Overlap Interpretation’. Wewill adopt pattern recognition classification technology to achieve the automatic processing oflogging data.The thesis mainly includes the following three aspects:1. In order to use the curve characters of RCCL to identify the type of down hole tools,we put forward a logging curve encoding and representation method that we combined curveshape type with included angle chain of changeable length. We use the sequence of curveshape to make dislocation match and skip the local perturbation. So we solve the problem thatcurve of the overall shape is similar but local curvature, amplitude, wave number are different.A feature weighted KNN algorithm is trained for classification. As a result, the automatic stepof ‘Tool Mark’is accomplished.2. In order to use data characteristics to make a layer’s water absorption interpretationdecision, we put forward an improved ID3decision tree algorithm. We simplify the formulaof information entropy, and combined category discrimination of attribute with the thought ofMID3decision tree to improve its execution efficiency, and solve its multiple-valued biasproblem. At last, we use the improved ID3decision tree algorithm to identify layer’s waterabsorption interpretation, the automatic step of ‘Overlap Interpretation’is accomplished.3. We apply the above two studies to the isotope injection profile data interpretationsystem.So automatic processing about isotope water injection profile logging is accomplished.