针对现有能耗模型对动态工作负载波动具有低敏感性和低精度的问题,该文基于卷积长短期记忆(convolutional long short-term memory, ConvLSTM)神经网络,提出了用于移动边缘计算的服务器智能能耗模型(intelligence server energy consump...针对现有能耗模型对动态工作负载波动具有低敏感性和低精度的问题,该文基于卷积长短期记忆(convolutional long short-term memory, ConvLSTM)神经网络,提出了用于移动边缘计算的服务器智能能耗模型(intelligence server energy consumption model,IECM),用于预测和优化服务器的能量消耗。通过收集服务器运行时间参数,使用熵值法筛选和保留显著影响服务器能耗的参数。基于选定的参数,利用ConvLSTM神经网络训练服务器能耗模型的深度网络。与现有的能耗模型相比,IECM在CPU密集型、I/O密集型、内存密集型和混合型任务上,能够适应服务器工作负载的动态变化,并在能耗预测上具有更好的准确性。展开更多
In order to study the major performance indicators of the twin-rotor piston engine(TRPE), Matlab/simulink was used to simulate the mathematical models of its thermodynamic processes. With consideration of the characte...In order to study the major performance indicators of the twin-rotor piston engine(TRPE), Matlab/simulink was used to simulate the mathematical models of its thermodynamic processes. With consideration of the characteristics of the working processes in the TRPE, corresponding differential equations were established and then simplified by period features of the TRPE. Finally, the major boundary conditions were figured out. The changing trends of mass, pressure and temperature of working fuel in the working chamber during a complete engine cycle were presented. The simulation results are consistent with the trends of an actual working cycle in the TRPE, which indicates that the method of simulation is feasible. As the pressure in the working chamber is calculated, all the performance parameters of the TRPE can be obtained. The major performance indicators, such as the indicated mean effective pressure, power to weight ratio and the volume power, are also acquired. Compared with three different types of conventional engines, the TRPE has a bigger utilization ratio of cylinder volume, a higher power to weight ratio and a more compact structure. This indicates that TRPE is superior to conventional engines.展开更多
文摘针对现有能耗模型对动态工作负载波动具有低敏感性和低精度的问题,该文基于卷积长短期记忆(convolutional long short-term memory, ConvLSTM)神经网络,提出了用于移动边缘计算的服务器智能能耗模型(intelligence server energy consumption model,IECM),用于预测和优化服务器的能量消耗。通过收集服务器运行时间参数,使用熵值法筛选和保留显著影响服务器能耗的参数。基于选定的参数,利用ConvLSTM神经网络训练服务器能耗模型的深度网络。与现有的能耗模型相比,IECM在CPU密集型、I/O密集型、内存密集型和混合型任务上,能够适应服务器工作负载的动态变化,并在能耗预测上具有更好的准确性。
基金Project(7131109)supported by the National Defense Pre-research Foundation of ChinaProject(51175500)supported by the National Natural Science Foundation of China
文摘In order to study the major performance indicators of the twin-rotor piston engine(TRPE), Matlab/simulink was used to simulate the mathematical models of its thermodynamic processes. With consideration of the characteristics of the working processes in the TRPE, corresponding differential equations were established and then simplified by period features of the TRPE. Finally, the major boundary conditions were figured out. The changing trends of mass, pressure and temperature of working fuel in the working chamber during a complete engine cycle were presented. The simulation results are consistent with the trends of an actual working cycle in the TRPE, which indicates that the method of simulation is feasible. As the pressure in the working chamber is calculated, all the performance parameters of the TRPE can be obtained. The major performance indicators, such as the indicated mean effective pressure, power to weight ratio and the volume power, are also acquired. Compared with three different types of conventional engines, the TRPE has a bigger utilization ratio of cylinder volume, a higher power to weight ratio and a more compact structure. This indicates that TRPE is superior to conventional engines.
