Hybrid IoT and AI-based Solution for Energy Management in Data Centres under Various Climate Conditions

Authors : Alireza Esmaili Jobani, Şükrü Mustafa Kaya

Pages : 107-124

View : 61 | Download : 100

Publication Date : 2025-12-12

Article Type : Research Paper

Abstract :The rapidly increasing demand for data processing and storage has made data centers one of the largest global energy consumers. This study proposes a hybrid energy management model that integrates diesel generators, solar panels, and wind turbines to optimize energy consumption in data centers. The developed system utilizes the Internet of Things (IoT) infrastructure and Artificial Intelligence (AI)-based machine learning algorithms to adapt to varying climatic conditions. Real-time data collected from IoT sensors—such as weather parameters, battery charge levels, and energy production rates—are processed using algorithms including Support Vector Classifier (SVC), Random Forest (RF), AdaBoost, Logistic Regression (LR), and Naive Bayes (NB), enabling autonomous and highly accurate energy source management. Simulations conducted in MATLAB and Python show up to a 20% reduction in energy consumption, a 15% decrease in operational costs, and a 50.1% increase in renewable energy utilization. In addition to supporting environmental sustainability, the proposed system enhances reliability under variable weather conditions, offering an intelligent and practical solution for data centers located in regions with renewable energy potential, such as Istanbul.
Keywords : Yapay Zekâ, Nesnelerin İnterneti, Veri Madenciliği, Enerji Tüketimi, Makine Öğrenimi, Yenilenebilir Enerji Kaynakları.

ORIGINAL ARTICLE URL