Machine Reading Comprehension(MRC) aims to enable machines to independently reason and extract information and answer questions. This study proposes improvements based on the Bidirectional Attention Flow(BiDAF) model to enhance the accuracy and efficiency of the question-answering system.Firstly, the BiDAF model imposes constraints on text length during modeling, which may lead to the loss of answers when processing long texts. To address this issue, a sliding window mechanism is introduced to retain excessive information within the same text. Secondly, due to the use of Long Short Term Memory (LSTM) in the model, it is difficult for the model to capture distant time step information, resulting in long-term dependence issues and poor parallelization capability; Based on this, an Encoder model based on self-attention mechanism is used to extract text information. Finally, for the limited length, the method of matching within the group and external sorting outside the group is designed to obtain the position information of model training. Test results of the improved BiDAF model on the public dataset SQuAD show that the F1 score and Exact Match (EM) rate have increased by 2.48 percentage points and 11.86 percentage points respectively compared with the traditional BiDAF model.

The low?permeability reservoir in Qiuling Oilfield of Tuha shows that the characteristics of early water discovery and rapid rise of water cut in some wells during water injection development, and dynamic analysis shows that there are dominant channels or the possibility of fractures in the reservoir. In order to further improve the development effect and enhance oil recovery, it is necessary to verify and determine the development of micro?fractures in the reservoir at the later stage of reservoir development. In this paper, based on the description of the core of a water washing inspection well, the development and distribution of reservoir fractures are studied, and a method of identifying the fracture development horizon based on logging curves are proposed. Through the analysis of logging data of two wells and the comparison of water absorption profiles, the method is proved to be feasible and reliable, and the study provides theoretical support for improving the development effect of water drive, effective utilization and recovery of remaining oil in Qiuling Oilfield of Tuha.

In view of the large displacement production of marine sandstone FOURD reservoir, it is found that the water drive pore volume multiples at different locations are different during the development process of bottom water reservoir, while the experiments measured in the conventional industry standard are insufficient to express the seepage characteristics of the reservoir after the high flooding multiple. The high?power water drive experiment shows that the half logarithm of core displacement multiples with different clay content has a linear relationship with oil displacement efficiency, and the water drive sweep volume of bottom water reservoir increased slowly. Based on the actual data of FOURD reservoir, the time?varying characteristics of the reservoir were characterized and studied by numerical simulation technology. According to the different north?south dynamics of FOURD reservoir, the water flooding degree in the south area is high and the sweep degree is difficult to improve. The oil displacement efficiency of the reservoir is improved by liquid extraction, and in the north area and where the impact degree is not high, the later potential tapping mode of well network encryption is adopted. After adopting the limit tapping potential strategy described in this paper, according to the results of numerical simulation and practical mine application, the swept volume increased by 11.1%, and the calibration recovery increased by 4.5%.

In view of the current situation that independent automobile brands are big but not strong in the domestic market, this paper took Chinese independent automobile brands as the research object, analyzed the consumption psychology of Chinese consumers, and discusseed the influence of brand cognition, customer experience, perceived novelty, perceived practicality and friends' recommendation on brand advocacy, and with customer purchase intention as the mediating variable, the structural equation theory was used to construct, test and revise the model, and the hypothesis test was carried out. The results show that customer experience, perceived novelty and friends' recommendation can directly promote customers' purchase intention and brand advocacy, while perceived practicality can indirectly promote brand advocacy, while brand cognition can only directly promote brand advocacy. Purchase intention plays a mediating role among customer experience, perceived novelty, perceived practicality, friends' recommendation and brand advocacy,but does not play an intermediary role between brand cognition and brand advocacy..