A secure data center optimized coldness prediction method and system

By employing supervised autoencoders and IoT devices in data centers, a PUE optimization framework with security assurance is constructed, solving the problems of temperature security and accuracy of cooling capacity prediction in data center cooling system optimization, and achieving efficient and stable cooling capacity prediction and PUE optimization.

CN117313351BActive Publication Date: 2026-06-19NANJING UNIV
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NANJING UNIV
Filing Date
2023-09-25
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

Existing technologies struggle to accurately capture complex characteristics during data center cooling system optimization, resulting in suboptimal, unstable optimization results that neglect security issues, particularly the security boundary constraints of data center temperature.

⚗Method used

By employing a neural network based on supervised autoencoders, combined with wireless sensors and IoT devices, a global monitoring system for data center environment and water cooling system is implemented. A PUE optimization framework with security guarantees is constructed. The relationship between IT load and cooling capacity is determined through multinomial fitting or machine learning methods, and a global optimization method is used to meet equipment constraints to achieve high-accuracy and low-complexity cooling capacity prediction.

🎯Benefits of technology

It ensures safety during the data center PUE optimization process, keeps the computer room temperature within a suitable range, reduces the energy consumption of the cooling system, and improves the accuracy of cooling capacity prediction and the stability of optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

This invention discloses a secure method and system for predicting cooling capacity in data centers. It proposes a PUE optimization framework with security guarantees for data centers, minimizing the data center's PUE by optimizing the controllable parameters of water-cooled system equipment while satisfying actual physical constraints, safe operation constraints, data center temperature constraints, and theoretical cooling capacity requirements to ensure security. Based on the fitted or learned correspondence between IT load and cooling capacity, the theoretical cooling capacity required to ensure security is obtained. Using a cooling capacity prediction model based on a supervised autoencoder and a PUE prediction model, a global optimization method is employed to obtain a set of water-cooled unit control parameters that meet cooling capacity requirements and minimize PUE. This invention enables the prediction of PUE and cooling capacity in data centers, offering advantages of low complexity and high accuracy. It provides accurate PUE and actual cooling capacity estimates for PUE optimization algorithms, thereby effectively ensuring the security of the data center PUE optimization process.
Need to check novelty before this filing date? Find Prior Art