An economic room-level thermal management of air-cooled cloud data centers based on human brain emotional intelligence

Cloud Data Centers (CDCs) are consistently recognized as pivotal in the digital economy due to their provision of flexible and cost-effective solutions for storing and processing large volumes of data. These facilities enable businesses to adapt to evolving demands, support advanced technologies, and facilitate global collaboration by offering instant access to resources. Consequently, ensuring their reliable and uninterrupted operation remains a primary concern for technology enterprises and service providers. Among the various factors influencing the performance of CDCs, thermal conditions are of particular significance. In this context, a novel Brain Emotional Learning-Based Intelligent Controller (BELBIC) is proposed in this paper for room-level temperature regulation of an air-cooled CDC. Unlike other established controllers, the proposed BELBIC is both intelligent and adaptive, thereby exhibiting superior performance and adaptability to the dynamic conditions of CDCs. To evaluate the efficacy of the proposed approach, it has been implemented in an air-cooled CDC comprising two server clusters across four distinct weather seasons. The performance of the proposed BELBIC is subsequently benchmarked against four existing controllers, with comparisons based on heat cost. In Cluster 1, Brain Emotional Learning-Based Intelligent Controller (BELBIC) reduces costs by approximately 94.00% to 95.72% compared to Fractional Order Proportional Integral controller (FPI) and Fractional Order Proportional Integral Derivative controller (FPID), and by about 71.13% to 72.47% compared to Proportional Integral Derivative controller (PID) and Proportional Integral controller (PI). Similarly, in Cluster 2, BELBIC achieves cost reductions of around 93.84% to 95.65% versus FPI and FPID, and approximately 70.51% to 72.26% versus PID and PI. The results demonstrate that the proposed model surpasses the others by achieving reduced heat costs and faster thermal regulation, attributed to its intelligent and adaptive features. • Novel human-brained thermal regulator with intelligence and adaptability proposed for a CDC. • Multi-cluster analytics model enables distinct control for each server cluster. • Economic analytics framework established for performance evaluation. • Superior performance in faster regulation and lower costs versus PID, PI, FPID, and FPI controllers.