AI Powered Data Center Optimization

“RDA’s software seamlessly model, visualize, analyze and optimize your energy systems, encompassing data-driven intelligence with deep learning engineering innovations.

We tailor cutting-edge AI applications that significantly reduce energy costs and CO2 emissions in data centers.” 

01.

Auto-calibration

The physics-guided model is able to auto-calibrate and achieve data quality within 10.5°C, providing an accurate and holistic overview.

02.

Data Augmentation

Generate massive amount of synthesized data with high quality and diversity (including "black swan" events such as emergencies and anomalies, peak/off peak loads and seasonal dynamics) within a short period of time.

03.

Action Validation

An automated cyber-physical control loop process allows control policies derived from well-trained Al models be quickly validated to mitigate deployment risk.

04.

Al Engine

A machine learning-based engine that utilizes deep reinforcement learning approach whereby system behaviours are learnt by iteratively interacting with the "environment" (data-driven neural network model) to derive an optimized. control policy for efficient energy management.

5. Autonomous: from modeling to updating, each step is automated

100%

4. Prescriptive: combined with Al, can perform control optimization and recommend optimized policy

100%

3. Predictive: what-if analysis

100%

2. Descriptive: connect with sensor for system state visualization, can perform anomaly detection

100%

1. Geometric: only geometrical information, can perform space positioning

100%

Complete Solution for DC Lifecycle

01.

Design

Build Digital Twin from Initial design

02.

Build

Power Usage Efficiency (PUE)
Water Usage Efficiency (WUE)
Carbon Usage Effectiveness (CUE)
Data Centre Power Density (DCPD)

03.

Operate

Availability
Efficiency
Thermal Safety

04.

Maintain

Asset Condition
Management
Early Threat
Detection

Use Case 1- Alibaba

Industry-grade Sandbox System based on Auto Calibration improved manageability of data centre

Reduced cost of sensors by

90%

Reduced stress testing time from

1 month to 1 week

“This reflects an effort to move ‘from zero to one’, as it marks the first completion of a real-time, high- precision sandbox system to help data center operators validate whether a change will cause issues, thus reducing the chance of outages.”

Alibaba Tech, 2019