Optimization of energy efficiency and integration of renewable energy of data centers in Kenya

Abstract

There has been exponential growth of ICT industry such as artificial intelligence (AI), social media (Netflix, Tiktok) and cloud computing which requires enormous amount of data storage, power consumption and cooling. This has become the challenge for the data centers industry which offers this storage, cooling, and power demand for this expanding industry. Moreover, data centers are currently facing a significant energy efficiency problem, which is contributing to an increase in power consumption that directly affects the operation cost. This study sought to optimize energy efficiency and integration of renewable energy for Data centers in Kenya. The methodology involved site survey (with the aim of taking measurement of power usage effectiveness (PUE) parameters for analysis and optimization), numerical simulation using CFD software for thermal environmental optimization and HOMER software for cost benefit analysis of power sources. The findings showed that data center three PUE was improved by 37% when centralized chiller system was used in place of direct expansion units. Similarly, the PUE values for the other two data centers were improved by an average of 28% when modular power conditioning units (UPS equipment) was used in place of fixed frame. Additionally, the selected data center for analysis of integration of renewable energy as alternative source of energy realized a saving of up to Ksh.31,000,000.00 annually with a payback period of 5 years. On thermal environmental analysis using CFD software, tile opening ration of 10% and 70% tile opening ration was analyzed. An optimized design was achieved using 10% tile opening ration which increased the airflow of the cold air. Energy performance in air management system was also analyzed using return temperature index (RTI) index. The RTI for DC1and DC2 was 71.48% and 94.79% respectively, implying the cold air was by-passed without first being properly directed into IT equipment, resulting in wasted energy while that for DC3 was 114.86% implying that air was recirculated, resulting to hot air from the IT equipment mixing with the cold air from the air conditioning equipment. It is recommended that metering and monitoring systems be established for all energy-consuming devices to provide environmental data and data related to energy efficiency, including end-user power breakdown by the various load components, power chain, and infrastructure power usage to enable real time monitoring of PUE parameters to identify gaps for energy efficiency improvements. Data center operator should also consider integrating their power sources with renewable energy sources such as solar photovoltaic system to reduce their operation cost and increase their PUE value, hence the efficiency of the data center.