Improving Peak Power Shaving in Data Centers using Deep Neural Networks Public Deposited

http://ir.library.oregonstate.edu/concern/graduate_projects/1v53k295k

Descriptions

Attribute NameValues
Creator
Abstract or Summary
  • Data centers have been charged a great amount of electric bill by the power company and demand charge can contribute up to 40% of the electric bill due to the "random" workload. This phenomenon can be avoided by using the existing Uninterrupted Power Supply (UPS) as the assistant power source to supply the servers. The UPS will shave the peak power by supporting the servers while the peak power exceeds threshold during the on-peak duration. For using the UPS to compensate the extra power, the problems are predicting the future data center resource usage and predicting the remaining capacity in the UPS battery. Since both problems involve in a highly correlated temporal relationship among their data, adopting Long-Short-Term-Memory(LSTM) Neural Network to perform the prediction solves the problems. After the experiment, the Long-Short-Term-Memory(LSTM) Neural Network solves the Resources Usage Prediction problem by having 6% Root Mean Square Error(RMSE) which outperforms the others and UPS Remaining Capacity achieves 7% RMSE.
License
Resource Type
Date Available
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Rights Statement
Publisher
Language

Relationships

Parents:

This work has no parents.

Last modified

Downloadable Content

Download PDF

Items