Collaborative Optimal Operation Of Data Center Workload And Power Supply System

With the advent of the information age and the development of the industrial Internet of Things,the important role of big data and cloud computing technology has become increasingly prominent,which has attracted widespread attention from all countries in the world and has become one of the core industries for the key development of China’s 14th Five-Year Plan.As the hardware foundation of the big data industry,data center has received extensive attention,and has become an important part of the national new infrastructure strategy,with broad development prospects.In recent years,the scale of the data center industry has been expanding,and the number and scale of data centers have been increasing.Followed by the rapid growth of data center energy consumption,energy consumption costs accounted for a gradual increase in the proportion of data center operating costs.How to reduce the energy consumption of data center equipment rooms and improve the power efficiency of each data center has become an urgent problem to be solved in this field.At the same time,due to the flexibility of scheduling in the existence time and space of data center loads,the effective use of local new energy and the saving of electricity costs in multiple data centers can be realized through the flexible scheduling of computing loads,so the collaborative scheduling research between multiple data centers and data center computing load and power supply side can be carried out,so as to further use the difference between new energy and electricity prices to reduce operating costs and power loss.In addition,in the existing research,there is little research on the pricing method of data center power purchase price,and the current power purchase price cannot guarantee the optimal electricity price.Through the game research between the data center and the power sales side,the most appropriate real-time electricity price can be formulated,which can effectively restrict and guide the operation mode of the data center and achieve a win-win situation between the data center and the power grid.On the power supply side,there are a large number of multi-stage power conversion links in the data center,and most of them operate in coordination with new energy,energy storage equipment,generators and other microgrid equipment.On the load side,data center loads have high scheduling flexibility in the time dimension and space dimension.In view of the above characteristics,this paper studies the two-layer data center load scheduling method considering the efficiency of data center power supply equipment,the distributed multi-data center microgrid collaborative scheduling method considering power supply efficiency,and the data center power purchase pricing method based on cooperative game and Steinberg game,and the specific research content is as follows:(1)A virtual machine placement method considering the nonlinear characteristics of energy consumption of uninterruptible power supply equipment in data center is proposed.Firstly,the energy consumption of data center power supply equipment(Uninterruptible Power Supply equipment)is modeled,and the placement of data center load(Virtual Machine)is modeled as a nonlinear discrete one-dimensional packing problem,so as to ensure the accuracy of modeling and the effectiveness of scheduling.Subsequently,in order to optimize the calculation and simplify the solution,the traditional adaptive gradient descent method is combined with the genetic algorithm to propose a two-step virtual machine placement method.In order to further verify the proposed method,the most widely used data center simulation software CloudSim software was used to conduct simulation experiments,and in view of the lack of power supply equipment model of the software,the model of the data center power supply part was added to the software and the effectiveness of the device was tested.Finally,the load data of PlanetLab public platform is used on the simulation platform for verification,and the results show that the proposed method can effectively reduce the energy consumption of data center compared with the four traditional methods,and the solution time is greatly reduced.(2)A geographically distributed multi-data center microgrid collaborative scheduling method considering the efficiency of data center power supply equipment is proposed.Firstly,the server load,uninterruptible power supply equipment energy consumption,energy storage and microgrid equipment such as energy storage,photovoltaic,wind power,and generator in the multi-data center microgrid are modeled.Subsequently,in order to solve the problem of large scale and long solution time,a scheduling solution method combining day-ahead scheduling and real-time scheduling was proposed,and the Taguchi experimental method was used to parameterize.Finally,six cases were set up for comparative analysis,and simulated and solved in Gurobi software.The simulation results show that the proposed method can better reduce the energy consumption and operating cost of multiple data centers.(3)A data center electricity pricing method based on game theory is proposed,and the cooperative game and non-cooperative game between data center and electricity retailers are studied respectively.In the research of cooperative game methodology,the cooperative game model of data center and electricity retailer is first established,and then the direct load control method is used to realize the cooperation between data center and electricity retailer under actual working conditions,and finally the Shapley value is used to distribute the profits of cooperative game.The simulation results show that compared with the traditional electricity price,after adopting the optimization game method,the operating cost of the data center is greatly reduced,and the profit of electricity retailer is further increased.In the research of non-cooperative game method,a two-layer optimization model based on the Stackelberg game method is first established to model the game between retailer data center as an upper-layer electricity price game and the lower data center operation optimization.Subsequently,a two-tier solution method based on genetic algorithm and linear programming is proposed.Finally,the model is solved and analyzed.The results show that compared with the traditional electricity price,the electricity price formulated by the non-cooperative game method can effectively reduce the operating cost of the data center,and significantly increase the revenue of electricity retalier,achieving a win-win situation for both.