Pipelines are the critical link between major offshore oil and gas developments and the mainland. Any inadequate on-bottom stability design could result in disruption and failure, having a devastating impact on the ec...Pipelines are the critical link between major offshore oil and gas developments and the mainland. Any inadequate on-bottom stability design could result in disruption and failure, having a devastating impact on the economy and environment. Predicting the stability behavior of offshore pipelines in hurricanes is therefore vital to the assessment of both new design and existing assets. The Gulf of Mexico has a very dense network of pipeline systems constructed on the seabed. During the last two decades, the Gulf of Mexico has experienced a series of strong hurricanes, which have destroyed, disrupted and destabilized many pipelines. This paper first reviews some of these engineering cases. Following that, three case studies are retrospectively simulated using an in-house developed program. The study utilizes the offshore pipeline and hurricane details to conduct a Dynamic Lateral Stability analysis, with the results providing evidence as to the accuracy of the modeling techniques developed.展开更多
Dynamic neural network(NN)techniques are increasingly important because they facilitate deep learning techniques with more complex network architectures.However,existing studies,which predominantly optimize the static...Dynamic neural network(NN)techniques are increasingly important because they facilitate deep learning techniques with more complex network architectures.However,existing studies,which predominantly optimize the static computational graphs by static scheduling methods,usually focus on optimizing static neural networks in deep neural network(DNN)accelerators.We analyze the execution process of dynamic neural networks and observe that dynamic features introduce challenges for efficient scheduling and pipelining in existing DNN accelerators.We propose DyPipe,a holistic approach to optimizing dynamic neural network inferences in enhanced DNN accelerators.DyPipe achieves significant performance improvements for dynamic neural networks while it introduces negligible overhead for static neural networks.Our evaluation demonstrates that DyPipe achieves 1.7x speedup on dynamic neural networks and maintains more than 96%performance for static neural networks.展开更多
基金supported by the Research Development Awards of University of Western Australia,Australia-China Natural Gas Technology Partnership Fund and Lloyd's Register Foundationsupports the advancement of engineering-related education and funds research and development that enhance the safety of life at sea,on land,and in the airforms part of the activities of the Centre for Offshore Foundation Systems(COFS) above,currently supported as a primary node of the Australian Research Council Centre of Excellence for Geotechnical Science and Engineering
文摘Pipelines are the critical link between major offshore oil and gas developments and the mainland. Any inadequate on-bottom stability design could result in disruption and failure, having a devastating impact on the economy and environment. Predicting the stability behavior of offshore pipelines in hurricanes is therefore vital to the assessment of both new design and existing assets. The Gulf of Mexico has a very dense network of pipeline systems constructed on the seabed. During the last two decades, the Gulf of Mexico has experienced a series of strong hurricanes, which have destroyed, disrupted and destabilized many pipelines. This paper first reviews some of these engineering cases. Following that, three case studies are retrospectively simulated using an in-house developed program. The study utilizes the offshore pipeline and hurricane details to conduct a Dynamic Lateral Stability analysis, with the results providing evidence as to the accuracy of the modeling techniques developed.
基金supported by the Beijing Natural Science Foundation under Grant No.JQ18013the National Natural Science Foundation of China under Grant Nos.61925208,61732007,61732002 and 61906179+1 种基金the Strategic Priority Research Program of Chinese Academy of Sciences(CAS)under Grant No.XDB32050200the Youth Innovation Promotion Association CAS,Beijing Academy of Artificial Intelligence(BAAI)and Xplore Prize.
文摘Dynamic neural network(NN)techniques are increasingly important because they facilitate deep learning techniques with more complex network architectures.However,existing studies,which predominantly optimize the static computational graphs by static scheduling methods,usually focus on optimizing static neural networks in deep neural network(DNN)accelerators.We analyze the execution process of dynamic neural networks and observe that dynamic features introduce challenges for efficient scheduling and pipelining in existing DNN accelerators.We propose DyPipe,a holistic approach to optimizing dynamic neural network inferences in enhanced DNN accelerators.DyPipe achieves significant performance improvements for dynamic neural networks while it introduces negligible overhead for static neural networks.Our evaluation demonstrates that DyPipe achieves 1.7x speedup on dynamic neural networks and maintains more than 96%performance for static neural networks.
