The Research And Implementation On Segmented Bond Tool Logging Imaging Processing

Cementing is an important part in the process of oil wells and water wellsconstruction.The cementing quality will directly affect the life of the well, will affect thesmoothly production during the entire injection and production. Thus cementing qualitydetection is very important at the time of well completion and in the process of production. Inthis paper, use the Segmented Bond Tool Logging to test the quality of cementing.The segmented bond logging data imaging processing will involve related imageinterpolation algorithm. The traditional image interpolation algorithms for image processingwill appear some shortcomings, such as mosaic, fuzzy edge and sawtooth phenomenon. Inorder to overcome the shortcomings of the traditional image interpolation algorithm, thispaper proposes an image interpolation algorithm based on regional consistency recursivepartitioning. The method is applied to image processing of logging data, the original imagedue to noise in the measurement signal glitches and pitting has been avoided, so as to be ableto more accurately reflect the cementing quality. Meanwhile, studied cement bond recognitionmethod based on feedforward neural network, and carried on the related experiment. Theexperimental results show that this method is superior to the relative amplitude method, andthe effect is significant.The major research work includes the following three aspects: Firstly, Segmented BondTool Logging principle is analyzed and the advantages and disadvantages of the traditionalimage interpolation algorithm, and then proposed a regional consistency recursivepartitioning-based image interpolation algorithm for image interpolation processing.Application effect is better.Secondly, according to the data of the oilfield Segmented BondTool Logging, the recognition method based on feedforward neural network is applied to thesegmented cement bond recognition process. By experiment indicated that the method canbetter analyze the cement bond quality, recognition accuracy has improved significantly, theeffect is significant, can help detect the cementing quality more accurately.Finally, byanalyzing and researching the data of cement bond logging as well as the actual demand, dealwith relevant data, and use the proposed image interpolation algorithm to process the imagedata of the segmented bond logging. The treatment variable density logging image can moreaccurately reflect the cementing quality.