Well Logging Based Lithology Identification Model Establishment Under Data Drift: A Transfer Learning Method.

Recent years have witnessed the development of the applications of machine learning technologies to well logging-based lithology identification. Most of the existing work assumes that the well loggings gathered from different wells share the same probability distribution; however, the variations in sedimentary environment and well-logging technique might cause the data drift problem; i.e., data of different wells have different probability distributions. Therefore, the model trained on old wells does not perform well in predicting the lithologies in newly-coming wells, which motivates us to propose a transfer learning method named the data drift joint adaptation extreme learning machine (DDJA-ELM) to increase the accuracy of the old model applying to new wells. In such a method, three key points, i.e., the project mean maximum mean discrepancy, joint distribution domain adaptation, and manifold regularization, are incorporated into extreme learning machine. As found experimentally in multiple wells in Jiyang Depression, Bohai Bay Basin, DDJA-ELM could significantly increase the accuracy of an old model when identifying the lithologies in new wells.

Sedimentological constraints on the initial uplift of the West Bogda Mountains in Mid-Permian.

The Late Paleozoic is considered to be an important stage in the evolution of the Central Asian Orogenic Belt (CAOB). The Bogda Mountains, a northeastern branch of the Tianshan Mountains, record the complete Paleozoic history of the Tianshan orogenic belt. The tectonic and sedimentary evolution of the west Bogda area and the timing of initial uplift of the West Bogda Mountains were investigated based on detailed sedimentological study of outcrops, including lithology, sedimentary structures, rock and isotopic compositions and paleocurrent directions. At the end of the Early Permian, the West Bogda Trough was closed and an island arc was formed. The sedimentary and subsidence center of the Middle Permian inherited that of the Early Permian. The west Bogda area became an inherited catchment area, and developed a widespread shallow, deep and then shallow lacustrine succession during the Mid-Permian. At the end of the Mid-Permian, strong intracontinental collision caused the initial uplift of the West Bogda Mountains. Sedimentological evidence further confirmed that the West Bogda Mountains was a rift basin in the Carboniferous-Early Permian, and subsequently entered the Late Paleozoic large-scale intracontinental orogeny in the region.

Trace metals in the surface sediments of the intertidal Jiaozhou Bay, China: Sources and contamination assessment.

The major (Al) and trace metal (Cu, Pb, Zn, Cr, Cd, and As) concentrations in 29 surface sediment samples from the intertidal Jiaozhou Bay (JZB) are evaluated to assess the contamination level. The results show that the overall sediment quality in the area has been obviously impacted by trace metal contamination. The geoaccumulation index and the enrichment factor values indicate that no Cr or Cu contamination has occurred on the whole, only a few stations have been polluted by As, and some areas have been polluted by Cd, Pb, and Zn. Principal component analysis suggests that the Cu, Pb, Zn, and Cd are derived from anthropogenic inputs and that Cr, As, Cu, and Zn are influenced by natural weathering processes. Cu and Zn may originate from both natural and anthropogenic sources. The contamination in the northeastern JZB is higher than that in other areas of the bay.

Probing the influential factors of NMR T1-T2 spectra in the characterization of the kerogen by numerical simulation.

The low field nuclear magnetic resonance (NMR) spectroscopy has been widely used to characterize the longitudinal and transversal relaxation (T1-T2) spectrum of unconventional resources such as shale gas and tight oil containing significant proportions of kerogen and bitumen. However, it requires exquisite design of the acquisition model and the inversion algorithm due to the fast relaxation nature of the kerogen and bitumen. A new direct two dimensional (2D) inversion algorithm combined the iterative truncated singular value decomposition (TSVD) and the Akaiake Information Criterion (AIC) is presented to perform the data inversion efficiently. The fluid component decomposition (FCD) is applied to construct the forward T1-T2 model of the kerogen, and numerical simulations are conducted to investigate factors which may influence inversion results including echo spacing, recovery time series, signal to noise ratio (SNR), and the maximal iteration time. Results show that the T2 component is heavily impaired by the echo spacing, whereas the T1 component is influenced by the recovery time series but with limited effects. The inversion precision is greatly affected by the quality of the data. The inversed spectrum deviates from the model seriously when the SNR of the artificial noise is lower than 50, and the T2 component is more sensitive to the noise than the T1 component. What's more, the maximal iteration time can also affect the inversion result, especially when the maximal iteration time is smaller than 500. Proper acquisition and inversion parameters for the characterization of the kerogen are obtained considering the precision and the computational cost.