Study on formation weathering degree based on unsupervised learning t-SNE and logging while drilling technique
A method to predict the degree of formation weathering based on unsupervised learning and measurement while drilling(MWD)technique is introduced in this paper.MWD technique provides a means for real-time assessment of the formations and could reflect the continuous changes in the strata.These multidimensional drilling parameters contain rich information about the formation,and the t-distributed stochastic neighbor embedding(t-SNE)algorithm is capable of uncovering hidden patterns and structures within the data,making it suitable for exploratory data analysis.Through the MWD system,the movement and operating parameters of the drilling rig are monitored in real-time.After processing the MWD data,the data from the pure drilling process are selected,followed by segmentation and normalization of these data.Finally,the data are imported into the t-SNE algorithm to calculate the similarity of data points in high-dimensional space and map them to a lower-dimensional space.The study results indicate that the t-SNE algorithm can effectively identify the degree of formation weathering through drilling parameters,which coincide with actual conditions.This method provides a new intelligent approach and perspective for engineering identification of rock formation weathering degrees.
measurement while drilling(MWD)unsupervised learningt-SNE
万方数据
维普
