Quantitative characterization of shale gas reservoir properties based on BiLSTM with attention mechanism

Evaluating the potential of shale gas reservoirs is inseparable from reservoir properties prediction.Accurate characterization of total organic carbon,porosity and permeability is necessary to understand shale gas reservoirs.Seismic data can help to estimate these parameters in the area crossing-wells.We develop an improved deep learning method to achieve shale gas reservoir properties estimation.The rela-tionship between elastic attributes and reservoir properties is built up by training a deep bidirectional long short-term memory network,which is suitable for time/depth sequence prediction,on the logging and core data.Except some commonly used technologies,such as layer normalization and dropout,we also introduce attention mechanism to further enhance the prediction accuracy.Besides,we propose to carry on the normal scores transform on the input features,which aims to make the relationship between inputs and targets clear and easy to learn.During the training process,we construct quantile loss function,then use Adam algorithm to optimize the network.Not only the characterization results,but also the confidence interval can be output that is meaningful for uncertainty analysis.The well exper-iment indicates that the method is promising for reducing prediction errors when training samples are insufficient.After analyzing in wells,the established model is acted upon seismic inverted elastic attri-butes to characterize shale gas reservoirs in the whole studied area.The estimation results coincide well with the actual development results,showing the feasibility of the novel method on the characterization for shale gas reservoirs.

The earthquake precursor detected in a granular medium and a proposed model for the propagation of precursive stress-strain signal

A way to detect the seismic precursor in granular medium is described and a model of propagation for precursive stress-strain signals is proposed.A strain sensor buried in a sandpit is used to measure a seismic precursor signal.The signal has been investigated and confirmed to originate from a specific earthquake.A comparison of simulated and experimental signals indicates that the signal results from the strain in the earth's strata.Based on the behavioral characteristics of granular materials,an analysis of why this method can be so sensitive to the seismic strain signal is undertaken and a model for the propagation of this stress-strain signal is proposed.The Earth's lithosphere is formed of tectonic plates,faults and fault gouges at their boundaries.In the case of the quasi-static mechanics of seismic precursory stress-strain propagation,the crustal lithosphere should be treated as a large-scale granular system.During a seismogenic event,accumulated force generates the stick-slip motion of adjacent tectonic plates and incrementally pushes blocks farther apart through stick-slip shift.The shear force released through this plate displacement causes soil compression deformation.The discrete properties of the sand in the sandpit lead to the sensitive response of the sensor to the deformation signal which enables it to detect the seismic precursor.From the analysis of the mechanism of the stress-strain propagation in the lithosphere,an explanation is found for the lack of signal detection by sensors installed in rocks.The principles and method presented in this paper provide a new technique for investigating seismic precursors to shallow-source earthquakes.