| With the improvement of social informatization,the scale and energy consumption of data centers increase rapidly,and the cooling energy consumption accounts for about26.7% of the total energy consumption.Meanwhile,data centers generate a large amount of waste heat and discharge it into the environment.Therefore,the research on the cooling system and waste heat utilization of data center is very important for its energy efficiency optimization and ecological environment protection.In this paper,a waste heat recovery system of data center based on ground source heat pump is proposed,which is used to provide heating for surrounding buildings.Then,a renewable energy generation and consumption type energy system is constructed,which endows data center with the identity of "producer of consumption" and improves the thermodynamic energy efficiency and perfection of the system.Integrating data centers into energy systems also provides more scope for efficiency improvements.First of all,this paper constructed a data center as the core of the production and consumption of energy system,design the structure and operation mode of the system,given the operating conditions under different operation modes,the structure design of the evaporator arranged in the cabinet,and completed the selection of other equipment.Secondly,this paper constructs the energy consumption model of the above production and consumption type energy system,and carries on the simulation analysis to the system operation.Aiming at the typical data center energy consumption scenario,the conventional plan of room-level data center cooling and building heating and the conventional plan of this system are constructed.The results show that this system can achieve 53.80% energy consumption,4.12 times exergy efficiency and 190.21% annual net income of the conventional system.Then,in this paper,wind power and photovoltaic are used together with the power grid to supply energy for the above system,and the energy consumption model of the energy consumption system with high proportion of renewable energy is constructed.In order to improve the overall economy and energy efficiency of the system,three measures,including building heating temperature regulation,data center computational load scheduling and battery charge and discharge volume regulation,are proposed,and corresponding optimization strategies are formulated.After calculation,the annual total power consumption,renewable energy waste,power consumption and operation cost of the system can be reduced by 4.18%,11.39%,8.89% and 6.96%,respectively.Finally,this paper uses reinforcement learning algorithm to optimize the operation of the energy system with high proportion of renewable energy.Under different average occupancy rates of data centers,the amount of renewable energy waste and operating cost can be further reduced by 40.40%-102.54% and 8.11%-60.88% by using algorithm optimization compared with using strategy optimization.By comparing and analyzing the characteristics of the two optimization schemes,the applicable scenarios of the two optimization schemes are given. |
