风光水互补发电系统与需求侧数据中心联动的优化调度方法

风光水互补发电系统中,积极利用需求侧的数据中心参与响应,在平衡风光间歇性出力、减轻水电机组调节压力以及提高源端经济效益等方面能够产生明显效果。数据中心作为高能耗负荷类型,能耗峰谷差大,资源分配和数据处理方式灵活,减少一定的用户满意度将有效地实现电量需求转移。根据批处理负荷区间性处理要求,提出数据中心对工作负荷的并发处理模式,在此基础上,计及IT设备处理能耗、空调散热能耗以及网络通信能耗3类主要电力消耗,建立数据中心资源分配、延迟决策与能耗转移模型。针对梯级水电之间出力、流量、水头等相互约束关系,并计入风光的不确定性,制定以经济效益、负荷追踪系数及风光消纳程度的多重目标,建立含风、水、光的多能互补发电端与负荷端数据中心的优化调度模型。最后通过算例仿真得到在不同季节、天气和时段下风光水的最佳出力及数据中心的最小负荷转移量,验证了所提方法的有效性和正确性。

A Optimal Scheduling Method for Hybrid Wind-solar-hydro Power Generation System With Data Center in Demand Side

In hybrid wind-solar-hydro power generation system, actively using data centers to participate in demand response can produce positive effects in balancing intermittent output of wind and solar power, relieving stress of hydro turbine and enhancing economic interests of the source side. Of a high energy consumption load type, data centers have a big difference in electricity demand between peak and valley. Resource allocation and the way to deal with workloads for data centers are flexible. Reducing a certain degree of user satisfaction can effectively achieve transfer of electricity demand. According to requirement of batch load to finish in a period of time, a concurrent processing mode of data center to handle data is put forward in this paper. On this basis, three main types of power consumption, including IT equipment, air conditioning heat dissipation and network communication, are taken into account, and a model of data center for resource allocation, delaying decision and energy transfer is established. Considering the constraints of cascade hydropower, such as output, flow, water head, and so on, and uncertainty of wind and solar power, multiple objectives of economic efficiency, load tracking coefficient and accommodation degree of wind and solar power are formulated, and an optimal scheduling model for hybrid wind-solar-hydro power generation system and data center in load side is established. Finally, the optimal output of the source side and the minimum load transfer volume of the data center in different seasons, weather conditions and time intervals are obtained with simulation examples, and validity and correctness of the proposed method are verified.