Evaluating Coupling Models for Cloud Datacenters and Power Grids

Abstract

The rapid growth of datacenter (DC) loads can be leveraged to help meet renewable portfolio standard (RPS, renewable fraction) targets in power grids. The ability to manipulate DC loads over time (shifting) provides a mechanism to deal with temporal mismatch between non-dispatchable renewable generation (e.g. wind and solar) and overall grid loads, and this flexibility ultimately facilitates the absorption of renewables and grid decarbonization. To this end, we study DC-grid coupling models, exploring their impact on grid dispatch, renewable absorption, power prices, and carbon emissions. With a detailed model of grid dispatch, generation, topology, and loads, we consider three coupling approaches: fixed, datacenterlocal optimization (online dynamic programming), and grid-wide optimization (optimal power flow).Results show that understanding the effects of dynamic DC load management requires studies that model the dynamics of both load and power grid. Dynamic DC-grid coupling can produce large improvements: (1) reduce grid dispatch cost (-3%), (2) increase grid renewable fraction (+1.58%), and (3) reduce DC power cost (-16.9%). It also has negative effects: (1) increase cost for both DCs and non-DC customers, (2) differentially increase prices for non-DC customers, and (3) create large power-level changes that may harm DC productivity.

Reducing Datacenter Operational Carbon Emissions Effectively by Cooperating with the Grid

“…Amazon, Microsoft, Google, Alibaba) are building larger and more datacenters to meet the needs of digitalization, which accelerated since the Covid-19 pandemic [24, 62,69]. The rapid growth of hyperscale cloud industry is reflected in its power consumption: 1% of worldwide power consumption today, projections suggest it could exceed 10% by 2030 [38,46,56]. Similar rapid growth rates are also supported by corporate renewable power purchase data [39,63].…”

“…Datacenter load flexibility is the basis of power adaptation. We follow the datacenter load flexibility model in [46] wherein every datacenters adjust load within a dynamic range and can defer workload (backlog), but must catch up within a 24-hour day (Figure 5). Formally,…”

“…We evaluate the coupling approaches under a realistic grid model. Grid operation is simulated by solving the direct-current optimal power flow (DC-OPF) problem in [41,46] and Appendix A, which minimizes the grid dispatch cost in one-day time horizon with hourly intervals, subject to typical grid constraints. The grid metrics for carbon optimization (ACI, GPrice, LMPrice) are derived from the OPF solutions.…”

“…As of 2021, the total power consumption of these three leading cloud providers is over 64 TWh, equivalent to the power consumption of approximately 6.4 million American homes. Some estimates project datacenter power consumption growing to more than 10% of global electric power use by 2030 [38,46,56]. The largest cloud datacenter sites consist of multiple buildings with power footprints from 200 MW to 1 GW [2,6,29,59,67].…”

“…21.7% increase in datacenter carbon emissions from large loads' overshifting in Section 3). As a result, it's important for datacenter power adaptation to be studied with power grid models [46]. One consequence of this new insight is that there is no known solution to coupled management of large-scale datacenter power adaptation and power grids.…”