Sub-layer Multi-source Data Fusion Based on Pattern Recognition

Abstract

AbstractMassive well logging, mud logging and dynamic running data generated during petroleum exploration, development and production have different scales. Because of the multiple factors such as sedimentary microfacies, stratum structure, lithology, drilling, and the effects from upper and lower layers, the complex situations appear on the underlying logging curves. How to fuse logging data and running data in sub-layer data and make fine portraits for sub-layers are important to the analysis of sub-layer data. This paper uses big data analysis and pattern recognition method to describe the logging data and characterize each sub-layer, which can make fusion of multi-source data such as logging data, mud logging data and running data on the sub-layer scale. First, for each logging parameter, we calculate the cosine similarity of any two sub-layers to identify their patterns. According to the pattern recognition results, we assign mode labels for each log parameter of each sub-layer. Then several logging mode labels, the logging data and running data are also fused into sub-layers which are used for sub-layer portraits. Using our method in this paper, a series of experiments and the underlying analysis were carried out on the relevant data of a typical block in an east oil field. The results show that the sub-layer multi-source data fusion method based on pattern recognition can obtain the multi-dimensional features of each sub-layer, which saves a lot of manual analysis, reduces the subjectivity and one-sidedness of manual analysis, and provides technical and data supports for the accurate identification of oil-bearing layers. Query ID="Q1" Text="As per Springer style, the name of an author is presented with the abbreviated initials first and then the expanded names. Accordingly, we have transposed author names in author group. Please confirm if this is fine." KeywordsMulti-source data fusionPattern recognitionCosine similarityData portrait