Energy Management and Operational Optimization of Data Centers in a Smart Grid with Wind Power Integration and Service Delay Constraints

Tayseer Karam  Alshekly; Royida A.Ibrahem  Alhayali; Raed Abdulkareem  Hasan; Khelil Yaseen  Ahmed; Omar K.  Ahmed

doi:10.37934/ard.138.1.162174

Authors

Tayseer Karam Alshekly Department of Computer Science, Al-Imam Al-A’tham University, Baghdad, Iraq
Royida A.Ibrahem Alhayali Department of Computer Engineering, College of Engineering, University of Diyala, Diyala, Iraq
Raed Abdulkareem Hasan Renewable Energy Research Unit, Northern Technical University, Mosul, Iraq
Khelil Yaseen Ahmed Department of Computer Engineering, College of Engineering, University of Diyala, Diyala, Iraq
Omar K. Ahmed Department of Computer Engineering, College of Engineering, University of Diyala, Diyala, Iraq

DOI:

https://doi.org/10.37934/ard.138.1.162174

Keywords:

Energy Management, Operational Optimization, Wind Power Integration, Service Delay Constraints, Smart Grid

Abstract

The paper will present a framework for management in energy and operational optimization of smart grid integrated data centers (DCs) under the operation of wind power generation, with diesel generators as a backup source. The generated wind power is not reliable due to its desirable intermittent nature, and it will require energy to be drawn from the grid so that the DC could operate without interruption. This paper considers the nature of mix to be supplied, where the total maximum peak loading of 35% of the total consumption of the DC may be made out of the wind and when in normal operation while in grid outage it is required to work in an islanded mode and it would use diesel to generate power for the fulfillment of this demand. The paper provides an elaborate analysis of DC workload and service delay including penalties for delay under Amazon Elastic Compute Cloud (EC2) Service Level Agreements (SLA). It further relates to power consumption modeling that incorporates server utilization and Power Usage Effectiveness (PUE), which indicates variations in power consumptions during the day concerning workload demands. This work proposes strategies to balance the trade-offs between energy costs, service reliability, and delay penalties to optimize DC operations in both grid-connected and islanded modes.