Research On Key Technologies Of Datacenter Power Consumption Regulation Towards Specific Targets In Smart Grid Environment

In recent years,with the maturity of new energy power generation technology,the scale of renewable energy is increasing.Due to the intermittent and random characteristics of these new energy sources,it is difficult to connect new energy power generation to the grid.The instability of new energy power generation will have a greater impact on the existing grid structure,and the resulting instability of the power system has become increasingly prominent.As a basic platform for hosting cloud applications,datacenters are mainly used for storage,computing and provisioning of various services.Cloud computing datacenters generally have hundreds to hundreds of thousands of servers,which have the characteristics of the large resources,strong heterogeneity,high energy consumption,wide distribution and so on.According to the characteristics of the datacenter,we can manage and adjust the power consumption of the datacenter to achieve the purpose of effectively utilizing and stabilizing the power system after renewable energy connected into the grid.Therefore,this thesis analyzes the main power consumption components of the datacenter server and cooling system,studies and analyzes the adjustable method of server power and cooling system power,and establishes the corresponding models about datacenter power adjustment,including the task scheduling model,dynamic voltage frequency scaling model,air conditioning cooling model and direct airside free cooling model.On this basis,by reusing the uninterruptible power supply equipment of the datacenter,the power source of the datacenter is changed and the power controllability of the datacenter could be further increased.In addition,combined with the demand response signal characteristics of smart grid,a dynamic optimal scheduling method is proposed.The method precisely adjusts the power consumption of the datacenter by comprehensively utilizing various means to achieve the goal of minimizing the cost of power adjustment.Aiming at the power management problem of geo-distributed datacenters,we further propose a task redistribution model,a solar-powered absorption cooling model and a chilled liquid storage system model.By considering various factors such as electricity price changes,operating costs,renewable energy constraints and demand response,synthetic management of server power and cooling system power in geodistributed datacenters and the minimization of total operating costs could be better achieved.