数据中心末端空调系统运行策略确定方法及装置
By constructing a Markov decision process model and a temperature field distribution model for the terminal air conditioning system of a data center, and combining reinforcement learning and sequential optimization methods, multiple strategies are generated and the optimal strategy is selected in a real environment. This solves the problem of high energy consumption in the terminal air conditioning system of the data center, ensures the thermal safety of server IT equipment, and reduces energy consumption.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TSINGHUA UNIVERSITY
- Filing Date
- 2022-12-08
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies struggle to effectively reduce the energy consumption of data center terminal air conditioning systems while ensuring the thermal safety of server IT equipment. Furthermore, traditional methods suffer from problems such as complex model building and unstable training strategy performance during the strategy optimization process.
A Markov decision process model for a data center terminal air conditioning system is constructed using reinforcement learning algorithms to generate multiple operating strategies. By building a temperature field distribution model and a strategy library, the optimal strategy is selected in a real environment using the ordinal optimization method. The model is then trained and evaluated using a neural network and a linearly weighted policy function based on basis functions.
It achieves the goal of minimizing the energy consumption of terminal air conditioning while ensuring the thermal safety of server IT equipment, and reduces the number of policy evaluations through sequence optimization methods, saving manpower, material resources and financial resources.
Smart Images
Figure CN115983438B_ABST