Environment data processing for a Data Centre (1)

• Main Author: Tan, Eric Jiawei
• Other Authors: Tan Rui
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/171912; https://doi.org/10.21979/N9/R1KU6R
• Institution: Nanyang Technological University

In the modern digital era, the increasing number of Data Centers (DCs) is a global trend driven by technological advancements. Alongside the escalating need for data storage and processing, the energy consumption of DCs is also on a steep rise. DCs are the one of the world’s greatest energy consumers. In tropical regions DC, substantial energy wastage is observed due to the continuous operation of electrical cooling systems aimed at controlling the DC’s temperature. Therefore, it becomes critical to investigate the environmental variables within DCs in these climates in order to enhance their management. This study focuses on the examination of the Tropical Dataset TDC1.0 to explore the relationships between environmental variables and energy/power consumption in tropical air free-cooled DCs. Additionally, we conducted a comprehensive analysis of the dataset to uncover trends and correlations among these variables. Our investigation involved the identification of relevant features associated with the target variables, followed by an exploration and optimisation of six distinct Machine Learning (ML) models: Linear Regression (LR), Decision Tree (DT), Random Forest (RF), XGBoost, as well as Neural Network (NN) models including Multi-Layer Perceptron (MLP) and TensorFlow-Neural Network (TF-NN). The results of the study offer promising insights into the efficacy of regression models for accurate predictions. This suggests that leveraging environmental variables within the DC can potentially enhance the management and monitoring of energy and power consumption in DCs.