Design and feasibility analysis of a new completion monitoring technical scheme for natural gas hydrate production tests

As a prerequisite and a guarantee for safe and efficient natural gas hydrates(NGHs)exploitation,it is imperative to effectively determine the mechanical properties of NGHs reservoirs and clarify the law of the change in the mechanical properties with the dissociation of NGHs during NGHs production tests by depressurization.Based on the development of Japan’s two offshore NGHs production tests in vertical wells,this study innovatively proposed a new subsea communication technology-accurate directional connection using a wet-mate connector.This helps to overcome the technical barrier to the communication between the upper and lower completion of offshore wells.Using this new communication technology,this study explored and designed a mechanical monitoring scheme for lower completion(sand screens).This scheme can be used to monitor the tensile stress and radial compressive stress of sand screens caused by NGHs reservoirs in real time,thus promoting the technical development for the rapid assessment and real-time feedback of the in-situ mechanical response of NGHs reservoirs during offshore NGHs production tests by depressurization.

Visual commonsense reasoning with directional visual connections

To boost research into cognition-level visual understanding,i.e.,making an accurate inference based on a thorough understanding of visual details,visual commonsense reasoning(VCR)has been proposed.Compared with traditional visual question answering which requires models to select correct answers,VCR requires models to select not only the correct answers,but also the correct rationales.Recent research into human cognition has indicated that brain function or cognition can be considered as a global and dynamic integration of local neuron connectivity,which is helpful in solving specific cognition tasks.Inspired by this idea,we propose a directional connective network to achieve VCR by dynamically reorganizing the visual neuron connectivity that is contextualized using the meaning of questions and answers and leveraging the directional information to enhance the reasoning ability.Specifically,we first develop a GraphVLAD module to capture visual neuron connectivity to fully model visual content correlations.Then,a contextualization process is proposed to fuse sentence representations with visual neuron representations.Finally,based on the output of contextualized connectivity,we propose directional connectivity to infer answers and rationales,which includes a ReasonVLAD module.Experimental results on the VCR dataset and visualization analysis demonstrate the effectiveness of our method.