This paper studies the current available options for floating production platforms in developing deepwater oil fields and the potential development models of future oil and gas exploration in the South China Sea. A de...This paper studies the current available options for floating production platforms in developing deepwater oil fields and the potential development models of future oil and gas exploration in the South China Sea. A detailed review of current deepwater platforms worldwide was performed through the examples of industry projects, and the pros and cons of each platform are discussed. Four types of platforms are currently used for the deepwater development: tension leg platform, Spar, semi-submersible platform, and the floating production system offloading. Among these, the TLP and Spar can be used for dry tree applications, and have gained popularity in recent years. The dry tree application enables the extension of the drilling application for fixed platforms into floating systems, and greatly reduces the cost and complexity of the subsea operation. Newly built wet tree semi-submersible production platforms for ultra deepwater are also getting their application, mainly due to the much needed payload for deepwater making the conversion of the old drilling semi-submersible platforms impossible. These platforms have been used in different fields around the world for different environments; each has its own advantages and disadvantages. There are many challenges with the successful use of these floating platforms. A lot of lessons have been learned and extensive experience accumulated through the many project applications. Key technologies are being reviewed for the successful use of floating platforms for field development, and potential future development needs are being discussed. Some of the technologies and experience of platform applications can be well used for the development of the South China Sea oil and gas field.展开更多
由于主机负载具有短期突变性和非线性等特点,主机负载中短期突变的信息难以被捕获。为提高主机负载预测的准确性,设计并实现一种基于Zoneout的LSTM(long short term memory with zoneout,LSTM-Z)主机负载预测方法。该方法能适应具有波...由于主机负载具有短期突变性和非线性等特点,主机负载中短期突变的信息难以被捕获。为提高主机负载预测的准确性,设计并实现一种基于Zoneout的LSTM(long short term memory with zoneout,LSTM-Z)主机负载预测方法。该方法能适应具有波动性特点的主机负载预测模式,通过遗传算法在迭代进化过程中探索最优的历史窗口权重向量,充分利用历史数据依赖关系,提高预测的准确性。通过在谷歌和阿里云两个真实的云平台数据上进行单步和多步预测实验,验证了其有效性。展开更多
文摘This paper studies the current available options for floating production platforms in developing deepwater oil fields and the potential development models of future oil and gas exploration in the South China Sea. A detailed review of current deepwater platforms worldwide was performed through the examples of industry projects, and the pros and cons of each platform are discussed. Four types of platforms are currently used for the deepwater development: tension leg platform, Spar, semi-submersible platform, and the floating production system offloading. Among these, the TLP and Spar can be used for dry tree applications, and have gained popularity in recent years. The dry tree application enables the extension of the drilling application for fixed platforms into floating systems, and greatly reduces the cost and complexity of the subsea operation. Newly built wet tree semi-submersible production platforms for ultra deepwater are also getting their application, mainly due to the much needed payload for deepwater making the conversion of the old drilling semi-submersible platforms impossible. These platforms have been used in different fields around the world for different environments; each has its own advantages and disadvantages. There are many challenges with the successful use of these floating platforms. A lot of lessons have been learned and extensive experience accumulated through the many project applications. Key technologies are being reviewed for the successful use of floating platforms for field development, and potential future development needs are being discussed. Some of the technologies and experience of platform applications can be well used for the development of the South China Sea oil and gas field.
文摘由于主机负载具有短期突变性和非线性等特点,主机负载中短期突变的信息难以被捕获。为提高主机负载预测的准确性,设计并实现一种基于Zoneout的LSTM(long short term memory with zoneout,LSTM-Z)主机负载预测方法。该方法能适应具有波动性特点的主机负载预测模式,通过遗传算法在迭代进化过程中探索最优的历史窗口权重向量,充分利用历史数据依赖关系,提高预测的准确性。通过在谷歌和阿里云两个真实的云平台数据上进行单步和多步预测实验,验证了其有效性。
