A Scalable Digital Twin Framework for Energy Optimization in Data Centers
A Scalable Digital Twin Framework for Energy Optimization in Data Centers
Authors: Raphael Hendrigo de Souza Gonçalves, Wendel Marcos dos Santos Date: 2026-05-07 Paper ID: openalex:2605.05581
Summary
This paper presents a scalable Digital Twin framework designed to optimize energy efficiency in data centers by integrating IoT sensor data with machine learning-based forecasting. The authors utilize Long Short-Term Memory (LSTM) models to predict energy demand, enabling proactive operational decision-making. Experimental results in a controlled, small-scale environment confirm that the framework reduces power consumption and improves overall Power Usage Effectiveness (PUE). The approach offers a cost-effective path toward more sustainable data center management through intelligent energy monitoring and optimization.
Key Contributions
- Proposes a scalable Digital Twin framework integrating IoT data acquisition and LSTM-based forecasting for data center energy management.
- Demonstrates improved Power Usage Effectiveness (PUE) and reduced power consumption in a controlled small-scale data center environment.
- Validates the framework’s effectiveness in real-time monitoring and operational decision-making using LSTM models for energy demand prediction.
Open Questions & Future Work
Archivist Review
The paper describes a standard application of LSTM for energy forecasting within a digital twin context, which does not introduce a novel algorithmic contribution or unique temporal representation suitable for a permanent vault note. I approved a refined version of the open question regarding the scalability gap, as this represents a recurring challenge for time-series-based digital twins in industrial infrastructure.
Approved Open Questions
- Scaling Digital Twins for Data Centers: The transition from small-scale experimental success to robust performance in large-scale infrastructure remains a critical barrier to the industrial adoption of digital twin technologies for energy optimization.
Rejected Candidates
- [open_question] Scaling Digital Twins for Data Centers (
scaling-digital-twins-for-large-data-centers) - duplicate_existing: Renamed to be more precise regarding production environments.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2605.05581
- paper_id
- 2605.05581
- paper_source
- openalex
- domain
- time-series
- tags
-
- architectures
-
- datasets
-
- concept_slugs
-
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-05-10T06:05:14Z