The Tie-line Power Fluctuation Smoothing Algorithm Based On Demand Side Resources Perception For Green Data Centers

With the rapid development of information technologies,such as cloud computing and big data,as the main carrier of data storage,operation and interaction,data centers have been gradually deployed in the world.As a consequence,high pollution and high energy consumpution become much more serious than ever before.In order to cope with the severe challenge of energy saving in data centers,renewable energy power generation technologies,such as wind power and photovoltaic power generation,have pointed out a new direction for the green sustainable development of data centers.While unlike the traditional power supply model using fossil fuel combustion,renewable energy is characterized by randomness,intermittency and instability.When such renewable energy is incorporated into the data center energy supply system,it will cause tie-line power fluctuations,which will not only affect the normal operation of data center facilities,but also have a negative impact on the main grid stable operation.The idea of green data centers provides an effective solution for energy efficiency optimization.According to the variable renewable energy output in green data centers,the intelligent load control technology is used to provide high efficiency and stable power supply for data centers.By analyzing the architecture of green data centers and the characteristics of user request task,this paper expores the controllable margin of data center server cluser load,and then combined with the redundant UPS energy storage system,a tie-line power control method for data centers based on demand side response is proposed in this paper.The contribution of this paper is to establish a controllable load model of the server cluster based on the task delay mechanism and the demand side response model of the UPS battery pack,and then consider them as demand side resources to participate in tie-line power smoothing control.At the same time,taking the server cluster control margin changes and UPS battery operation economy into account,on one hand,a variable time constant control strategy is proposed in this paper: by adjusting the time constants of the two grade Butterworth filte,the response power targets of the server cluster and UPS battery group can be changed.On the other hand,on the basis of effective perception of demand side resources,a heuristic algorithm based on greedy strategy is proposed to realize the optimization of the load of the server cluster and the output of the UPS battery pack in this paper.The experimental part verifies the effectiveness of the control method.By analyzing the results of the basic calculation example,it is found that the power control method with variable time constant strategy can achieve better smoothing effects,meanwhile it can improve the operation economy of UPS battery pack.Furthermore,to verify the universality of the control method,experiments of the influence factor is added in this paper.The effects of new energy penetration,server cluster resource utilization range,delay-tolerant task categoris and proportion are discussed in turn.The results showed that in different experimental conditions,the proposed tie-line power control method can effectively smooth power fluctuations.