Automatic mutation feature identification from well logging curves based on sliding t test algorithm

• Published in: Cluster computing 2019-11, Vol.22 (Suppl 6), p.14193-14200
• Main Authors: Du, Ruishan, Shang, Fuhua, Ma, Nan
• Format: Article
• Language: English
• Subjects: Algorithms; Computer Communication Networks; Computer Science; Data processing; Earthquakes; Fractals; Identification methods; Lithology; Microelectrodes; Morphology; Mutation; Operating Systems; Permeability; Processor Architectures; Sandstone; Sliding; Test methods; Time series; Wavelet transforms
• ISSN: 1386-7857; 1573-7543
• DOI: 10.1007/s10586-018-2267-z

The mutation detection is an effective identification method of well logging curves, which can be utilized to detect mutations of the correlation time series by a sliding window technology. The computational complexity and accuracy of mutation detection are very important for detecting the series change points, however, some of the existing calculation methods take none of these into account. By sliding window technology, in the present paper we put forward a new method, the sliding t test for detecting a series of dynamic mutations. The principle is that mutation morphology of the spontaneous potential curve and microelectrode that are related to sandstone recognition contrast layers are described. In order to prove the performance of the method, the mutation analyses of time series are carried out by selecting different sliding windows. The test results are shown that the method can quickly and accurately detect the mutation change points and intervals. It has robust stability, and depends less on the sliding window length, which has some advantages in the large data processing. Finally, the method is utilized to detect the mutation of numerous experiments in well logging curves. The experimental results indicate that the mutation interval is consistent with the abrupt change, which further is verified the validity of the method.