Research On Optimal Resource Scheduling Of Data Center Microgrid In Uncertain Environment

With the rapid development of information age,the number of global data centers has increased dramatically.Data center operators are facing great challenges such as huge power consumption and serious environmental pollution.Due to the strict requirement of power supply reliability,data centers are often equipped with various backup power sources such as energy storage,diesel generators and so on.Considering corporate image and social responsibility,operators of large data centers tend to make use of renewable energy to build green data centers.From the point of load characteristics,the unique batch workload of data center can be transferred to different time periods on the premise of satisfying the quality of service.Thus,data centers have strong demand response ability.In the operation of data center microgrid,the fluctuation and intermittence of renewable energy generation,the uncertainty of spot market price and the randomness of workload distribution will all affect the rationality and economy of dispatching results.How to flexibly dispatch multiple microgrid resources under uncertain environment is a challenging problem.The main research work of this paper is as follows:Firstly,the operation characteristics of data center workload and various power supply in microgrid are analyzed,and the energy management framework of microgrid in data center is proposed.This paper described the power consumption characteristics of interactive workload and batch workload in data center,established the mathematical model of multiple power supply in microgrid,and put forward the basic principle of optimal scheduling from the perspective of data center operators.Secondly,an optimal dispatching model of data center microgrid is established by integrating various energy consumption optimization methods.Under the power market environment,the lowest operating cost and carbon emission cost can be achieved through batch load distribution,energy storage charging and discharging operation,and traditional unit economic dispatching.The effectiveness of the model is verified by solving mixed integer linear programming problem and comparing it with the results of partial optimization method.Finally,the uncertainties in the operation of data centers are analyzed and the conditional value at risk(CVaR)method is selected as the measure of operation risk.A stochastic optimization model of data center microgrid is established to realize the trade-off among operation cost,carbon emission and operation risk.A simulation example of data center microgrid is constructed,and the dayahead resource scheduling scheme driven by economy and environment for risk-neutral operators is given.By adjusting the risk preference coefficient,the influence of risk preference coefficient on resource selection and operation cost of data center scheduling is studied.This paper has a positive promotion to the research theory of data center energy consumption optimization.