AI Lite
AI Lite adopts the minimalist design concept, fully integrates the local edge computing power, deeply integrates the air conditioning control strategy, has strong safety and reliability, and truly brings a light use experience of one click delivery and quick effect. In order to ensure the adaptability and high availability of the system, AI Lite takes a single micro module as the minimum control unit and adopts the algorithm scheme of reinforcement learning + regression prediction
Saved This Product to Your Dashboard
You just saved this product to your dashboard to view at a later time. You can easily remove the item from your dashboard when you no longer wish to have it saved.
Please login or create an account to save this for later
- Banking, Financial and Insurance
- Data Center/Colocation/Hosting
- Education
- Government
- Manufacturing
- Telecom
- Transportation
AI Lite
AI Lite adopts the minimalist design concept, fully integrates the local edge computing power, deeply integrates the air conditioning control strategy, has strong safety and reliability, and truly brings a light use experience of one click delivery and quick effect. In order to ensure the adaptability and high availability of the system, AI Lite takes a single micro module as the minimum control unit and adopts the algorithm scheme of reinforcement learning + regression prediction
- Banking, Financial and Insurance
- Data Center/Colocation/Hosting
- Education
- Government
- Manufacturing
- Telecom
- Transportation
Specifications
With the rapid development of AI technology, the intelligent energy-saving application scheme of AI + computer room group control has gradually become popular. As the proponent of deep reinforcement learning, Google has successfully applied its algorithm to the energy saving of chilled water system. However, due to the large dependence on historical data, high computational power requirements and long commissioning cycle, it has always been difficult to land. Aiming at this technical bottleneck, Weidi technology innovation has developed a lightweight and group control scheme AI Lite based on air cooling system. In order to ensure the adaptability and high availability of the system, the scheme takes a single micro module as the minimum control unit and adopts the algorithm scheme of reinforcement learning + regression prediction, which reduces the dependence on historical data and computing power, speeds up the convergence speed, and can automatically adapt to various modular computer room scenarios. The scheme is built in the micro module control unit of Vertiv, which can be built together with the module and deployed with one click.
Benefits
Sa3.0 hardware management platform integrated AI computing power
Adapt to frequently changing environment
Keep pue optimal in a rapidly changing environment
Adjust the safety margin of the algorithm on demand
Optimize pue while avoiding local hot spots
Adjustment without repeated manual intervention
Fully automatic adjustment, eliminating manual parameter adjustment
Features
Reinforcement learning + regression prediction, high algorithm efficiency
Convergence rate block
The optimal convergence can be achieved by an average of about one week
Stable and adjustable temperature
Seeking optimal energy saving on the premise of achieving safe temperature
Environmental parameter adaptation
The algorithm automatically adjusts the response parameters through learning
