No abstract

No abstract

Accurate prediction of shale gas well production and estimated ultimate recovery (EUR) is always a difficult and hot spot in shale gas development. In particular, the production and EUR prediction of shale gas wells in new production blocks are faced with the lack of field gas well data and the difficulty of model development. In view of the above problems, this study proposes a new deep transfer learning strategy, which uses transfer component analysis (TCA) and deep neural network (DNN) to achieve shale gas well production and EUR prediction across formations/blocks. The feature extractor based on TCA can narrow the input feature distribution of the source and the target domains. The neural network model can be used to establish a domainadaptive transfer learning model without the prediction performance degradation caused by distribution offset. Validity and accuracy of the model were analyzed using gas well data from Weiyuan and Luzhou blocks in Sichuan Basin, China. The results appear that the reasonable application of TCA can greatly improve the prediction performance of shale gas well transfer learning model. For data sets of the same size, compared with the transfer learning model developed by classical machine learning algorithms, the proposed neural network-based transfer learning model can significantly improve the accuracy of production prediction across formations/ blocks. In addition, the proposed model can also be extended to other types of oil and gas production prediction tasks cross formations/blocks.

Imbibition and flowback of fracturing fluid usually occur in the shale matrix after hydraulic fracturing, which significantly impacts shale gas production and environmental protection. The rocks of deep shale gas reservoirs are under high-temperature and high-temperature conditions. There are rich micro-nano pores with various pore structures in deep shale. In addition, the flowback behavior is significantly affected by the imbibition behavior because the flowback begins after the end of the imbibition. Therefore, an accurate pore-scale description of the coupled imbibition-flowback behavior is crucial to understand the flowback mechanism and its impacts. In this paper, a pseudo-potential lattice Boltzmann method is employed to simulate the coupled imbibition-flowback behavior in a digital shale core, where the digital core is reconstructed by Markov Chain-Monte Carlo method based on scanning microscope images of deep shale cores. The microcosmic mechanism of the imbibition and flowback is studied under deep shale conditions. The influence of some factors, such as pore structure, fluid viscosity, wettability, and flowback pressure difference, on the flowback behavior of fracturing fluid is investigated. It is found that the fracturing fluid advances almost uniformly throughout the pore space during the imbibition process. The fracturing fluid is easy to adsorb on the pore wall, and the shale gas is located in the middle of the pore space. The viscous fingering is clearly observed during the flowback process, where shale gas flows through large pores to form a flow channel, and the fracturing fluid stays in tiny pores. The flowback rate increases gradually with the flowback time and eventually tends to be almost constant. The wettability, flowback pressure difference, and pore structure significantly influence the flowback behavior, while the fracturing fluid viscosity has a smaller effect on the flowback process.

The H oilfield is currently in the development planning stage. There were two problems that were to be resolved during the study. First, there is no clear relationship between porosity and permeability; if using single relationship to calculate permeability, then strong heterogeneity of the reservoir will be concealed, and the development plan will be affected. Second, how is it possible to classify reservoir rock type on wells and in 3D models? Through analysis of core and laboratory data, rock types are divided into four categories, and each type gets a good porosity-permeability relationship, the formula of each type has been calculated the permeability. For the purpose of reservoir classification on wells and in 3D model, the cutoffs of electrical logs and petrophysical properties were analyzed. Vshale and DEN were used to divide reservoir on wells, and FZI and pore throat radius were used to divide reservoir in 3D model. In the end reservoir is in quantitative characterization.