Channeling analysis of wavelet threshold processing based on K-means clustering algorithm
Lixiong Gan,Ming Li,Wenyuan Cai,Jian Li,Zhanglong Chen,Jian Sun,Rui Deng
DOI: https://doi.org/10.1007/s11600-023-01019-2
IF: 2.293
2023-01-31
Acta Geophysica
Abstract:Through the spectrum noise logging technology, the oil field is dynamically monitored, and according to its simple logging instrument and convenient operation, the position of the outer channeling of the casing can be qualitatively judged by the abnormal noise of the measurement record, and the downhole production status of the water injection well can be accurately diagnosed. Fully grasp the problems of oil casing leakage, outer channeling and packer leakage in water injection wells, and enrich downhole operations. In this paper, the downhole noise signal data are standardized, and the K-means clustering algorithm is used to classify the downhole noise signal according to the correlation coefficient of different frequencies to obtain the low-frequency noise signal, and the low-frequency noise signal is clustered twice to obtain the channeling frequency band and the reservoir fluid frequency band. The accurate channeling frequency range is determined and conforms to the domestic and foreign research data. The channeling frequency band is processed by wavelet threshold, and the useless noise in the channeling frequency band is eliminated. The channeling noise signal curve after processing is analyzed, and the main output layers have an obvious amplitude back channeling. The K-means clustering algorithm is used to analyze the channeling frequency band, and the channeling noise is processed by wavelet threshold. It is a new noise signal curve processing method, which provides a new idea for the spectrum noise logging technology to master the problem of channeling outside the pipe in the water injection well.
geochemistry & geophysics