Data-Driven Prediction of Power Usage Effectiveness: A Machine Learning Case Study

This paper presents the results of a comprehensive six-month case study on the operations of an IT company’s data center. Environmental and operational parameters such as temperature, humidity, voltage, energy consumption, and power readings were meticulously collected. The Power Usage Effectiveness (PUE) metric was derived using real-time data on consumed versus total available energy. We applied machine learning models, including Support Vector Machine (SVM), Linear Regression, and Random Forest, to predict the PUE metric. Each model’s performance was evaluated using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). The results demonstrated that the Random Forest model outperformed the other models, indicating its potential as an effective tool for predicting data center energy efficiency. This paper is important as it provides actionable insights and a data-driven approach to improving energy efficiency in data centers, which is critical for reducing operational costs and minimizing environmental impact.